Credit: “Brewing Beer with Robots” Brew Age/YouTube

What tasks in a craft brewery can you do with a collaborative robot? Here are five options to improve the efficiency, consistency and safety of manufacturing in the beverage industry.

In a previous post, we showed how craft brewers can benefit from collaborative robots (aka cobots). These new robots can help to overcome the challenge of scaling up a brewery, without compromising the integrity of the beer. They work alongside you and don’t need safety fences.

But, which tasks should you give to a craft brewing cobot? Here are five great applications to get you started. These applications are also relevant to other beverage manufacturers, so read on even if you’re not a brewer.

Application 1: Packaging

Packaging beer is a perfect task for a cobot. Simple packaging applications only require a robot and an off-the-shelf gripper (you may have to make a custom fingertip to hold your bottles or cans).

In this video you can see a demo of a FANUC LR MATE 200iD robot packing beer. It was developed by students and it could be improved quite a lot (e.g. better part presentation than their mock conveyor-belt and a more effective glueing method), but you get the idea.

• It is a non-value added task. In no way does it affect the quality of the beer.
• It involves repetitive, physical movements so there is a risk of repetitive strain injury (RSI).
• It is boring, tiring and wastes time you could use for more important tasks (you’ll know what I’m talking about if you’ve ever lost the best part of a day packing 4000 bottles into boxes).

Programming this type of task is really simple. In fact, I would be willing to bet that a 13-year-old could program it.

If you wanted a more advanced packing application, you could easily add some extra parts to make it even quicker, like in this video:

By the way, the reason that the robot is inside a glass case is that the LR MATE is not a collaborative robot. However, there is a collaborative version of it: the FANUC CR-7iA, which was released earlier this year. It’s exactly the same robot, but with added safety features so you can use it without a safety fence.

Application 2: Palletizing and depalletizing

These days, most things are shipped on pallets. Palletizing involves arranging items onto a pallet; depalletizing involves taking items off. This is a very popular task for brewery robots, for the same reasons that we listed above.

Programming a collaborative robot to do these tasks is really easy when you use a palletizing wizard. Palletizing in breweries often involves lifting heavy objects, which is also possible with modern collaborative robots.

Heavy Lifting with Collaborative Robots

Heavy lifting? Collaborative robots? You’re joking, right? … Nope.

A lot of brewery tasks involve heavy lifting: moving beer kegs, lifting boxes, dragging bags of grain, etc. These tasks are not just hard work, they are actually dangerous. Repeated heavy lifting can lead to musculoskeletal disorders, which account for 35% of the work days lost to injury.

If you had asked me this time last year, I would have said that collaborative robots can’t do heavy lifting (most have a payload capacity of around 5-10kg). That all changed this year. Collaborative robots are starting to appear with some impressive lifting capability. For example, Comau’s AURA robot can lift up to 110kg.

Here are the weights of common brewery objects. I have included examples of collaborative robots which can lift these weights and are available this year:

For most small tasks in your brewery, you will need a normal sized collaborative robot, of around 10kg payload. There are many collaborative robots within this range. However, if you want to use robots for heavy lifting, it is now an option.

If you need to go even heavier — and you have some extra space to spare — you could also go for a high-payload, traditional robot inside a secure area.

Application 3: Bottle capping

Manual bottle capping is an arduous task. Even if you are using a bench capper, your arm will get very tired after a few hundred bottles. That means a high risk of musculoskeletal disorders. It also means that the muscles of your capping arm are much stronger than the other one, which is never good.

Automatic, or semi-automatic, bottle capping machines are quite common in the industry. Nobody but the most hard-line artisan brewer will think less of you for using one.

The problem with the semi-automatic machines is that you still waste time loading and unloading the bottles. This is the perfect job for a collaborative robot!

Imagine a robot loading the semi-automatic bottle capper in this video, using a dual gripper configuration like the one shown to hold two bottles at once:

Application 4: Labeling

Labeling is a similar application to bottle capping because you would use an automatic labeling machine to do it with a robot. We’ve already seen companies using collaborative robots for this type of task, like Australian manufacturer Prysm Industries who used a UR5 to label plastic tubs.

Many craft breweries already use these automatic labeling machines. One popular type looks a bit like a desktop printer — you just place the bottle and press a big button to apply the label. If you already use this type of machine, you won’t need to change it when you introduce the robot. We often recommend that you use existing interfaces for this type of task: simply program the robot to press the button.

Application 5: Serving beer

Okay, so maybe you don’t serve much beer in your brewery (or maybe you do). We just had to include this application because we love beer serving robots here at Robotiq. Two of our highlights this year came from our visit to the Automatica 2016 trade fair:

Two robots were cooking sausages and pouring beer at the SAR Elektronic booth.

KUKA were using a beer pouring robot bartender at their booth.

We learned one thing from these demos: beer serving robots really draw a crowd!

Credit: “Brewing Beer with Robots” Brew Age/YouTube

Craft brewing could be the perfect industry for collaborative robots. But, does automation mean losing your artisan status? We find how craft breweries can use robotics to scale-up their business without compromising on quality.

Robotics and craft beer don’t sound like they go together, right?

