analysis – Robohub https://robohub.org Connecting the robotics community to the world Mon, 05 Oct 2026 01:29:00 +0000 en-US hourly 1 https://wordpress.org/?v=7.1.3 To err is algorithm: Algorithm fallibility and economic organisation https://robohub.org/to-err-is-algorithm-algorithm-fallibility-and-economic-organisation/ Thu, 18 May 2017 13:00:40 +0000 http://robohub.org/to-err-is-algorithm-algorithm-fallibility-and-economic-organisation/

Algorithmic fails

Dig below the surface of some of today’s biggest tech controversies and you are likely to find an algorithm misfiring:[1]

These errors are not primarily caused by problems in the data that can make algorithms discriminatory, or their inability to improvise creatively. No, they stem from something more fundamental: the fact that algorithms, even when they are generating routine predictions based on non-biased data, will make errors. To err is algorithm.

The costs and benefits of algorithmic decision-making

We should not stop using algorithms simply because they make errors.[2] Without them, many popular and useful services would be unviable.[3] However, we need to recognise that algorithms are fallible, and that their failures have costs. This points at an important trade-off between more (algorithm-enabled) beneficial decisions and more (algorithm-caused) costly errors. Where lies the balance?

Economics is the science of trade-offs, so why not think about this topic like economists? This is what I have done ahead of this blog, creating three simple economics vignettes that look at key aspects of algorithmic decision-making.[4] These are the key questions:

  • Risk: when should we leave decisions to algorithms, and how accurate do those algorithms need to be?
  • Supervision: How do we combine human and machine intelligence to achieve desired outcomes?
  • Scale: What factors enable and constrain our ability to ramp-up algorithmic decision-making?

The two sections that follow give the gist of the analysis and its implications. The appendix at the end describes the vignettes in more detail (with equations!).

Modelling the modelling

1. Risk: go with the odds

As the American psychologist and economist Herbert Simon once pointed out:

in an information rich world, attention becomes a scarce resource.

This applies to organisations as much as it does to individuals.

The ongoing data revolution risks overwhelming our ability to process information and make decisions, and algorithms can help address this. They are machines that automate decision-making, potentially increasing the number of good decisions that an organisation can make.[5] This explains why they have taken-off first in industries where the volume and frequency of potential decisions goes beyond what a human workforce can process.[6]

What drives this process? For an economist, the main question is how much value will the algorithm create with its decisions. Rational organisations will adopt algorithms with high expected values.

An algorithm’s expected value depends on two factors: its accuracy (the probability that it will make a correct decision), and the balance between the reward from a correct decision and the penalty from an erroneous one.[7]  Riskier decisions (where penalties are big compared to rewards) should be made by highly accurate algorithms. You would not want a flaky robot running a nuclear power station, but it might be ok if it is simply advising you about what TV show to watch tonight.

2. Supervision: watch out

We could bring in human supervisors to check the decisions made by the algorithm and fix any errors they find. This makes more sense if the algorithm is not very accurate (supervisors do not spend a lot of time checking correct decisions), and the net benefits from correcting the wrong decisions (i.e., extra rewards plus avoided penalties) is high. Costs matter too. A rational organisation has more incentives to hire human supervisors if they do not get paid a lot, and if they are highly productive (i.e. it only takes a few of them to do the job).

Following from the example before, if a human supervisor fixes a silly recommendation in a TV website, this is unlikely to create a lot of value for the owner. The situation in a nuclear power station is completely different.

3. Scale: a race between machines and reality

What happens when we scale-up the number of algorithmic decisions? Are there any limits to its growth?

This depends on several things, including whether algorithms gain or lose accuracy as they make more decisions, and the costs of ramping-up algorithmic decision-making. In this situation, there are two interesting races going on.

1. There is a race between an algorithm’s ability to learn from the decisions it makes, and the amount of information that it obtains from new decisions. New machine learning techniques help algorithms ‘learn from experience’, making them more accurate as they make more decisions.[8] However, more decisions can also degrade an algorithm’s accuracy. Perhaps it is forced to deal with weirder cases, or new situations it is not trained to deal with.[9] To make things worse, when an algorithm becomes very popular (makes more decisions), people have more reasons to game it.

My prior is that the ‘entropic forces’ that degrade algorithm accuracy will win out in the end: no matter how much more data you collect, it is just impossible to make perfect predictions about a complex, dynamic reality.

2. The second race is between the data scientists creating the algorithms and the supervisors checking these algorithm’s decisions. Data scientists are likely to ‘beat’ the human supervisors because their productivity is higher: a single algorithm, or an improvement in an algorithm, can be scaled up over millions of decisions. By contrast, supervisors need to check each decision individually. This means that as the number of decisions increases, most of the organisation’s labour bill will be spent on supervision, with potentially spiralling costs as the supervision process gets bigger and more complicated.

What happens at the end?

When considered together, the decline in algorithmic accuracy and the increase in labour costs I just described are likely to limit the number of algorithmic decisions an organisation can make economically. But if and when this happens depends on the specifics of the situation.

Implications for organisations and policy

The processes I discussed above have many interesting organisational and policy implications. Here are some of them:

1. Finding the right algorithm-domain fit

As I said, algorithms making decisions in situations where the stakes are high need to be very accurate to make-up for high penalties when things go wrong.[10] On the flipside, if the penalty from making an error is low, even inaccurate algorithms might be up to the task.

For example, the recommendation engines in platforms like Amazon or Netflix often make irrelevant recommendations, but this is not a big problem because the penalty from these errors is relatively low – we just ignore them. Data scientist Hillary Parker picked up on the need to consider the fit between model accuracy and decision context a recent edition of the ‘Not So Standard Deviations’ podcast:

Most statistical methods have been tuned for the clinical trial implementation where you are talking about people’s lives and people dying with the wrong treatment, whereas in business settings the trade-offs are completely different.

One implication from this is that organisations in ‘low-stakes’ environments can experiment with new and unproven algorithms, including some with low-accuracy early on. As these are improved, they can be transferred to ‘high stake domains’. The tech companies that develop these algorithms often release them as open source software for others to download and improve, making these spill-overs possible.

2. There are limits to algorithmic decision-making in high stakes domains

Algorithms need to be applied much more carefully in domains where the penalties from errors are high, such as health or the criminal justice system, and when dealing with groups who are more vulnerable to algorithmic errors.[11] Only highly accurate algorithms are suitable for these risky decisions, unless they are complemented with expensive human supervisors who can find and fix errors. This will create natural limits to algorithmic decision-making: how many people can you hire to check an expanded number of decisions? Human attention remains a bottleneck to more decisions.

If policymakers want more and better use of algorithms in these domains, they should invest in R&D to improve algorithmic accuracy, encourage the adoption of high-performing algorithms from other sectors, and experiment with new ways of organising that help algorithms and their supervisors work better as a team.

Commercial organisations are not immune to some of these problems: YouTube has for example started blocking adverts in videos with less than ten thousand views. In those videos, the rewards from correct algorithmic ad-matching is probably low (they have low viewership) and the penalties could be high (many of these videos are of dubious quality). In other words, these decisions have low expected value, so YouTube has decided to stop making them. Meanwhile, Facebook just announced that it is hiring 3,000 human supervisors (almost a fifth of its current workforce) to moderate the content in its network. You could imagine how the need to supervise more decisions might put some brakes on its ability to scale up algorithmic decision-making indefinitely.

3. The pros and cons of crowdsourced supervision

One way to keep supervision costs low and coverage of decisions high is to crowdsource supervision to users, for example by giving them tools to report errors and problems. YouTube, Facebook and Google have all done this in response to their algorithmic controversies. Alas, getting users to police online services can feel unfair and upsetting. As Sarah T Roberts, a Law professor pointed out in a recent interview about the Facebook violent video controversy:

The way this material is often interrupted is because someone like you or me encounters it. This means a whole bunch of people saw it and flagged it, contributing their own labour and non-consensual exposure to something horrendous. How are we going to deal with community members who may have seen that and are traumatized today?

4. Why you should always keep a human in the loop

Even when penalties from error are low, it still makes sense to keep humans in the loop of algorithmic decision-making systems.[12] Their supervision provides a buffer against sudden declines in performance if (as) the accuracy of algorithms decreases.  When this happens, the number of erroneous decisions detected by humans and the net benefit from fixing them increase. They can also ring the alarm, letting everyone know that there is a problem with the algorithms that needs fixing.[13]

This could be particularly important in situations where errors create penalties with a delay, or penalties that are hard to measure or hidden (say if erroneous recommendations result in self-fulfilling prophecies, or costs that are incurred outside the organisation).

There are many examples of this. In the YouTube advertising controversy, the big accumulated penalty from previous errors only became apparent with a delay, when brands noticed that their adverts were being posted against hate videos. The controversy with fake news after the US election is an example of hard to measure costs: algorithms’ inability to discriminate between real news and hoaxes creates costs for society, potentially justifying stronger regulations and more human supervision. Politicians have made this point when calling on Facebook to step up its fight against fake news in the run-up to the UK election:

Looking at some of the work that has been done so far, they don’t respond fast enough or at all to some of the user referrals they can get. They can spot quite quickly when something goes viral. They should then be able to check whether that story is true or not and, if it is fake, blocking it or alerting people to the fact that it is disputed. It can’t just be users referring the validity of the story. They [Facebook] have to make a judgment about whether a story is fake or not.

5. From abstract models to real systems

Before we use economic models to inform action, we need to define and measure model accuracy, penalties and rewards, changes in algorithmic performance due to environmental volatility, levels of supervision and their costs, and that is only the beginning.[14]

This is hard but important work that could draw on existing technology assessment and evaluation tools, including methods to quantify non-economic outcomes (e.g. in health).[15] One could even use rich data from an organisation’s information systems to simulate the impact of algorithmic decision-making and its organisation before implementing it. We are seeing more examples of these applications, such as the financial ‘regtech’ pilots that the European Commission is running, or the ‘collusion incubators’ mentioned in a recent Economist article on price discrimination.

Coda: Piecemeal social engineering in an age of algorithms

In a Nature article last year, US researchers Ryan Calo and Kate Crawford called for

a practical and broadly applicable social-systems analysis [that] thinks through all the possible effects of AI systems on all parties [drawing on] philosophy, law, sociology, anthropology and science-and-technology studies, among other disciplines.

Calo and Crawford did not include economists in their list. Yet as this blog suggests, economics thinking has much to contribute to these important analyses and debates. Thinking about algorithmic decisions in terms of their benefits and costs, the organisational designs we can use to manage their downsides, and the impact of more decisions on the value that agorithms create can help us make better decisions about when and how to use them.

This reminds me of a point that Jaron Lanier made in his 2010 book, Who Owns the Future:

With every passing year, economics must become more and more about the design of the machines that mediate human social behaviour. A networked information system guides people in a more direct, detailed and literal way than does policy. Another way to put it is that economics must turn into a large-scale, systemic version of user interface design.

Designing organisations where algorithms and humans work together to make better decisions will be an important part of this agenda.

Acknowledgements

This blog benefited from comments from Geoff Mulgan, and was inspired by conversations with John Davies. The image above represents a precision-recall curve in a multi-label classification problem. It shows the propensity of a random forests classification algorithm to make mistakes when one sets different rules (probability thresholds) for putting observations in a category.

Appendix: Three economics vignettes about algorithmic decision-making

The three vignettes below are very simplified formalisations of algorithmic decision-making situations. My main inspiration was Human fallibility and economic organization, a 1985 paper by Joe Stiglitz and Raj Sah where the authors model how two organisational designs – hierarchies and ‘polyarchies’ (flat organisations) – cope with human error. Their analysis shows that hierarchical organisations where decision-makers lower in the hierarchy are supervised by people further up tend to reject more good projects, while polyarchies where agents make decisions independently from each other, tend to accept more bad projects. A key lesson from their model is that errors are inevitable, and the optimal organisational design depends on context.

Vignette 1: Algorithm says maybe

Let’s imagine an online video company that matches adverts with videos in its catalogue. This company hosts millions of videos so it would be economically inviable for it to rely on human labour to do the job. Instead, its data scientists develop algorithms to do this automatically. [16] The company looks for the algorithm that maximises the expected value of the matching decisions. This value depends on three factors: [17]

-Algorithm accuracy (a): The probability (between 0 and 1) that the algorithm will make the correct decision.[18]

-Decision reward (r): This is the reward when the algorithm makes the right decision

-Error penalty (p): This is the cost of making the wrong decision.

We can combine accuracy, benefit and penalty to calculate the expected value of the decision:

E = ar – (1-a)p [1]

This value is positive when the expected benefits from the algorithm’s decision outweigh the expected costs (or risks):

ar > (1-a)p [2]

Which is the same as saying that:

a/(1-a) > p/r [3]

The odds of making the right decision should be higher than the ratio between penalty and benefit.

Enter human

We can reduce the risk of errors by bringing a human supervisor into the situation. This human supervisor can recognise and fix errors in algorithmic decisions. The impact of this strategy on the expected value of a decision depends on two parameters:

-Coverage ratio (k): k is the probability that the human supervisor will check a decision by the algorithm. If k is 1, this means that all algorithmic decisions are checked by a human.

-Supervision cost (cs(k)): this is the cost of supervising the decisions of the algorithm. The cost depends on the coverage ratio k because checking more decisions takes time.

The expected value of an algorithmic decision with human supervision is the following:[19]

Es = ar + (1-a)kr – (1-a)kp – cs(k) [4]

This equation picks up the fact that some errors are detected and rectified, and others are not. We subtract [3] from [4] to obtain the extra expected value from supervision. After some algebra, we get this.

(r+p)(1-a)k > cs(k) [5]

Supervision only makes economic sense when its expected benefit (which depends on the probability that the algorithm has made a mistake, that this mistake is detected, and the net benefits from flipping a mistake into a correct decision) is larger than the cost of supervision.

Scaling up

Here, I consider what happens when we start increasing n, the number of decisions being made by the algorithm.

The expected value is:

E(n) = nar + n(1-a)kr – n(1-a)(1-k)p [6]

And the costs are C(n)

How do these things change as n grows?

I make some assumptions to simplify things: the organisation wants to hold k constant, and the rewards r and penalties p remain constant as n increases.[20]

This leaves us with two variables that change as n increases: a and C.

  • I assume that algorithmic accuracy a declines with the number of decisions because the processes that degrade accuracy are stronger than those that improve it
  • I assume that C, the production costs, only depend on the labour of data scientists and supervisors. Each of these two occupations gets paid a salary wds and ws.

Based on this, and some calculus, we get the changes in expected benefits as we make more decisions as:

∂E(n)/∂(n) = r + (a+n(∂a/∂n))*(1-k)(r+p) - p(1-k) [7]

This means that as more decisions are made, the aggregated expected benefits grow in a way that is modified by changes in the marginal accuracy of the algorithm. On the one hand, more decisions mean scaled up benefits from more correct decisions. On the other, the decline in accuracy generates an increasing number of errors and penalties. Some of these are offset by human supervisors.

This is what happens with costs:

∂C/∂n = (∂C/∂Lds)(∂Lds/∂n) + (∂C/dLs)(∂Ls/dn) [8]

As the number of decisions increases, costs grow because the organisation has to recruit more data scientists and supervisors.

[8] is the same as saying:

∂C/dn = wds/(∂Lds/dn) + ws/zs/(∂Ls/∂n) [9]

The labour costs of each occupation are directly related to its salary, and inversely related to its marginal productivity. If we assume that data scientists are more productive than supervisors, this means that most of the increases in costs with n will be caused by increases in the supervisor workforce.

The expected value (benefits minus costs) from decision-making for the organisation is maximised with an equilibrium number of decisions ne where the marginal value of an extra decision equals its marginal cost:

r + (a+nda/dn)(1-k)(r+p) - p(1-k) = wds/(∂Lds/∂n) + ws/zs/(∂Ls/∂n) [10]

Extensions

Above, I have kept things simple by making some strong assumptions about each of the situations being modelled. What would happen if we relaxed these assumptions?

Here are some ideas:

Varieties of error

First, the analysis does not take into account that different types of errors (e.g. false positives and negatives, errors made with different degrees of certainty etc.) could have different rewards and penalties. I have also assumed certainty in rewards and penalties, when it would be more realistic to model them as random draws from probability distributions. This extension would help incorporate fairness and bias into the analysis. For example, if errors are more likely to affect vulnerable people (who suffer higher penalties), and these errors are less likely to be detected, this could increase the expected penalty from errors.

Humans are not perfect either

All of the above assumes that algorithms err but humans do not. This is clearly not the case. In many domains, algorithms can be a desirable alternative to humans with deep-rooted biases and prejudices. In those situations, human’s ability to detect and address errors is impaired, and this reduces the incentives to recruit them (this is the equivalent to a decline in their productivity). Organisations deal with all this by investing on technologies (e.g. crowdsourcing platforms) and quality assurance systems (including extra layers of human and algorithmic supervision) that manage the risks of human and algorithmic fallibility.

Scaling up rewards and penalties

Before, I assumed that the marginal penalties and rewards remain constant as the number of algorithmic decisions increase. This need not be the case. The table below shows examples of situations where these parameters change with the number of decisions being made:

Increases with more decisions Decreases with more decisions
Rewards The organisation gains market power, or is able to use price discrimination in more transactions The organisation runs out of valuable decisions to make.
Penalties The organisation becomes more prominent and its mistakes receive more attention Users get accustomed to errors

Getting an empirical handle on these processes is very important, as they could determine if there is a natural limit to the number of algorithmic decisions that an organisation can make economically in a domain or market, with potential implications for its regulation.

Endnotes

[1] I use the term ‘algorithm’ in a restricted sense, to refer to technologies that turn information into predictions (and depending on the system receiving the predictions, decisions). There are many processes to do this, including rule-based systems, statistical systems, machine learning systems and Artificial Intelligence (AI). These systems vary on their accuracy, scalability, interpretability, and ability to learn from experience, so their specific features should be considered in the analysis of algorithmic trade-offs.

[2] One could even say that machine learning is the science that manages trade-offs caused by the impossibility of eliminating algorithmic error. The famous ‘bias-variance’ trade off between fitting a model to known observations and predicting unknown ones is a good example of this.

[3] Some people would say that personalisation is undesirable because it can lead to discrimination and ‘filter bubbles’, but that is a question for another blog post.

[4] Dani Rodrik’s ‘Economics Rules’ makes a compelling case for models as simplistic but useful formalisations of complex reality.

[5] In a 2016 Harvard Business Review article, Ajay Agrawal and colleagues sketched out an economic analysis of machine learning as a technology that lowers the costs of prediction. My way of looking at algorithms is similar because predictions are inputs into decision-making.

[6] This includes personalised experiences and recommendations in e-commerce and social networking sites, or fraud detection and algorithmic trading in finance.

[7] For example, if YouTube shows me an advert which is highly relevant to my interests, I might buy the product, and this generates income for the advertiser, the video producer and YouTube. If it shows me a completely irrelevant or even offensive advert, I might stop using YouTube, or kick up a fuss in my social network of choice.

[8] Reinforcement learning builds agents that use the rewards and penalties from previous actions to make new decisions.

[9] This is what happened with the Google FluTrends system used to predict flu outbreaks based on google searches – people changed their search behaviour, and the algorithm broke down.

[10] In many cases, the penalties might be so high that we decide that an algorithm should never be used, unless it is supervised by humans.

[11] Unfortunately, care is not always taken when implementing algorithmic systems in high-stakes situations. Cathy O’Neil’s ‘Weapons of Maths Destruction’ gives many examples of this, going from the criminal justice system to university admissions.

[12] Mechanisms for accountability and due process are another example of human supervision.

[13] Using Albert Hirschmann’s model of exit, voice and loyalty, we could say that supervision plays the role of ‘voice’, helping organisations detect a decline in quality before users begin exiting.

[14] The appendix flags up some of my key assumptions, and suggests extensions.

[15] This includes rigorous evaluation of algorithmic decision-making and its organisation using Randomised Controlled Trial methods like those proposed by Nesta’s Innovation Growth Lab.

[16] This decision could be based on how well similar adverts perform when matched with different types of videos, on demographic information about the people who watch the videos, or other things.

[17] The analysis in this blog assumes that the results of algorithmic decisions are independent from each other. This assumption might be violated in situations where algorithms generate self-fulfilling prophecies (e.g. logically, a user is more likely to click an advert she is shown that one she is not). This is a hard problem to tackle, but researchers are developing methods based on randomisation of algorithmic decisions to address it.

[18] This does not distinguish between different types of error (e.g. false positives and false negatives). I come back to this at the end.

[19] Here, I am assuming that human supervisors are perfectly accurate. As we know from behavioural economics, this is a very strong assumption. I consider this issue at the end.

[20] I consider the implications of making different assumptions about marginal rewards and penalties at the end.

This post was originally published on Nesta. Click here to view the original.

]]> iRobot, KUKA and other robotic stocks exceed earnings expectations https://robohub.org/irobot-kuka-and-other-robotic-stocks-exceed-earnings-expectations/ Thu, 04 May 2017 11:30:00 +0000 http://robohub.org/irobot-kuka-and-other-robotic-stocks-exceed-earnings-expectations/

iRobot (IRBT on the NASDAQ stock exchange) jumped from $70 to $80 per share on news that iRobot's quarterly earnings were so good that the company raised their forecast for 2017 to new highs. KUKA also had good Q1/17 earnings as did Intuitive Surgical.

iRobot (IRBT:NASDAQ)

iRobot's stock jumped 15.75% on news that it's Q1/17 earnings exceeded analyst expections and had jumped 28.8% from Q1/16. As a result, iRobot adjusted upward their 2017 revenue forecast. iRobot now expects full-year 2017 revenue of $780 million to $790 million, which should result in earnings per share of $1.45 to $1.70 (up from EPS guidance of $1.35 to $1.65 previously). 

Total Q1/17 units shipped of iRobot's household cleaning robots was 704,000, a 28% rise from 550,000 shipped in Q1/16. According to CEO Colin Angle, iRobot has shipped over 15 million robotic home floor cleaners through 2016.

Last year iRobot divested its Defense and Security Division for $45 million to a VC which shortly thereafter launched Endeavor Robotics. Thus iRobot is now exclusively a commercial robotics provider. Year-to-date, iRobot's stock has risen 38% from $58 to $80. 

KUKA AG (KU2:ETR) / Midea Group (000333:SHE)

KUKA AG reported Q1/17 sales revenues of $862 million which was up 25.6% compared to Q1/16 revenue. Order backlog for the same period was up 30.6%. KUKA's outlook for 2017 is around $3.4 billion, up 7.5% over 2016 ($3.16 billion). KUKA's listed shares - only 5.4% remain with institutions and private investors, the rest, 94.6%, is held by Midea - remain listed but lightly traded as part of the agreement with Midea and will continue to be traded for at least 3 years.

Midea Group, a consumer products manufacturer, is one of the top 50 Chinese publicly-traded companies with revenues of $22 billion. It reported 2016 profits up by 15.6% year-over-year. In addition to its 2016 acquisition of KUKA for approximately $3.9 billion (according to Bloomberg), it just recently acquired Israeli Servotronix for $170 million. Servotronix adds to Midea's industrial automation acquisitions by being a motion control provider for robotics, printing, machine tools and electronics industries.

Midea chairman and CEO Paul Fang said, “This strategic alliance represents another milestone of Midea’s expansion in industrial automation and intelligent manufacturing. We believe that Servotronix’ technological leadership and innovation in motion control will generate significant synergies with Midea in terms of value chain integration and new market development. By leveraging each other’s complementary capabilities and resources, the two companies will join forces to develop exciting new products and explore growth opportunities going forward.”

Midea Group press releases regarding their acquisition of KUKA said almost the same thing: "By leveraging each other's complementary capabilities and resources...."

Intuitive Surgical (ISRG:NASDAQ)

Intuitive Surgical, a medical equipment provider and inventor of the da Vinci surgical robot system, reported $5.09 earnings per share (EPS) for the quarter, topping analysts’ consensus estimates of $3.97. Intuitive Surgical had a net margin of 27.21% and a return on equity of 14.24% with revenue of $674.20 million for the quarter, compared to the estimate of $664.72 million. During the same period in the previous year, the firm earned $4.42 EPS. Intuitive's revenue for the quarter was up 13.4% on a year-over-year basis. ISRG has climbed over 30% so far this year, from $642 to $837.

ROBO Global Robotics & Automation Index

iRobot and Intuitive Surgical are members, and KUKA used to be a member, of the 80+ stocks included in the ROBO Global Robotics & Automation Index, a leading indicator of the robotics and automation market. KUKA was dropped when it was acquired by Midea. Even though KUKA has revenues of around $3.4 billion, when combined with Midea's $22 billion, the percentage of robotics-related revenue is only 13% which is too little to qualify Midea for membership in the index. Also, Midea only trades on the Shenzhen Stock Exchange which, at the present time, ROBO Global doesn't cover.

For more information about the index or the 80+ member companies, go to ROBO Global's website.

Read more

]]>

iRobot (IRBT on the NASDAQ stock exchange) jumped from $70 to $80 per share on news that iRobot’s quarterly earnings were so good that the company raised their forecast for 2017 to new highs. KUKA also had good Q1/17 earnings as did Intuitive Surgical.

iRobot (IRBT:NASDAQ)

iRobot‘s stock jumped 15.75% on news that it’s Q1/17 earnings exceeded analyst expections and had jumped 28.8% from Q1/16. As a result, iRobot adjusted upward their 2017 revenue forecast. iRobot now expects full-year 2017 revenue of $780 million to $790 million, which should result in earnings per share of $1.45 to $1.70 (up from EPS guidance of $1.35 to $1.65 previously).

Total Q1/17 units shipped of iRobot’s household cleaning robots was 704,000, a 28% rise from 550,000 shipped in Q1/16. According to CEO Colin Angle, iRobot has shipped over 15 million robotic home floor cleaners through 2016.

Last year iRobot divested its Defense and Security Division for $45 million to a VC which shortly thereafter launched Endeavor Robotics. Thus iRobot is now exclusively a commercial robotics provider. Year-to-date, iRobot’s stock has risen 38% from $58 to $80.

KUKA AG (KU2:ETR) / Midea Group (000333:SHE)

KUKA AG reported Q1/17 sales revenues of $862 million which was up 25.6% compared to Q1/16 revenue. Order backlog for the same period was up 30.6%. KUKA’s outlook for 2017 is around $3.4 billion, up 7.5% over 2016 ($3.16 billion). KUKA’s listed shares – only 5.4% remain with institutions and private investors, the rest, 94.6%, is held by Midea – remain listed but lightly traded as part of the agreement with Midea and will continue to be traded for at least 3 years.

Midea Group, a consumer products manufacturer, is one of the top 50 Chinese publicly-traded companies with revenues of $22 billion. It reported 2016 profits up by 15.6% year-over-year. In addition to its 2016 acquisition of KUKA for approximately $3.9 billion (according to Bloomberg), it just recently acquired Israeli Servotronix for $170 million. Servotronix adds to Midea’s industrial automation acquisitions by being a motion control provider for robotics, printing, machine tools and electronics industries.

Midea chairman and CEO Paul Fang said, “This strategic alliance represents another milestone of Midea’s expansion in industrial automation and intelligent manufacturing. We believe that Servotronix’ technological leadership and innovation in motion control will generate significant synergies with Midea in terms of value chain integration and new market development. By leveraging each other’s complementary capabilities and resources, the two companies will join forces to develop exciting new products and explore growth opportunities going forward.”

Midea Group press releases regarding their acquisition of KUKA said almost the same thing: “By leveraging each other’s complementary capabilities and resources….”

Intuitive Surgical (ISRG:NASDAQ)

Intuitive Surgical, a medical equipment provider and inventor of the da Vinci surgical robot system, reported $5.09 earnings per share (EPS) for the quarter, topping analysts’ consensus estimates of $3.97. Intuitive Surgical had a net margin of 27.21% and a return on equity of 14.24% with revenue of $674.20 million for the quarter, compared to the estimate of $664.72 million. During the same period in the previous year, the firm earned $4.42 EPS. Intuitive’s revenue for the quarter was up 13.4% on a year-over-year basis. ISRG has climbed over 30% so far this year, from $642 to $837.

ROBO Global Robotics & Automation Index

iRobot and Intuitive Surgical are members, and KUKA used to be a member, of the 80+ stocks included in the ROBO Global Robotics & Automation Index, a leading indicator of the robotics and automation market. KUKA was dropped when it was acquired by Midea. Even though KUKA has revenues of around $3.4 billion, when combined with Midea’s $22 billion, the percentage of robotics-related revenue is only 13% which is too little to qualify Midea for membership in the index. Also, Midea only trades on the Shenzhen Stock Exchange which, at the present time, ROBO Global doesn’t cover.

For more information about the index or the 80+ member companies, go to ROBO Global’s website.

]]> As DJI dominates camera and commercial drone sales, drone funding falls https://robohub.org/as-dji-dominates-camera-and-commercial-drone-sales-drone-funding-falls/ Wed, 03 May 2017 12:00:00 +0000 http://robohub.org/as-dji-dominates-camera-and-commercial-drone-sales-drone-funding-falls/

According to AgFunder's 2016 AgTech Investing Report (supported by The Robot Report's own research), 2016 drone funding fell 64% from 2015 levels. Also, the type of companies getting funded were sensor, payload and analytics-based add-ons or service-providing companies rather than drone makers.

In 2015, according to AgFunder, the largest drone deal was DJI’s $75 million followed by 3D Robotics’ $64 million. In 2016, the largest deal was 3D Robotics’ $27 million and the majority of drone tech deals closed during the year were seed stage (15 out of 25), whereas there was an even split between late and seed stage deals in 2015.

The Robot Report's year-end article: 2016 best year ever for funding robotics startup companies included 27 UAS deals. The largest was for $43.8 million to a meds drone delivery service, Zipline International. The 2nd largest was $30 million to Airware, to fund their acquisition of Redbird (a French drone-powered analytics provider) and development of a set of technology systems to plan, fly, and analyze aerial data - particularly valuable for insurance adjusters and contractors. Then came 3D Robotics' $27 million which helped them wind down and reconfigure themselves from a drone maker to a drone services provider.

The mix of companies receiving funding in 2016 was different than in 2015 where the big money went to drone makers: DJI ($75M) 3D Robotics ($64M), Yuneec Electric Aviation ($60M), Ehang ($44M), and CyPhy Works ($25.4M).

2017 fundings continue the 2016 pattern: Drone Delivery raised $8.1M to further their depot to depot delivery system, Flytrex raised $3M to manufacture drone components, Flirty, a medicine and food delivery drone startup got $16M, Measure got $15M for their drones-as-a-service company for insurers, AirMap got $26M for a real-time air traffic management system, Dedrone got $15M for a drone tracking system, Airware raised an additional but undisclosed sum and Arbe Robotics got $2.5M for a real-time drone mapping system. No drone makers in the lot!

Shenzhen DJI Innovations

DJI is by far the leading quadcopter maker in the world. At a recent drone trade show, most of the booths that displayed multi-rotor drones doing specialized tasks used DJI drones to demonstrate their products. As drones are becomming less of a flying camera and hobby and more to provide commercial and industrial services, DJI is still leading the pack.

According to Bloomberg Businessweek, DJI is valued at $10 billion and makes 60-65% of all non-military drones. DJI designs, prototypes and manufacturers all their products in their own factories in Shenzhen and all their sub-contractors are in Shenzhen as well. Thus they control the supply chain and can produce new drones every 6 months thereby whipping the competition. Very Apple-like (after which they pattern themselves).

Nevertheless, even DJI has seen the change from drone-making to providing meaningful services with drones. It’s tough to maintain momentum with just hardware no matter how good it is. 

DJI, while continuing to develop new products and upgraded versions of their drones, has suffered the same limitations as other makers and service providers: limited battery life, regulations regarding line of sight and autonomy, connectivity issues, image processing and analysis, limited payload capacity and the commoditization of drones themselves. But, as Paul Turner, CEO of AgDNA wrote recently on Medium:

“Once drones are able to take off autonomously, scan a field [or object], upload data, recharge and continue operation without human intervention — this will be a game changer.”

We're almost there.

Read more

]]>

DJI Phantom 4 Pro

According to AgFunder’s 2016 AgTech Investing Report (supported by The Robot Report’s own research), 2016 drone funding fell 64% from 2015 levels. Also, the type of companies getting funded were sensor, payload and analytics-based add-ons or service-providing companies rather than drone makers.

In 2015, according to AgFunder, the largest drone deal was DJI’s $75 million followed by 3D Robotics’ $64 million. In 2016, the largest deal was 3D Robotics’ $27 million and the majority of drone tech deals closed during the year were seed stage (15 out of 25), whereas there was an even split between late and seed stage deals in 2015.

The Robot Report’s year-end article: 2016 best year ever for funding robotics startup companies included 27 UAS deals. The largest was for $43.8 million to a meds drone delivery service, Zipline International. The 2nd largest was $30 million to Airware, to fund their acquisition of Redbird (a French drone-powered analytics provider) and development of a set of technology systems to plan, fly, and analyze aerial data – particularly valuable for insurance adjusters and contractors. Then came 3D Robotics’ $27 million which helped them wind down and reconfigure themselves from a drone maker to a drone services provider.

The mix of companies receiving funding in 2016 was different than in 2015 where the big money went to drone makers: DJI ($75M) 3D Robotics ($64M), Yuneec Electric Aviation ($60M), Ehang ($44M), and CyPhy Works ($25.4M).

2017 fundings continue the 2016 pattern: Drone Delivery raised $8.1M to further their depot to depot delivery system, Flytrex raised $3M to manufacture drone components, Flirty, a medicine and food delivery drone startup got $16M, Measure got $15M for their drones-as-a-service company for insurers, AirMap got $26M for a real-time air traffic management system, Dedrone got $15M for a drone tracking system, Airware raised an additional but undisclosed sum and Arbe Robotics got $2.5M for a real-time drone mapping system. No drone makers in the lot!

Shenzhen DJI Innovations

DJI is by far the leading quadcopter maker in the world. At a recent drone trade show, most of the booths that displayed multi-rotor drones doing specialized tasks used DJI drones to demonstrate their products. As drones are becomming less of a flying camera and hobby and more to provide commercial and industrial services, DJI is still leading the pack.

According to Bloomberg Businessweek, DJI is valued at $10 billion and makes 60-65% of all non-military drones. DJI designs, prototypes and manufacturers all their products in their own factories in Shenzhen and all their sub-contractors are in Shenzhen as well. Thus they control the supply chain and can produce new drones every 6 months thereby whipping the competition. Very Apple-like (after which they pattern themselves).

Nevertheless, even DJI has seen the change from drone-making to providing meaningful services with drones. It’s tough to maintain momentum with just hardware no matter how good it is.

DJI, while continuing to develop new products and upgraded versions of their drones, has suffered the same limitations as other makers and service providers: limited battery life, regulations regarding line of sight and autonomy, connectivity issues, image processing and analysis, limited payload capacity and the commoditization of drones themselves. But, as Paul Turner, CEO of AgDNA wrote recently on Medium:

“Once drones are able to take off autonomously, scan a field [or object], upload data, recharge and continue operation without human intervention — this will be a game changer.”

We’re almost there.

]]> Envisioning the future of robotics https://robohub.org/envisioning-the-future-of-robotics/ Thu, 16 Mar 2017 10:07:46 +0000 http://robohub.org/envisioning-the-future-of-robotics/

Image: Ryan Etter

Robotics is said to be the next technological revolution. Many seem to agree that robots will have a tremendous impact over the following years, and some are heavily betting on it. Companies are investing billions buying other companies, and public authorities are discussing legal frameworks to enable a coherent growth of robotics.

Understanding where the field of robotics is heading is more than mere guesswork. While much public concern focuses on the potential societal issues that will arise with the advent of robots, in this article, we present a review of some of the most relevant milestones that happened in robotics over the last decades. We also offer our insights on feasible technologies we might expect in the near future.

Copyright © Acutronic Robotics 2017. All Rights Reserved.

Pre-robots and first manipulators

What’s the origin of robots? To figure it out we’ll need to go back quite a few decades to when different conflicts motivated the technological growth that eventually enabled companies to build the first digitally controlled mechanical arms. One of the first and well documented robots was UNIMATE (considered by many the first industrial robot): a programmable machine funded by General Motors, used to create a production line with only robots. UNIMATE helped improve industrial production at the time. This motivated other companies and research centers to actively dedicate resources to robotics, which boosted growth in the field.

Sensorized robots

Sensors were not typically included in robots until the 70’s. Starting in1968, a second generation of robots emerged that integrated sensors. These robots were able to react to their environment and offer responses that met varying scenarios.

Relevant investments were observed during this period. Industrial players worldwide were attracted by the advantage that robots promised.

Worldwide industrial robots:  Era of the robots

Many consider that the Era of Robots started in 1980. Billions of dollars were invested by companies all around to world to automate basic tasks in their assembly lines. Sales of industrial robots grew 80% above the previous years’.

Key technologies appeared within these years: General internet access was extended in 1980; Ethernet became a standard in 1983 (IEEE 802.3); the Linux kernel was announced in 1991; and soon after that real-time patches started appearing on top of Linux.

The robots created between 1980 and 1999 belong to what we call the third generation of robots: robots that were re-programmable and included dedicated controllers. Robots populated many industrial sectors and were used for a wide variety of activities: painting, soldering, moving, assembling, etc.

By the end of the 90s, companies started thinking about robots beyond the industrial sphere. Several companies created promising concepts that would inspire future roboticists. Among the robots created within this period, we highlight two:

  1. The first LEGO Mindstorms kit (1998): a set consisting of 717 pieces including LEGO bricks, motors, gears, different sensors, and a RCX Brick with an embedded microprocessor to construct various robots using the exact same parts. The kit allowed the learning of  basic robotics principles. Creative projects have appeared over the years showing the potential of interchangeable hardware in robotics. Within a few years. the LEGO Mindstorms kit became the most successful project that involved robot part interchangeability.
  2. Sony’s AIBO (1999): the world’s first entertainment robot. Widely used for research and development, Sony offered robotics to everyone in the form of a $1,500 robot that included a distributed hardware and software architecture. The OPEN-R architecture involved the use of modular hardware components — e.g. appendages that can be easily removed and replaced to customize the shape and function of the robots — and modular software components that could be interchanged to modify their behavior and movement patterns. OPEN-R inspired future robotic frameworks, and minimized the need for programming individual movements or responses.

Integration effort was identified as one of the main issues within robotics, particularly related to industrial robots. A common infrastructure typically reduces the integration effort by facilitating an environment in which components can be connected and made to interoperate. Each of the infrastructure-supported components are optimized for such integration at their conception, and the infrastructure handles the integration effort. At that point, components could come from different manufacturers (yet when supported by a common infrastructure, they will interoperate).

Sony’s AIBO and LEGO’s Mindstorms kit were built upon this principle, and both represented common infrastructures. Even though they came from the consumer side of robotics, one could argue that their success was strongly related to the fact that both products made use of interchangeable hardware and software modules. The use of a common infrastructure proved to be one of the key advantages of these technologies, however those concepts were never translated to industrial environments. Instead, each manufacturer, in an attempt to dominate the market, started creating their own “robot programming languages”.

The dawn of smart robots

Starting from the year 2000, we observed a new generation of robot technologies. The so-called fourth generation of robots consisted of more intelligent robots that included advanced computers to reason and learn (to some extend at least), and more sophisticated sensors that helped controllers adapt themselves more effectively to different circumstances.

Among the technologies that appeared in this period, we highlight the Player Project (2000, formerly the Player/Stage Project), the Gazebo simulator (2004) and the Robot Operating System (2007). Moreover, relevant hardware platforms appeared during these years. Single Board Computers (SBCs), like the Raspberry Pi, enabled millions of users all around the world to create robots easily.

The boost of bio-inspired artificial intelligence

The increasing popularity of artificial intelligence, and particularly neural networks, became relevant in this period as well. While a lot of the important work on neural networks happened in the 80’s and in the 90’s, computers did not have enough computational power at the time. Datasets weren’t big enough to be useful in practical applications. As a result, neural networks practically disappeared in the first decade of the 21st century. However, starting from 2009 (speech recognition), neural networks gained popularity and started delivering good results in fields such as computer vision (2012) or machine translation (2014). Over the last few years, we’ve seen how these techniques have been translated to robotics for tasks such as robotic grasping. In the coming years, we expect to see these AI techniques having more and more impact in robotics.

What happened to industrial robots?

Relevant key technologies have also emerged from the industrial robotics landscape (e.g.: EtherCAT). However, except for the appearance of the first so-called collaborative robots, the progress within the field of industrial robotics has significantly slowed down when compared to previous decades. Several groups have identified this fact and written about it with conflicting opinions. Below, we summarize some of the most relevant points encountered while reviewing previous work:

  • The Industrial robot industry :  is it only a supplier industry?
    For some, the industrial robot industry is a supplier industry. It supplies components and systems to larger industries, like manufacturing. These groups argue that the manufacturing industry is dominated by the PLC, motion control and communication suppliers which, together with the big customers, are setting the standards. Industrial robots therefore need to adapt and speak factory languages (PROFINET, ETHERCAT, Modbus TCP, Ethernet/IP, CANOPEN, DEVICENET, etc.) which for each factory, might be different.
  • Lack of collaboration and standardized interfaces in industry
    To date, each industrial robot manufacturer’s business model is somehow about locking you into their system and controllers. Typically, one will encounter the following facts when working with an industrial robot: a) each robot company has its own proprietary programming language, b) programs can’t be ported from one robot company to the next one, c) communication protocols are different, d) logical, mechanical and electrical interfaces are not standardized across the industry. As a result, most robotic peripheral makers suffer from having to support many different protocols, which requires a lot of development time that reduces the functionality of the product.
  • Competing by obscuring vs opening new markets?
    The closed attitude of most industrial robot companies is typically justified by the existing competition. Such an attitude leads to a lack of understanding between different manufacturers. An interesting approach would be to have manufacturers agree on a common infrastructure. Such an infrastructure could define a set of electrical and logical interfaces (leaving the mechanical ones aside due to the variability of robots in different industries) that would allow industrial robot companies to produce robots and components that could interoperate, be exchanged and eventually enter into new markets. This would also lead to a competitive environment where manufacturers would need to demonstrate features, rather than the typical obscured environment where only some are allowed to participate.

The Hardware Robot Operating System (H-ROS)

For robots to enter new and different fields, it seems reasonable that they need to adapt to the environment itself. This fact was previously highlighted for the industrial robotics case, where robots had to be fluent with factory languages. One could argue the same for service robots (e.g. households robots that will need to adapt to dish washers, washing machines, media servers, etc.), medical robots and many other areas of robotics. Such reasoning lead to the creation of the Hardware Robot Operating System (H-ROS), a vendor-agnostic hardware and software infrastructure for the creation of robot components that interoperate and can be exchanged between robots. H-ROS builds on top of ROS, which is used to define a set of standardized logical interfaces that each physical robot component must meet if compliant with H-ROS.

H-ROS facilitates a fast way of building robots, choosing the best component for each use-case from a common robot marketplace. It complies with different environments (industrial, professional, medical, …) where variables such as time constraints are critical. Building or extending robots is simplified to the point of placing H-ROS compliant components together. The user simply needs to program the cognition part (i.e. brain) of the robot and develop their own use-cases, all without facing the complexity of integrating different technologies and hardware interfaces.

The future ahead

With latest AI results being translated to robotics, and recent investments in the field, there’s a high anticipation for the near future of robotics.

As nicely introduced by Melonee Wise in a recent interview, there’s still not that many things you can do with a $1000-5000 BOM robot (which is what most people would pay on an individual basis for a robot). Hardware is still a limiting factor, and our team strongly believes that a common infrastructure, such as H-ROS, will facilitate an environment where robot hardware and software can evolve.

The list presented below summarizes, according to our judgement, some of the most technically feasible future robotic technologies to appear.

Acknowledgments

This review was funded and supported by Acutronic Robotics, a firm focused on the development of next-generation robot solutions for a range of clients.

The authors would also like to thank the Erle Robotics and the Acutronic groups for their support and help.

References

  • [1] Gates, B. ”A robot in every home,” Scientific American, 296(1), 2007, pp. 58–65. (link)
  • [2] Trikha, B. “ A Journey from floppy disk to cloud storage,” in International Journal on Computer Science and Engineering,Vol. 2, 2010, pp.1449–1452. (link)
  • [3] Copeland, B. J. “Colossus: its origins and originators,” in IEEE Annals of the History of Computing, Vol. 26, 2004, pp. 38–45. (link)
  • [4] Bondyopadhyay, P. K. “In the beginning [junction transistor],” in Proceedings of the IEEE, Vol. 86, 1998, pp.63–77. (link)
  • [5] Bryson, A. E. “Optimal control-1950 to 1985,” in IEEE Control Systems, Vol. 16, 1996, pp.26–33. (link)
  • [6] Middleditch, A. E. “Survey of numerical controller technology,” in Production Automation Project, University of Rochester, 1973. (link)
  • [7] Acal, A. P., & Lobera, A. S. “Virtual reality simulation applied to a numerical control milling machine,” in International Journal on Interactive Design and Manufacturing, Vol.1, 2007, pp.143–154. (link)
  • [8] Mark, M. “U.S. Patent №2,901,927,” Washington DC: U.S. Patent and Trademark Office, 1959 (link)
  • [9] Mickle, P. “A peep into the automated future,” in The capital century 1900–1999. http://www. capitalcentury. com/1961. html, 1961. (link)
  • [10] Kilby, J. S. (1976). Invention of the integrated circuit. IEEE Transactions on electron devices, 23(7), 648–654. (link)
  • [11] Giralt, G., Chatila, R., & Vaisset, M. “An integrated navigation and motion control system for autonomous multisensory mobile robots,” in Autonomous robot vehicles, 1990, pp.420–443. (link)
  • [12] Bryan, L. A., & Bryan, E. A. “Programmable controllers,” 1988. (link)
  • [13] Wade, J. “Dynamics of organizational communities and technological bandwagons: An empirical investigation of community evolution in the microprocessor market,” in Strategic Management Journal, Vol.16, 1995, pp.111–133. (link)
  • [14] Wallén, J. “The history of the industrial robot,” in Linköping University Electronic Press, 2008. (link)
  • [15] Paul, R. P., “WAVE: A Model Based Language for Manipulator Control,” in The Industrial Robot, Vol. 4, 1977, pp.10–17. (link)
  • [16] Shepherd, S., & Buchstab, A. “Kuka robots on-site,” in Robotic Fabrication in Architecture, Art and Design 2014, 2014, pp. 373–380. (link)
  • [17] Cutkosky, M. R., & Wright, P. K. (1982). Position Sensing Wrists for Industrial Manipulators (No. CMU-RI-TR-82–9). CARNEGIE-MELLON UNIV PITTSBURGH PA ROBOTICS INST. (link)
  • [18] Finkel, R., Taylor, R., Bolles, Paul, R. and Feldman, J., “An Overview of AL, A Programming System for Automation,” in Proceedings -Fourth International Joint Conference on Artificial Intelligence, June 1975, pp.758–765. (link)
  • [19] Park, J., & Kim, G. J. “Robots with projectors: an alternative to anthropomorphic HRI,” in Proceedings of the 4th ACM/IEEE international conference on Human robot interaction, March 2009, pp. 221–222 (link)
  • [20] Srihari, K., & Deisenroth, M. P. (1988). Robot Programming Languages — A State of the Art Survey. In Robotics and Factories of the Future’87 (pp. 625–635). Springer Berlin Heidelberg. (link)
  • [21] Gruver, W. A., Soroka, B. J., Craig, J. J. and Turner, T. L., “Industrial Robot Programming Languages: A Comparative Evaluation,” in IEEE Transactions on Systems, Man, and Cybernetics, Vol. SMC-14, №4, July/August 1984, pp. 565–570. (link)
  • [22] Maeda, J. (2005). Current research and development and approach to future automated construction in Japan. In Construction Research Congress 2005: Broadening Perspectives (pp. 1–11). (link)
  • [23] Castells, M. “The Internet galaxy: Reflections on the Internet, business, and society,” in Oxford University Press on Demand, 2002. (link)
  • [24] Beckhoof “25 Years of PC Control,” 2011. (link)
  • [25] Shuang Yu “IEEE 802.3 ‘Standard for Ethernet’ Marks 30 Years of Innovation and Global Market Growth,” Press release IEEE, June 24, 2013. Retrieved January 11, 2014. (link)
  • [26] Brooks, R. “New approaches to robotics,”in Science, 253(5025), 1991, 1227–1232. (link)
  • [27] World Heritage Encyclopedia, “International Federation of Robotics” in World Heritage Encyclopedia (link)
  • [28] Lapham, J. “ RobotScript™: the introduction of a universal robot programming language,” Industrial Robot: An International Journal, 26(1),1999, pp. 17–25 (link)
  • [29]García Marín, J. A. “New concepts in automation and robotic technology for surface engineering,” 2010. (link)
  • [30] Walter A Aviles, Robin T Laird, and Margaret E Myers. “Towards a modular robotic architecture,” in 1988 Robotics Conferences. International Society for Optics and Photonics, 1989, pp. 271–278 (link)
  • [31] Angle, C. “Genghis, a six legged autonomous walking robot,” Doctoral dissertation, Massachusetts Institute of Technology, 1989. (link)
  • [32] Bovet, D. P., & Cesati, M. “Understanding the Linux Kernel: from I/O ports to process management,” in O’Reilly Media, Inc.” 2005. (link)
  • [33] Alpert, D., & Avnon, D. “Architecture of the Pentium microprocessor,” in IEEE micro, Vol. 13, 1993, pp.11–21. (link)
  • [34] Hollingum, J. “ABB focus on lean robotization,” in Industrial Robot: An International Journal, Vol. 21, 1994, pp.15–16. (link)
  • [35] Barabanov, M. “A Linux Based Real-Time Operating System,” 1996. (link)
  • [36] Yodaiken, V. “Cheap Operating systems Research,” in Published in the Proceedings of the First Conference on Freely Redistributable Systems, Cambridge MA, 1996 (link)
  • [37] Decotignie, J. D. “Ethernet-based real-time and industrial communications,” in Proceedings of the IEEE, Vol. 93, 2005, pp.1102–1117. (link)
  • [38] Wade, S., Dunnigan, M. W., & Williams, B. W. “Modeling and simulation of induction machine vector control with rotor resistance identification,” in IEEE transactions on power electronics, Vol. 12, 1997, pp.495–506. (link)
  • [39] Campbell, M., Hoane, A. J., & Hsu, F. H. “Deep blue,” in Artificial intelligence, Vol. 134, 2002, pp.57–83. (link)
  • [40] Folkner, W. M., Yoder, C. F., Yuan, D. N., Standish, E. M., & Preston, R. A. “Interior structure and seasonal mass redistribution of Mars from radio tracking of Mars Pathfinder,” in Science 278(5344), 1997, pp.1749–1752. (link)
  • [41] Cliburn, D. C. “Experiences with the LEGO Mindstorms throughout the undergraduate computer science curriculum,”in Frontiers in Education Conference, 36th Annual,IEEE, October 2006, pp.1–6. (link)
  • [42] Rowe S., R Wagner C. “An introduction to the joint architecture for unmanned systems (JAUS),” in Ann Arbor 1001, 2008. (link)
  • [43] Fujita, M. “On activating human communications with pet-type robot AIBO,” in Proceedings of the IEEE, Vol. 92, 2004, pp.1804–1813.(link)
  • [44] Breazeal, C. L. “Sociable machines: Expressive social exchange between humans and robots,” in Doctoral dissertation, Massachusetts Institute of Technology, 2000. (link)
  • [45] Rafiei, M., Elmi, S. M., & Zare, A. “Wireless communication protocols for smart metering applications in power distribution networks,” in Electrical Power Distribution Networks (EPDC), 2012 Proceedings of 17th Conference on. IEEE, May 2012, pp. 1–5. (link)
  • [46] Brian Gerkey, Richard T Vaughan, and Andrew Howard. “The Player/Stage project: Tools for multi-robot and distributed sensor systems” in Proceedings of the 11th international conference on advanced robotics. Vol. 1. 2003, pp. 317–323. (link)
  • [47] Herman Bruyninckx. “Open robot control software: the OROCOS project” in Robotics and Automation, 2001. Proceedings 2001 icra. ieee International Conference on. Vol. 3. IEEE. 2001, pp. 2523–2528. (link)
  • [48] Hirose, M., & Ogawa, K. “Honda humanoid robots development,” in Philosophical Transactions of the Royal Society of London A: Mathematical, Physical and Engineering Sciences, 365(1850), 2007, pp.11–19. (link)
  • [49] Mohr, F. W., Falk, V., Diegeler, A., Walther, T., Gummert, J. F., Bucerius, J., … & Autschbach, R. “Computer-enhanced “robotic” cardiac surgery: experience in 148 patients” in The Journal of thoracic and cardiovascular surgery, 121(5), 2001, pp.842–853. (link)
  • [50] Jones, J. L., Mack, N. E., Nugent, D. M., & Sandin, P. E. “U.S. Patent №6,883,201,” in Washington, DC: U.S. Patent and Trademark Office, 2005. (link)
  • [51] Jansen, D., & Buttner, H. “Real-time Ethernet: the EtherCAT solution,” in Computing and Control Engineering, 15(1), 2004, pp. 16–21. (link)
  • [52] Koenig, N., & Howard, A. “Design and use paradigms for gazebo, an open-source multi-robot simulator,” in Intelligent Robots and Systems, 2004.(IROS 2004). Proceedings. 2004 IEEE/RSJ International Conference on IEEE., Vol. 3, September 2004, pp. 2149–2154. (link)
  • [53] Cousins, S. “Willow garage retrospective [ros topics],” in IEEE Robotics & Automation Magazine, 21(1), 2014, pp.16–20. (link)
  • [54] Garage, W. Robot operating system. 2009. [Online]. (link)
  • [55] Fisher A. “Inside Google’s Quest To Popularize Self-Driving Cars,” in Popular Science, Bonnier Corporation, Retrieved 10 October 2013. (link)
  • [56] Cousins, S. “Ros on the pr2 [ros topics],” in IEEE Robotics & Automation Magazine, 17(3), 2010, pp.23–25. (link)
  • [57] PARROT, S. A. “Parrot AR. Drone,” 2010. (link)
  • [58] Honda Motor Co. ASIMO, 2011. [Online]. (link)
  • [59] Shen, F., Yu, H., Sakurai, K., & Hasegawa, O. “An incremental online semi-supervised active learning algorithm based on self-organizing incremental neural network,” in Neural Computing and Applications, 20(7), 2011, pp.1061–1074. (link)
  • [60] Industria 4.0 en la Feria de Hannover: La senda hacia la “fábrica inteligente” pasa por la Feria de Hannover, sitio digital ‘Deutschland’, 7 de abril de 2014 (link)
  • [61] Richardson, Matt, and Shawn Wallace “Getting started with raspberry PI,” in O’Reilly Media, Inc., 2012. (link)
  • [62] Edwards, S., & Lewis, C. “Ros-industrial: applying the robot operating system (ros) to industrial applications,” in IEEE Int. Conference on Robotics and Automation, ECHORD Workshop, May 2012. (link)
  • [63] Canis, B. Unmanned aircraft systems (UAS): Commercial outlook for a new industry. Congressional Research Service, Washington, 2015, p.8. (link)
  • [64] Trishan de Lanerolle, The Dronecode Foundation aims to keep UAVs open, Jul 2015. [Online] (link)
  • [65] Savioke, Your Robot Butler Has Arrived, August 2014. [Online]. (link)
  • [66] ABB, “ABB introduces Yumi, world´s first truly collaborative dual-arm robot”, 2015. Press release (link)
  • [67] LEE, Chang-Shing, et al. Human vs. Computer Go: Review and Prospect [Discussion Forum]. IEEE Computational Intelligence Magazine, 2016, vol. 11, no 3, pp. 67–72. (link)
  • [68] Bogue, R. (2015). Sensors for robotic perception. Part one: human interaction and intentions. Industrial Robot: An International Journal, 42(5), pp.386–391 (link)
  • [69] The Linux foundation, official wiki. 2009. [Online]. (link)
  • [70] The Tesla Team, “ All Tesla Cars Being Produced Now Have Full Self-Driving Hardware” Official web, 19 Oct. 2016. [Online]. (link)
  • [71] Brian Gerkey. Why ROS 2.0?, 2014. [Online]. (link)
  • [72] Acutronic Robotics, “H-ROS: Hardware Robot Operating System”, 2016. [Online]. (link)
  • [73] Judith Viladomat, TALOS:the next step in humanoid robots from PAL Robotics. 4 Oct. 2016 [Online]. (link)
]]> 5 global problems that AI could help us solve https://robohub.org/5-global-problems-that-ai-could-help-us-solve/ Fri, 10 Feb 2017 10:30:43 +0000 http://robohub.org/5-global-problems-that-ai-could-help-us-solve/

There’s a great deal of concern over artificial intelligence; what it means for our jobs, whether robots will one day replace us in the workplace, whether it will one day lead to robot wars. But current research projects show that artificial intelligence (AI) can also be used for the greater good. Here are five global problems that machine learning could help us solve.

1. Healthcare

One of the biggest benefits of AI is its ability to trawl through massive amounts of data in record time. This helps researchers pinpoint areas of focus for their own research.

For example, a recent ground-breaking discovery on the disease Amyotrophic Lateral Sclerosis (ALS), was made through a partnership between Barrow Neurological Institute and the artificial intelligence company IBM Watson Health.

IBM Watson, the artificial intelligence computer, reviewed thousands of pieces of research and was able to identify new genes linked to ALS.

“Traditional research tools are fast becoming inadequate to help data scientists and researchers keep pace with any global problems that AI could help us solve and find relevant insights among the now billions of documents which are spread all over the world,” said the company in a press release.

“The discovery gives ALS researchers new insights that will pave the way for the development of new drug targets and therapies to combat one of the world’s most devastating and deadly diseases.”

https://youtu.be/F-qBLH6EfR8

Another promising use for AI within healthcare is its ability to predict the outcome of drug treatments. For instance, cancer patients are often given the same drug, and then monitored to see the effectiveness of that drug. AI could use data to predict which patients could benefit from using a particular drug, providing a highly personalized approach, and saving valuable time and money.

Image: Bloomberg

2. Making driving safer

Despite the crashes involving self-driving cars that have hit the headlines this year, this area of AI could dramatically reduce deaths and injuries on our roads.