Think of craft beer and you think of a lone artisan brewer, stirring a huge pot of boiling malt with a long wooden spoon. Think robotics, on the other hand, and you probably think of a huge factory with automated conveyor belts, rivers of bottles and an army of robot arms.

In the past, automation meant mass produced beer — not anymore. Robots are coming to craft brewing.

Craft beer’s popularity is a problem for brewers

Credit: Robotiq

We’ve been thinking a lot about beer here at Robotiq because we recently tried our hand at home brewing. We made a limited edition batch of DoF Lager especially for members of the DoF robotics community. Yeah, we like a challenge at Robotiq — lager is one of the trickiest beer styles to get right.

Our brewing experience got us thinking about the craft brewing industry. Craft beers are now popular in many countries. This is great because it means that quality beer is everywhere. However, it’s popularity also causes a big problem for brewers. How can you scale up a craft brewery while staying true to your artisan spirit?

Ben Dobler is the Brewing Innovation Manager for Windmer Brothers. He explains a familiar problem for brewers:

“You build your perfect brew house and then, all of a sudden, you can’t make enough beer so you increase your fermentation capacity. Then you have to increase your cellar capacity. Well now you can brew it, you can cellar it, but you can’t package it fast enough. So, now you need to increase your packaging efficiencies, your racking line efficiencies, speeds, size. Then you start that cycle back over again.”

The more popular your beer becomes, the more people want to buy it. Sooner or later, all manufacturing businesses hit a bottleneck. For brewers, that bottleneck could, quite literally, be in the neck of a bottle.

How to use robots to scale-up your brewery

Earlier this year, I visited a craft brewery where they do everything by hand. It’s a small brewery, but not tiny; they can brew over 1500 litres per day. One thing that struck me, as an engineer, was that they did all their bottle capping with a bench capper (which caps two bottles at a time by pushing down hard on a lever). If this were my brewery, I thought, I’d automate that task.

Credit: “Why I Automate – Widmer Brothers” A3/YouTube

There’s a real danger of musculoskeletal disorders when performing this type of repetitive action hundreds of times a day. Also, the task doesn’t add any value to the beer, it’s just a necessary step in the bottling and packaging process. Surely, a perfect candidate for automation.

Craft brewers are often suspicious of any type of automation, which I understand. There’s a worry that using robots will mean that you can no longer call your beer an artisan product; however, automation doesn’t mean you have to make generic beer. In fact, automation can free you up to make even better beer.

“When we talk about automation, in some circles within the craft beer world it’s almost a taboo word, which to me is kind of baffling.” Ben Dobler, Windmer Brothers

In 2015, Windmer Brothers made a brave choice to add robotics to their craft brewery. They added a KUKA manipulator to help with palletizing, which is also a non-value-added process.

Ben makes a good point in this video. If you give the repetitive tasks to a robot, you have more time to focus on other aspects of the brewing process — like coming up with new beer recipes!

As we are also starting to see in the artisan bread, farming and coffee industries, robotics is converging with the artisan food and beverage sector.

The robot used by Windmer Brothers was big — requiring a lot of space and restricted access — but collaborative robots are much smaller and are designed to work alongside humans. They can fit right alongside your existing equipment.

How to start using robots in your craft brewery

As with any investment, it’s a good idea to consider specific applications before deciding to buy a robot. A good place to start: How to Get Started With Collaborative Robots and the eBook How to Shop for a Robot. Use the worksheet in the learning kit to assess some possible tasks for the robot.

Most craft brewers who have started using robots have applied them to packaging processes. Packaging applications are a great place to start because:

• Packaging is a non-value-added process: The beer is already made at this stage so there is no chance of the automation compromising on quality.
• Packaging has a high risk of musculoskeletal disorders: As anyone who has packed beer knows, it involves a lot of repetitive movements. This might not be risky when your production is low; but the more beer you make, the higher your risk of repetitive strain injury (RSI).
• Hand packaging produces breakages: Packing bottles into a box is a simple task, but after you have done it over 300 times (and you’re only halfway through the batch) you are much more likely to drop a bottle or two. One big advantage of robots is that they don’t get tired.
• Packaging usually doesn’t require extra technology: Some of the tasks which you can automate in craft breweries will require extra machines (like automatic labeling machines or semi-automatic bottle cappers). However, packing applications are usually going to be palletizing or simple pick-and-place tasks, which require no extra sensors or equipment. This makes them a great first robot application.

In the next article, we will discuss specific applications in a craft brewery which are great for robots. None of these tasks involve touching beer, so your artisan status is safe. However, as the guys from Dirty Robot Brewworks show in this video, you could also use robots in the cooking if you really wanted to.

Julius robot. (Source: Mining-ROX)

Robots, like Julius from the Mining-ROX project, are a promising option to keep mining economical and safe. In our interview, we spoke with Prof. Bernhard Jung from TU Freiberg about why robots are important for the future of mining.
Mining is a dangerous, unpleasant job. The atmosphere underground is hot, humid and hypoxic (which means there’s not a lot of oxygen down there). Humans have been going down mines for centuries, but it has always been risky. Working conditions are bad and disasters like tunnel collapse or explosions are real dangers. What’s more, as surface level minerals are used up, mines continue to get deeper and deeper with the deepest mines going over 4km underground (that’s the same distance as the height of 10 Empire State Buildings!).