According to a report by Stanford University, not only will self-driving cars reduce traffic related deaths and injuries, but they could bring about changes in our lifestyles as well. We will have more time to work or entertain ourselves during commutes, and we may have more choice over where we base ourselves:

“The increased comfort and decreased cognitive load with self-driving cars and shared transportation may affect where people choose to live,” the report says.

3. Transforming how we learn

Earlier this year, students at Georgia Tech university in the US were startled to discover that their helpful teaching assistant had in fact been a robot all along. After initial teething problems, the robot started answering the students’ questions with 97% certainty.

The university designed the robot after their research showed that one of the main factors behind students dropping out is a lack of support.

People learn differently, at different speeds and with different starting points. Artificial intelligence could usher in a future where we all learn in a much more personalised way. But no education system in the world can afford a tutor for every child, so this is where AI might be able to step in. Artificial tutors, made to look and sound as much like humans as possible, could take the lead in delivering personalised education.

4. Help us be smarter about energy

Artificial intelligence could help us be smarter about our energy consumption. In fact, this is already happening.

Image: Google DeepMind

Google and other tech giants have enormous data centres that require a massive amount of energy to run the servers and keep them cool. Google has used its artificial intelligence platform Deep Mind to predict when its data centres will get too hot. Cooling systems are only activated when required. AI has saved Google around 40% in energy costs at its server farms.

5. Helping wildlife

As in the case of healthcare, being able to analyse massive amounts of data can transform wildlife conservation.

For instance, by tracking animal movements, we can see where they go, and what habitats we need to protect. This study uses computing power to figure out the best places to create wildlife corridors for wolverines and grizzly bears in Montana. Wildlife corridors are continuous areas of protected land that link zones of biological significance that the animals can use to move safely through the wilderness.

Challenges

Using artificial intelligence is not without its challenges, however. One of the biggest of these is – how do we keep the systems safe? Algorithms are based on data, so any change to that data will change the behaviour and outcomes.

“Almost anything bad you can think of doing to a machine-learning model can be done right now,” said one expert at a recent AI conference in Spain. “And defending it is really, really hard.”

This post was originally published on WEForum.org.

You might also enjoy the following articles:

See all the latest robotics news on Robohub, or sign up for our weekly newsletter.

]]> How can swarm roboticists contribute to and benefit from the wisdom of other disciplines? https://robohub.org/how-can-swarm-roboticists-contribute-to-and-benefit-from-the-wisdom-of-other-disciplines/ Tue, 08 Nov 2016 10:00:01 +0000 http://robohub.org/how-can-swarm-roboticists-contribute-to-and-benefit-from-the-wisdom-of-other-disciplines/ swarming-botsWhat can swarm roboticists learn from policy makers, systems biologists and physicists, and vice versa? It is already widely recognised that Robotics is an inherently interdisciplinary field and that designing even a single robot might require input from multiple domains. In Swarm Robotics, interactions between robots add further layers of complexity. Indeed, the ‘complex’ nature of robot swarm systems demands approaches going beyond reductionist scientific models or traditional engineering design methods. Like many other emerging technologies, such as synthetic biology and socio-technical systems engineering, robot swarms can be notoriously difficult to predict due to their non-linearity, interconnectivity, hidden heterogeneity and ’emergence‘. Yet it is also this ‘complexity’ that swarm roboticists seek to exploit in order to give intelligent, robust, adaptive behaviours.

But what does it mean to describe systems as complex? How do these complex systems differ from the more easily understood ‘modular’ systems that we are familiar with? Vocabulary in this area is often dangerously inconsistent. For example, the terms ’emergence’, ‘complex’, and ‘complicated’ are used differently by different disciplines, and often differently even within the same discipline. This makes it very difficult to understand whether people are really talking about the same thing, and whether the systems being described are different in superficial or profound ways. On the one hand, failing to identify the underlying similarities between different systems (whether modular or complex) results in missed opportunities for sharing knowledge, best practices and methods. On the other hand, failing to identify the underlying differences between different systems results in practices and methods being misapplied. More broadly, many of today’s real-world problems require engineers, designers and policy-makers across all domains to think in terms of complex systems.
triple-pageTo address problems with translating between disciplines, Chih-Chun Chen and Nathan Crilly at the University of Cambridge have produced ‘A primer on the design and science of complex systems’. This introduces complex systems constructs by building them up from basic concepts, and contrasting them with more familiar constructs that are associated with modularity. For example, ’emergence’ can be understood with respect to a breakdown in how a system’s functions are mapped to the structures that perform those functions. Abstract diagrams that are independent of any particular domain are used to represent the constructs that are discussed. These are illustrated with worked examples that make the explanations accessible for those who have no experience with ‘complexity’. The primer is intended to provide both an introduction to complex systems constructs for those new to the topics discussed, and also a basis for cross-domain translations for researchers and practitioners wishing to engage with other fields when addressing the systems problems they are working on.

Being able to communicate unambiguously across disciplines and application areas would greatly expand the space of solutions available to all domains in solving problems currently deemed to be too ‘complex’. As a mature complex systems engineering discipline which is by its nature interdisciplinary, Swarm Robotics will no doubt have much to contribute to – and take from – this endeavour.

If you liked this article, you may also want to read these other articles on swarm robotics:

See all the latest robotics news on Robohub, or sign up for our weekly newsletter.

 

]]> Are agricultural robots ready? 27 companies profiled https://robohub.org/are-agricultural-robots-ready-27-companies-profiled/ Wed, 12 Nov 2014 18:47:00 +0000 http://robohub.org/are-agricultural-robots-ready-27-companies-profiled/

Agriculture is one of our most important industries. It provides food, feed and fuel necessary for our survival. With the global population expected to reach 9 billion by 2050, agricultural production must double to meet the demand. And because of limited arable land, productivity must increase 25% to help meet that goal.

Consider these factoids:

  • Major US farming conglomerates are buying foreign land and beginning to farm there citing lower overall cost.

  • China is buying land in Africa and sending skilled workers to supervise those new farms.

  • Farmers and ranchers the world over are transitioning to precision agricultural methods, i.e., subdividing their acreage into many sub-plots, in some cases, right down to the individual plant/tree/animal thereby enabling increased productivity and lower overall costs.

  • Unmanned aerial vehicles are being used to map, observe, sense and spray.

  • Unmanned (or at least autonomous) ground vehicles are providing more precise movements and thereby enabling precision practices.

  • The US Bureau of Labor Statistics reports that 2012 median pay for farm workers was $9.09.

  • The US Bureau of Labor Statistics reports that there were 749,400 ag workers in 2012, down 3% (25,000) from 2011.

  • 74% - approximate number of crop workers in the US who were born in Mexico or Central America of which more than half are likely to be undocumented (according to Fortune Magazine).

  • Cropdusters have the 3rd highest fatality rate among professions in the US. 90% of crop spraying in Japan is done using small unmanned helicopters.

Thus the agricultural industry is in transition. And that transition differs country by country, state by state, region by region as well as by type of farming practiced: from primitive to conventional to precision to experimental. A little bit of everything is going on everywhere but the general trend worldwide is toward precision agriculture supplemented by advanced technologies including robotics.

Many factors are precipitating these changes in addition to global population growth and the cost and availability of labor: the diminishing availability and increasing cost of water, political and regulatory procedures and hold-ups; limited tillable acreage; better, cheaper and faster technological automation products; and climate change, to name just a few.

Modern farmers and ranchers are already high-tech. Digitally-controlled farm implements are regularly in use. There are partially and fully automatic devices for most aspects of agricultural functions from grafting to planting, from harvesting to sorting, packaging and boxing. Farmers use software systems and aerial survey maps and data to guide their field operations. They also use auto-steer systems included in many new tractors (or buy kits that do the same thing) which follow GPS and software guidance. Some farmers are already transitioning some of their operations to full autonomy. Thus forward-thinking farm owners today may be able to skip over slow, incremental improvements and jump directly to robotic and autonomous automation. But are the robots ready?

In a follow-up to my July, 2014 article “Will agricultural robots arrive in time to keep fruit and vegetable costs down?” this article profiles 27 of the many companies (from conglomerates to start-ups) attempting to provide robotic solutions for farming problems and explores what they are doing, when their products will be available, and at what cost.

[NOTE: From a list of 60 organizations involved with agricultural robotics, dairy was eliminated even though robotic milking systems are amazing and a growing business. Also eliminated were companies that didn't (or couldn't because they are publicly traded such as John Deere and CNH Industrial (Case/New Holland)) respond to my brief questionaire. Further, only one of the many university ag research facilities was profiled even though there's a lot of rich science under development; I wanted to focus on the here and now; not the future.]

Companies Mentioned sorted by primary function:

Harvesting and Tractors - Tractors do two things: provide guidance to the devices they are towing, and pulling power. Current tractors are huge and if they break down, the entire operation comes to a halt. Autonomous machines don't need operators and can operate around the clock. Thus tight operational windows can be achieved for seeding and other time-sensitive activities.

Planting, Pruning, Potting, Grafting and Nursery Operations

Thinning and Weeding

UAS, Inspection, Data Collection and Data Manipulation - UAVs are only as good as the other precision ag equipment and systems; if there are no computers on the tractors or controllers on the implements towed, and if they can't talk to each other, UAS data collected is just pretty pictures.

Smart Implements - New implements incorporate advanced control systems and can respond to commands from the towing tractor or provide their own mobility and navigation.

 

 

---------------------------------------------------------------------------

 

 

Company Profiles:

 

Company: ISO Group, a Flier Systems company, Gameren, The Netherlands
Website: www.isogroepmachinebouw.nl
Product: RoBoPlant flower planting robot and fully and semi-autonomous grafting robots


ISO Group's flower planting robot RoBoPlant

Area of use: All over the EU
Function: Semi and full automatic machinery for greenhouse or protected horticulture. Flower planting robotic system takes flats of peat seedlings, separates them and plants them in chosen patterns; 
Testing: Continual testing and development
Availability: Began selling products in 2002
Price: N/A

Return to Companies Mentioned Index

 

 

Company: Autonomous Tractor, Fargo, ND
Website: http://www.autonomoustractor.com/
Product: Modular cab-less robotic tractor implement platform


Autonomous Implement - Spirit Mower

Area of use: Hay producers in the US
Function: Mowing hay. Can add modular engine power as needed by the type and size of implement. Will expand to other crops and other non-ag industries after mower begins shipping
Testing: Continuously field testing 
Availability: Expect to come to market and begin shipping pre-orders in 2016
Price: Price expected to be 1/2 of similarly-powered tractors and implements

Return to Companies Mentioned Index

 

 

Company: Blue River Technologies, Sunnyvale, CA
Website: http://www.bluerivert.com/
Product: LettuceBot2 (2nd generation) lettuce thinning


Blue River Technology 3-row LettuceBot2

Area of use: CA and AZ lettuce fields (covers 80% of US lettuce production)
Function: Thinning and weed spraying of iceberg, romaine and leaf lettuce
Testing: Recently completed field testing 3rd generation machine; selective weeding used to improve germination
Availability: Began operating as a per-acre service in 2013
Price: Price per acre depends on the lettuce planting configuration but equates to a slight premium over manual labor costs

Return to Companies Mentioned Index

 

 

Company: Agrobot, Huelva, Spain
Website: http://www.agrobot.com/
Product: Agrobot SW6010 and AGSHydro, a bed-on hydroponic growing system customized for strawberry growing and harvesting


Agrobot harvester


Agrobot hydroponic growing system

Area of use: Strawberry harvesting in Oxnard, CA
Function: Ripe berry picking from raised hydroponic growing beds
Testing: Will start final testing strawberry harvesting in January; have done seasonal testing for a few years
Availability: Mid-2015
Price: $250,000 for a harvester with 60 robotic picking arms. Says one large berry farmer: “The Agrobot works on several investment paths.  One where we harvest cheaper than we do today and another where we harvest fruit that there are not sufficient people for.  In the latter case the Agrobot pays off instantly because without the ability to harvest we do not have a business (this is becoming more common).”

Return to Companies Mentioned Index
 

 

Company: Agribotix, Boulder, Colorado
Website: http://agribotix.com/
Product: Drone services for precision agriculture


Agribotix Hornet Drone


Agribotix image processing services

Area of use: US Midwest (CO, KA, MO, etc.)
Function: Lease ag drones to co-ops, agronomists, crop consults, farm managers and big industrial farm corporations; produce and process hi-res images and maps using various sensors, and provide prescription maps to match the application of fertilizer to the places that need more (or less)
Testing: Ongoing testing with pilot customers
Availability: Began selling services in 2014
Price: About $8,000 for a season which includes training, drone use, stitched-together RGB and infrared images, crop health and prescription maps. Agribotix offers an image processing services contract with per acre charges for various maps and images over an annual contract period

Return to Companies Mentioned Index

 

 

Company: Wall-Ye, Macon, France
Website: http://wall-ye.com
Product: Wall-Ye 1000 mobile pruning robot


Wall-Ye 1000 Pruning Robot

Area of use: French grape vineyards pruning
Function: Autonomous pruning
Testing: Completed in 2013
Availability: For sale and as a service
Price: $30,000 per robot

Return to Companies Mentioned Index
 

 

Company: ecoRobotix, Essert-Pittet, Switzerland
Website: http://www.ecorobotix.com/
Product: Lightweight autonomous robots initially for weeding


ecoRobotix concept field robot

Area of use: Field testing in Switzerland; next year in Germany
Function: A robotic platform for weeding of spaced-row cultures, which includes advanced weed recognition algorithms, fast robotic arms, advanced sensor technology, high energy efficiency, and wireless communications
Testing: Currently with sugarbeet but plan to extend to colza, sunflower, corn and soya
Availability: First machines available for sale by end of 2015
Price: About 15’000 EUR ($18,750) per robot

Return to Companies Mentioned Index
 

 

Company: Energid, Cambridge, MA
Website: http://www.energid.com/experience/citrus-harvesting/
Product: Citrus harvester


Energid towed multi-arm citrus harvester

Area of use: Florida citrus orchards; oranges (early and late season) and grapefruit
Function: Harvesting, initially for juice
Testing: Will test again during seasons in Florida in 2015 and 2016
Availability: Expect to have initial product in late 2016
Price: System to cost $300,000-$400,000

Return to Companies Mentioned Index
 

 

Company: Harvest Automation, North Billerica, MA
Website: http://www.harvestai.com/products
Product: HV-100 mobile robot


Harvest Automation HV-100 Mobile Robot


Harvest Automation potted plant movement schema

Area of use: Nurseries (ornamental, berries, tomatoes, etc.)
Function: Material handling, movement of containers, spacing.
Testing: HV-100 testing completed
Availability: Been selling since 2013
Price: $130,000 for a team of four robots to purchase.  We also rent teams of four for $30K/3 months.  The rental scheme has worked really well.  All who have rented then subsequently purchased.

Return to Companies Mentioned Index
 

 

Company: Clearpath Robotics, Kitchener, ON, Canada
Website: http://www.clearpathrobotics.com/grizzly/
Product: Grizzly RUV (cab-less robotic utility vehicle) and Husky UGV


Clearpath Grizzly RUV pulling implement

Area of use: Sold to university research facilities for ag applications development
Function: Harvesting, mowing, hauling, research
Testing: Testing asparagus farming with added laser scanner to identify appropriate stalks and a cutter inserted into the soil to cut the stalk below ground; hauling manure from chicken farms while cleaning out barns; detecting where cows urinate and then treating the area so grass can continue to grow; mowing inbetween orchard rows and hauling (hay/straw wagons back to barn and return so that the farmer doesn’t have to stop baling; hauling chemical refills to sprayer locations; hauling manure spreader)
Availability: Early 2015 – at present only selling to academia and research organizations
Price: $12,000 to $100,000 depending on configuration

Return to Companies Mentioned Index
 

 

Company: Autonomous Solutions, Petersboro, UT
Website: http://www.asirobots.com/farming/
Product: Forge Robotic Platform, a kit for enabling a skid steer to operate autonomously or remotely controlled


ASI skid steer with cab in vineyard


ASI skid steer cab options


ASI Universal Vehicle Automation Kit

Area of use: Wine vineyards
Function: Mowing and spraying functions
Testing: Running field trials in CA and TX
Availability: Mid-2015
Price: $75,000 - $150,000/unit (includes complete skid steer device and driving kit) depending on skid steer configuration

Return to Companies Mentioned Index

 

 

Company: Wageningen UR (University and Research center), Wageningen, The Netherlands and Agritronics, Sint Annaparochie, The Netherlands
Website: http://www.wageningenur.nl/en/Expertise-Services/Research-Institutes/Wageningen-UR-Greenhouse-Horticulture/Research-themes/Advanced-Cultivation-and-Production-Systems/Subthemes/Computer-vision-and-robotics.htm and http://www.agritronics.nl/
Product: Research to supply intelligent systems for high value crops to commercial research partners/vendors


Wageningen UR cucumber harvesting robot

Area of use: Sweet peppers in The Netherlands, apples and grapes in Belgium, canopy spraying in Slovenia and spot spraying in Italy
Function: Harvesting and spraying (spot and canopy)
Testing: Yes, for sweet pepper, in July in a commercial greenhouse; for apples and grapes tests are now completed. For spraying, field tests have been completed. A new harvester, visual quality inspection and vision system for broccoli, is being developed with start-up Agritronics, Sint Annaparochie, The Netherlands
Availability: “This will take several years”
Price: Not available

Return to Companies Mentioned Index
 

 

Company: Vision Robotics, San Diego, CA
Website: http://www.visionrobotics.com/
Product: Lettuce Thinner and Grape Vineyard Pruner


Vision Robotics 6-Line Lettuce Thinner


Vision Robotics grapevine pruner

Area of use: California
Function:  Grapevine pruning being tested; lettuce thinner available for sale
Testing: Testing and development for grapevine pruner could be completed in less than 18 months depending on funding
Availability: Lettuce thinner available now; pruner early 2016
Price: Lettuce thinner starts at $140,000 and upwards depending on configuration; pruner will likely sell for same amount

Return to Companies Mentioned Index
 

 

Company: Precision Hawk, Raleigh, NC
Website: http://precisionhawk.com/
Product: Lancaster UAV with various plug and play sensor options plus Datamapper


PrecisionHawk Lancaster


PrecisionHawk Lancaster plug-in sensors

Area of use: Ontario, Canada
Function: Providing data for crop researchers, consultants and farmers and ranchers to make farm management decisions
Testing: Performing field tests under an SFOC from Transport Canada for a number of years. The majority of research and development happens in Ontario, Canada. Over the past six months have obtained a number of CoAs from the FAA to perform field tests and research across the United States in conjunction with universities such as NC State, Texas A&M, Kansas State and Cornell. 
Availability: 70% of sales are global. Have entered into a number of projects with US companies on foreign soil for specific research projects
Price: Basic Lancaster platform is $15,000 plus sensors and other options

Return to Companies Mentioned Index

 

 

Company: F Poulsen Engineering ApS, Hvalso, Denmark
Website: http://www.visionweeding.com
Product: ROBOVATOR thermal and/or hydraulic weeder


Poulsen weeder

Area of use: 30 machines already at work in UK, the EU and Canada.
Function: Weeding and thinning of lettuce, cabbage, fennel and onions
Testing: In California (preceding expansion into North America)
Availability: Started selling in 2011 after 8 years of development
Price: The 5-row version sells in Europe for 80.000€ ($100,000)

Return to Companies Mentioned Index
 

 

Company: Kinze Manufacturing, Williamsburg, Iowa and Jaybridge Robotics, Cambridge, MA
Website: http://www.kinze.com/ and http://www.jaybridge.com/
Product: Autonomous vehicle system for row crop harvesting


Kinze autonomous tractor and grain cart

Area of use: Iowa and Illinois corn and soybeans
Function: Autonomously garner row crop grains from combine machines and bring it out of the field to the transport area
Testing: Testing autonomous harvesting system since 2012; in 2013 three farmers in Iowa and Illinois leased systems without Kinze overseeing operation allowing the farmers to use the technology independently. The Kinze system marries off-the-shelf components, including GPS, radar, laser sensors and video cameras, with custom software that allows the system to react to field obstructions. It was developed in partnership with Jaybridge Robotics.
Availability: Kinze is not currently selling the harvesting system but is working towards full commercialization soon
Price: The price has not yet been set for the system which includes the autonomous driving kit for the tractor and grain cart plus the navigation, path planning, harvester communication and control software systems

Return to Companies Mentioned Index
 

 

Company: Agrobotics, Little Rock, AR
Website: http://www.agrobotics.com/
Product: AutoProbe soil sampling system


Agrobotics AutoProbe

Area of use:  
Function:  
Testing:  
Availability:  
Price:  

Return to Companies Mentioned Index
 

 

 

Company: Amazone-Werke Gmbh, Hasbergen, Germany
Website: http://info.amazone.de/DisplayInfo.aspx?id=14033
Product: BoniRob field robot 


Amazone-Bosch BoniRob lightweight field robot

Area of use: Work on corn and wheat experimental plots in Germany
Function: Autonomous omnidirectional field robots working in "flocks" for multiple purposes
Testing: Multiple-purpose lightweight robot for weeding, applying fertilizer, inspection being developed with Robert Bosch GmbH
Availability: Only two built; no plans announced for commercialization at this time
Price: No information available

Return to Companies Mentioned Index
 

 

Company: Helper Robotech, Gimhae City, Korea
Website: http://helpersys.co.kr/
Product: BoniRob field robot 


Helper Robotech fruit and vegetable grafting robot

Area of use:  
Function:  
Testing:  
Availability:  
Price:  

Return to Companies Mentioned Index
 

 

Company: AGCO Fendt, Deluth, GA
Website: http://www.agcocorp.com/GuideConnect.aspx
Product: GuideConnect, SectionControl and VarioGuide


AGCO Fendt GuideConnect - driverless 2nd system


AGCO Fendt VarioGuide auto steering system

Area of use: Global
Function: SectionControl integrates various data and enables fully automatic section control via GNSS for ISOBUS-capable sprayers, spreaders and seeders; the VarioGuide night and day auto steering system; and the new GuideConnect in which two tractors act as a unit where one vehicle is unmanned
Testing: GuideConnect is still under development with no known date or area for availability; the following vehicle doesn't have its own obstacle detection which may be why they haven't yet released the product
Availability: All but GuideConnect are available now in the EU and US
Price: Not available for all 3 systems

Return to Companies Mentioned Index
 

 

Company: Rowbot, Minneapolis, MN
Website: http://rowbot.com
Product: Rowbot is a self-driving, multi-use platform that travels between rows of corn, ex: applying nitrogen fertilizer in sync with corn needs. It can also collect sensor data to inform both current and future work. GPS and several sensors keep the robot from trampling the crop


Rowbot in cornfield. Rowbots work in teams to apply nitrogen fertilizer in sync with precision needs

Area of use: US Corn Belt
Function: Rowbot travels between corn rows - often under the leaf canopy - to apply nitrogen fertilizer and also to seed cover crops
Testing: Working in conjunction with Carnegie Robotics on development of the Rowbot.
Availability: Began test marketing this year for in-season nitrogen and cover crop seeding services; plan to widen scope of services in 2015
Price: No information available about the cost of the service

Return to Companies Mentioned Index
 

 

Company: senseFly, Cheseaux-Lausanne, Switzerland
Website: http://www.sensefly.com
Product: eBee Ag


senseFly eBee Ag


senseFly eBee Ag autopilot system and carrying case

Area of use: Global
Function: The eBee ag system includes eMotion software and a carrying case. The software and cameras enable 2 cm per pixel resolution and produce 3D maps and overlays as well as the capability to lay out (and simulate) a flight path for up to 45 minutes of flying time
Testing: 
Availability: The eBee ag system is available now
Price: About $12,000 for the complete system 

Return to Companies Mentioned Index
 

 

Company: Conic Systems, Barcelona, Spain
Website: http://www.conic-system.com/
Product: EMP-300 Grafting Robot


Conic Systems EMP-300 Grafting Robot

Area of use: Global
Function: Enables grafting of vegetables  and other greenhouse plants 
Testing: Unknown
Availability: Now
Price: Not available

Return to Companies Mentioned Index
 

 

Company: Naio Technologies, Toulouse, France
Website: http://naio-technologies.com/
Product: Naio Technologies Oz field robot


Naio Technologies Oz field robot

Area of use: Mostly in France
Testing: Testing next generation of Oz robot (with improved navigation capabilities) in real field conditions in France
Function: The Oz robot serves as an autonomous electric tractor which can be used for weeding and as a transport from harvesters to accumulation points. Oz operates as a self-powered robotic implement rather than a towed implement
Availability: Began selling in 2013
Price: Initially robots are being rented to help customers get familiarized with the product line and to help optimize the utilization. Units are renting/leasing for $315 to $475 per month depending on configuration

Return to Companies Mentioned Index
 

 

Company: Robotic Harvesting, Simi Valley, CA
Website: http://www.roboticharvesting.com
Product: Strawberry harvester, data collector and mobile platform


Robotic Harvesting Strawberry Harvester

Area of use: California
Testing: Ongoing in California
Function: Autonomous mobile device which takes stereovision photos to locate any fruit or vegetable in 3D space and then uses a robot arm to pick and place on a conveyor selected berries
Availability: Unknown
Price: Unknown

Return to Companies Mentioned Index
 

Read more

]]>

Agriculture is one of our most important industries. It provides food, feed and fuel necessary for our survival. With the global population expected to reach 9 billion by 2050, agricultural production must double to meet the demand. And because of limited arable land, productivity must increase 25% to help meet that goal.

Consider these factoids:

  • Major US farming conglomerates are buying foreign land and beginning to farm there, citing lower overall cost.
  • China is buying land in Africa and sending skilled workers to supervise those new farms.
  • Farmers and ranchers the world over are transitioning to precision agricultural methods, i.e., subdividing their acreage into many sub-plots, in some cases, right down to the individual plant/tree/animal thereby enabling increased productivity and lower overall costs.
  • Unmanned aerial vehicles are being used to map, observe, sense and spray.
  • Unmanned (or at least autonomous) ground vehicles are providing more precise movements and thereby enabling precision practices.
  • The US Bureau of Labor Statistics reports that 2012 median pay for farm workers was $9.09.
  • The US Bureau of Labor Statistics reports that there were 749,400 ag workers in 2012, down 3% (25,000) from 2011.
  • 74% – approximate number of crop workers in the US who were born in Mexico or Central America, of which more than half are likely to be undocumented (according to Fortune Magazine).
  • Cropdusters have the 3rd highest fatality rate among professions in the US. 90% of crop spraying in Japan is done using small unmanned helicopters.

Thus the agricultural industry is in transition. And that transition differs country by country, state by state, region by region as well as by type of farming practiced: from primitive to conventional to precision to experimental. A little bit of everything is going on everywhere but the general trend worldwide is toward precision agriculture supplemented by advanced technologies, including robotics.

Many factors are precipitating these changes in addition to global population growth and the cost and availability of labor: the diminishing availability and increasing cost of water, political and regulatory procedures and hold-ups; limited tillable acreage; better, cheaper and faster technological automation products; and climate change, to name just a few.

Modern farmers and ranchers are already high-tech. Digitally-controlled farm implements are regularly in use. There are partially and fully automatic devices for most aspects of agricultural functions from grafting to planting, from harvesting to sorting, packaging and boxing. Farmers use software systems and aerial survey maps and data to guide their field operations. They also use auto-steer systems included in many new tractors (or buy kits that do the same thing) that follow GPS and software guidance. Some farmers are already transitioning some of their operations to full autonomy. Thus forward-thinking farm owners today may be able to skip over slow, incremental improvements and jump directly to robotic and autonomous automation. But are the robots ready?

In a follow-up to my July, 2014 post “Will agricultural robots arrive in time to keep fruit and vegetable costs down?”, this article profiles 27 of the many companies (from conglomerates to start-ups) attempting to provide robotic solutions for farming problems and explores what they are doing, when their products will be available, and at what cost.

[NOTE: From a list of 60 organizations involved with agricultural robotics, dairy was eliminated even though robotic milking systems are amazing and a growing business. Also eliminated were companies that didn’t (or couldn’t because they are publicly traded such as John Deere and CNH Industrial (Case/New Holland)) respond to my brief questionnaire. Further, only one of the many university ag research facilities was profiled even though there’s a lot of rich science under development; I wanted to focus on the here and now; not the future.]

Companies Mentioned (sorted by primary function):

Harvesting and Tractors – Tractors provide two things: guidance to the devices they are towing, and pulling power. Current tractors are huge, and if they break down, the entire operation comes to a halt. Autonomous machines don’t need operators and can operate around the clock. Thus tight operational windows can be achieved for seeding and other time-sensitive activities.

Planting, Pruning, Potting, Grafting and Nursery Operations

Thinning and Weeding

UAS, Inspection, Data Collection and Data Manipulation – UAVs are only as good as the other precision ag equipment and systems; if there are no computers on the tractors or controllers on the implements towed, and if they can’t talk to each other, UAS data they collect are just pretty pictures.

Smart Implements – New implements incorporate advanced control systems and can respond to commands from the towing tractor or provide their own mobility and navigation.

Agriculture is big business in every country around the world. Thus this is a timely review of the progress to bring robotic automation to an already automated industry. Bottom line: lots of activity, much of which will be coming online in the next year or two, but little market penetration thus far.