What is Mining-ROX?

The Mining-ROX project is a research project headed by the Technische Universität Bergakademie Freiberg, Europe’s only university with a research and teaching mine. The research team are looking at ways to apply autonomous robotics to mining, to remove dangerous tasks from human workers and also aid in rescue scenarios.

The project is part of the larger research initiative called Robots in Saxony (ROX), a collaboration between German universities to apply autonomous service robots to unstructured environments. Their partner project T-ROX (Teaching Robots in Saxony) at Dresden University was a finalist for the KUKA Innovation Award at Hannover Messe this year, presenting as Team WEIR.

Mining-ROX currently has two robotic research platforms: Alexander (great name!) and Julius. Both are named after famous German scientists.

• Alexander is a four wheel drive electric mobile platform running ROS Indigo. The platform is based on a Clearpath robotics Husky. It has two laser scanners, a lighting system, ethernet cameras and also RGB-D cameras.
• Julius is also a four wheel drive electric platform, this time based on an Innok Robotics vehicle. It also runs ROS Indigo and has fewer sensors than Alexander. It has a UR5 manipulator with our Robotiq 3-Fingered Gripper.
• There is also a third robot, Elisabeth, which is currently under development.

Alexander. (Source: Mining-ROX)

We spoke to Prof. Bernhard Jung to learn more about the project.

Robotiq: Why do you think that autonomous robots, like Julius, are important to the future of mining?

Prof. Bernhard Jung: Mines can be hostile environments for human workers, but robots can go to places that are too dangerous for humans. They are a promising option, particularly for mines in remote areas where no skilled workforce is available. The deep mines of the future will be very hot places and ventilation and cooling systems will be economically prohibitive. In fact, a long-term vision in the mining research community is the fully automated “man-less mine.”

Julius as an assistant (Source: Mining-ROX)

Robotiq: What mining scenarios are you investigating with Julius?

Bernhard: One role of Julius is to act as an assistant during mine surveying tasks. The robot accompanies a human surveyor, carrying the heavy equipment and collecting sensory data. It can hold and operate handheld measurement devices, whose weight can easily become cumbersome when manually operating them for several hours in the strenuous mining environment.

Another scenario is the exploration of mine areas that are unsafe for humans, for example in disaster cases, abandoned mines or the fully man-less mine of the future.

Robotiq: You currently have two research robots. How are the research goals different for each of them?

Bernhard: With Julius, we are exploring the opportunities for mobile manipulators in underground mining. This includes collecting sensory data, which are automatically geo-referenced thanks to the robot’s visual localization capabilities. It operates handheld devices that have been designed for use by humans. In the exploration scenario, Julius is typically tele-operated. The robot has to establish a data communication link to the base station, which it does by placing WLAN relay stations throughout the mine.

Alexander (Source: Mining-ROX)

Alexander is another mobile robot that we use for 3D mapping of mines. It features several cameras and Lidar sensors but does not need a manipulator. By the way, Alexander is named after Alexander Von Humboldt, an alumnus of our university who investigated the vegetation of the Freiberg underground mines when he was a student here.

Robotiq: Who is the Julius robot named after?

Bernhard: It is named after Julius Weisbach, who was a famous German Mathematician and Engineer. He was a professor at the Bergakademie Freiberg in the middle of the 19th century. Today he is maybe best known for his contributions to fluid dynamics including the so-called Darcy-Weisbach equation for modeling pressure loss of fluids in pipe systems due to friction.

Weisbach made significant contributions to the theory and practice of mine surveying. He pioneered the application of the theodolite for mine surveying, a precursor to the total stations commonly used for the task today. Weisbach represents the advancement of the state of art in mining through new technology, and this is what we strive to achieve with our mining robots.

Robotiq: What made you choose the Robotiq 3-Finger Gripper for Julius’s Manipulator?

(Source: Mining-ROX)

Bernhard: We were looking for a versatile Gripper that is capable of grasping very differently shaped objects. Given our requirements, the choices of commercially available options are actually limited. The robot should be able to operate handheld scanning devices made for human operation, such as a XRF-scanner for on-site material analysis. Only a hand-like gripper can firmly operate these devices, so the 3-Finger Gripper design was a natural choice. Also, in our underground mining environment, a further requirement is the protection against acidic water drops. Luckily, Robotiq offers a protective glove for that purpose as an accessory.

Robotiq: You are using ROS Indigo to control Julius. Did ROS compatibility influence your choice of Gripper?

ROS integration is absolutely essential. All of our control software are based on ROS: the mobile robotic platform, the UR5 arm and the Gripper. Without ROS support we would not have opted for the Robotiq Gripper.

Robotiq: How is the collaborative nature of the UR5 important for a mining robot?

Bernhard: The light-weight design of the UR5 is clearly desirable considering the limited payload and energy supply of the mobile robotic platform. However, even more critical in mining environments are the safety aspects. The fact that the UR5’s safety system has been certified by the TÜV (the German Technical Inspection Association) was key in our decision to use it with the Julius robot.

Robotiq: What’s next for the Mining-ROX project and the Julius robot?

Bernhard: There are plenty of things on our todo-list. For Julius, a next goal is to have the robot set up a wireless communication network in an abandoned mine. Another goal of the Mining-ROX project is to investigate creative use-cases of the 3D mine models which have been scanned by the mobile robots. For example, in customized Virtual Reality training simulators for miners and rescue teams.

Back when I was in university, I was part of a team trying to design a robot to play Jenga. One of my tasks was to develop the computer vision algorithms that would detect the tower and the blocks within it. Intuitively, this seemed to me to be quite a simple task: the tower was made of many parallel straight lines; we only needed to detect if a block was actually present, not model the accurate positions of the blocks; and the environment lighting was completely controlled by us.

However, it turned out that the task was far more complicated than I could have imagined. The straight lines between the blocks were not as clear to the camera as they were to my eyes. A hole containing a block looked exactly the same as a hole without a block when the image was reduced to detected edges. The lighting, which I had thought was controlled, turned out to be sensitive to changes in daylight.

Although the task seemed very easy to me, it was a big challenge for machine vision. This is not uncommon. When designing computer vision setups, you have to forget what your eyes see and start by thinking what the camera and algorithms actually detect.

What a Robot Doesn’t See

The human visual system is quite amazing. We are able to recognize the same object under remarkably different conditions. There can be changes in lighting, texture, size, distance, shading, color, motion, and many more visual properties. We will still be able to recognize the object.

As this robotics thesis explains, our visual system is so good partly because “the hardware on the average human is excellent – high resolution color stereo cameras with built-in preprocessing.” It is also effective because “this low-level information somehow gets combined with high-level domain knowledge about what [the object] generally looks like (a rough sense of a “template”).”

Robotic vision systems also use templates to detect known objects. However, they are nowhere near as advanced as the templates used in the human visual system. We are able to see that something is “a car” even if we have never seen the exact model of car before, because we are able to generalize what “carness” looks like. A robot can only see what is actually in the image.

How a Vision Sensor Detects an Object

Let’s introduce the basics of a computer vision algorithm. Now, modern vision systems often use advanced combinations of different machine vision techniques. Some are impressively robust to changes in lighting, distortion, low contrast and many of the other issues that have traditionally plagued vision setups.

In this article, we’re just looking at one basic method – edge detection and template matching. Even if you’re using a system with more advanced capability, you’ll design a much more robust setup if you imagine your system is basic.

Edge Detection: Is There Any Object Here?

Edge detection is the first step for many machine vision systems. It involves tracing the edges within the image by detecting the borders between dark and light areas.

You’ve probably seen this image before. It’s an optical illusion which demonstrates how our own visual system is good at correcting for lighting changes. Squares A and B are exactly the same shade of grey (see here for the proof) but we see B as lighter because our brain corrects for the shadow cast by the cylinder. A computer vision algorithm working on the raw grayscale image would see the squares as exactly the same. However, an edge detection algorithm would still be able to detect the lines by detecting pixels where the brightness changes sharply compared to their neighbors. In this instance, edge detection would allow the algorithm to correct for lighting inconsistencies. It would draw lines between the squares.

However, in other parts of the image, an edge detection algorithm might have problems. Consider the lines around the edge of the checkerboard, along the bottom left and right edges. The differences in brightness between the two faces are very small, so the algorithm might not detect them as edges. This is most clear if you look at the corners which I’ve circled in blue. The gray in both faces is almost exactly the same – just 3 brightness levels (out of 255) at the corner. We see an edge because we look at the whole shape, but the algorithm would not. Similarly, an algorithm might falsely detect an edge down the middle of the cylinder, due to the harsh shadow.

When designing vision setups for a robot, it’s important to pay attention to what is actually present in the image, rather than what your own visual system sees. In future articles, we’ll discuss specific steps you can take to avoid these traps. For now, just remember – don’t trust your own eyes.

Template Matching: Is this Really My Object?

The next step taken by many robotic vision systems is template matching. This involves taking a template image of the object and trying to find areas of the current image which are similar to the template. Template matching algorithms allow you to detect the position of the object within the current image. Some algorithms can also detect rotated objects, scaled objects and even objects with distortion or occlusion.

In robot vision systems, setting the template image is often as simple as putting the object under the camera then dragging a box around the object using a graphical user interface. The system will save the template image and use it as a reference to find the object in future images.

The most basic method of template matching is to directly compare the grayscale images, without using edge detection. For example, if you were trying to detect a circuit board in a collaborative pick-and-place operation, you would take a cropped image of a single circuit board as the template. During operation, the template matching algorithm would analyze the current camera image to find areas which are similar to the template. This basic approach is quite limited. For one thing, it is not robust to inconsistent changes in brightness within the image. Using a vision sensor with an integrated light source would help this by providing consistent lighting.

A slightly more advanced form of template matching is to extract the edges in the image, as described above. The object edges are extracted from the template image. This outline is then matched to the edges in the current camera image. This is more robust to changes in lighting, although the algorithm can still be fooled by harsh shadows, lines in the background and distortion.