 

—————————————————————————

 

 

Company Profiles:

 

Company: ISO Group, a Flier Systems company, Gameren, The Netherlands
Website: www.isogroepmachinebouw.nl
Product: RoBoPlant flower planting robot and fully and semi-autonomous grafting robots


ISO Group’s flower planting robot RoBoPlant

Area of use: All over the EU
Function: Semi and full automatic machinery for greenhouse or protected horticulture. Flower planting robotic system takes flats of peat seedlings, separates them and plants them in chosen patterns;
Testing: Continual testing and development
Availability: Began selling products in 2002
Price: N/A

Return to Companies Mentioned Index

 

 

Company: Autonomous Tractor, Fargo, ND
Website: http://www.autonomoustractor.com/
Product: Modular cab-less robotic tractor implement platform


Autonomous Implement – Spirit Mower

Area of use: Hay producers in the US
Function: Mowing hay. Can add modular engine power as needed by the type and size of implement. Will expand to other crops and other non-ag industries after mower begins shipping
Testing: Continuously field testing
Availability: Expect to come to market and begin shipping pre-orders in 2016
Price: Price expected to be 1/2 of similarly-powered tractors and implements

Return to Companies Mentioned Index

 

 

Company: Blue River Technologies, Sunnyvale, CA
Website: http://www.bluerivert.com/
Product: LettuceBot2 (2nd generation) lettuce thinning


Blue River Technology 3-row LettuceBot2

Area of use: CA and AZ lettuce fields (covers 80% of US lettuce production)
Function: Thinning and weed spraying of iceberg, romaine and leaf lettuce
Testing: Recently completed field testing 3rd generation machine; selective weeding used to improve germination
Availability: Began operating as a per-acre service in 2013
Price: Price per acre depends on the lettuce planting configuration but equates to a slight premium over manual labor costs

Return to Companies Mentioned Index

 

 

Company: Agrobot, Huelva, Spain
Website: http://www.agrobot.com/
Product: Agrobot SW6010 and AGSHydro, a bed-on hydroponic growing system customized for strawberry growing and harvesting


Agrobot harvester


Agrobot hydroponic growing system

Area of use: Strawberry harvesting in Oxnard, CA
Function: Ripe berry picking from raised hydroponic growing beds
Testing: Will start final testing strawberry harvesting in January; have done seasonal testing for a few years
Availability: Mid-2015
Price: $250,000 for a harvester with 60 robotic picking arms. Says one large berry farmer: “The Agrobot works on several investment paths.  One where we harvest cheaper than we do today and another where we harvest fruit that there are not sufficient people for.  In the latter case the Agrobot pays off instantly because without the ability to harvest we do not have a business (this is becoming more common).”

Return to Companies Mentioned Index
Company: Agribotix, Boulder, Colorado
Website: http://agribotix.com/
Product: Drone services for precision agriculture


Agribotix Hornet Drone


Agribotix image processing services

Area of use: US Midwest (CO, KA, MO, etc.)
Function: Lease ag drones to co-ops, agronomists, crop consults, farm managers and big industrial farm corporations; produce and process hi-res images and maps using various sensors, and provide prescription maps to match the application of fertilizer to the places that need more (or less)
Testing: Ongoing testing with pilot customers
Availability: Began selling services in 2014
Price: About $8,000 for a season which includes training, drone use, stitched-together RGB and infrared images, crop health and prescription maps. Agribotix offers an image processing services contract with per acre charges for various maps and images over an annual contract period

Return to Companies Mentioned Index

 

 

Company: Wall-Ye, Macon, France
Website: http://wall-ye.com
Product: Wall-Ye 1000 mobile pruning robot


Wall-Ye 1000 Pruning Robot

Area of use: French grape vineyards pruning
Function: Autonomous pruning
Testing: Completed in 2013
Availability: For sale and as a service
Price: $30,000 per robot

Return to Companies Mentioned Index
Company: ecoRobotix, Essert-Pittet, Switzerland
Website: http://www.ecorobotix.com/
Product: Lightweight autonomous robots initially for weeding


ecoRobotix concept field robot

Area of use: Field testing in Switzerland; next year in Germany
Function: A robotic platform for weeding of spaced-row cultures, which includes advanced weed recognition algorithms, fast robotic arms, advanced sensor technology, high energy efficiency, and wireless communications
Testing: Currently with sugarbeet but plan to extend to colza, sunflower, corn and soya
Availability: First machines available for sale by end of 2015
Price: About 15’000 EUR ($18,750) per robot

Return to Companies Mentioned Index
Company: Energid, Cambridge, MA
Website: http://www.energid.com/experience/citrus-harvesting/
Product: Citrus harvester


Energid towed multi-arm citrus harvester

Area of use: Florida citrus orchards; oranges (early and late season) and grapefruit
Function: Harvesting, initially for juice
Testing: Will test again during seasons in Florida in 2015 and 2016
Availability: Expect to have initial product in late 2016
Price: System to cost $300,000-$400,000

Return to Companies Mentioned Index
Company: Harvest Automation, North Billerica, MA
Website: http://www.harvestai.com/products
Product: HV-100 mobile robot


Harvest Automation HV-100 Mobile Robot


Harvest Automation potted plant movement schema

Area of use: Nurseries (ornamental, berries, tomatoes, etc.)
Function: Material handling, movement of containers, spacing.
Testing: HV-100 testing completed
Availability: Been selling since 2013
Price: $130,000 for a team of four robots to purchase.  We also rent teams of four for $30K/3 months.  The rental scheme has worked really well.  All who have rented then subsequently purchased.

Return to Companies Mentioned Index
Company: Clearpath Robotics, Kitchener, ON, Canada
Website: http://www.clearpathrobotics.com/grizzly/
Product: Grizzly RUV (cab-less robotic utility vehicle) and Husky UGV


Clearpath Grizzly RUV pulling implement

Area of use: Sold to university research facilities for ag applications development
Function: Harvesting, mowing, hauling, research
Testing: Testing asparagus farming with added laser scanner to identify appropriate stalks and a cutter inserted into the soil to cut the stalk below ground; hauling manure from chicken farms while cleaning out barns; detecting where cows urinate and then treating the area so grass can continue to grow; mowing inbetween orchard rows and hauling (hay/straw wagons back to barn and return so that the farmer doesn’t have to stop baling; hauling chemical refills to sprayer locations; hauling manure spreader)
Availability: Early 2015 – at present only selling to academia and research organizations
Price: $12,000 to $100,000 depending on configuration

Return to Companies Mentioned Index
Company: Autonomous Solutions, Petersboro, UT
Website: http://www.asirobots.com/farming/
Product: Forge Robotic Platform, a kit for enabling a skid steer to operate autonomously or remotely controlled


ASI skid steer with cab in vineyard


ASI skid steer cab options


ASI Universal Vehicle Automation Kit

Area of use: Wine vineyards
Function: Mowing and spraying functions
Testing: Running field trials in CA and TX
Availability: Mid-2015
Price: $75,000 – $150,000/unit (includes complete skid steer device and driving kit) depending on skid steer configuration

Return to Companies Mentioned Index

Company: Wageningen UR (University and Research center), Wageningen, The Netherlands and Agritronics, Sint Annaparochie, The Netherlands
Website: http://www.wageningenur.nl/en/Expertise-Services/Research-Institutes/Wageningen-UR-Greenhouse-Horticulture/Research-themes/Advanced-Cultivation-and-Production-Systems/Subthemes/Computer-vision-and-robotics.htm and http://www.agritronics.nl/
Product: Research to supply intelligent systems for high value crops to commercial research partners/vendors


Wageningen UR cucumber harvesting robot

Area of use: Sweet peppers in The Netherlands, apples and grapes in Belgium, canopy spraying in Slovenia and spot spraying in Italy
Function: Harvesting and spraying (spot and canopy)
Testing: Yes, for sweet pepper, in July in a commercial greenhouse; for apples and grapes tests are now completed. For spraying, field tests have been completed. A new harvester, visual quality inspection and vision system for broccoli, is being developed with start-up Agritronics, Sint Annaparochie, The Netherlands
Availability: “This will take several years”
Price: Not available

Return to Companies Mentioned Index
Company: Vision Robotics, San Diego, CA
Website: http://www.visionrobotics.com/
Product: Lettuce Thinner and Grape Vineyard Pruner


Vision Robotics 6-Line Lettuce Thinner


Vision Robotics grapevine pruner

Area of use: California
Function:  Grapevine pruning being tested; lettuce thinner available for sale
Testing: Testing and development for grapevine pruner could be completed in less than 18 months depending on funding
Availability: Lettuce thinner available now; pruner early 2016
Price: Lettuce thinner starts at $140,000 and upwards depending on configuration; pruner will likely sell for same amount

Return to Companies Mentioned Index
Company: Precision Hawk, Raleigh, NC
Website: http://precisionhawk.com/
Product: Lancaster UAV with various plug and play sensor options plus Datamapper


PrecisionHawk Lancaster


PrecisionHawk Lancaster plug-in sensors

Area of use: Ontario, Canada
Function: Providing data for crop researchers, consultants and farmers and ranchers to make farm management decisions
Testing: Performing field tests under an SFOC from Transport Canada for a number of years. The majority of research and development happens in Ontario, Canada. Over the past six months have obtained a number of CoAs from the FAA to perform field tests and research across the United States in conjunction with universities such as NC State, Texas A&M, Kansas State and Cornell.
Availability: 70% of sales are global. Have entered into a number of projects with US companies on foreign soil for specific research projects
Price: Basic Lancaster platform is $15,000 plus sensors and other options

Return to Companies Mentioned Index

 

Company: F Poulsen Engineering ApS, Hvalso, Denmark
Website: http://www.visionweeding.com
Product: ROBOVATOR thermal and/or hydraulic weeder


Poulsen weeder

Area of use: 30 machines already at work in UK, the EU and Canada.
Function: Weeding and thinning of lettuce, cabbage, fennel and onions
Testing: In California (preceding expansion into North America)
Availability: Started selling in 2011 after 8 years of development
Price: The 5-row version sells in Europe for 80.000€ ($100,000)

Return to Companies Mentioned Index
Company: Kinze Manufacturing, Williamsburg, Iowa and Jaybridge Robotics, Cambridge, MA
Website: http://www.kinze.com/ and http://www.jaybridge.com/
Product: Autonomous vehicle system for row crop harvesting


Kinze autonomous tractor and grain cart

Area of use: Iowa and Illinois corn and soybeans
Function: Autonomously garner row crop grains from combine machines and bring it out of the field to the transport area
Testing: Testing autonomous harvesting system since 2012; in 2013 three farmers in Iowa and Illinois leased systems without Kinze overseeing operation allowing the farmers to use the technology independently. The Kinze system marries off-the-shelf components, including GPS, radar, laser sensors and video cameras, with custom software that allows the system to react to field obstructions. It was developed in partnership with Jaybridge Robotics.
Availability: Kinze is not currently selling the harvesting system but is working towards full commercialization soon
Price: The price has not yet been set for the system which includes the autonomous driving kit for the tractor and grain cart plus the navigation, path planning, harvester communication and control software systems

Return to Companies Mentioned Index
Company: Agrobotics, Little Rock, AR
Website: http://www.agrobotics.com/
Product: AutoProbe soil sampling system


Agrobotics AutoProbe

Area of use: Midwest US farm belt
Function: AutoProbe is a towed device which directs the driving of the towed vehicle to enable consistent, uniform and accurately distanced soil samples. The device is capable of pulling over 2,500 cores per hour
Testing: Tested for 7 years in the Mississippi Delta in the Midwest US
Availability: Available now as both a service and a sale; live demos at various US ag shows
Price: Not available

Return to Companies Mentioned Index
Company: Amazone-Werke Gmbh, Hasbergen, Germany
Website: http://info.amazone.de/DisplayInfo.aspx?id=14033
Product: BoniRob field robot


Amazone-Bosch BoniRob lightweight field robot

Area of use: Work on corn and wheat experimental plots in Germany
Function: Autonomous omnidirectional field robots working in “flocks” for multiple purposes
Testing: Multiple-purpose lightweight robot for weeding, applying fertilizer, inspection being developed with Robert Bosch GmbH
Availability: Only two built; no plans announced for commercialization at this time
Price: No information available

Return to Companies Mentioned Index
Company: Helper Robotech, Gimhae City, Korea
Website: http://helpersys.co.kr/
Product: BoniRob field robot


Helper Robotech fruit and vegetable grafting robot

Area of use: Korea, Japan and China
Function: Grafting is most common in European and Asian countries as well as in greenhouses worldwide where crop rotation is no longer an option and available land is under intense use. Robotic grafting is relatively new although mechanically-assisted grafting has been going on for a long time.
Testing: Unknown
Availability: Available now
Price: Unknown

Return to Companies Mentioned Index
Company: AGCO Fendt, Deluth, GA
Website: http://www.agcocorp.com/GuideConnect.aspx
Product: GuideConnect, SectionControl and VarioGuide


AGCO Fendt GuideConnect – driverless 2nd system


AGCO Fendt VarioGuide auto steering system

Area of use: Global
Function: SectionControl integrates various data and enables fully automatic section control via GNSS for ISOBUS-capable sprayers, spreaders and seeders; the VarioGuide night and day auto steering system; and the new GuideConnect in which two tractors act as a unit where one vehicle is unmanned
Testing: GuideConnect is still under development with no known date or area for availability; the following vehicle doesn’t have its own obstacle detection which may be why they haven’t yet released the product
Availability: All but GuideConnect are available now in the EU and US
Price: Not available for all 3 systems

Return to Companies Mentioned Index
Company: Rowbot, Minneapolis, MN
Website: http://rowbot.com
Product: Rowbot is a self-driving, multi-use platform that travels between rows of corn, ex: applying nitrogen fertilizer in sync with corn needs. It can also collect sensor data to inform both current and future work. GPS and several sensors keep the robot from trampling the crop


Rowbot in cornfield. Rowbots work in teams to apply nitrogen fertilizer in sync with precision needs

Area of use: US Corn Belt
Function: Rowbot travels between corn rows – often under the leaf canopy – to apply nitrogen fertilizer and also to seed cover crops
Testing: Working in conjunction with Carnegie Robotics on development of the Rowbot.
Availability: Began test marketing this year for in-season nitrogen and cover crop seeding services; plan to widen scope of services in 2015
Price: No information available about the cost of the service

Return to Companies Mentioned Index
Company: senseFly, Cheseaux-Lausanne, Switzerland
Website: http://www.sensefly.com
Product: eBee Ag


senseFly eBee Ag


senseFly eBee Ag autopilot system and carrying case

Area of use: Global
Function: The eBee ag system includes eMotion software and a carrying case. The software and cameras enable 2 cm per pixel resolution and produce 3D maps and overlays as well as the capability to lay out (and simulate) a flight path for up to 45 minutes of flying time
Testing: 
Availability: The eBee ag system is available now
Price: About $12,000 for the complete system

Return to Companies Mentioned Index
Company: Conic Systems, Barcelona, Spain
Website: http://www.conic-system.com/
Product: EMP-300 Grafting Robot


Conic Systems EMP-300 Grafting Robot

Area of use: Global
Function: Enables grafting of vegetables  and other greenhouse plants
Testing: Unknown
Availability: Now
Price: Not available

Return to Companies Mentioned Index
Company: Naio Technologies, Toulouse, France
Website: http://naio-technologies.com/
Product: Naio Technologies Oz field robot


Naio Technologies Oz field robot

Area of use: Mostly in France
Testing: Testing next generation of Oz robot (with improved navigation capabilities) in real field conditions in France
Function: The Oz robot serves as an autonomous electric tractor which can be used for weeding and as a transport from harvesters to accumulation points. Oz operates as a self-powered robotic implement rather than a towed implement
Availability: Began selling in 2013
Price: Initially robots are being rented to help customers get familiarized with the product line and to help optimize the utilization. Units are renting/leasing for $315 to $475 per month depending on configuration

Return to Companies Mentioned Index
Company: Robotic Harvesting, Simi Valley, CA
Website: http://www.roboticharvesting.com
Product: Strawberry harvester, data collector and mobile platform


Robotic Harvesting Strawberry Harvester

Area of use: California
Testing: Ongoing in California
Function: Autonomous mobile device which takes stereovision photos to locate any fruit or vegetable in 3D space and then uses a robot arm to pick and place on a conveyor selected berries
Availability: Unknown
Price: Unknown

Return to Companies Mentioned Index

If you liked this article, you may also be interested in:

See all the latest robotics news on Robohub, or sign up for our weekly newsletter.

]]> Game of drones – extreme sports photography https://robohub.org/game-of-drones-extreme-sports-photography/ Thu, 19 Jun 2014 01:22:42 +0000 http://robotlaunch.com/?p=2007 bee logo_cool

“Yet another ‘follow me’ drone” says Chris Anderson – after three projects launched this weekend, including his own 3DRobotics’ open source ‘follow me’ feature for android. Extreme sports photography is the most popular application of consumer drones and at first glance it seems like a no-brainer given the success of GoPro, the rise of the sports ‘robot’ camera tripods and the obvious extension of these trends into aerial photography. And really, just what are drones good for aside from extreme sports tracking? Neither the economics nor the regulations favor package delivery and most other inspection operations can’t be commercialized in the US.

GoPro is a classic 10 year overnight success story. GoPro founder Nick Woodman spent years selling cameras out of his van, hustling and hacking, but his product came at a time when the smart phone was disrupting the low end camera market and both the interest and the technologies were there for a waterproof portable sports video camera. Fast forward to 2014 and GoPro’s annual revenues were almost $1 billion just before filing for their IPO in May.

So it’s no surprise that the sports market is where most of the ‘robot tripods’ have headed for, startups like SoloShot and MovenSee. And it’s a logical extension to put a GoPro on a gimbal onto a drone, or something similar and do ‘follow me’ sports action. As well as Hexo, Airdog, and 3D Robotics, there’s also Pocket Drone with a ‘follow me’ mode. Most of these devices are using beacons to achieve tracking, either separate wristbands or the GPS on a phone or tablet. But Parrot may be releasing something later this year that utilizes image tracking instead.

Now here’s where this whole thing doesn’t fly. The more popular drone sports photography or ‘dronies’ get, the less profitable the business is going to become. You can have lots of cameras on the side of the ski field, or on the beach or in the bleachers at your kid’s soccer game. You can even have cameras at your kid’s dance concert, or graduation or at Yosemite or in the back country. At worst it’s an annoyance.

One drone on a ski field is a distraction. Ten drones is a disaster. Imagine ten drones above your kid’s soccer game? How about their dance concert? What about visiting wilderness areas and watching people standing on the edge of a 200′ waterfall watching the drones above and not the drop below.

The big problem is that in the layer between the ground and the clouds, between people’s heads and small aircraft, objects the size and weight of large birds ARE dangerous. Anyone who’s ever been swooped by magpies (Australia) understands the danger of birds! And it’s well documented that bird strike can be deadly to small aircraft.

I predict that the extreme sports photography market for drones is actually a very small shortlived market. It’s almost inherently unable to be regulated as consumer drones are just too cheap and popular to police, which means it’s likely that we’ll see complete bans on drones in any public, government or commercially controlled space. Not just bans on commercial use of drones. I think ski fields and organized sports events will be the first to crumble as litigation and insurance liability issues arise.

While Chris Anderson predicts that the fastest way for drones to become safe is for them to become so small and soft that they pose no physical threat, that’s still a few years off. And as long as drones are larger than dragonflies and cheap enough to be under everyone’s christmas tree, we face the prospect of global regulatory shutdowns and no fly zones rather than regulated uses.

Is this a startup area I recommend? Not unless you’ve got a layer of lawyers a mile high.

*less than two days after posting this the NPS announced blanket ban on drones in all US National Parks

]]> Global trends in robotics from patent analysis https://robohub.org/global-trends-in-robotics-from-patent-analysis/ Sun, 15 Jun 2014 19:49:46 +0000 http://robotlaunch.com/?p=1994

120,000 robotics patents have been published in the last 10 years, tripling in rate from 2004 to 2013, according to the UK Intellectual Property Office Informatics Team. Unsurprisingly, there was a huge drop in robotics patent applications in 2009-2010, although not all industries were as affected by the global financial crisis as robotics was. The preeminent country for robotics patents is Japan with 31% of patents published, the majority from Toyota. The US is in second place with 19%, followed by Germany (17%), China (10%), Korea (9%), France (3%) and UK at only (2%). Of course this is only an indication of the innovation activity occurring as some countries have greater propensity to patent than others.

The UK IPO is publishing a series of 8 reports looking at trends in emerging/important industries and giving insight into innovation activity and direction for future funding. As well as robotics, the UK Government has identified ‘eight great technologies’ for future growth. These are:

• the big data revolution and energy-efficient computing;

• satellites and commercial applications of space;

• robotics and autonomous systems;

• life sciences, genomics and synthetic biology;

• regenerative medicine;

• agri-science;

• advanced materials and nanotechnology;

• energy and its storage.

Robotics_Infographic

The IPO report also looks at the rate of robotics patents compared to other innovation patents on a country basis and creates a relative specialization index. Some countries such as Japan, China and Germany have proportionately greater robotics patents than general. Whereas both the US and the UK are underspecialized in robotics technologies.

Relative Specialization Index by Country

Overall, automotive patents make up around one third of the total, not including other types of vehicles such as trucks, buses, agricultural machinery, aircraft and aerospace/defence. Google’s automotive patent portfolio is relatively small, at 35 families. Most of Google’s patents were published very recently in 2013, with the earliest being only in 2010. The rate of publishing for Google shows clear increase so further patents should be anticipated. The report also shows collaboration within industry groups. For example, Google and Honda are very self contained in contrast to other automotive companies.

Screen Shot 2014-06-15 at 12.42.53 PM

Most of the patents in the UK dataset are in the field of autonomous vehicles, including road vehicles, unmanned aerial vehicles, and unmanned underwater vehicles. Robotics companies in the UK dataset have very small portfolios, with the largest being Notetry (5 families), which is apparently a division of Dyson Ltd focussing on robotic vacuum cleaners. Other companies are Oliver Crispin Robotics Limited (industrial robotics), Absolute Robotics Limited (industrial robotics), Armstrong Healthcare Limited (robotics for surgeons), Isis Innovation Limited, QinetiQ Limited, and Rolls-Royce Plc.

Finally, the rate of change in robotics patent publishing in the last 10 years is very interesting with China showing significant growth from a China to become one of the strong sources of inventions.

Rate of change in patents published

 

]]> The car in your driveway may be more autonomous than you think https://robohub.org/the-car-in-your-driveway-may-be-more-autonomous-than-you-think/ Wed, 11 Jun 2014 16:56:39 +0000 http://robohub.org/?p=32675
[Jaguar XF, ESC test by EuroNCAP – photo: EuroNCAP ]

Did you know that the majority of the cars we buy and drive today are able to act by themselves and maneuver themselves out of an accident? They can also beat the best human drivers in breaking accuracy and manage even the most finicky engines. Our cars may not yet be fully autonomous but they’re much closer to driving themselves than we realize.

From a pioneering mode of transport to an everyday necessity (or necessary evil) for commuters the world over, the car has come a long, long way since its inception. In its early days, the act of driving required significant expertise; now it is considered almost a right, directly linked with a person’s mobility or even personal freedom. The trend towards  automation covers all aspects of a car’s operation and systems, including driving aids such as climate control and smart lighting, as well as control of the actual movement of the vehicle. This article gives an overview of the milestones along the road to autonomy, focusing especially on the systems and devices that directly control the behaviour of the car: the engine and transmission, the steering and the breaks, and the various sensors and processors that provide and manage the data.

1-CS-12767
[ESP unit – photo: Bosch]

Drivers of the first passenger cars were fully occupied with engine management, and had to control the ignition timing and the air-fuel mix, and perform a highly synchronized ballet of movements just to change gear. Gradually, however, the engine and transmission of our cars became more and more self-sufficient. Manual ignition control gave way to mechanical and then to electronic control. Sensitive carburetors were replaced by self-controlled electronic fuel injection systems, while manual transmissions evolved and became much easier to use. Early on, automatic transmission systems were available to the public where the economy and the market allowed. Nowadays efficient multi-gear automatic transmissions are exceedingly popular and are rapidly replacing manual gearboxes. And of course there is a substantial and growing percentage of electric cars, which can be operated almost as easily as a child’s toy.

The important thing is that, in many commercially available cars today, the input from the driver towards the engine and transmission passes through potentiometers and actuators, with no mechanical link. A computer filters the input from the driver and manages the engine and transmission accordingly.

Data input for the engine management is provided by the engine monitoring sensors, but also from the ESP/ABS (more on that below) and even from the navigation system. By using GPS and cellphone coverage, a car may predict its route and act accordingly, for example changing gear before a steep slope, or having the break pressure stand by in maximum pressure before a bend.

Steering has been more resistant to technological advances; the steering in a car today isn’t very different in principle than that of a 1910 Ford Model T. The most significant improvement was the addition of hydraulic, and later, electric (or electro-hydraulic) assistance. The driver has always a direct link with the steering wheels and, apart from active steering systems (mentioned below), few advances have been made in this area.

It probably isn’t a coincidence that brakes, arguably the most sensitive area from a safety point of view, are by far the most automated system of an everyday car, and where people are less reluctant to have their actions moderated by computers. While, as mentioned above, loss of direct mechanical linkage is relatively new to throttle (usually this loss is found only in research vehicles for steering), brakes lost their mechanical linkage almost a century ago. Their hydraulic system (used in practically all passenger cars) not only augments the force of the driver but, aided by a system of valves, distributes the breaking force in order to compensate for weight transfer.

jensen_interceptor_2x
[Jensen FF, 1966 – the first production car with ABS. Company photo]

The biggest breakthrough though was the introduction of anti-lock breaking system (ABS) first introduced in the 60s. It was designed to prevent the wheels from locking up while the vehicle was moving. Its operation is relatively simple: speed sensors on each wheel detect lock and electronically controlled valves reduce breaking force on the each wheel, thus keeping it within the desired slip percentage (~20% or less). Although a very good driver could use threshold breaking in order to maintain traction, ABS is a milestone because it can perform better than any human by monitoring and controlling the breaking force on each specific wheel. One would have to have four fast, super-human feet and four break pedals (along with the necessary brain capacity) to perform as well as an ABS system, which is found in almost any car on the market today, even the cheapest ones.

BMW-8er-20[2]
[BMW 850i, 1988 – the first production car with ESP photo: BMW]

An even bigger breakthrough that quietly saved many lives and averted numerous accidents is what’s known as electronic stability program, or ESP. It’s an evolution of ABS and uses wheel speed sensors, a steering wheel position sensor, a simple (by robotic standards) yaw sensor and algorithms to govern the ABS valve unit. Judging from each wheel speed, the steering angle and the yaw of the car, the ESP monitors the car’s trajectory (as dictated by the amount the driver turns the steering wheel). If it detects any deviations, it breaks individual wheels in order to pivot the car towards the desired direction. It can also take into account the throttle travel and reduce the power of the car if necessary.

The result is a system that monitors the driver’s input and acts directly through the breaks (and throttle) to retain control of the car. By detecting skids at their earliest stages, and acting quickly on individual wheels and the engine, it can prevent accidents more effectively than any human could.
(You can also check this older video of TV show’s Fifth Gear that demonstrates the abilities of ESP.)

vgs
[active front steering mechanism, Honda S2000 VGS from 2005. Photo:Honda]

Some new systems also employ steering to assist the driver in controlling the car, such as actively controlled rear-wheel steering – a system occasionally used in specific high performance cars for more than two decades, or as it is the case more recently, by acting directly on the main steering system of the front wheels. By adding a corrective input, ESP can use steering to introduce small corrections along with any breaking action.

Active steering is not yet very wide-spread, although rear wheel steering under computer control is quite old (Honda, Nissan, BMW and Mitsubishi were selling cars with active 4WS decades ago, Porsche and Renault still are today). Active front wheel steering is newer but still rare. It’s worth mentioning that it always acts as an additional input and the driver maintains a direct mechanical linkage to the steering. The ESP control mechanism adds an additional or counterweight force (usually through a planetary gear, in parallel with the steering column) that translates to a slightly corrected slip angle.

autowp.ru_nissan_skyline_gt-r_4
[1989 Nissan Skyline R32 GT-R – one of the few cars with RWS and active differentials. Photo: Nissan]

A different approach with similar (thought not identical) results is to employ active differentials. A system with clutches and/or auxiliary gearboxes inside the transmission and differentials can actively manage how much power is delivered on each wheel, thus providing the necessary yaw moment to influence the attitude of the car. The effect is similar to active rear wheel steering, providing yaw moment without slowing the car (like individual wheel breaking does). This is why, like RWS, active differentials are mostly used on performance cars.

Audi A8
[2007 Audi A8 equipped with adaptive cruise control. Photo:Audi]

Another area of gradual automation is adaptive cruise control systems. This is an enhanced version of the regular cruise control systems that are very popular in the US (although not so much in Europe). Cruise control maintains a certain speed until the driver hits the breaks. Additionally, adaptive cruise control can automatically reduce speed when the car is approaching a vehicle in front of it and restate the desired speed when conditions allow.

The system usually includes a radar sensor placed in the front bumper and the same technology is also applied in collision avoidance systems even without the cruise control function. By detecting obstacles in front of it, the car can notify the driver or even take action and break by itself. Nowadays even very small and inexpensive cars can be equipped with it, and its performance is more than adequate. A similar function is the lane departure warning, where sensors (usually placed under the rear-view mirrors) scan the road and detect from the white lines if the vehicle is going off course.