In future posts, we will look at how to design vision setups which get the most from template matching algorithms. For now, just remember that clear template images, with well defined edges and consistent lighting, will improve your object detection. By remembering how computer vision works, you can ensure that your vision setup is as robust as possible.

Do you have any questions about template matching? How could you integrate vision into your robot application? Which vision systems have you used? Tell us in the comments below or join the discussion on LinkedIn, Twitter or Facebook.

Robot in assembly. Source: bmwgroup

What’s the biggest benefit for using collaborative robots? It’s not better efficiency. It’s not the extra hours the robot can work in a shift. It’s not even having improved consistency across your product. Whilst these are all great bonuses the biggest benefit of robots is their impact on reducing workplace injury.

Workplace injury is an issue that affects millions of workers worldwide, each year. It costs businesses billions in revenue. Although it’s not possible to avoid all injuries completely, many workplace injuries are avoidable. Musculoskeletal disorders are often preventable, as they are usually caused by bad workplace ergonomics. Collaborative robots are a great way to solve this problem.

In a previous article, we found how collaborative robots are designed to be ergonomic products in themselves. In this article, we’ll look at how collaborative robots can be used to reduce ergonomic problems across your workplace. We’ll also show how you can pick the best tasks for your collaborative robot by putting on your “ergonomics glasses.”

Musculoskeletal Disorders: Too important to ignore

Musculoskeletal disorders (or MSDs) refer to a set of injuries and disorders that affect the human body’s movement or musculoskeletal system, for example, the muscles, joints, tendons, nerves and ligaments. According to FitForWork Europe, MSDs are a huge problem in the modern world. 21.3% of disabilities worldwide are due to MSDs, and they are estimated to cost the European Union around 240 billion euros each year in lost productivity and absence, due to sickness. MSDs accounted for 35% of all work days lost in 2007 in Austria.

As they are caused by repetitive physical stresses on the human body some industries are more affected by it than others. Manufacturing and food processing are classed as high-risk, as outlined this report on the impact of MSDs in the USA. Some jobs are prone to specific injuries due to the type of work involved, such as industrial inspection and packaging jobs, which are prone to upper extremity MSDs. In 2012, the manufacturing industry had the fourth highest number of MSDs, with 37.4 incidents per 10,000 workers.

All this injury costs your business money; a lot of money. Ergo-Plus shows you would have to generate $8 million worth of extra sales to cover the cost associated with the most common MSDs. This is crazy, as the injuries are preventable by simply applying basic ergonomic principles.

How robots can reduce workplace injury

In 2013, we reported on a case study at Volkswagen, who had applied the UR5 collaborative robot to their facility. In it, Jürgen Häfner explained their reasons for introducing the robot:

“We would like to prevent long ­term burdens on our employees in all areas of our company with an ergonomic workplace layout. By using robots without guards, they can work hand ­in ­hand together with the robots. In this way, the robot becomes a production assistant in manufacture and as such can release staff from ergonomically unfavorable work.”

Source: Robotiq

You can improve a task’s ergonomics in two ways:

1. Use ergonomic principles to redesign the task, to reduce the physical stress on the worker.
2. Find a way to complete the stressful part of the task differently, without using a human worker at all. This can hugely reduce the chance of MSDs.

The use of collaborative robots falls squarely into the second category. However, you still need to have a knowledge of ergonomic principles to know if a task can cause MSDs.

How to tell if a task can cause MSDs

Ergonomic professionals sometimes talk about having “ergonomics glasses” to mean viewing the workplace from an ergonomics perspective. Before you learned about ergonomics, you could have easily failed to notice that a task could injure a worker. After learning about them, ergonomics issues will jump out at you as you walk through your workplace. There are two different types of ergonomics:

1. Proactive Ergonomics – This involves solving ergonomics issues before they arise, either by walking around your workplace whilst “wearing your ergonomics glasses” or by incorporating ergonomic principles into the initial design of processes.
2. Reactive Ergonomics – This is what usually happens. It’s solving a problem when it is already a problem. A worker suffers an injury as a result of the task and so we retrospectively try to improve the ergonomics of the task.

At the very least, we should adopt a more proactive approach to ergonomics. In an ideal world, all ergonomics issues would be solved proactively, before they arise. However, being realistic, we’re likely to end up with a combination of the two approaches.

Three steps to improve ergonomics with collaborative robots

Applying ergonomics principles is really quite simple. It just involves a slight change of mindset and three easy steps.

Step 1: Learn what to look out for

The first step to proactive ergonomics is to learn how to spot bad ergonomics. There are some great resources online, particularly the free resources and blogs from Ergo-Plus, the International Labor Organization, Dan MacLeod and FitForWork Europe.

There are a few fundamental principles of ergonomics, which can vary slightly in wording depending on which resource you consult. Tasks which violate one or more of these will need to be redesigned or passed off to a robot, to avoid inflicting injury on the worker over time:

1. Workers must maintain a neutral posture without putting their body in awkward positions.
2. Reduce excessive forces and vibrations on the human body.
3. Keep everything in easy reach and at the proper height, to allow the worker to operate in the natural power/comfort zones of the human body.
4. Reduce excessive motions in the task. This is especially important in repeated motions.
5. Minimize fatigue, static load and pressure points on the human body.
6. Provide adequate clearance and lighting in the workplace.
7. Allow the worker to move, exercise and stretch. After all, the human body is “designed to move” not stay in the same position for long periods of time.