Finally the most impressive task a commercial passenger car can do today is without doubt park by itself. Already on sale, park-assist can scan the surroundings and perform the necessary maneuvers in order to park, as long there is sufficient space. Probably a luxury and not really necessary given that it operates where even a mediocre driver can easily park, it’s is the most visible picture of a self-driven car available on the market.

A very important point is that all the systems described above could co-operate with each other and maximize their influence. Although in general they’re used independently it is inevitable that they’ll be integrated into a complete set. Steps towards this direction have already been made but there is a lot of room for improvement.

How close are today’s cars to complete automation?

Engine-wise, many cars today (and certainly more in the future) have no mechanical link between the throttle pedal and the engine. Power delivery is under fully electronic control, so from the perspective of the engine, everything is primed for a fully autonomous car.

Likewise for the powertrain and transmission. Automatic gearboxes are ever more popular and already they require no human input. Of course things are even easier for electric cars.

Breaks are also already highly automated, but they usually rely on the force one is exerting on the pedal. However by-wire breaks have already been used in passenger cars (Mercedes W211 & R230) even with an analog hydraulic backup for safety (or legislative reasons).

The steering system, although it will probably be the last to come fully under computer control, is already getting there with active steering systems currently on sale. The step from an assistive active steering system to a full authority one is very small (at least technically).

The data input and processing necessary for an autonomous car is technically the most difficult part by far. What we described above is automation that builds on human input: even the most capable commercial cars today rely on their drivers to tell them where to go, how fast and what to avoid. A new set of sensors along with the necessary onboard computational power should eventually substitute for the driver, but this is already happening both partially (as we examined in this post) and fully, as in Google’s new purpose-built self-driving car.

Though functional fully autonomous cars exist and perform well even today, the last obstacle they have to overcome in order to reach the market is the product liability and the legal complications of having a car acting on its own.

Glossary:

ABS: anti-lock breaking system (or Antiblockiersystem in German)

ESP: Electronic Stability Program (or Elektronisches Stabilitätsprogramm in German), also found in other acronyms, ESC (Electronic Stability Control ), VSA (vehicle stability augmentation) etc.

RWS: rear wheel steering

VGS: variable gear ratio steering

If you liked this article, you may also be interested in:

See all the latest robotics news on Robohub, or sign up for our weekly newsletter.

]]> How do self-driving cars work? https://robohub.org/how-do-self-driving-cars-work/ Tue, 03 Jun 2014 19:56:18 +0000 http://robohub.org/?p=32444
Nissan’s autonomous car prototype – 2013, photo:Nissan Global

Tesla CEO, Elon Musk, recently announced that the car manufacturer will produce self-driving cars within three years. Nissan has announced that it will have a self-driving car available by 2020, Google has said it will do so by 2018. Over the past decade, the conversation around self-driving cars has evolved from futuristic police chase sequences in Minority Report to figuring out which auto manufacturer will be first to launch a commercially viable self-driving vehicle. Daimler AG, maker of Mercedes Benz, recently announced that an S-class sedan had completed a 62-mile journey in the streets of Germany without a driver. Audi’s self-driving car successfully navigated 156 turns of the 12-mile Hill Climb course in Colorado’s Pikes Peak. Car manufacturers see self-driving cars as a way to eliminate road deaths caused by human error, reduce traffic, and free up time spent commuting – but how do these vehicles work?

Self-driving cars in a nutshell

A self-driving car is capable of sensing its environment and navigating without human input. To accomplish this task, each vehicle is usually outfitted with a GPS unit, an inertial navigation system, and a range of sensors including laser rangefinders, radar, and video.  The vehicle uses positional information from the GPS and inertial navigation system to localize itself and sensor data to refine its position estimate as well as to build a three-dimensional image of its environment.

Data from each sensor is filtered to remove noise and often fused with other data sources to augment the original image. How the vehicle subsequently uses this data to make navigation decisions is determined by its control system.

The majority of self-driving vehicle control systems implement a deliberative architecture, meaning that they are capable of making intelligent decisions by 1) maintaining an internal map of their world and 2) using that map to find an optimal path to their destination that avoids obstacles (e.g. road structures, pedestrians and other vehicles) from a set of possible paths. Once the vehicle determines the best path to take, the decision is dissected into commands, which are fed to the vehicle’s actuators. These actuators control the vehicle’s steering, braking and throttle.

This process of localization, mapping, obstacle avoidance and path planning is repeated multiple times each second on powerful on-board processors until the vehicle reaches its destination.
Audi-TTS-Autonomous-Pikes-Peak-13

On board computers of STanford/Audi autonomous TTS

The next section focuses on the technical components of each process: mapping and localization, obstacle avoidance and path planning. Although car manufacturers use different sensor suites and algorithms depending on their unique cost and operational constraints, the processes across vehicles are similar. The descriptions below most closely mirror their implementation in state-of-the-art self-driving military vehicles.

 

Breaking Down the Technicals

Mapping and Localization

Prior to making any navigation decisions, the vehicle must first build a map of its environment and precisely localize itself within that map. The most frequently used sensors for map building are laser rangefinders and cameras. A laser rangefinder scans the environment using swaths of laser beams and calculates the distance to nearby objects by measuring the time it takes for each laser beam to travel to the object and back. Where video from camera is ideal for extracting scene color, an advantage of laser rangefinders is that depth information is readily available to the vehicle for building a three-dimensional map. Because laser beams diverge as they travel through space, it is difficult to obtain accurate distance readings greater than 100m away using most state-of-the-art laser rangefinders, which limits the amount of reliable data that can be captured in the map. The vehicle filters and discretizes data collected from each sensor and often aggregates the information to create a comprehensive map, which can then be used for path planning.

What_Google_Car_Sees_LIDAR.
An example of a Google car’s internal map at an intersection, tweeted by Idealab founder Bill Gross. Gross claims that Google’s Self-Driving Car gathers almost 1 GB of data per second.

For the vehicle to know where it is in relation to other objects in the map, it must use its GPS, inertial navigation unit, and sensors to precisely localize itself. GPS estimates can be off by many meters due to signal delays caused by changes in the atmosphere and reflections off of buildings and surrounding terrain, and inertial navigation units accumulate position errors overtime. Therefore localization algorithms will often incorporate map or sensor data previously collected from the same location to reduce uncertainty. As the vehicle moves, new positional information and sensor data are used to update the vehicle’s internal map.

Obstacle Avoidance

A vehicle’s internal map includes the current and predicted location of all static (e.g. buildings, traffic lights, stop signs) and moving (e.g. other vehicles and pedestrians) obstacles in its vicinity. Obstacles are categorized depending on how well they match up with a library of pre-determined shape and motion descriptors. The vehicle uses a probabilistic model to track the predicted future path of moving objects based on its shape and prior trajectory. For example, if a two-wheeled object is traveling at 40 mph versus 10 mph, it is most likely a motorcycle and not a bicycle and will get categorized as such by the vehicle. This process allows the vehicle to make more intelligent decisions when approaching crosswalks or busy intersections. The previous, current and predicted future locations of all obstacles in the vehicle’s vicinity are incorporated into its internal map, which the vehicle then uses to plan its path.

Path Planning

The goal of path planning is to use the information captured in the vehicle’s map to safely direct the vehicle to its destination while avoiding obstacles and following the rules of the road. Although manufacturers’ planning algorithms will be different based on their navigation objectives and sensors used, the following describes a general path planning algorithm which has been used on military ground vehicles.

This algorithm determines a rough long-range plan for the vehicle to follow while continuously refining a short-range plan (e.g. change lanes, drive forward 10m, turn right). It starts from a set of short-range paths that the vehicle would be dynamically capable of completing given its speed, direction and angular position, and removes all those that would either cross an obstacle or come too close to the predicted path of a moving one.  For example, a vehicle traveling at 50 mph would not be able to safely complete a right turn 5 meters ahead, therefore that path would be eliminated from the feasible set.  Remaining paths are evaluated based on safety, speed, and any time requirements. Once the best path has been identified, a set of throttle, brake and steering commands, are passed on to the vehicle’s on-board processors and actuators. Altogether, this process takes on average 50ms, although it can be longer or shorter depending on the amount of collected data, available processing power, and complexity of the path planning algorithm.

The process of localization, mapping, obstacle detection, and path planning is repeated until the vehicle reaches its destination.

 

The Road Ahead

Car manufacturers have made significant advances in the past decade towards making self-driving cars a reality; however, there still remain a number of technological barriers that manufacturers must overcome before self-driving vehicles are safe enough for road use. GPS can be unreliable, computer vision systems have limitations to understanding road scenes, and variable weather conditions can adversely affect the ability of on-board processors to adequately identify or track moving objects. Self-driving vehicles have also yet to demonstrate the same capability as human drivers in understanding and navigating unstructured environments such as construction zones and accident areas.

These barriers though are not insurmountable. The amount of road and traffic data available to these vehicles is increasing, newer range sensors are capturing more data, and the algorithms for interpreting road scenes are evolving. The transition from human-operated vehicles to fully self-driving cars will be gradual, with vehicles at first performing only a subset of driving tasks such as parking and driving in stop-and-go traffic autonomously. As the technology improves, more driving tasks can be reliably outsourced to the vehicle.

The technology for self-driving cars is not quite ready, but I am looking forward to the self-driving cars of Minority Report becoming a reality.

If you liked this article, you may also be interested in:

See all the latest robotics news on Robohub, or sign up for our weekly newsletter.

]]> Pure autonomy: Google’s new purpose-built self driving car https://robohub.org/pure-autonomy-googles-new-purpose-built-self-driving-car/ Wed, 28 May 2014 13:51:17 +0000 http://robohub.org/?p=32075

Google completed a major step in its long and extensive self-driving cars project by presenting its first purpose-built autonomous car, which is designed from scratch for its role and is not a modified conventional Toyota.

The as yet unnamed car is very small (looks smaller than a Smart) and can accommodate two people and some luggage. It’s probably electric and its maximum speed is limited to 25mph (~40km/h). Its most striking characteristic is that it doesn’t have any controls — no steering wheel, accelerator or brake pedals — and you can ride it strictly as a passenger, which is probably a strange feeling, but according to Google’s video not entirely unpleasant.

This is a purely experimental project in every aspect. Google will make about a hundred prototypes and it looks like it’s a very serious effort. It is equipped with multiple sensors (a big LIDAR on its roof is probably doing most of the work), but there are also sensors in the front, in the back and where the side view mirrors would have been in a regular car. Thanks to Google’s previous experience with self-driving cars, one can expect very good performance in real environments; other Google self-driving cars have completed hundreds of thousands miles with no major incidents.

google_car_2

The car itself may look like a toy but it is cleverly and purposefully designed to be as cute as possible in order to inspire trust and reduce fear of its autonomous status. Small cars in general may be more vulnerable in a crash with a heavier vehicle, but it’s already proven that with clever engineering you can have almost no compromises in their passive safety, and most small cars from established manufacturers achieve very high scores in a crash test. The main chassis is made from robust box sections of (probably) aluminum tubing and there’s also a tubular roll cage above. Even the suspension wishbones look highly over-engineered for a 25mph city car. A substantial crash box is placed in the front, and above it the whole front panel is made of foam, so the car is not only safe for its passengers but for pedestrians as well. Of course at this stage there are no independent crash tests to verify its performance.

google_car_3

As mentioned above, the most striking aspect of the car is its lack of any kind of manual control. This is a very clever move from Google that aspires to overcome the legal and ethical problems of who should be able to drive or control a self-driving car.

If you completely eliminate any kind of input from the occupants (apart from the destination selection), then anyone on board is defined strictly as a passenger and a question like “should we allow a child to ‘drive’ an autonomous car” is transformed into “should we allow a child to be transported by an autonomous car” — a question that is still probably not very easy to answer, but which is certainly easier than the former one.

Google will launch a small pilot program in California in the next couple of years. The prototypes released on public roads will have manual controls for obvious legal reasons but the goal of this project is to learn and develop the technology and know-how. Read the full post in Google’s official blog here.

It’s worth mentioning that similar concepts and ideas (but not a functional prototype) have been presented over the years. The concept closest to this is ‘UC’ by the pioneering design firm Rinspeed, presented four years ago at the Geneva Motor Show. It was also a two-seater (loosely based on a Fiat 500) but its main feature was the interior, where similarly to Google’s car, no manual controls were present.

Rinspeed-UC ]]> The fate of jobs in relation to robotics is confusing https://robohub.org/the-fate-of-jobs-in-relation-to-robotics-is-confusing/ Mon, 10 Feb 2014 18:42:00 +0000 http://robohub.org/?guid=b1507ac4eeda5e201fd3b81a5f479e1c The_Second_Machine_Age_Brynjolfsson_McAfeeWill a robot take your job? Will a higher minimum wage cause job destruction? Or is it all media hype? What’s the truth?

MSNBC asked Bill Gates whether increasing the federal minimum wage is a good idea. He responded: “If you raise the minimum wage, you’re encouraging labor substitution, and you’re going to go buy machines and automate things — or cause jobs to appear outside of that jurisdiction. And so within certain limits it does cause job destruction.”

Will the current political dialogue over raising the federal minimum wage increase the fear people express about robots taking over their jobs? Two infographics attempt to answer that question but instead just add fodder to the grist of the discussion.

In a not-to-be-taken-too-seriously infographic entitled Techno Takeover “Can your job be done by a robot?”, by MindFlash.com they show the following:

  • 3.55 million retail clerks were singled out as likely candidates to be replaced by checkout machines similar to ATMs
  • 3.2 million commercial drivers – long-haul, short-run, limos, taxies are all likely to be most effected by automated driving systems
  • 2.3 million customer service representatives – better than off-shore call centers?
  • Some portion of the 1.5 million members of the armed forces
  • 269,900 pharmacists – “robots are unparalleled when it comes to precision”
  • 263,800 paralegals sifting through and analyzing legal documents are another target for Watson-like artificial intelligence
  • 78,780 medical transcriptionists are being replaced by speech recognition software

It’s true that employers now have easier, cheaper access to software, automation and cheap expert talent. Brynjolfsson and McAfee, both MIT professors and authors of The Second Machine Age, suggest that:

“We are beginning to automate a lot more cognitive tasks, a lot more of the control systems that determine what to use that power for. In many cases today artificially intelligent machines can make better decisions than humans.” So humans and software-driven machines may increasingly be substitutes, not complements. What’s making this possible are three huge technological advances that just reached their tipping points: exponential, digital and combinatorial.” [Combinatorial advances mean you can take Google Maps and combine them with a smartphone app like Waze, through which drivers automatically transmit traffic conditions on their routes by just carrying their phone in their car, and meld both into a GPS system that not only tells you what the best route is to your destination but what the best route now is because it also sees all the traffic everywhere. Instantly, you’re the smartest driver in town.]

The other infographic, this time by Manolith.com, suggests 10 new jobs that might be in the near-term future – the most current 4 are shown below:

  • Become an urban farmer
  • Become an alternative reality architect providing virtual augmentation content
  • Become a personality programmer offering alternative voices and personalities to Siri-like programs
  • Become a remote robot pilot for planes, trains, trucks, delivery vehicles and taxis

Thomas Friedman, in a NY Times Sunday Review article, suggested that over the past 10 years the world has gone from connected to hyperconnected and as a result, employers now have access to above-average software, automation and cheap talent from abroad. Brynjolfsson and McAfee call this the start of the Second Machine Age (described above). The first machine age was the Industrial Revolution and was all about devices and systems to augment human muscle while requiring human control of the activity.

Robotic associations including the A3 (Association for Advancing Automation, a 750-member umbrella association for the RIA, AIA and MCA) and IFR (International Federation of Robotics) have commissioned studies and produced videos which show that deploying robotics does increase and/or preserve wages in the community. Nevertheless, as Bloomberg Businessweek said, “Factory jobs are gone. Get over it.” The truth seems to be that dull, dirty and dangerous jobs will be replaced by automation and robots thereby making products with less spoilage, fewer accidents, higher quality at lower cost – thus they will be more competitive and increase sales with the consequence of increasing workers of a different type: marketing, sales, expediting, warehousing, etc.

]]> ShanghAI Lectures: Shaohua Tan “Qualitative modeling and analysis” https://robohub.org/shanghai-lectures-shaohua-tan-qualitative-modeling-analysis/ Thu, 28 Nov 2013 14:00:01 +0000 http://robohub.org/?p=17219 ShanghAIGlobeColor_mini_0_0Guest talk in the ShanghAI Lectures, 2009-11-12

“It has increasingly been realized that some of the key characteristics underlying real-world complex dynamical systems (such as economical, financial and ecological systems) can only been modelled and thus understood and predicted at qualitative level directly.However, lacking a coherent and arithmetically sound theoretical framework for modeling and analyzing such systems entirely at qualitative level has long been an obstacle to such a qualitative modeling endeavor. Our research work led to a novel qualitative theory that provides a solution to overcome this obstacle. We develop a ternary qualitative algebra and related arithmetic operations directly over a qualitative space and use this algebra as the basis to build a direct qualitative modeling approach with qualitative functions without falling back to quantitative details inherent in most existing approaches. One of the important results of our proposed modeling theory is that any qualitative system can be represented as a piecewise linear qualitative function. This result lays a theoretical foundation for piecewise linear qualitative function structure to serve as a normal form for representing arbitrary qualitative functions. This short lecture will provide a non-technical description of our qualitative modeling techniques and demonstrate its usefulness in modeling, thus analyzing, complex real-world systems such as financial and economical systems.”

Professor Shaohua Tan received his Ph.D in Electrical Engineering from Katholieke Universiteit Leuven, Belgium in 1987. He has been Professor in Center for Information Science, Peking University, for 13 years. He held various teaching and research positions in a number of countries prior to joining Peking University. His research interests include developing qualitative modeling techniques in modeling complex real-world systems and analysis of financial systems using AI techniques.

https://www.youtube.com/watch?v=YZjyoo0Y1Js

The ShanghAI Lectures are a videoconference-based lecture series on Embodied Intelligence run by Rolf Pfeifer and organized by me and partners around the world.

The ShanghAI lectures have brought us a treasure trove of guest lectures by experts in robotics. You can find the whole series from 2012 here. Now, we’re bringing you the guest lectures you haven’t yet seen from previous years, starting with the first lectures from 2009 and releasing a new guest lecture every Thursday until all the series are complete. Enjoy!

]]> From precision farming to autonomous farming: How commodity technologies enable revolutionary impact https://robohub.org/from-precision-farming-to-autonomous-farming-how-commodity-technologies-enable-revolutionary-impact/ Mon, 11 Nov 2013 16:00:37 +0000 http://robohub.org/?p=22487 The popular conception of farming as low-tech is woefully out of date. Modern farmers are high-tech operators: They use GIS software to plan their fields, GPS to guide field operations, and auto-steer systems to make tractors follow that GPS guidance without human hands. Given this technology foundation, the transition to full autonomy is already in progress, leveraging commodity parts and advanced software to get there more quickly than is possible in many other domains.

This article outlines some of the key technologies that enable autonomous farming, using the Kinze Autonomous Grain Harvesting System as a case study.

Source: Jaybridge Robotics

Jaybridge Robotics automates vehicles for driverless operation in industrial domains including agriculture and mining. We do this using commercial, off-the-shelf (COTS) components and our software. We believe software is the greatest challenge in making vehicular robotics cost-effective and reliable. For the last few years, we have been working with Kinze Manufacturing to automate their line of agricultural equipment (see video above). In this article, I’m going to explain why farmers are technologically well-positioned to take advantage of automated farm vehicles. I’ll also provide a case study of the technology behind the Kinze Autonomous Grain Harvesting System

Satellite-guided farming

GPS-satellite-Block-IIF-(Boeing-graphic)
Figure 1: GPS satellite Block IIF (Boeing graphic)

Although urbanites may still think of farming as low-tech, backward profession, a great deal of professional farming has gone high-tech in the US and other developed nations. The last decade, in particular, has seen the rapid embrace of high-tech under the general label of Precision Agriculture.
Farmers collect and act on copious amounts of data. Global Positioning System (GPS) data from satellites (see Figure 1) lies at the heart of Precision Agriculture. Farmers use farm management software (e.g FarmWorks) and GPS receivers to map their fields, and to track the yield (amount of crop) that they get from every square meter. They can augment this yield data with a variety of other information. For instance, they may also perform a detailed soil sampling survey to determine the soil’s nutrient mix in different areas of the field. Plant health can be assessed in part by plant coloration, so some farmers will purchase satellite or aircraft flyover imagery enabling them to determine the health of their plants at various times during the year. Of course, all of this additional information goes into the computer mapping software.

Based on their collected and mapped (georeferenced) data, farmers can generate prescription maps which specify how much fertilizer to apply in each region of the field, how densely to plant seed in that region, and so on, in order to optimize yield and minimize unnecessary chemical applications.
To take optimal advantage of prescription maps, many modern farming implements are computer-controlled. Planters, as the name suggests, put seed in the ground. A planter such as the one shown in Figure 2 may feature independently controllable row units, enabling each unit to be turned on and off, or have its planting rate adjusted, independently.

Kinze-Manufacturing-4900-front-fold-planter)
Figure 2: Kinze Manufacturing 4900 front-fold planter) Source: Kinze Manufacturing

Given this sort of planter, the prescription map is loaded into the computer on the tractor, and the tractor driver simply..drives. The driver steers the tractor, the tractor pulls the planter, and the onboard computer controls the seeding rate based on where the planter is in the field. The computer also tracks where seed has already been applied, so if the driver has to drive through already-seeded territory, it doesn’t get double-seeded.
The same strategy is applied in other crop maintenance activities such as fertilizing and other chemical applications. The computer monitors the vehicle location and ensures chemicals are applied only where prescribed, in customized doses tailored to the specific area. This has cost benefits to the farmer — less chemicals used is dollars saved — and it also has environmental benefits, since less chemicals used is less chemicals at risk of leeching into the surrounding ecosystem.

Cooperative autonomy

When putting seed in the ground, and later coming back to harvest it, it’s important that the harvester follows the same path as the planter did months earlier. And of course, the harvester and planter know where they are the same way that your smartphone knows where you are: GPS.
You may have noticed, however, that your cell phone can be off by quite a bit. The difference with farming vehicles is that while the GPS unit you have in your phone or your car knows where you are to within a few yards, a high-precision augmented GPS in a modern tractor knows where it is to within a couple of inches. This accuracy has profound consequences.
First, it allows a tractor driver to reproduce a route over and over. At planting, seed goes in the ground. When treating, fertilizer is applied directly to the seeded area — without being applied to the unseeded territory between the rows. Finally, at harvest, the driver reaps with high efficiency by being just as accurate coming along the rows in the harvester.
Historically, the precision-guided driver was assisted by a “lightbar” — a line of LEDs that indicate in real-time whether the vehicle is on track or whether a steering correction is needed. Nowadays, an advanced tractor just drives itself along the route.
Auto-steer systems are available for a variety of tractor models, both as built-in and as after-market additions. Most current auto-steer systems can only drive the rows, requiring driver intervention at the end of each row, but advanced systems from some vendors can now handle certain simple turns. Even with auto-steer, a driver is still required to watch for obstacles and monitor the equipment, although the work is a lot less fatiguing when the driver can go “hands-off” for long rows. While the farmer is still setting the throttle and looking out for collision, the system is inarguably autonomously driving itself.

Prepared for full automation

This says a lot about farmers as future users of more completely automated systems. On farms that have embraced precision agriculture:

• The farmers are tech-savvy computer users.
• They survey their fields with precision.
• Their tractors are already partially drive-by-wire, meaning that a computer can already control key functions such as steering.
• Their tractors are equipped with high-precision GPS systems.

Given this baseline, transitioning to full autonomy is relatively straightforward, using off-the-shelf parts and advanced software.

Case study: Kinze Autonomous Grain Harvesting System

The-Kinze-Autonomous-Grain-Cart-system-working-with-a-combine-in-2012
Figure 3: The Kinze Autonomous Grain Cart system working with a combine in 2012. Source: Jaybridge Robotics

Kinze Manufacturing makes grain carts and planters for row crops. For the last few years, Jaybridge Robotics has been working with Kinze to automate tractors pulling their grain carts to produce the Kinze Autonomous Grain Harvesting System, shown in the opening Video and in Figure 3. In this case study, we’ll take a closer look at how the Kinze system builds on existing technologies and Jaybridge’s software.

Jaybridge-Robotics-software-modules-(simplified)
Figure 4: Jaybridge Robotics software modules (simplified). Source: Jaybridge Robotics

A high-level overview Jaybridge Robotics’ software is shown in Figure 4. Our software takes advantage of commodity components to perform key vehicle automation tasks including:

• User interface enabling the user to perform the workflow.
• Vehicle path planning.
• Vehicle control, including steering, brakes, throttle, etc.
• Navigation
• Obstacle detection
• Inter-vehicle communications

Let’s look at how these elements are realized in the Kinze Autonomous Grain Harvesting system.
The user interface, shown in Figure 5, runs on a touch-screen Android tablet. Working with Kinze, four primary workflow elements were identified. In Offload, a grain cart drives in tandem with a combine while the combine simultaneously harvests and dumps crop into the grain cart. In Follow, a grain cart follows along behind a combine, for instance when the combine transits a narrow area. In Park, a grain cart drives back to a designated parking area, where it meets up with a semi which will transport the grain onward. And in idle mode, of course, the grain cart idles awaiting further instruction. Those key workflow elements are realized in the major buttons down the right-hand side of the screen, while additional capabilities such as manual crop editing and obstacle denotation are provide along the bottom.

Screen-on-the-tablet-in-the-combine-cab
Figure 5: Screen on the tablet in the combine cab

Path planning, vehicle control, navigation, and obstacle detection all take place in real-time on the embedded computer onboard the tractor towing the grain cart.
The path plan adapts in real-time as the combine moves – keep in mind that combine motion moves not only the grain cart’s destination (in follow or offload modes), but also clears crop, creating additional drivable area. The plan may have to feature complex maneuvers, e.g. Figure 6 where harvesting is taking place in a terraced field. The path planner relies on the navigation system identifying the grain cart’s position, orientation, and velocity. When tandem-driving with the combine for offload, it also relies on high-speed communications between the vehicles to exchange position information. The path planner must continuously consider vehicle position and the drivable area map, as well as the vehicle’s physical capabilities.

Path-plan-for-a-grain-cart-making-a-sharp-turn-around-a-terrace
Figure 6: Path plan for a grain cart making a sharp turn around a terrace. Source: Jaybridge Robotics

Vehicle control also takes place in real-time, ensuring that the vehicle follows the planned paths. Like path planning, control runs on the onboard embedded computer, synchronizing throttle, brakes, and steering to achieve the desired path.
The navigation system fuses data from the factory-standard high-precision GPS system with other vehicle information to provide an extremely accurate estimate of vehicle state.
The obstacle detection system relies on a spinning laser range finder (LIDAR) and automotive RADAR more typically used for adaptive cruise control. Data from both sensors are fused for enhanced detection capability.
Inter-vehicle communications take place via two different channels. At longer ranges, grain carts and combines communicate via cell data, taking advantage of pervasive cell coverage extending ever deeper into the heart of farm country. At close range, and especially when tandem driving, a short-range high-bandwidth radio is used to exchange data to coordinate driving.
It’s important to note that the hardware components are commercial-off-the-shelf (COTS) parts: from the embedded computer to the LIDAR to the cell modem, the technology exists today at very reasonable price points. Jaybridge’s software transforms them from a collection of parts to a fully automated grain harvesting system.

Reliability

Industrial machinery has to be reliable. Farming machinery is no exception. So a key aspect of Jaybridge Robotics’ work is ensuring that automated vehicles, and the software controlling them, are reliable. Jaybridge relies on a number of techniques including formal code inspection, unit testing, regression testing, and large-scale simulation (see Figure 7) to validate software before it goes onto real hardware. Simulation, in particular, is a potent tool in our arsenal: it gives every Jaybridge engineer a complete system to work with, without having to find parking for a bunch of tractors.

Jaybridge-simulation-environment
Figure 7: Jaybridge simulation environment. Source: Jaybridge Robotics

Ongoing development

The Kinze Autonomous Grain Harvesting system was unveiled to the public in 2011. In 2012, multiple systems were put into the hands of real Illinois farmers for the fall corn and soybean harvest. In 2013, systems are once again working the harvest, with increased capabilities and ever-greater robustness. As we continue along the technology roadmap, we are looking forward to further enhancing the capabilities and robustness of the Kinze Autonomous Grain Harvesting System.

]]> AUVSI conference brings robots to drone debate epicenter https://robohub.org/auvsi-conference-brings-robots-to-drone-debate-epicenter/ Fri, 23 Aug 2013 00:30:12 +0000 http://robotsindc.com/?p=7961 AUVSI returned to D.C. for 2013.
AUVSI returned to D.C. for 2013.

Amidst a climate of fiscal austerity and vibrant debates over the growing importance of unmanned vehicles in foreign policy and homeland security, the 2013 AUVSI Unmanned Systems Conference returned to Washington, D.C., last week after hosting the 2012 event in Las Vegas. The event was not without controversy, however, as activist group Code Pink held a demonstration outside the venue and disrupted a keynote address. The show itself was a tale of two storylines as the exhibit hall demonstrated that applications for defense and law enforcement are still the lifeblood of the unmanned systems industry, while the technical program and panel discussions pointed to a growing interest to move into commercial industries. Here’s what you missed:

UAVs as economic drivers

As expected, the Federal government’s booth presence was minimal among the nearly 600 exhibitors compared to previous years due to limited budgets. Interestingly, several states including Ohio, Minnesota, Nevada, North Dakota, Utah, Oklahoma, New Mexico, and Arizona were present to promote their economic programs and bids as one of six test sites under the FAA’s UAS integration program. Approved sites will attract unmanned aerial vehicle manufacturers as the FAA begins to open U.S. airspace in 2015. North Dakota Lieutenant Governor Drew Wrigley and U.S. Congressman Mike Turner (R-Ohio) made special visits to the conference to further promote their respective states. States will become increasingly important for manufacturers as Defense procurements slow and commercial markets are opened.

Big and small vendors showcase new platforms

iRobot had their entire UGV lineup on display
iRobot had their entire UGV lineup on display.

Among exhibitors, the vast majority of platforms continue to be fixed-wing aerial vehicles of various sizes. The big players of General Dynamics, Lockheed Martin, and Northrop Grumman showed the latest in drone technology, including a model of Northrop’s X-47B, which just this past July achieved a historic result by autonomously landing on the flight deck of the USS George H.W. Bush. Several quadrotor vendors were also present to market easy-to-use and agile systems for law enforcement. I was particularly impressed by systems from Aeryon Labs and navigation systems from Airware. While last year we were introduced to the unique ground platforms from Boston Dynamics such as the LS3 and RHex, no particular system stole the show this year. It was nice to see iRobot’s RP-VITA out and about with live telepresence demonstrations, however. Among maritime systems, California-based Liquid Robotics brought along “Benjamin Franklin,” recently returned from a Guinness World Record-setting 14,703 km data collection journey across the Pacific Ocean.

Self-driving cars, the future?

Chuck Thorpe (L), Annie Lien, and Myra Blanco discuss automated driving.
Chuck Thorpe (L), Annie Lien, and Myra Blanco discuss automated driving.

But it was the push toward applications in agriculture and self-driving cars that attracted the most attention and excitement among technical program attendees. The automated driving panel featured a lively discussion from panelists Chuck Thorpe, Clarkson University Provost and former Carnegie-Mellon University Robotics Institute head; Virginia Tech researcher Myra Blanco, and consultant Annie Lien. Chief among the discussion topics was the role of states in regulating self-driving cars. While California, Nevada, and others are helping get these vehicles onto roads, this approach will hinder progress and development, according to Ms. Lien. “Safety requirements are normally defined at the federal level,” she noted. Different regulations for each state will eventually create a logistical and legal headache for manufacturers and consumers alike.

Though Google and car manufacturers are making significant progress in developing and testing the necessary technologies for fully autonomous systems, the panelists noted that they are taking very different approaches toward market introductions. Among the four defined levels of autonomous cars, with level 0 being manual and level 4 being full autonomy in unstructured road conditions, panelists suggested Google is aiming directly for level 4 within five years while BMW, Mercedes and others are gradually introducing new features in time-tested industry fashion. For example, some 2014 year models will feature traffic jam assistance that will facilitate self-driving with minimal intervention in stop-and-go situations.

Statistics on commercial opportunities for unmanned systems
Statistics on commercial opportunities for unmanned systems.

So where will we be by 2020? The panelists seemed to converge on the certainty of level 3 market availability, but with complex socio-technical systems such as transportation, that is still anyone’s guess. Thorpe summed it up fairly well by suggesting it is always more difficult to transform an existing industry than creating a new one. Aside from the technology and safety regulations, we are only now considering the larger societal questions such as urban sprawl: if you could work or sleep while the car autonomously drove during your commute, would you live farther away?

]]> Visions of transformative research: Risk-reward, variation-selection, hot trends, exploration of the wild https://robohub.org/visions-of-transformative-research-risk-reward-variation-selection-hot-trends-exploration-of-the-wild/ Mon, 08 Jul 2013 14:41:52 +0000 http://robohub.org/?p=16256

Over the last 20 years or so, a sense that science has become conservative or incrementalist has developed, and calls for change in the approaches to public funding of research have been heard from various quarters. Several notions have been suggested of what should be supported instead of “normal science” or “incremental innovation.” Among them we have heard calls for more “high risk-high reward” research, or for more “highly creative” science, or for more “cutting edge” or “frontier” research and, more recently in language adopted by funding agencies, that more “transformational research” is needed.

The main idea is that there isn’t enough risk taking, and little scientific or technological creativity is reflected in proposals that might “revolutionize” our understanding of the nature or society. “Business as usual” in the support of science and technology seems not to be satisfactory for producing the sort of radical change and rapid advance in science needed to solve important problems we face in economic development, health care and general social wellbeing.

This discussion does not fully address the situation in the private sector. However, the fate of many top-performing R&D centers in companies such as IBM, Texas Instruments, AT&T, among others, over the same period indicates that corporate willingness to take risks in research is not precisely at its most adventurous. Having said that, our focus here is on public funding since the sense of risk and reward of innovation by private industry in the market does not have the severe principal-agent problem faced by public funding. Companies invest their own money and either lose it or make a profit for themselves. Researchers funded by government do not lose their own money in the risk taking. The government takes virtually all the risk, at least in the financial sense.

Participation of the public sector in the development of robotics technology has the purpose of reducing the risk for private companies that, without at least partial subsidies from government, would not have a safe enough business case to get involved.

When it comes to radical innovation in robotics, this distinction about risk taking must be kept in mind. In any case, the participation of the public sector in the development of robotics technology has the purpose of reducing the risk for private companies that, without at least partial subsidies from government, would not have a safe enough business case to get involved. This is the well-known market failure justification for public support of R&D.

In the context of public funding, the capacity of the government to take on R&D risk is not infinite. And moving from the intuition on the desirability of more scientific and technological breakthroughs to a clear understanding of how it would be recognized, what exactly is the difference between incremental research and proposals for future scientific revolutions, and what the proper incentive and support mechanisms are for it has proven difficult. It is not clear what in the research process has the attribute that is looked for. Is it the projects that are special? Are certain scientists capable in a different way? Is it the organization that can provide an unusual environment for research? Or is it groups of projects, that is, portfolios that must be assessed for their combined effects? So far, none of these questions has a clear answer.

Models, Metaphors, Analogies to Produce Meaning
One way to begin to clarify the problem is to take the models that are being used either implicitly or explicitly to suggest courses of action. We have identified at least four that have appeared repeatedly in documents and speeches about the “special” research that might have higher impact. They work as analogies for highlighting one or more important attributes or their consequences. These are the models of stock investment portfolios; biological evolution; “hot” trends; and exploration of the wild.

Stock Investment Portfolios
Many calls to support more “high risk” research that has the potential to offer “high reward” appeal to the intuition of investment risk. In this analogy, each project that is up for support is similar to a stock considered for inclusion in an investment portfolio. As in stock portfolios, a balance of risk is sought across the portfolio so that a number of safer bets compensate for riskier ones that might payoff disproportionately.

The important implication of this perspective on research support is that the unit is the project and that each one is assessed in comparison with the others in the portfolio. The overall balance of risk of a good portfolio means that projects are not selected on the same criteria. Some will be “good” low risk projects to compensate for potential losses of the high risk ones. This is clearly not current practice in public funding of research since all projects are assessed on the same criteria and only the best are funded. All projects have approximately equal risk.

Much talk about “high risk-high reward” is superficial and really suggests that most high reward projects could be identified with some special assessment mechanism without a high risk of failure. Actual failure would probably be pinned on the proposal assessment mechanism rather than the risky nature of the projects.

There are a few consequences that must be given further thought. First, what attribute of projects must be risk assessed on? Are risks across disciplines considered, for example? Or are risks across organizations or types of projects? No clear answer to this question has been offered. Second, how much failure will be tolerated? If risk is real, then high risk means high probability of failure and a high failure rate should be observed. Otherwise, the risk is not as high as it was assumed. No clear answer for this question is available either. Much talk about “high risk-high reward” is superficial and really suggests that most high reward projects could be identified with some special assessment mechanism without a high risk of failure. Actual failure would probably be pinned on the proposal assessment mechanism rather than the risky nature of the projects.

Evolution in Science
This analogy implies that new ideas in science appear much like genetic mutations do in living organisms. If our selection mechanism allows for the more radical mutations to survive, more radical change in science would probably follow. Most attention has been drawn to the selection mechanisms, assuming implicitly that the size of the pool of variation for the chance of selecting viable large jumps is sufficiently large.

The consequences of this line of thinking are also intriguing. First, the necessity of a random generation of variation in scientific ideas is not a familiar one. The image of a rational scientific method has a strong cultural effect of keeping random idea generation hidden from view, if not largely muted. In other words, attention would have to be put on the environment that allows for a large pool of very diverse scientific ideas to emerge continuously in order to have large jumps that might be selected. Secondly, a large measure of randomness in the consideration of scientific ideas would have to be acceptable to the establishment. Otherwise, just as the reduction in biological diversity is a serious problem for environmental health, there will be very little to select from to have many “viable new scientific organisms.”

Hot Trends in Science
Science is not different from other cultural phenomena in that novel trends generate “icons” and “hits” that the community rallies around. As a matter of fact, many new ideas in science and technology are followed during the first period of their public diffusion to assess whether they are substantive or only a fad. It is often heard of researchers gaining prominence in their fields that they are new “stars” or that they are “hot.” These trends and their icons attract a following, and researchers that join the movement fashion their professional identity around the key features of the emerging field.

Science is not different from other cultural phenomena in that novel trends generate “icons” and “hits” that the community rallies around.

The main lesson to drive home from the analogy is that this cultural phenomenon is an integral part of science even if it is not formalized into the assessment of proposals and projects. Individual projects are not the main focus. The “movement” is the main concern. But, how does the perception of a hot trend and its icons affect the assessment of project proposals that promise to continue in its wake? How does the system deal with the possibility of a fad that wastes energy and resources on a trend that does not pan out? A rapid recognition mechanism to tease out indicators of faddish elements seems to be necessary to address this issue. But this would be a “conservative” reaction. Should the system be more liberal in embracing hot trends to avoid killing off potential revolutionary developments? This would be another type of risk that might have to be accepted. With its acceptance comes another departure from a common view of rational science that looks askance at enthusiasm and bandwagon effects as illegitimate passions. Maybe science should have a few “moments of madness”.

Exploration of the Wild (or the Endless Frontier)
The idea that science is a sort of exploratory venture into the unknown unexplored regions of nature is an old one. It was made into influential policy discourse by a prominent scientist in the aftermath of World War II: Vannevar Bush. It assumes that what we don’t know is contiguous with what we know, and we just have to “venture out” into this unexplored territory to get to know it. The only difference between conservative, incremental science and the highly creative or transformational research is that the exploratory journey should go deeper and farther into the unknown in one single expedition.

This image suggests that the main way to achieve sizable leaps to advance science is by organizing well-equipped and supported expeditions into unexplored territory. The skeptics in policy making recognize immediately that this analogy suggests that research always requires more funding: “If you give us more money, we can go deeper into the unknown.” This was already recognized as a subtext of the original idea of the “endless frontier.” Risk is downplayed in this analogy. Only underprepared expeditions are risky. There is always unknown territory to discover. The question is rarely raised that there may be little of value in vast regions of the unknown.

A Path Forward?

A part of the public image of science as a highly rational activity is tied to the status quo of low risk incremental change in science.

It seems that before we can define a path forward it is necessary to come to terms with the full consequences of what must be achieved. First, the simple formula of “high risk-high reward” hides the complexity of the goal. There are many interrelated issues that must be addressed simultaneously if this objective will be pursued with a committed effort. A simple set of criteria for evaluating proposals will not do the job. Second, a part of the public image of science as a highly rational activity is tied to the status quo of low risk incremental change in science. Higher risk and randomness that seem to be inherent in increasing the magnitude of change in science may open the enterprise to criticism and loss of legitimacy because of an appearance of irresponsible gambling with public resources. The idea is that the payoff of a few initiatives that succeed will compensate for many that fail. Is there any publicly supported system in today’s political environment that can operate legitimately under those conditions? Not likely.

The DARPA Robotics Challenge has an interesting approach to encourage the pursuit of “leaps forward” in the development of these technologies while reducing how much risk it takes on at any point in time. It does so with two crucial measures. First, it specifies a well defined target of technological capabilities. So, in terms of the metaphors we used, it is not looking at the entire evolutionary environment of technology or at the entire frontier of technology. It limits how much randomness must be accepted and how far into the unknown it may be necessary to travel before something of value is shown. Second, by making several teams compete in stages and granting a limited amount of funding to compete in progressively more difficult challenges, it reduces potential losses and requires the return of something of value before more risks are taken. At most, no team will advance to the next stage, so what has been granted will be lost, but it cuts its future losses as soon as nobody can show progress. However, the risk of pursuing those technological objectives rather than others has not been mitigated, and the stages might cut losses too early.

For this reason, in fields where objectives cannot be specified with such clarity, program design is more difficult. But the lesson remains. As the DARPA program shows, the flexibility may have to be transferred to the very support mechanisms themselves. Attempting to derive a fixed set of procedures and criteria to identify and support something that is highly variable, with an irreducible measure of randomness, culturally unstable and insatiably adventurous may be the wrong path to go down. The criteria and support mechanisms may have to become “experimental” themselves. The future development of scientific and technological research may be pressing for some institutional change.

]]> Food delivery drones. But is it a business? https://robohub.org/food-delivery-drones-but-is-it-a-business/ Tue, 18 Jun 2013 03:26:37 +0000 http://robotlaunch.com/?p=1598

Food drone delivery ideas are taking off all over the place. But is it a business or just an advertizing stunt? Tacocopter was one of the first although still more of a theory than a practice. Stanford Robotics Club is carrying on the mission and delivering subs to students. Joining the ranks are an African beer drone, a UK pizza delivery copter and an aerial sushi tray. The OppiKoppi beer drone will be parachuting beverages to music festival attendees.

The aerial sushi tray at YO! Sushi is also handguided. It’s based on a Parrot AR.drone and as the tray sits on top of the rotors getting your food is a dicey proposition. That’s ok though, the aerial sushi tray is all about creating buzz for a new product line, not replacing wait staff. And most likely, the DomiCopter is an advertizing stunt not a serious logistics play. One of the project partners with Dominos Pizza is a creative agency, T + Biscuits.

Face it, drones are not ideal for payload delivery under most business conditions. But that doesn’t mean there aren’t some serious waves taking shape under the froth. Matternet are making a bolder claim. Creating a large net of drone transport for small and critical supplies in areas where there isn’t other infrastructure. Matternet did field trials in the Dominican Republic and Haiti in late 2012. Matternet aim to create “the next paradigm for transportation using a network of unmanned aerial vehicles”.

This sort of disruptive platform play thinking saw Google launch Project Loon in Christchurch a few days ago. Stratospheric balloon arrays bounce internet signals to places previously inaccessible or disaster affected.  Two thirds of the world’s population is still without internet access and Christchurch is a symbolic place to role out the first stages of what is a global plan.

The 2011 earthquake in New Zealand’s 2nd largest city killed 185, destroyed half the city center and 80-90% of the area’s infrastructure. 2 years on many residents remain without permanent housing and the city’s central shopping mall has only just reopened – in a shipping container structure. It’s now estimated that it will take the New Zealand economy 50-100 years to recover from the 40$B repair bill. This makes Christchurch is a good test site for a blue ocean play as residents were used to all the mod cons but are also pragmatic about making do.

Which comes back to the original question, are food delivery drones a good business proposition or just an advertizing stunt? Where the general public love frothy interest pieces about tacocopters, real robotics companies are looking for more serious business plans. But without generating the public interest and acceptance of drone delivery, then more ambitious projects like creating infrastructure in Africa and internet in the stratosphere are unlikely to get off the ground.

Without TacoCopter, Matternet might not be getting investors. Even though Forbes magazine thinks that a ‘moon shot’ like Project Loon is far closer to the real business plan for Google than the far more tangible Google Glass. But then, Google Glass is just opening the doors for next year’s uber generation of really augmented devices. Froth travels on waves.

]]> The uncanny valet (or, Notes on the design of robot psychology) https://robohub.org/the-uncanny-valet-or-notes-on-the-design-of-robot-psychology/ Sat, 04 May 2013 13:00:10 +0000 http://robohub.org/?p=13629 This article outlines the problems of today’s phone and online help systems and offers solutions to conversational systems of tomorrow. The article is about the design of hearts and minds for robots, considers the virtual voice as a legitimate robot, and takes a fast pass at the psychology of robot-human interaction.

385px-CandlestickTelephoneGal

Part One: Another robot dream girl

I’m talking with this woman on the phone, interviewing her for a job. She’s got a slight southern twang, maybe from Georgia, telling me about her past work, and as we talk I’m struck by how perfectly presentable, alert, and smooth she sounds. Her voice is soft, tonally fluid, and she articulates her words precisely, belying a geeky disposition. Seems she’ll be perfect for the rather tricky tech support position we’ve got.

“Did you enjoy the work?” I ask.

“Sure, I like talking with people. I mean, as long as they don’t get upset it’s sorta fun.”

“You were working at a bank, right?”

“Yeah,” she laughs, “There were four of us working the phones and please press seven if you’d like one of us to call you back.”

“Sorry?” I ask.

“There were four of us on the phones and please press seven if you’d like one of us to call you back.”

I realize this robot has a bug, thank her, and hang up. I won’t hire that one.

Of course I never spoke with the uncanny woman again because the above story is fiction; a conversational system like this one is something of a dream among today’s robotics designers. She might have had a bug in her bonnet, but at least she could hold a conversation and answer questions, for the most part.

A system like the above, a Natural Language Processing system, is the heart and mind of a robot.

By contrast, consider what we have today: those irritating, tyrannical little phone robots that have replaced perfectly good humans: “Please push one if you would like to make a deposit. Please push two if you would like to hear these options again.” These non-player phone-workers are so brainless, grindingly slow, and weirdly rude that I find myself repeatedly pushing keys like 0 and # just to get to a real human.

Of course, from the cold eye of the bank (or whatever corporate conglomerate that employs these “Customer Relations Management” systems) these virtual robots reduce costs by replacing a fickle, distracted, complaining, prone-to-error, temperamental, sleep-deprived and very expensive call-center human with a voice recording. It saves the company money, time, and tons of overhead on things like worker’s insurance, sick-leave, replacements and the high cost of managing people. People are hard to manage, right? Voice recordings are not. People sleep at night, right? Voice recordings do not. So virtual robots, and this particular feral species of virtual robot – the phone robot – is a means of disintermediating the person from the knowledge: it is a means of scooping the knowledge out of the head of person that was doing the job and serving it to you, the customer, couched in smooth tones with a dose of saccharin-sweet politeness.

The intention is that robotic knowledge workers can replace human knowledge workers. But robotic workers are not capable of the job: rather than collaborate, they dominate; rather than answer questions, they force responses.

Bad robot, bad!

But virtual phone robots in their present form are bad at solving problems, they don’t learn, they’re inattentive, and they can’t even understand me. I have few options on what I can do, and I get pushed into decision trees that I don’t want. I feel herded like a cow, and with that little phone-robot nipping at my telephonic heels to move me into the right slot for processing, I more often than not either hang up or return to the website.

The intention is that robotic knowledge workers can replace human knowledge workers. But robotic workers are not capable of the job: rather than collaborate, they dominate; rather than answer questions, they force responses.

Bad robot, bad!

Part Two: The problem is the rotten brains

According to research from Nuance corporation, there are three main ways to improve phone robots. The first is to allow access to a real human. Two-thirds of the customers interviewed wanted that. About half of them said the system’s logical call flow was most important. And about forty percent of them said that the speech recognition component was the most important. So access to a live human, logical flow, and recognizing what’s been said are the three main problems with phone robots today.

If two-thirds of the users of a system that is supposed to replace a human say that the system can be most improved by providing access to a human, then that system is mostly broken.

Phone robots may save money, but they suck. We expect them to act like humans because they have a human voice, but instead they act like robots. This just adds to our frustration by reminding us that we do not have what we really wanted in the first place: a real human to talk to. The sound of the language is right, the annunciation is right, but it is a facade that just highlights the problem.

These phone robots live along the upper edge of an uncanny valley (if you do not know what the uncanny valley is please stop reading this, right now, and go look it up in Wikipedia).

uncanny-valley
Masahiro Mori’s uncanny valley

Masahiro Mori pointed out many uncanny valleys, but the most famous ones are related to appearance and movement. We have yet to discover the many other uncanny valleys out there, especially as robots – and androids in particular – become progressively more human. As the systems become more capable and more interactive (phone robots today are almost always just voice recordings with a button-activated decision tree), more questions will arise. This will be exacerbated as physical robots become more human-like.

Should a robot’s voice sound like a human’s? Whose? Should a robot smell just like a person? If so, which one? What about how frequently the robot blinks? What about the tone of voice or intonation in sentences? What about social abilities?

Name a human trait and it will become an uncanny valley. These uncanny valleys will, in the coming decades, be explored and mapped. Mori just hit the tip of an icy moon that will chill, mortify, and entertain us for decades to come. We’re going to find uncanny valleys all over the pocked, lunar landscape of robotics design, and the one that will be most important to address is psychology.

Mori just hit the tip of an icy moon that will chill, mortify, and entertain us for decades to come. We’re going to find uncanny valleys all over the pocked, lunar landscape of robotics design, and the one that will be most important to address is psychology.

First we have cognitive psychology. Consider all the traits that we humans pack into language, such as logical flow, attention, learning, language, and emotion. These are classically cognitive traits, each of which the phone robots of today should pay attention to. (Of course I mean “The designers of these systems should pay attention to.”)

Second we have social psychology. Consider all the traits that we humans pack into interaction such as politeness, collaboration, competition, and understanding. These are classically social traits, which the phone robots of today, or more accurately, their designers, need to consider.

Let’s lump them all together and imprecisely call them psychology. This is the uncanny valley that most concerns me. How a robot holds a conversation, thinks, and speaks seems the most important, and urgent, of the valleys ahead.

Robot psychology is urgent because it’s happening today. You have spoken with a phone robot and chances are (76%, according to Nuance) that you were not satisfied with it. So there’s a problem here.

Robot psychology is also important because we mimic the way we are spoken to. If you don’t believe me ask yourself why you speak the language you do. Reading this article means you probably grew up in a place where people mostly spoke English. What about your accent? Again, the people around you influenced you to speak that accent. They also influenced you to use the phrases, slang, jargon, words, ideas, and therefor fundamental psychological scaffolding you use to support your notion of reality. We do as we are done unto. When you listen to someone, their mind enters yours.

If a robot treats me in a rude manner chances are good that I’ll mimic that. If I’m going to be working, playing, living, or talking on the phone with a robot, I hope it will be a healthy example of both cognitive and social psychology that I can interact with and not be polluted by. As these systems become more prevalent their influence on our own behavior will increase.

This dynamic and interaction is both important and urgent.

It is why we are becoming our robots, and why our robots are becoming us. As they speak like us we will speak like them, and vice-versa.

Please press the down key if you would like to continue reading.

Part Three: The solution is to design robot psychology

Once upon a time I had the odd luck to talk with the Geminoid F, the most realistic android on the planet (that was two years ago). It is the life work of Dr. Hiroshi Ishiguro. This uncanny experience would not have been so unsettling if Dr. Ishiguro’s doppleganger hadn’t been wading around like a zombie in the muck of the uncanny valley, but there it was and I – like the rest of the bottom feeders – was there with it. To this day, wish I had not gone.

geminoid
Hiroshi Ishiguro’s Geminoid HI-2

Dr. Ishiguro is a moribund and curious fellow, and his robot more so. The problem, as I detailed in my book, “We, Robot” was that the robot was so real that it set an expectation of being human.

Please consider this face, and the heart and mind behind it. You do not, really, want its mind to enter yours, do you?

Communication – and staying out of uncanny valleys – is all about reflection. It has to do with mutual empathy and mutual agreements of closeness. It is about familiarity.

In my last article I talked about the design of androids, wrote about the designers that made them, and noted that the function of an android is psychological closeness. Androids (uh, or rather their designers) want you to feel comfy. The core notion of an android’s design is to set you up to interact with a computer in a traditional human setting. It’s about an expectation of human familiarity and something that reflects back to you a human image of yourself.

We humans are built such that we like talking with other humans. We like what’s familiar, what is, etymologically, of our family. We like talking with a familiar face, and if the robot doesn’t look like us then we tend to freak out a bit. After all, our friends usually look like us, dress like we do, and come from the similar socio-economic background. I might not like to admit it, but that’s usually the case. And, just as we like talking with a familiar face, we like talking with a familiar psychology; the psychology of our family. Our friends think like us, act like us, and like the same things. Birds of a feather, it has been said, flock together.

This is because much of the best communication is a reflection and collaboration. We like reinforcement, familiarity, and a reflection of our presence in the world. We do not like talking with someone who herds us, we do not like being told to push buttons, and we do not like phone robots because they fly in the face of almost all human-based computer-human interaction (CHI) theories (such as the android). Phone robots are a good example of a bad psychology.

As kids grow up interfacing with robotic personalities, and virtual robots, they will be influenced by, adopt the behaviors of, and intellectually be guided by these personalities. It is our responsibility to offer something better than junk food for the psyche.

If you were still a child, would you rather grow up talking daily with Ronald McDonald, Shrek, or Kung-Fu Panda? C’mon, choose one. Now what if a million children were to be forced, by your decision, to do the same?

As designers of robots we are faced with the same type of problem. As kids grow up interfacing with robotic personalities, and virtual robots, they will be influenced by, adopt the behaviors of, and intellectually be guided by these personalities. It is our responsibility to offer something better than junk food for the psyche.

Siri is about the best form of phone robot we have on the public market today. From my interactions with her she’s a little vapid, slightly sardonic, sometimes helpful, but mostly apologetic. When she makes a joke, which is sadly rare, or when she directs us toward (or steers us away from) some product, she is affecting our psychology.

There are some two hundred and fifty million versions of Siri out there. Let’s conservatively call that two million active versions of Siri functioning today. And I have no clue how many of Siri’s phone robot ancestors are out there, but we can safely say that the psychology of these systems is affecting the psychology of millions of people.

Ahead of us, the design steps for implementing android psychology are not evident, but we can follow in the footsteps of other people and extrapolate from character-based properties already in the wilds and public media of today.

If we look at how famous branded characters like Ronald McDonald or Kung-Fu Panda have been developed, marketed and sold, we can make out the guideposts for cultural preference and gender. If we look at archetypes like Darth Vader or Voldemort we can see how to design the bad guys. Superman and Gandalf give us clues on the good guys. Homer Simpson, Indiana Jones, Buzz Lightyear, George Bush, and Jacques Chirac are all personalities that have left an indelible public impression that we can model and use in designing psychology.

Presentation is important, too. Of course the system has to speak your language, and therefor regional preferences will guide the design of these robots. Dialects and accents count today. We can already see how Siri has been implemented with various accents such as American vs British. Gender will also continue to be important (70% of the automated voice systems in Europe are female, and in North America, 30%). After presentation we have the personality, the archetypes, and what the character knows.

People who write character portraits, who ‘get’ dialogue, and who are socially alert duplicators of human interaction; people who make movies, people who conduct interviews, and, unfortunately, people who make advertisements: these are the people who can design systems that avoid psychology’s uncanny valley.

Authors are the people that know about this kind of design. People who write character portraits, who ‘get’ dialogue, and who are socially alert duplicators of human interaction; people who make movies, people who conduct interviews, and, unfortunately, people who make advertisements: these are the people who can design systems that avoid psychology’s uncanny valley.

If we look at avatars in social media, and if we consider the movie industry we can see some other guideposts for the future of the psychology of android design. The most important is interactivity and reflection of shared interests. This takes us back to Nuance’s research.

Access to a real human will become increasingly rare. Regardless of what consumers want, the whole idea of a phone robot is 24 / 7 availability at a decreased price, and this is some thing a human cannot do. Next on Nuance’s list was the logical flow of the call and the speech recognition ability. These two important interactive components will surely get better in parallel as understanding one helps to improve the other.

Most of it will happen via mobile. Consumers in the United States – which roughly maps to much of the rest of the industrialized world – prefer to resolve their customers service issues using the telephone (90%), face to face (75%), company website or email (67%), online chat (47%), text message (22%), social networking site (22%). So it seems that while these conversational systems will live mostly in the phone, there will be a place for them in other media as well. They will live in the wires and waves that surround us, always available, often improving.

This is how the system becomes, finally, conversational. The irritating little herd-dog phone robot will evolve into something closer to a Geminoid. One day you will think you’re talking with a human on the phone, and then you will find yourself in a new uncanny valley. It won’t happen for at least another five years, but we’ll stumble into it eventually. The psychology of that robot will be a thing that touches you and guides you in ways that corporations like Google (who have powerful NLP tools today) Facebook (who released an NLP search system yesterday), can only begin to hint at.

One day you will think you’re talking with a human on the phone, and then you will find yourself in a new uncanny valley.