Neutral and awkward back postures (source)

Become familiar with these principles by reading over the linked resources and looking at example images of good and bad ergonomic practices.

Step 2: Stand up and walk!

The second step is to stand up off your chair and walk around your workplace, noticing everything you can about the ergonomics of the tasks. Ask yourself (and then ask your colleagues) the following two questions:

1. How could this task be improved to make it more comfortable (less physically stressful) to the worker?
2. Which parts of the task could we give to a collaborative robot to solve the ergonomic issues?

We recommend that you take photos and videos of the tasks, to document the ergonomics improvement process. This will be useful in two ways, as you can then use the same photos to help you to design your collaborative robot process.

Step 3: Apply collaborative robots to improve ergonomics

Once you have identified problem areas in the workplace, you are likely to have a list of tasks which need improvement. Some of the tasks will be possible to carry out with a collaborative robot. Others will not, so you will need to improve their ergonomics in other ways.

Best practices for looking at tasks ergonomically

With so much information available about ergonomics, you might be thinking that you’d have to invest big in training to become a proactive ergonomics business. However, this is not necessarily the case. You can start small and still see big benefits. The International Labor Organization gives three simple things to remember when applying ergonomics to your workplace:

1. It is most effective to examine tasks on a case-by-case basis. Ergonomics issues can be very specific to a task, so don’t think that doing exactly the same everywhere in the workplace will be effective. Start small, looking at just one or two tasks, and build up gradually.
2. Even minor changes to ergonomics can drastically reduce injury. It might seem trivial to have a robot pick up a part and move it 50cm to be closer to a human operator, but after an 8 hour shift this simple movement can mount up to a huge physical strain on the worker. Simply knowing that the optimal working radius of a table is 25cm allows you to understand this and put it into practice.
3. Staff should be involved in making any ergonomics changes to the workplace. People themselves are the best source of insight into how to improve their work tasks. Get your workers involved right from the start to identify problem areas and apply collaborative robots.

SMErobot invention by ABB: Lead-Through-Programming (Image credit: ABB AG)

Robotics is going to be one of the biggest job creators in modern history. But are there enough qualified people to fill those jobs? As various sectors announce skill shortages, we looked at how robotics is affected and which skills are the most important for employers.

Despite scaremongering from media, studies have shown robotics is going to be a huge job creator. However those jobs may require particular skills that the current job market is struggling to provide.

Sources within the industry say that there aren’t enough skilled applicants to fill jobs in robotics and manufacturing. Plotr, a careers advice website, found that 52% of robotics engineering jobs in the UK are hard to fill, and almost all of these are due to skill shortages. A recent study by Deloitte predicted that 57% of US manufacturing jobs will be left unfilled over the next decade for the same reason.

What are skill gaps and skill shortages?

Skill gaps and skill shortages both refer to situations where there are not enough skilled people to fill the requirements of a job. Although they are similar concepts they are not exactly the same. A “skill gap” happens when a company’s existing workforce does not have the required skills to fulfill the job. A “skill shortage” happens when the job can’t be filled because there is a shortage of skilled applicants in the external job market.

Robotics can be affected by both. For example, bringing new robotic technology into the workplace can mean that existing workers don’t have the skills needed to operate them. This produces a skill gap, which can be filled by training and/or hiring new employees. However, a skill shortage in the wider job market may prove difficult to find the right applicant.

Source: Robotiq.com

Is there a skill shortage in robotics?

Jobs in robotics require quite a wide range of skills. Robotics spans several different fields, including mechanical engineering and computer science. To be effective, applicants need skills like systems analysis, critical thinking and complex problem solving. As a result of this, working with robots is a “skilled job”, whereas in the past, jobs in the same workplace may have been “unskilled”. Employers have reported finding it difficult to fill jobs, because their applicants don’t have the necessary level of technical skill needed to operate the robots.

A 2013 report by the International Federation of Robotics found that companies have been cutting back on low-paid, low-skill jobs since the beginning of the current global recession. This is apparently quite common during recessions. To replace these jobs, employers have been trying to improve the efficiency of their existing workforce, and robotics is one way to do this. However, this has led to the creation of skilled robotics jobs in many industries including manufacturing, electronics and the food industry. Unfortunately, this is often also combined with a skills gap in their existing workforce.

Skill shortages in the job market mean that many of these new positions are not being filled. A recent study by Deloitte Consulting and the Manufacturing Institute predicted that, of the 3.5 million US manufacturing jobs projected for 2015 to 2025, a huge 2 million of these positions will be left unfilled due to skill shortages.

Who is affected?

The skills shortage is likely to affect everyone, including employers, job seekers and the existing workforce.

• Employers need to provide training for their existing workforce, to ensure that a skills gap doesn’t develop which can’t be filled by hiring. The Deloitte study also found that involvement with local schools and community colleges can be effective, to ensure a future supply of engaged, skilled professionals.
• Job seekers need to ensure that they are continually developing. Training can be costly for employers, so coming to a job interview with existing technical skills could greatly increase the chances of landing the job.
• The existing workforce also needs to continually reeducate itself. Although robots are not replacing jobs directly, it does seem that many unskilled jobs are mutating into more skilled positions.

As Dan Kara, Practice Director of robotics at ABI Research, is quoted as saying

You don’t want to be in a place where your job can be replaced by automation or technology. You want to be in a place where you’re the master of automation technology.

The answer, then, seems to be continual training for everyone involved. No longer is it the case that you can stop learning after you have finished your degree, apprenticeship or other training course. Robotics technology is continually developing and it is the responsibility of the entire workforce to keep up with technology, or risk skill shortages and skill gaps such as we are seeing now.

The top skill robotics employers want

Although there are many technical skills required of robotics workers, there is one skill which comes up again and again and again and again.

So, what is that skill? …. Programming.

Today, programming is required even for jobs which would previously have been low skilled. Factory workers who operate robots need to have at least a basic understanding of programming, even if they won’t be writing lines of code.

Wrist singularity. photo: GPA546/youtube

If you are interested in science, “singularity” probably makes you think of a black hole. Black holes are popular in the media right now since the LIGO lab in America proved the existence of gravitational waves. At the center of a black hole, physicists theorize there is a “gravitational singularity”. This means the force of gravity is so big, it goes to infinity. Robotic singularities use exactly the same concept as black holes.

Hang on, let’s back up for a second. What are robotic singularities? And how can they be anything like black holes?

My robot is going crazy

Imagine you want to use your robot to paint a line using a spray painting gun. To paint a perfect line, the robot needs to move at a constant velocity. If the robot changes velocity, some parts of the line will have more paint than others. This won’t look very good. If the robot slows down too much, we’ll get an unsightly blob of paint. Obviously, it’s important that the robot moves along the line at a constant velocity. Robots are precise. Usually they can handle this without a problem. However, if there are any kinematic singularities in the line, the paint job could be ruined.

https://www.youtube.com/watch?v=zlGCurgsqg8

This short video demonstrates robot singularities. It shows various examples of the three types of singularity for a 6-axis robot, which we’ll discuss in a moment. The second example of a “wrist singularity” shows the robot trying (and failing) to follow a straight line at a constant velocity. What’s causing this failure? A singularity right in the middle of the line. You could solve this problem in a couple of ways, but first it’s important to understand what’s happening.

Singularities tend to infinity (whether you want them to or not)

Remember I said that the gravity at the center of a black hole “goes to infinity?” This means that the gravitational force gets stronger and stronger the closer you get to the center. At the center of the black hole the gravitational force is theoretically infinite. This might not actually be true (nobody knows), but it’s a property of maths. Mathematics can easily handle the concept of infinity. The real world cannot.

Lots of mathematical equations tend to infinity. As this physicist explains, theoretically you should create a singularity in your bathtub every time you pull the plug. The basic equation for the spiraling water says that the closer you get to the center of the plughole, the faster the water spins. According to the equation, the water should be moving infinitely fast right at the center. In reality that’s not what happens. Physical systems cannot move infinitely fast (as far as we know).

Robot singularities happen because robots are controlled by math (which is okay with infinity), but are made of real, physical moving parts (which isn’t). If an equation commands one robot joint to “rotate 180 degrees at infinite angular velocity”, the robot joint will say “I don’t’ think so!”

What are robot singularities?

Singularities are caused by the inverse kinematics of the robot. When placed at a singularity, there may be an infinite number of ways for the kinematics to achieve the same tip position of the robot. If the optimal solution is not chosen, assuming there is one, the robot joints could be commanded to move in an impossible way. Infinite velocity is not the only type of singularity that causes problems and certain types of singularities can be more problematic than others. Some robots can be put in such a bad position that they need to be turned off, moved, and restarted manually.

The Stewart Platform is a parallel robot which has several difficult singularities. The following video shows one singularity which causes the robot to lose rigidity in two actuators and fall down completely. This happens because the mathematics requires that the (linear) actuator becomes infinitely stiff, which is obviously impossible.

The follow-up video shows how the platform can’t be moved back to the correct position, because the actuator controllers have an infinite gain at the singularity.

The 3 types of robot singularities

Thankfully, industrial robots are not as problematic as the Stewart Platform. 6-axis industrial robots only suffer from three types of singularities. However, these singularities can still cause havoc.

The American National Standard for Industrial Robots and Robot Systems defines singularities as “a condition caused by the collinear alignment of two or more robot axes resulting in unpredictable robot motion and velocities”. Therefore, the three types of singularities are defined by which joint alignments cause the problem:

1. Wrist singularities – These happen when two of the robot’s wrist axes (joints 4 and 6) line up with each other. This can cause these joints to try and spin 180 degrees instantaneously.
2. Shoulder singularities – These happen when the center of the robot’s wrist aligns with the axis of joint 1. It causes joints 1 and 4 to try and spin 180 degrees instantaneously. A subset of this is an Alignment Singularity, where the first and last joints of the robot (joints 1 and 6) line up with each other.
3. Elbow singularities – These happen when the center of the robot’s wrist lies on the same plane as joints 2 and 3. Elbow singularities look like the robot has “reached too far”, causing the elbow to lock in position. This video shows a good example of an elbow singularity which causes the robot to get stuck.