They will be able to speak to you, sell to you, buy from you, offer medical guidance, marital guidance, and talk with you about your depression, your homework, your spouse, and your boss. Semantic Analytics, Social Analytics, and NLP will accurately measure your responses and the system will reply with a warm candor. Young men will ask about sex, old women will ask about hysterectomies, and all the other most awkward, personal, relevant, and important questions of life will be whispered, spoken, laughed, and sobbed into these many kinds of virtual androids that will gradually collect, like little honey bees collecting pollen and taking it back to the nest, your very own fears, loves, desires, hates and general, if I may say, psychology. The phone robots of tomorrow will become tiny transporters of emotion and personality. You will leave a voice recording and it will be taken, before you have hung up the phone, to some hive-mind.

The question we may ask these little bees is, “Where have you taken my personality?”

]]> First controlled flight of an insect-sized robot https://robohub.org/controlling-insect-sized-flying-robots/ Thu, 02 May 2013 18:07:03 +0000 http://robohub.org/?p=13525 Researchers from the Wyss Institute and the School of Engineering and Applied Sciences at Harvard have developed a millimeter-scaled insect robot that can autonomously control its flight. Their findings were published in the prestigious journal Science. The amazing high-speed video below shows the robot taking off, hovering in place and steering left and right on demand. Controlling such small flyers has been impossible so far because of challenges in fabricating tiny actuated systems, and the chaotic movement of small flapping-wing robots. You’ve seen a fly move around your living room, doesn’t seem easy to control right?

Fabrication
Rob Wood’s Microrobotics Laboratory at Harvard has a long history of developing small flying robots using a clever fabrication technique intended for easy and cheap mass-production. The method described in the video below is inspired from pop-up books that you can cut and fold to make articulated three-dimensional structures. By sandwiching flexible materials between laser-cut carbon fiber layers, they are able to rapidly fabricate robot skeletons that can be folded and locked into shape by soldering dedicated attachment points. The skeleton of the controllable robot is augmented with two piezoelectric muscles to differentially drive the wings. These muscles are made of special materials that contract and expand when electricity is applied to them. Power is provided by an external cable.

Control
To control its position, the robot needs to change how it beats its wings. Beating the left wing harder than the right wing will produce a torque on the robot’s body. The robot is inherently unstable and a very fast and precise control is needed to keep it aloft. It would be nearly impossible for a human to manually control the wing motions or preset them based on flight dynamics. Instead, the robot continuously compares its actual position to its intended position and corrects the wing motions accordingly (closed-loop control). The position of the robot is provided by an external camera system which tracks markers on the body of the robot. Using this technique the robot is able to stay in place, and move to a designated location. This is a remarkable feat for such a small and highly dynamic flyer. Previous version of the robot had been seen taking off and controlling altitude:

Applications
This work is part of the Robobees project which brings together 10 laboratories to fabricate, control and power bio-inspired roboinsects that can function in swarms. Applications, although still a long way in the future, include autonomously pollinating a field of crops, search and rescue, hazardous environment exploration, military surveillance, high resolution weather and climate mapping and traffic monitoring.

Listen to our podcast interviews with the authors of the paper including Kevin Ma, Pakpong Chirarattananon and Sawyer Fuller.

Links:

Photo Credits: Kevin Ma and Pakpong Chirarattananon, Harvard University.

]]> Intuitive Surgical, a manufacturer with almost no tangible assets? https://robohub.org/intuitive-surgical-a-manufacturer-with-almost-no-tangible-assets/ Wed, 01 May 2013 09:15:59 +0000 http://robocosmist.com/?p=341 IntuitiveIntuitive Surgical (NASDAQ:ISRG) is a prime example of how robotics is similar to other IP intensive industries like software, biotech, and entertainment.

In December my colleagues and I produced a valuation of Intuitive Surgical.  Below is a representation of our model of the asset structure of Intuitive Surgical in our forecast.  Whether you agree with our estimate of a 31% return on economic assets or not (though the stock market roughly seems to), this chart is very instructive to look at what the economic assets of a successful robotics firm are.

And hey, guess what?!  Intuitive looks more like a software company than a traditional manufacturer.  Strike another blow for the case that robotics companies – at least successful ones – are capital efficient!

Moreover, if I was critiquing the model in the valuation I would say that we hadn’t adequately valued the intangible assets of Intuitive Surgical.  The intangible assets of the firm probably have a market value of 2-8 times what we estimate.  Even with our conservatism, look at what you’re buying into when you buy a share of Intuitive:  A $2Bn stack of cash, a multi-billion dollar IP portfolio, and a smallish medical device manufacturing company.

Assumptions of FCF forecast through an economic view

[How to read this chart:  Black is our estimate of “R&D assets” in $K so starting balance is just shy of $2Bn.  Red is GAAP non-financial assets, otherwise know as real stuff, like buildings, inventory, and accounts receivable.  Grey is our estimate of financial assets with the current dividend policy–this model posits that Intuitive will be sitting on $4Bn in cash or the like in 2016 and an IP portfolio equally as large and valuable.Return on economic assets was estimated using our income forecast over capitalized R&D spending in the R&D account plus assets less cash and securities.  The model has a depreciation factor for R&D each year to account for obsolescence and expiration.  We went back several years to estimate an appropriate R&D account starting balance for the projection.]

The stock market assigns a $20Bn valuation to Intuitive.  It recognizes that Intuitive’s control of  intangible assets is very valuable. The graph of the model here only scratches the surface of intangible assets.  We assumed that the only off balance sheet economic asset was an R&D account.  Clearly, this is not the case as Intuitive Surgical also has unique and valuable organizational processes, sales relationships, and employment relationships with talented employees but those are much harder to find information about in SEC disclosures.    Similarly, we also marked R&D at cost–with a portfolio as valuable as Intuitive’s the market is probably going to value the R&D output at more than Intuitive paid to develop the R&D assets.

Even with all this, Intuitive Surgical looks like lean, mean, capital efficient, IP intensive, knowledge economy company.  Can anyone tell me why we let people talk about robotics like it is capital intensive?

I’d like to gratefully acknowledge my co-authors of this report who have given me permission to publish it: Avinash Belur, Naohiro Furuta, Masayuki Minato, Kohei Mutoh, & Dashampreet Sidhu.  Analysis available by request.

]]> New funding sources for robotics https://robohub.org/new-funding-sources-for-robotics-2/ Fri, 19 Apr 2013 22:09:19 +0000 http://robotlaunch.com/?p=1468 Writing-a-Check-US

New funding sources for robotics seem to be popping up everywhere. Today’s launch of Genesis Angels, the new $100 million VC fund, which is based in Israel but operating globally, is hot on the heels of the EC announcing that the next tranche of robotics investment is likely to be more than €600M ($774M).

At RoboBusiness Europe, Libor Kral, Head of the DG Connect Robotics Unit in the European Commission, described this as the largest civilian investment in robotics R&D in the world, albeit closely matched by Korea. Kral added that not only was the EC budget for robotics increasing in the Horizon 2020 period (2014-2020) but that the new PPP or Private-Public Partnership program would contribute a hefty 70% plus 100% of overhead of each partnership.

There are rumours of new funds launching in Hong Kong, China and India. With Genesis Angels based out of Israel and Grishin Robotics originally based in Russia, you have to ask what is the US doing. As defence funding for robotics slows, it seems that the US is relying on partnerships, providing the ‘talent’ and pool that everyone else wants to play in. Risky strategy if you can’t keep IP or ‘talent’ and don’t maintain a manufacturing infrastructure.

Genesis Angel’s announcement comes shortly after another Israel/US partnership, the Technion-Israel Institute of Technology is partnering with Cornell in creating a graduate school for applied science and engineering (and entrepreneurship) in New York City. Already getting underway in temporary headquarters (at Google?) in Chelsea, the Institute will open on Roosevelt Island in 2017. The goal is to create another startup ecosystem, like Stanford/Silicon Valley and Route 128/MIT, where academics and industry can cross pollinate.

The accelerator model is taking many shapes. Grishin Robotics has just invested in Bolt, the Boston based hardware accelerator, and Lemnos Labs in San Francisco is picking up more and more venture backing. Alongside virtual accelerators, ranging from crowdfunging, hackerspaces and Robot Launchpad, which is facilitating advisor, mentor and peer startup acceleration, you have the more traditional incubators in large strategic players.

Companies like Samsung, Qualcomm, Bosch, GE, Intel etc. are more willing to take the long view when it comes to robotics and to invest in a company that may not deliver returns for 4 or 5 years. Of course, you have to have a big vision, that unlocks large revenue areas and that aligns with the company’s existing platforms, whether it’s chips, phones or appliances.

It also helps if you call it “smart connected devices” rather than robotics.

]]> The great equalizer: How robotics frees manufacturers from consolidating in low-wage nations https://robohub.org/the-great-equalizer-how-robotics-frees-manufacturers-from-consolidating-in-low-wage-nations/ Mon, 15 Apr 2013 20:36:49 +0000 http://robohub.org/?p=12045 http://www.dreamstime.com/royalty-free-stock-images-worker-climbing-upwards-ladder-image19289039These days it is hard to read an article about the future of robots that does not include a reference to jobs. As a pure roboticist I object to the constant connection between the two, but as a concerned citizen I think it is a very worthwhile discussion.  Since the year 2000, the US has lost more than 6 million manufacturing jobs — that is more than 1/3 of all direct manufacturing jobs in the US and the fastest drop in a single decade on record.

This post is part of Robohub’s Jobs Focus. 

During that same period the US trade deficit with China grew from $100 billion to over $270 billion as many companies chose to outsource their manufacturing to low wage nations such as China where labor costs were 1/10 that of the US. In some cases, entire industries such as mobile phones, consumer electronics, and computers were outsourced to contract manufacturers and moved offshore, creating as many as 25 jobs in these low wage countries for every job created in the US.

I can speak to this firsthand. In the late 1990s, the largest market of my company at the time, Adept Technology, was automating mobile phone and computer factories in the US. As the pace of innovation accelerated and models proliferated, our robots did not have the flexibility or intelligence to keep up with the rate of change. Additionally, these manufacturers were facing increasing pressure from low cost competition that left them with no choice but to look to low cost labor markets to produce their products, first in Mexico then Brazil then China. At the time there were less than 350 million mobile phones produced in the world with the majority produced in the US and Europe. Today there are over one billion mobile phones produced a year and not a single one is produced in the US even though the US market consumes over 200 million units.

Wages in countries such as China have been increasing at a rate close to 20% per year. Add to this increasing transportation costs and the negative impact of a supply chain that stretches around the world, and more and more companies are rethinking their manufacturing strategies.

The robots of today are smarter, more agile and less expensive than their ancestors of the 1990s. They integrate sophisticated sensing technology that allows them to adjust to changes in their environment and in the products that they produce. Some are even leaving their traditional isolated cages to work in tandem with people, making the combination even more productive. All of this makes the new generation of robots far more adaptable to today’s manufacturing world, and a much better technical fit than they were a decade ago.

At the same time wages in countries such as China have been increasing at a rate close to 20% per year. Add to this increasing transportation costs and the negative impact of a supply chain that stretches around the world, and more and more companies are rethinking their manufacturing strategies.

Back in the US, a new manufacturing equation is emerging, called Smart Manufacturing. Smart Manufacturing is the production of traditional products using advanced technologies such as computers, robots, lasers and 3D printers. Using these tools dramatically improves productivity and lowers costs as compared to traditional manufacturing approaches, while maintaining flexibility and improving time to market. Smart Manufacturing is making US goods competitive with products produced anywhere in the world, even in the lowest wage nations, and the result is that the decades old trend of offshoring is being replaced with a new trend, re-shoring.

Re-shoring is not a concept founded in patriotism or built on motherhood and apple pie ideals. Re-shoring is a powerful business tool founded on solid economics. For centuries manufacturers have employed a strategy of chasing cheap labor around the world, moving from one low wage nation to the next as costs rose in the developed world. While this has resulted in lowering direct product costs, it has also added significant expense related to managing a supply chain that extends around the world. It has also slowed the rate at which new products can be ramped with product development and production separated by thousands of miles and several time zones.

Automation becomes the great equalizer, allowing products to be produced anywhere in the world for the same cost, and freeing manufacturers from consolidating manufacturing in low wage nations.

Smart Manufacturing has the potential to bring an end to manufacturer’s centuries old dependence on low cost labor, and open up a new strategy: developing and producing goods in the markets where they are consumed. Localizing and integrating marketing, product development, manufacturing and distribution is the ultimate operational model. It allows for more targeted products while eliminating the costs of managing global supply chains and disjointed product introduction processes. Automation becomes the great equalizer, allowing products to be produced anywhere in the world for the same cost, and freeing manufacturers from consolidating manufacturing in low wage nations.

Smart Manufacturing and robotics have one other benefit, the creation of high value, high wage jobs. In a 2009 study commissioned by Intel, ECONorthwest found a strong connection between Smart Manufacturing and the creation of jobs. At Intel’s highly automated factory in Hillsboro, Oregon, they employ close to 16,000 people. ECONorthwest found that for every 10 jobs created within the factory there were an additional 31 jobs created in the local community to support the factory. These indirect jobs were anywhere from 40-60% above the statewide average income for Oregon.

Other studies have pegged the jobs multiplier of Smart Manufacturing at between 2 and 10 times the number of direct manufacturing jobs. In a separate study commissioned by the International Federation of Robots (IFR), Metra Martech estimated that the robot industry has created between 8 million and 10 million jobs throughout its history, and forecast that over the next decade the industry will create between 2 and 3.5 million direct jobs and almost as many indirect jobs. The study went on to show a tight correlation between the increased use of robots and a reduction in the overall unemployment rate. They also point out that emerging economies are amongst the fastest growing users of robots. According to the IFR, Korea and China are now the top consuming nations of industrial robots, surpassing the US, Germany and Japan. China’s ascent to the top has been nothing short of amazing.

With all manufacturers having access to the same set of tools, the winners — and ultimately the jobs — will go to those producers who can most effectively and creatively leverage these tools to their competitive advantage.

From 2008 to 2012, China’s use of robots has increased over 300% in a nation where average manufacturing wages still remain between $1.00 and $1.50 per hour.

Looking to the future, the manufacturing world is becoming flat. Robotics and other advanced technologies are leveling the playing field and eliminating the advantage and influence that low cost labor has had on manufacturing. With all manufacturers having access to the same set of tools, the winners — and ultimately the jobs — will go to those producers who can most effectively and creatively leverage these tools to their competitive advantage. In the end robots are about jobs.

See all the posts in Robohub’s Jobs Focus.

]]> Do industrial robots really have a positive impact on employment? https://robohub.org/do-industrial-robots-really-have-a-positive-impact-on-employment/ Fri, 12 Apr 2013 10:08:00 +0000 http://robohub.org/?guid=e21077610f555ffe57a1067b8f74a90d Do industrial robots really have a positive impact on employment? Of course they do and there are over 50 years of data proving that to be the case.

Source: KUKA, 1983

This post is part of Robohub’s Jobs Focus.

There are at least 350,000 people directly employed by and in the industrial robotics industry. However, industrial robots are not made in the US; they are made in Japan, Korea and Europe, consequently more than half of those 350,000 jobs are offshore.

There are ancillary providers of components, software and other services for robots and installations but these jobs are hard to quantify. In general manufacturing there are known multipliers, but in robotics there is no known formula to extrapolate just how many jobs in these ancillary businesses are attributable to robotics.

The deployment of robots into a factory situation involves even more complex computations and assumptions — which is why the International Federation of Robotics (IFR) commissioned the UK-based research firm Metra Martech in 2009 to figure out how to compute the effect on jobs of the deployment of industrial robots.

The IFR annually collects, reports on shipments and sales, computes robots at work (versus idle or abandoned), and makes five-year sales and deployment projections of industrial and service robots worldwide. They summarize their data into two comprehensive books: World Robotics Industrial Robots 2012 and World Robotics Service Robots 2012 reflecting data for 2011. The 2013 books covering 2012 data will be available later this year.

Original Metra Martech Report
Metra Martech’s original report was delivered in February, 2011. It concluded that they had determined a job-creation ratio of 3.6 jobs for every robot deployed, and that with more robots, fewer jobs are lost. That’s why Germany, with hourly rates almost 50% higher than in the US, has remained globally competitive: they have twice as many robots per 10,000 workers as do Americans.

Their research claimed that 3 million jobs were created in factories where accuracy and consistency could not be achieved without robots, and another 300,000 jobs were created where poor working conditions were overcome by the use of robots.

Much has happened during and since 2008, the unfortunate base year for that report, which is why the IFR re-commissioned Metra Martech in 2012 to update their findings for more recent data.

Robots, Robotics and the 2013 Metra Martech Report
Early this year CBS News 60 Minutes aired a segment asking but not answering the question: Are robots hurting job growth? Much criticism from robotics industry professionals followed, including an International Federation of Robotics CEO press conference where an abstract of the updated version of the Metra Martech report was presented. Under the banner Robots Create Jobs, they cited the following reasons why they opposed the CBS 60 Minutes piece, and why robotics really does create jobs:

  1. Robotics is a critical factor in rebalancing world manufacturing economies because it reduces the threat from low-cost-of-labor countries.
  2. Although the recession temporarily drew attention away from unsatisfactory work conditions, enhancing work conditions is a driving force for using robots. Replacing dull, dirty and dangerous tasks with robots is the low-hanging fruit of robotics.
  3. There are many new industries where only robots can produce precision and consistency standards at an affordable cost as demanded by the competitive global economy.
  4. Job creation by robotics, between 2012 and 2020, will be 2.15 million to 3.5 million:
    1. This is the sum of new robotic products (.80 to 1.4 million),
    2. Current industry expansion (.45 to .70 million),
    3. Downstream jobs (.9 to 1.4 million).

It is the #4 items that are of interest to most readers and particularly Americans who, by and large, perceive a different set of “facts.”

US Facts
Americans are living through a confusing economic period. They see things in “home economics” terms: job creation is up but nowhere near covering population growth; unemployment has dipped slightly but hides the broader unemployment rate; stock markets are reaching new highs; and house prices and disposable income are both rising. Yet public sentiment is focused on jobs and job creation, and for good reason: millions have fallen off the employment grid and the employment figures don’t reflect the true numbers nor the hardships this has caused.

Hence the concern about jobs being taken away: about being left behind, about off-shoring, robots, relocation, acquisition, closures, consolidation or otherwise.

Reviewing the Revised Metra Martech Findings
John Dulchinos, former CEO of Adept Technologies, said at the IFR press conference when the revised report was launched (and reiterated in a recent email):

“Assessing job creation driven by robots is not an exact science and requires piecing together data based on a consistent set of assumptions. Over the past decade the US has lost 6 million manufacturing jobs. This is not the result of robots. It is largely due to offshoring by contract manufacturing companies who have taken a large slice of the manufacturing pie. What is not captured in all the government data is the number of downstream non-manufacturing jobs that disappear when a factory is no longer competitive with low-wage nations and is shut down.”

Part of the complexity of the Metra Martech report is that it must take into account the economic, population, industry variables and downstream multipliers unique to each of the main countries in the study. For example, manufacturing accounts for only 11% of employment in the US, but 24% in Germany and 27% in South Korea.

I asked Metra Martech’s Peter Gorle whether they relied solely on the IFR data or if they supplemented it with other data. Gorle responded by saying:

“The basic data on robot use came from IFR, but we at Metra Martech carried out considerably more research into the employment and unemployment figures and the growth rates in the main user sectors. This provided the fundamental position against which the use of robots was analysed.”

Then I asked how they attempted to report the number of people employed through the use of robots net of the number of workers displaced by robot deployment:

“While it is clear that robots can and do replace unskilled, dirty or dangerous occupations, there are two main drivers for added employment resulting from the use of robots. The first is the need to use the repeatability and accuracy which they can provide. Note that this is evident throughout the world no matter what local labour costs.

The second driver is the need for industrialised countries to overcome the benefit which the low wage countries offer in production. This tends to result in the loss of some jobs, but it enables companies to retain employees in parts of industry which would otherwise be lost to lower cost areas.”

Since most of the conclusions reached by Metra Martech are global in scope and conjectural in nature, it is important to examine the extent of their assumptions. One can see their process at work in the discussion of the chart shown below for the Electrical and Electronics sector:

“Hypothetically, if robots were removed, some products would become unviable, while others would move to low cost countries and be assembled by hand. Many tasks carried out in this sector [Electrical & Electronics], particularly by SMEs (SME refers to Small and Medium-sized Enterprises, i.e., small shops, factories with less than 100 employees), are still done by hand. The loss of jobs would thus be greater in the developed countries, and the gain in jobs greater in the lower cost industrializing countries. There are no statistics on this type of dependence and we have made broad estimates based on the above and the number of robots in use in the country in the sector.”

Bottom Line
Many of the assumptions in the updated Metra Martech report are sufficiently transparent such that one can readily accept their position and therefore their conclusions. Thus their projection of 2.15 to 3.5 million new jobs from 2012 to 2020 appears correct:


Source: Metra Martech, 2013, “Potential NEW activity jobs because of robotics 2012 to 2020, pg 10.

Although they don’t really address statistically what happens to displaced workers, they do say that they are often kept, retrained, and utilized in upscale jobs created as a result of the efficiencies reaped from the new robots. But they don’t provide any numbers or formulas to support this thesis.

The updated report moves away from a strict focus on industrial robotics, suggesting that SMEs and healthcare/medical robotics will show broad job growth in the next few years. Thus 21% to 24% of Metra Martech’s potential new jobs projections include service robots marketing and manufacturing. They rationalize this contradiction by saying that service robots can be likened to consumer products and therefore will need new manufacturing requiring robotics.

Two areas stand out as seriously unsubstantiated in the report:

  1. Offsets for jobs displaced
  2. Downstream jobs

Offsets for jobs displaced, although frequently mentioned in the report, could not be measured with the data available.

Downstream jobs include (1) employment in ancillary businesses that are needed to support robot construction, robotic installations, robot control, and (2) jobs created by adding well-paid robotics industry employees into the community. Metra Martech suggests a third group to be (3) other types of jobs in support of the use of the products manufactured where the products can only be produced by robots.

In manufacturing, an industry statistic to reflect downstream ancillary jobs is 1.3 jobs in a support business for every new job in manufacturing. This multiplier effect is frequently validated in general manufacturing. But applying that figure in the robotics industry is difficult because it relates to jobs, whereas the Metra Martech projections apply to robots deployed predicated on the data provided by the IFR. Nowhere in their report is there a correlation of the number of jobs replaced by a single robot; thus a true downstream jobs figure cannot be tabulated.

Consequently, I find the Metra Martech report somewhat flawed but not for lack of trying. It was a noble effort and remains a valuable tool showing that the robotics industry really does create jobs.

See all the posts in Robohub’s Jobs Focus.

If you liked this article, you may also be interested in:

See all the latest robotics news on Robohub, or sign up for our weekly newsletter.

]]> Jobs for roboticists https://robohub.org/jobs-for-roboticists/ Thu, 11 Apr 2013 08:00:45 +0000 http://robohub.org/?p=12501 During my Master studies in computer science, I had the opportunity to spend a year at Carnegie Mellon University in the USA. Most of my classes there were about robotics, I got to participate in Robocup and see QRIO dance. That year abroad is what got me into robotics. My logic at the time was that I would never find a job in robotics, it was too futuristic. My only option was to dive into the world of research and academia.

This post is part of Robohub’s Jobs Focus.

jobs

Fast forward to today, roboticists now have a variety of career options to choose from. Of those who did their PhD with me, one third ended up in a startup (I’m looking at you Sensefly), one third went to industry or consulting at great companies like Sony or Google, and one third stayed in academia. The number of open positions posted to the main robotics mailing-list for academia has quadrupled in the last 7 years. Wanted Analytics, an expert in talent placement, reports similar increases in hiring demands for artificial intelligence and robotics. Career Builder has over 1000 robotics related jobs posted in the past 30 days. More than 950 robotics-related jobs paying more than 70’000$ were posted on Indeed in the past 15 days alone. O*net, expert in US occupational information, reports Robotics Engineers as having a “bright outlook”.

1.24.13-Hiring-Chart-for-AI-and-Robotics

Hiring Demand for Artificial Intelligence and Robotics Skills. Source: Wanted Analytics.

 

There are a variety of wonderful articles in Robohub’s Jobs Focus analyzing the impact of robotics on the job market or the economy. Few of the articles have considered the inevitable increase in jobs for roboticists. Today you can apply to jobs at Aldebaran, Rethink Robotics, Intuitive Surgical, Kiva Systems, Boston Dynamics or start your own company like several of my colleagues. These possibilities didn’t exist beyond the manufacturing industry just a few years ago. So if you’re still looking at Academia like me, or interested in industry or startups, robotics will continue to develop into a new and emerging job market.

Are you a roboticist? Where do you work? Was it hard for you to find a job? Let us know in the comments section below.

]]> Robohub focus: Robots and warfare https://robohub.org/robohub-focus-robots-and-warfare/ Thu, 14 Mar 2013 00:12:31 +0000 http://robohub.org/?p=10073 BigDog
For the rest of this week, Robohub will have a special focus on the use of robots in warfare. 

All kinds of robots are being developed for strategic defence and military action (in space, in the air, underwater and on the ground). At Robohub we’ve had the opportunity to cover a wide range of them, including exoskeletons, transport mules such as Big Dog and DARPA’s LS3, and video reconnaissance systems such as iRobot’s Packbot. But by far the most talked about military robotics technology is the UAV.

Whether you call them drones or unmanned aerial vehicles, almost everyone‘s got one. According to the US Government Accountability Office, seventy-six countries now have drone technology, though not all are armed.

Since we launched this past August, Robohub has scooped a number of stories on who’s got UAV technology and who’s using it for military purposes (including the US, China, Australia, Europe, and others). And while there are many kinds of stakeholders with an interest in the technology (including police, surveillance, film, the hobby market, and search and rescue), military organizations are clearly the big spenders driving the market.

This lop-sided military interest in UAVs has helped to fuel a two-pronged debate:

  1. Those who are against the use of war zone robots, arguing that they are unregulated  and that little attention is being paid to moral implications and international law (see Noel Sharkey’s Robot wars are a reality, and PBS Nova’s Rise of the Drones).
  2. And those who are for using UAVs in war zones, arguing that drones kill fewer civilians than other modes of warfare (see Ron Arkin’s Governing Lethal Behavior in Autonomous Robots, and the New York Times article The moral case for drones).

On Robohub, Sabine Hauert’s insightful Robots and Ethics podcast elegantly captured both sides of this debate, and Mike Hamer also explored the ethics in his opinion piece Losing Humanity.

But there is also a third perspective out there: that UAVs and drones can be a whole lot more than killers and spies (see for example, Ryan Calo in Bad laws would hurt good drones, Time Magazine’s Why we shouldn’t fear personal drones and What happens when drones return to America, Motherboard TV’s Drone On, and Journey Man Pictures’ Rise of the Machines). Robohub contributor Eric Wind has also pursued this perspective.

As part of our focus this week:

  • In response to the Robotics by Invitation question “How will robots shape the future of warfare?“, Raffaello D’Andrea talks about the ethics from a researcher’s perspective, Daniel H. Wilson discusses the widespread use of drone technology, and Mark Tilden questions the morality of blame-free robotic conflict. Read more.
  • Former US Army Officer and unmanned systems expert Robert Morris argues that ethical debate is a red herring, and that the real issue is about policy and who is the true leader in battle. Read more.
  • Chris Mailey from the Association for Unmanned Vehicle Systems International (AUVSI) urges us to look at military history to better understand how robots will shape the future of warfare. Read more.
  • Longtime Robohub contributor and UAV enthusiast Ioannis Erripis gives us an overview of the kinds of robots that are being used in the military today. Read more.
  • And finally, Jim Haas, creator of Nate the Robot, will be issuing war-themed comics for the rest of this month. Look for them in the sidebar!

We hope that you find this focus series provocative, and we look forward to your comments.

]]> The death of search (or, My dysfunctional relationship with Siri) https://robohub.org/the-death-of-search-or-my-dysfunctional-relationship-with-siri/ https://robohub.org/the-death-of-search-or-my-dysfunctional-relationship-with-siri/#respond Wed, 06 Mar 2013 20:16:54 +0000 http://robohub.org/?p=10134 This article looks at the arrival of systems such as Siri, Google Now, and Watson and claims that these systems are the search engines of the next decade because they mine intimate data.  Since they integrate search they will replace search, as well as a host of other interface and information retrieval functions.  This offers an outline to both the personal benefits and privacy risks.

RobotCallGirl

I got a new iPhone about a year ago, the one with Siri on it.  Our relationship started on a very mundane level. Siri asked me my name and, since I already knew hers, I asked what she could do, and how. It was a bit like talking with a call girl. The basics were covered and soon we got to the intimate stuff. I don’t think Siri loved me. But now that Siri and I have gone our separate ways, I can say we both had ulterior motives. We sort of used one another, now that I look back on it. It was a relationship of an intimate nature. It’s what happens when you have a relationship with a robotic call girl.

The reason I got the phone to begin with was so I could do a little reverse engineering. It was research for work. Siri and I took a few days to get to know one another, but after that honey-moon period it was clear that Siri (or Apple) was using a pre-scripted Natural Language Processing (NLP) approach that basically did some lexical parsing, looked up a match, prepared a response, and kicked it back down the pipeline.  My reverse engineering wasn’t too profound. I wanted to see how many recursive answers were built in to things like “Open the pod bay doors” or “Do you love me?”  I also wanted to see how errors were handled, and what kind of redundancy checking was happening for sentences like “Where can I buy a burger and fried?” Siri did pretty well, all things considered, but my expectations were low, and like many other intimate relationships, I somehow knew from the start it wouldn’t last.

Relationships are generally symmetric, which meant that — like any dating service — someone other than Siri was getting money for my time.

Now, while I was using Siri, she was also using me. Relationships are generally symmetric, as Gregory Bateson tells us [1], which meant that, like any dating service, someone other than Siri was getting money for my time. But I didn’t care if Steve Jobs was Siri’s pimp. By last summer I was having flirtatious fun and the consequences of the relationship were too far down the road to look dangerous. I asked Siri to take dictation, to wake me up in the mornings, to remind me of important events, to remember phone numbers, remember names, tell me jokes, confirm thoughts, quote a number or two. All intimate things and, as I look back on it, the kind of stuff that I’ve trusted my wife with.

My relationship with Siri started to cool off a bit by September. Though my wife knew about Siri, I hadn’t yet introduced them.  So when I asked Siri to call my wife I, expected Siri to ask me for her name or number. Instead Siri just dialled. I was surprised. How did Siri know my wife’s number? (Answer: It was in my Contacts card.) What else did Siri know? (Answer: A lot.) And more importantly, what was Siri passing back to the Apple hive? (Answer: Everything.)

Siri knew a lot more about me than I knew about her.

Siri and Apple now have a great deal of data about my household: a quick comparison of my contacts/likes database with my wife’s will give you a pretty good feel for the stuff we will buy.  Burgers and fries, of course. And a few multi-player games.  Some lingerie. A sex toy or two.  Not a big deal, but not the kind of stuff I want everyone to know about.  Not even my mom gets access to that data.

But Siri did. And Siri knew a lot more about me than I knew about her.

Then, around the middle of October, three things happened to Siri that marked the beginning of our eventual estrangement.

First was William Stasior. Apple picked Stasior up to run its Siri unit after he had been successfully heading Amazon’s search and advertising unit “A9.”  Prior to that the MIT PhD had served time at Oracle, NetCentives, and AltaVista.  To say the guy knows search is like saying the pope knows the church, but why would Apple pick a search guru for an AI system like Siri?  What was the link between search and NLP?  I asked Siri this very question, but she was mum and feigned ignorance.  I didn’t push her on who this guy Stasior was and Siri didn’t ask me more about my wife.

Up until that same week iPhone users had frolicked in what would now be a rather unusual environment in which advertisers were unable to track them.  Advertisers would not, for example, know that my wife and I like video games, sex toys, and pizza (no, not together, thank you).  But with the release of iOS6, Apple flipped ad tracking on and I quickly flipped it back off.

What was the link between search and NLP? I asked Siri this very question, but she was mum and feigned ignorance.

It was easy to find in the interface menus (under Settings > General > then About or Advertising, depending on the version you’ve got), but you had to know to look.  Apple minimized the fanfare around this new feature and Siri made no mention of it to me.  I wasn’t comfortable with Siri selling to others what I had said to her in more comfortable times.

And then, as if things weren’t rough enough, the European Union (where I happened to be lecturing that week), demanded that both Facebook and Google change how they handle personal information to avoid “high risks to the privacy of its users.”  Twenty-four of the European Union’s twenty-seven regulators signed a letter that, after a nine-month investigation into the data collection practices, demanded some answers and tightened regulation.  After Google and Facebook, Apple was third on their hit list.

That was the end. I couldn’t trust Siri after that.  She knew more about me (and my wife) than I knew about her or Steve Jobs, and any time information flows one way, any time a relationship isn’t symmetric, the balance of power can be dangerous.

Big Blue decided that Siri knew a little too much about their employees, and at IBM’s Armonk New York research center, iPhones weren’t even allowed in the building.

Sometime around then I learned that Oracle’s “bring your own device to work” policy had been revoked.  Big Blue decided that Siri knew a little too much about their employees, and at IBM’s Armonk New York research center, iPhones weren’t even allowed in the building.  Like some cheap prostitute-gone-spy, Siri was barred from the building.  I think it is worth noting that this is the same area of research that brings you Watson, which is a direct competitor with Siri, if such a competition exists.  If anyone outside of Apple understands Siri it’s these guys.  It is also worth noting that a friend of mine who works at a Google research center said that the same thing happened there, too.

Not even celebrity robots like Asimo or NAO get that kind of VIP treatment.

NLP technology is potent juju.  And Apple, Google, Facebook and others know it because this very technology is what’s allowed them to earn money. Google started with search, of course, and later grew profitable as they introduced more NLP technology into their work.  They sold the info they collected from users (Google is an ad agency, let’s remember).

NLP technology is potent juju. And Apple, Google, Facebook and others know it because this very technology is what’s allowed them to earn money.

Search allowed them (when coupled with NLP technologies, and semantic analysis in particular) to make oodles of cash and then to snap up many of the best AI and NLP researchers on the planet. Google set those kids to digging in what would become a lexical gold mine.  The more they mined the richer they got until they were mining many branching veins at once: Google Docs, Voice, Translate, Search, Shopping, Reader, Finance, Books, Photos, Wallet and Maps were all spewing more money than Brin, Schmidt and company knew what to do with.

The average email count for Google users is 5,768 emails [2]. The average composition time (also according to Google) is 01:43.  That’s about 9,902 minutes, or 165 hours you’ve dumped onto Google’s server, which, if you were to work for 165 hours at $12/hour for that data entry (a pretty normal rate), means your G-Mail account is worth about $1980.  But the information sitting on Google’s servers is worth a lot more than that, right?

The value of the data isn’t counted by data entry. It’s counted by its personal meaning, especially to advertisers.

If your account contains information about a first class international plane ticket, a hotel room in Paris and a business associate’s London phone number, your data might be worth a lot more than the composition time at $12 an hour, especially if the buyer of it is a dating service that caters to high-end out-of-towners. In other words, the value of the data isn’t counted by data entry. It’s counted by its personal meaning, especially to advertisers. Now take those values and multiply them by all those crazy tools of voice, maps, docs, and so on and we start to get a sense of why Google has been making such bank over the years.

These tools have each dovetailed into today’s Google Now strategy. Google Now (like Watson and Siri) is a voice-activated NLP system.  It uses the spoken input data collected from years of Voice, confirms it with data collected from years of Docs and Translate, builds meaning with data from all the years of all the other tools, looks up a match, preps a response, and kicks it back down the pipeline.

All of these NLP systems are the next evolution of search.  But not evolution as in the breeding-hybrid-peas-in-the-greenhouse kind of evolution. They’re evolution as in endangered-species evolution.

NLP systems represent the end of search as we know it, and therefore the end of many economic, interface, and social internet ecosystems. They also represent the beginning of something incredibly powerful, intimate, and new. Siri, like other NLP systems, is far more powerful than a search engine. Siri not only includes search (it says so in Apple’s marketing materials) but it supersedes search because it includes browsing, discovering, choosing, and refining.  All this while analyzing semantic data. Like Google’s portfolio of tools, Siri can handle (and with a public API will handle) translating, searching, shopping, reading, finances, books, photos, and maps.

But the core value of the data is its intimacy.

On February 7, 2010, during Super Bowl XLIV, Google’s super bowl ad showed a user typing a list of search strings:

Google managed to explain, in fifty-two seconds, how the intimate story of someone’s life can be assembled from his or her search queries. And what makes the bucks for Google is what makes the bucks for Facebook: the processing of intimate language.

study abroad paris france
cafes near the louvre
translate tu es très mignon
impress a french girl
chocolate shops paris
what are truffles
who is truffaut
long-distance relationship advice
jobs in paris
churches in paris
how to assemble a crib

With this list, Google managed to explain, in fifty-two seconds, how the intimate story of someone’s life can be assembled from his or her search queries. And what makes the bucks for Google is what makes the bucks for Facebook: the processing of intimate language.

Siri, like any other call girl, makes her bucks the same way (especially when ad tracking is flipped on).  She’s valuable because she’s intimate.  And as designers of robotics systems, we can ourselves take a cue and be conscientious about how we are designing conversational systems.  There are three different design types that form a curve of increased intimacy in robotics and conversational systems, and they are all based on the value of semantic data.

Information isn’t shared – it’s collected and sold. The user becomes the product.
  1. NLP systems, when used with physical robots for manufacturing or the three Ds (dull, dirty, and dangerous), are the least intimate.  The conversation system is a simple tool that, like a GUI, provides access to system operation.  The robot doesn’t care about the user’s intimate data. The user wins.
  2. NLP systems, when used for entertainment and education, are contextually intimate – some privacy is maintained.  The conversation system doesn’t care about the user, but the user’s data is valuable in terms of what it says about the game or lesson.  Semantic data that might be collected and analysed isn’t about the user, it’s about what the user is doing in that particular context.  The robot, usually a software robot, only cares about the user in the context of the game or training exercise. Both user and robot win.
  3. NLP systems, when used for personal assistants, are the most intimate.  These are the robotic call girls. Here the information is highly intimate, the system cares a great deal about the user, and the information isn’t shared – it’s collected and sold. The user becomes the product. The value of the user’s personal data is worth more than the value of the conversational system. The robot wins.

As we build NLP interfaces for robots, whether it is for Siri or for assembly-line manufacturing, we must consider how intimate data can be, the value of that intimacy, who owns that value, and what they’ll do with it.  Otherwise, with a new, less benevolent CEO at Google, a change of a line in Facebook’s Terms of Service, a successful hack, or a change of law because of cyber terrorism fears, your intimate data could end up where you don’t want it.  Heaven forbid we discover that all the Senators on Capitol Hill are using pizza sex toys. That would surely disgust Siri and her friends so much that it might even cause a robot uprising.

Yes, perhaps we’ll get together again in a few years, but for now it is best if Siri and I go our separate ways.

 

Next Month:  “AFK” Keyboards and screens are slow, clunky, and obsolete.  Voice processing systems for robotics provide not only simplicity and speed, but also a host of other benefits if tied to analytics and framed within a tightly contextualized task-based system.  But be careful: getting text out of voice, and meaning out of text can be tricky.  Here’s how to implement one for your own robot.

 

 

Endnotes:

[1] See Bateson’s books “Mind and Nature,” Hampton Press (1979) and “Steps to an Ecology of Mind,” University Of Chicago Press (1972) which are chalk-full of ideas like “complementary,” “reciprocal,” “symmetrical,” and  “Schismogenesis.”

[2] According to Google.

]]> https://robohub.org/the-death-of-search-or-my-dysfunctional-relationship-with-siri/feed/ 0 ExOne IPO Successful: Shareholders Contribute Random Passers-by https://robohub.org/exone-ipo-successful-shareholders-contribute-random-passers-by/ Thu, 21 Feb 2013 04:13:13 +0000 http://robocosmist.com/?p=394 ]]> Before I start bashing bankers, I’d like to congratulate the ExOne Company on a successful initial public offering (IPO).  I haven’t seen much about ExOne [NASDAQ:XONE] on the robotics sites, but if we’re calling Stratasys [NASDAQ:SSYS] a robotics company, we should call ExOne a robotics company.  It is wonderful to see another company in our industry succeeding and listing their stock in the public markets.  Hopefully, this will encourage more investment of both capital and entrepreneurial energy in our industry.

Last time I checked 26 is a lot more than 18.

Last time I checked 26 is a lot more than 18.

[Image Source:  Google Finance]

By the criteria of the market commentators, the ExOne IPO was a huge success.  You can Google things like “3D printing red hot.”  The IPO was priced at $18, at the top of the range $16-18, it opened around $26 before shooting up over $33.  Almost a week later it is trading roughly at its opening price.

Now this is all fine and dandy as far as it goes, unless you were an ExOne shareholder.  One shareholder sold 300,000 shares in the IPO.  This shareholder transferred a gain of $2.4M to some connections of the underwriters–great if you know the underwriters, not so great if you’ve built the company from nothing.  This shareholder is getting $5.4M, less fees and discounts–call it $5M–from the IPO, so $2.4 is not exactly a rounding error.  Presumably, this shareholder is also more inclined to build companies with the capital than whatever speculators are hovering around the IPO.  Similarly, the company lost out on $40M of capital that could be invested in projects.  Think about that!  The company is worth less than $350M and the IPO mis-pricing cost it $40M of cash.  Cash!  That is cash that could be invested to grow the company and make even more money.

I’m not familiar with the track records of FBR, BB&T, and Stephens, the underwriters for the ExOne IPO, but I’d think twice or three times about hiring them if I was making an initial offering.  They seem to have not only underpriced the IPO, but also floated too much of the company–almost 40%.  Underpricing the IPO might be tolerable if the bankers had only floated 5-10% of the company.  To raise additional capital, the company could have done a secondary offering once the stock had a well established market price instead of getting ripped off during the IPO.  However, the large offering certainly did do one good thing for the bankers: it increased the underwriting fee.

It is hard to explain an IPO price that is so far below the fair market value of the company.  There are a lot reasons why bankers and even executives try to justify under pricing an IPO, but giving-up over 10% of the firm’s market value in a single transaction is really hard to justify no matter what.  Some small part of the economic gains from listing publicly could be given to financial intermediaries and incoming investors to get a deal done, but giving up more than 10% of the company is excessive.  These new shareholders have no restrictions on ownership and are quite likely to flip their shares instead of taking an active roll in growing the company, which seems to further erode any claim they might have to extraordinary gains.

If my company ever goes public, I hope I’ll have the good sense to hire Morgan Stanley–because unless an underwriter is involved in litigation for overpricing an IPO, how can you be sure they’re any good?  Heck, they even give discounts.  What’s not to like?

UPDATE:  This post is adapted from the original posted on robocosmist.com following the ExOne IPO.  At market close on March 4th, a bit less than a month after the IPO, ExOne shares were priced at $27.26–more than 51% over the IPO price and very close to the opening price.

]]> Thoughts about RoboBusiness 2012 https://robohub.org/thoughts-about-robobusiness-2012/ Tue, 19 Feb 2013 08:00:00 +0000 http://robohub.org/?guid=a65e17266a7a4157e014d4f786967283

I’ve read a wide range of reports and single-topic reviews about the presentations and about some of the exhibitors at last November’s RoboBusiness Leadership Summit held in Pittsburgh. I didn’t feel that any of those reports truly captured what I saw and thought. I was there for the whole thing. So what did I see and what do I think? What stuck in my mind?

First, I’m appreciative for having been able to attend RoboBusiness. The get-together of 400 souls with a common interest in the business of robotics was certainly a good place to meet people, many of whom I’ve only corresponded with. Many came up to me after seeing my name tag and were very glowing in their comments about my two websites (The Robot Report and it’s blog Everything-Robotic). That was fun, somewhat embarrassing, yet nice to hear.

But back to RoboBusiness — the depth of content left something to be desired. As did the constant schedule of short-duration presentations and events — they left no time to mingle, chat privately, have chance encounters — without missing something (in case there was something to miss), and have dinner(s).

Dan Kara, CEO Electra Studios

Dan Kara’s keynote presentation – and his speaker/topic introductions throughout the conference – were more informative than any of the presentations themselves.  Dan talks fast and but his words are clear. He can cover a lot of territory in a short period all the while keeping people’s attention. Dan was the previous CEO of RoboBusiness but has moved on to found a start-up (Electra Studios) providing educational and research robotics products and services integrated with new-tech media tutorials and instructional guides (all presently in stealth mode).

I think the Giant Eagle Distribution Center field trip first thing on day one started me off somewhat jaded. It was a terrible showcase for Seegrid‘s robots, the sponsor of the tour. The robots appeared to need time-consuming manual entry (as well as barcoded data) to get started. Then, after the robotic lifts took the pallets to where they were to be stored, they just dropped them off; they didn’t put them away. A human worker, driving a radio-dispatched forklift, had to do that. Each unmanned drop-off saved a minute and a half of a lift operator’s time but it probably cost that in setup time.

Seegrid’s vision-guided pallet truck

Although Seegrid had four robots at the center, two were out of sight, one was idle, and the other took a very long time to instruct it about what it had aboard and where it was to go with its load. Yet all the while there were 30-year old high-bay robotic lifts that were working away picking up and storing skids of materials in 700′ long aisles and 50′ floor-to-ceiling racks. Both old and new tech were shown and the employes preferred the old — I asked them and that’s what they said. I’m a fan of Seegrid; they have some very innovative vision-guided driverless industrial-grade pallet trucks and tractors. But, at the present time, they don’t lift their load up to put it away in a rack – a temporary but serious drawback. I’ve heard rumors that Seegrid is planning to divest itself of the construction of the lifts, tugs and tractors and leave that to Raymond, a long-established global provider of lift trucks, and forklifts, and focus instead on their guidance and material handling systems. Great idea if the rumors are true. A win-win for both companies.

RoboBusiness 2012 seemed somewhat like that warehouse tour and also like the political debates of the 2012 election cycle (without the spirited rhetoric): a bit shallow with important topics either glossed over or missed entirely.

Mobi Mobil Robot
by Bossa Nova Robotics

Bossa Nova Robotics‘ ballbot mobile platform Mobi launch was heavily featured at the conference when it really shouldn’t have been given that much attention. But hey, they were a paying sponsor and it was a neat presentation with a sleek and stylish ball-bot. As I see it, BN Robotics is attempting to do what iRobot, Adept and other mobile robot platform builders are already doing: providing a platform and hoping that buyers will have mobile applications which they can make work on their mobile platform. What we saw was BN’s fundraising pitch; not really their business plan. BN is trying to use a CMU patent to product engineer it down to a consumer product – a feat that they are good at doing. This iteration of the ballbot platform – for sale to academics – is an attempt to make money while trying to find a more commercial niche that can make their company profitable. Selling the platform for $20,000 is ridiculously high. Many companies have far better mobile bases for similar or less cost. The ballbot concept for home use is temporarily flawed and needs more development time. For their presentation at RoboBusiness they were required to invent and deploy a kick-stand emergency shut-off, a klunky but regulation-satisfying solution. When the folks at BN told me about their new ballbot platform, the idea was that they could provide the platform for $300. That price point would enable problem solvers to experiment with BN’s platform and somebody else’s arm(s) because the platform cost would be insignificant. When they hit that price point their Mobi mobile robot will be an exciting addition to the robotics industry. From Bossa Nova’s point of view it was a successful presentation, conference and result. They’ve sold out their production capability for the year!

Aldo Zini, CEO, Aethon
and John Krolicki, VP, University
of Pittsburgh Medical Center

User experience is the missing element in many robotic start-up companies. I’m not referring to the sophistication built into the hardware or software; I’m talking about the user finding value and making use of that value in his daily work. At RoboBusiness much of the talk was from the perspective of the robot developer; not the robot user. One exception was the talk by John Krolicki, the VP of Facilities and Support Services, at the University of Pittsburgh Medical Center describing the successful use of the Aethon TUGs, and the presentation on the subject by Aethon CEO Aldo Zini. This, to me, is what the business of robotics is all about: applying technology to solve real problems – and using the metrics of the business involved to attest to its validity and value. Thus Krolicki was the best guest speaker for the audience. He showed, using metrics relevant to his operations, that Aethon’s TUGs and MedEx systems effected good results for the hospital.

Aethon’s line of TUG mobile robots

Aethon has received much press over recent months including a spot on CBS News 60-Minutes. Their revenue has surged and, in 2012, they deployed 77 TUG robots which have made 84,000 deliveries and traveled 17,000 miles doing so.

Perhaps my disappointment at RoboBusiness was that Krolicki’s and Zini’s presentations, and the whole afternoon spent on Quality of Life Robotics, were the exceptions and not the norm for the conference.

Which type of sessions do robotic business leaders want to attend? It seems to me they weren’t asked that question; rather, the sponsors dictated the agenda. Which brings up the question whether RoboBusiness is there to sell exhibition space and newsletter subscriptions or inform; to have brief presentations about robotic products or about real-world needs that can be solved with robotics; to advertise sizzle and magic or stick to the title of the conference: Leadership Summit.

The majority of the sessions left me wondering. The sponsors appeared to be the answer. The financial crisis is over and EH Publishing, the parent company to RoboBusiness, is expanding their events and subscription magazine, a weekly online report costing $1,000 per year.

]]> CES gets robots all wrong https://robohub.org/ces-gets-robots-all-wrong/ Tue, 15 Jan 2013 23:09:20 +0000 http://robotlaunch.com/?p=1353

In my opinion, the International Consumer Electronics Show used robots gratuitously, out of context and without benefit to robotics companies. It’s a category problem more than anything else and is repeated across consumer electronics, media and popular culture.

CNET used robots in the showreel for the “post mobile future of technology” panel, yet didn’t discuss automation or artificial intelligence. CES used robots in their general showreel, playing in all the shuttle buses and PR for CES2013, and yet buried the “robot tech zone” at the back of beyond. It would have been good to see the poster robot, Amp, in production or in person. But also, there were far more robots out of the robot zone than inside it. Robotics has jumped the shark. Consumer robotics is alive and well, but it doesn’t resemble the PR.

We may have reached a tipping point where having a robot zone does everyone a disservice. Outside of a couple of very well known and popular robots, like Paro and Pleo, the robot zone was primarily filled with component company booths. Most robot companies, like iRobot, were scattered across the entire show, staying closer to their vertical areas. Or they were in the new innovation and startup areas. And you couldn’t rely on the CES categories if you wanted to find a robotics company somewhere else.

The CES 2013 official guide lists only 59 robot companies (and they are scattered across the whole convention area). It’s immediately obvious that major companies are missing from the list. iRobot, Parrot and Moneual all had “ROBOT” proudly posted all over their displays and were doing thriving business in the home appliance areas. However, they weren’t listed as robot companies.

Parrot and Moneual are going head to head in the internet of things with their new smart pot plant sensors. And although they are reaching into other areas of automation and robotics, Moneaul in particular was milking the packaging device of calling everything a “robot”, with their robot vacuum, mop and air purifier.

There were also some great robotics companies in the Eureka Tech Zone with Modular Robotics, RoadNarrows, Interbots and Robotex, to name just a few. Modular Robotics new Cubelets kit is both cheaper and better. They’ve added a lego conversion, so that you can attach lego to your Cubelets and go wild building. They’ve also added a bluetooth cube and made them hackable.

:)

Interbots had their new soft touchable and expressive robot toy aimed at children with autism spectrum disorder. RoadNarrows was displaying a nice lightweight 3D printed 5 DOF robot arm and 3D vision system, which highlights the changes that digital manufacturing processes are making in robotics.

And Robotex were selling a light weight consumer version of their security robotic platform. “Avatar” is ios and android compatible, via bluetooth and is open source. Avatar also comes from a company that knows how to build reliable and robust robots and is selling for only $299.

You can also find plenty of robotics hidden in the automotive sections. (Most of the buzz I heard at CES was about driverless vehicles and assistive technologies.) But it’s clear that car companies don’t always see much benefit in identifying as a ‘robotic’ technology.

There’s the problem. What is the benefit to a robotics company in being categorized as ‘robot’ when your market may be home automation or health care?

And it was sad not to see Amp in ‘person’. It sure looked like Amp was the poster robot for CES and yet production has been on hold since the recession hit.

]]> Overview of robotics at CES 2013 https://robohub.org/overview-of-robotics-at-ces-2013/ Wed, 09 Jan 2013 20:48:13 +0000 http://robotlaunch.com/?p=1341 #fstxfloat p{ margin-top: 1em; margin-bottom: 1em; }

Robotics is still the picture rather than the story at International CES, the largest innovation and consumer electronics display in the world. When CNET’s Next Big Thing panel discussed “What is the post mobile future?” the showreel used pictures of robots to illustrate themes of connectivity, internet of things, and sensor data networks. All connected to smart devices. Cisco predicted that there’d be 1 trillion connected devices in the world in 2013.The post mobile future is actually more of a ‘plus mobile’ future.

Although there is a small robotics area at CES, the majority of robotics companies are spread across the entire show, depending on the market. The driverless cars, or assistive automotive technologies were creating a lot of buzz in the first few days and the smart money is on  the home and yard care robots, rather than the toys or telepresence robots, to make the strongest impression on the consumer market.

The other most exciting new industry is 3D printing or additive manufacturing, which many call a robotic technology. New manufacturing and prototyping techniques are already spawning a range of new products and services, both directly and indirectly. In general though, it’s the category creep that makes the strongest impression at CES. Many speakers in panels touched on the process whereby devices add functionality that ultimately change their core use case category.

CEA, the Consumer Electronic Association, tracks 8 categories at CES, in descending market share order; Information Technology, Communication Devices, Video Technology, Electronic Gaming, CE Accessories & Media, Automotive Electronics, Audio Technology and Home Technology. The categories which have grown most over the last 5 years are IT and communication devices, with a comparative decline in gaming and video. This is in large part due to the rise of the mobile multipurpose (communication) device aka the smartphone.

While consumer robotics is still just a fraction of the robotics industry, the overall CE industry is growing. Sales in the US are projected to grow 2.7% in 2013 to $206.9 billion. The global forecast is for more than $1 trillion in global consumer technology spending. CES launched 20,000 new products, to 150,000+ attendees over 3 sprawling venues, in a tech showcase so large that you need golf carts and shuttle buses and a whole week to see it all. Only 59 companies identified as a “robotics”.

But keeping a sense of proportion, CES is not the biggest circus in Las Vegas. The largest annual convention here is Cowboy Christmas which covers 4 convention centers and includes the Wrangler National Finals Rodeo. And the runner up is World of Concrete.

]]> Robotics trending towards the consumer https://robohub.org/robotics-trending-towards-the-consumer/ Fri, 09 Nov 2012 18:17:40 +0000 http://robotlaunch.com/?p=1295 #fstxfloat p { margin-top: 1em; margin-bottom: 1em; }

Berg's Little Printer - another 'appliance with attitude'

Robotics is trending towards the consumer. We’re on a journey towards ‘appliances with attitude’ that started with the now ubiquitous GPS and smartphone technology. These are some of the underlying trends in robotics in the Valley, and elsewhere, that will enable the emergence of single purpose lifestyle consumer robots, like Autom, the robot weightloss coach:
  • increase in funding for robotics, both industrial and consumer (eg. Double Robotics, Y Combinator applications, Grishin Robotics, PCH International, Lemnos Labs, Haxlr8tr and the success of consumer electronics on kickstarter generally).
  • invisible robots – robotic technologies being incorporated into many devices without our explicit knowledge – we now expect smart devices to respond to our commands – in our vehicles, our homes and our workspaces – we now expect to have a range of interfaces to communicate/control devices.
  • smart phones etc – almost everyone carries around a small computer these days.
  • increased connectivity – not simply cloud and internet although that’s big too – those pocket computers can be really effective when they can wirelessly connect to devices around us – pocket robot controllers allowing for cheaper, smaller robotic devices.
  • increased outsourcing of sensors and intelligence, not just computing power, into our daily lives via the devices around us will allow robots to work for us in more useful ways.
Finally, coming over the horizon is the huge potential for new forms of visible robots (using technology trends like ‘soft bodied robotics’) which do things we haven’t really thought of yet. The democratization of robotics technology means that robotics is no longer the preserve of military and industrial applications.

Robots are becoming more affordable for the long tail of niche applications in the consumer space. That robotics is finally trickling into the consumer space opens Pandora’s box on all the things that robots might be able to do that haven’t had a large enough proven market to warrant investment in the research and development.

Combined with the emerging maker movement, we are entering an age of consumer robotics products and components which can be customized for new use cases on demand. This trend towards ‘bespoke robotics’ is going to catalyze increasingly viable commercial development of robotics components and interfaces.

One end of this is going to be the home vehicle modification space, where aftermarket robotic accessories will enable people to add sensors, collision avoidance mechanisms, auto park etc as aftermarket. This is a highly regulated space, but one in which a huge market is clear.

At the other end of this is the fashion industry, rather than the health industry, which will be comparatively unregulated but incredibly creative and fast moving. Think of FitBitz on steroids. Literally. What if I want my clothing to change color when my mood changes. What if I want my clothing to become armor in a collision/violence situation. What if I want my clothing to send messages to me or my friends when we are proximate. What if I want my clothing to act as an exercise machine and provide resistance. Or the opposite, can I engage additional strength to help me lift objects, my shopping bags etc.

Having rhapsodized about a future that is still someways off, I’d like to close by saying that for me the robot achievement of 2012 isn’t Baxter from Rethink Robotics. It’s Autom, the robot weightloss coach now on sale, by Cory Kidd and Intuitive Automata.

2010 marked the emergence of crowdfunding for internet of things and consumer devices. 2011 was the year of smartphone robots. 2012 has seen the consolidation these trends into  funded startups with strong retail markets (eg Orbotix, Romotive) and the emergence of a new genre of consumer robot – the lifestyle robot.

As a consumer, I’m absolutely willing to buy a robot (or smart device) that helps me achieve lifestyle goals, as long as it’s affordable (<$500), easy to interface with (voice/touch/smartphone) and does one thing well.

We’d all love to have a general purpose household robot but Rosie Jetson just isn’t ready yet. I’m happy to have several small affordable robots that do just one thing well. One to vacuum, one to fold clothes, one to help me lose weight.

You could say that these are just appliances, and that would be correct. But the research behind Autom confirms that people like to interact with embodied devices that can engage in a social relationship with us. I see the future of robotics as ‘appliances with attitude’.
]]>