This video shows a useful simulation of these robot singularities. In the video, the joints are colored red when they are commanded to an infinite velocity, which shows the whole idea pretty clearly.

How to avoid singularities

Manufacturers program their robots so that singularities don’t break the robot. However, in the past this just meant that if one joint was commanded to move too fast, the robot stopped completely with an error message. This was not a very elegant solution.

These days, many robot manufacturers are improving their singularity avoidance. In the video above, the robot had been programmed with a maximum speed for each joint. When the wrist joint was commanded “to infinity”, this caused the software to reduce the velocity of the tip. The robot slowed down when it reached the middle of the line. When it passed the singularity, it was able to continue doing the rest of the line at the correct velocity. The paint job still would have been ruined, but the robot nevertheless functions correctly and doesn’t get stuck.

How programmers avoid singularities

Singularity avoidance has been a developing topic for many years. Various solutions have been proposed and some are starting to make their way into industrial robots. For an introduction to the basics, the ETS Control and Robotics Lab cites a good academic paper which explains the mathematics behind robot singularities and provides examples using an industrial robot.

The more axes that a robot has the more possibilities for singularities. This is because there are more axes which can line up with each other. However extra axes can also reduce the effect of singularities by allowing alternative positions to reach the same point. This paper gives a good introduction on the control of redundant robot arm manipulation, with discussions of how to deal with singularities through programming.

How technicians avoid singularities

Robot technicians have come up with many ingenious ways to avoid singularities over the years. Singularities usually occur when the robot’s links are lined up straight and/or when a joint approaches zero degrees. As a result, technicians started to add small angles to the tooling to reduce the chance that the robot moves into a singularity.

This technique is still a good way to avoid singularities. Mounting a spray painting gun at a very slight angle (5-15 degrees) can sometimes ensure that a robot avoids singularities completely. Not always, but it’s a cheap solution and easy to try.

Finally, another good technique is to move the task into a part of the workspace where there are no singularities. This is not always possible, but can be very effective.

Original blog post: ‘Why Singularities Can Ruin Your Day’

Force torque sensors are now ubiquitous in robotics, but where did they come from? What series of inventions and discoveries happened for robot force sensors to arrive where they are today? It turns out, this sensor was developed thanks to the hard work of many scientists, inventors and researchers.

Let’s take a look at the history of the force torque sensor.

Early history – scientists under pressure

The first step toward robot force sensors was taken by the classical scientists of antiquity. However, they weren’t interested in force, but rather pressure. We now know that pressure is defined by “force over area” but in the beginning, the only “area” people were interested in was a field of dry crops that needed to be watered.

1594: Galileo is confounded

It all started because Galileo Galilei designed and patented a machine for irrigating crops. It used a syringe-based suction pump to pull water out of a river. Unfortunately, there was a problem with the machine – the pump could never pull the water above 10m. Galileo couldn’t find the explanation why. The problem puzzled him, and other scientists, for years.

1644 – 1661: Filling the vacuum

Finally, the puzzling question of pressure was solved during a succession of discoveries in the 17th century. In 1644, Evangelista Torricelli explained Galileo’s problem being caused by the weight of the water and declared the existence of the empty vacuum, which annoyed supporters of the omni-present god. Two years later, Blaise Pascal walked up a mountain to prove pressure is the force of air. Then in 1656 Offo von Guerke developed a new vacuum pump and tested the vacuum by trying (and failing) to pull it apart with eight horses. Finally in 1661, Robert Boyle derived the relationship between pressure and volume of air which sets the stage for the first force sensor.

All in all, it was a busy 17 years! It gave us a fundamental understanding of the relationship between pressure and force. However, after such a busy period, activity died down for the development of force sensing. Nothing happened for almost 200 years.

Late antiquity – A sense of change

Old aneroid barometer. Image credit: David R. Ingham via Wikipedia

1843 – 1849: The first force sensor

In 1843, the first force sensor was developed – the aneroid barometer. This was followed by the Bourdon tube in 1849. These sensors used the principles discovered 200 years earlier to measure the atmospheric pressure. A fixed volume of air contracts or expands as the atmospheric pressure changes, moving the needle on the dial. Using the same principle, MEMS barometers have been developed into cheap robot force sensors in the form of the TakkTile sensor from Harvard Biorobotics Laboratory.

1843 – 1856: The Wheatstone Bridge – born 100 years too early

At exactly the same time a key theory was developed that would become a vital part of modern force sensors. Sir Charles Wheatstone designed an electric circuit that could precisely measure resistance of an unknown resistor (which would eventually be replaced by a strain gauge). In 1856, Lord Kelvin presented a report on the relationship between mechanical strain and electrical resistance.

Unfortunately these developments would have to wait for over a century to actually be applied to force sensors. Nothing much happened for 80 years. The theories of Sir Wheatstone and Lord Kelvin sat around in textbooks, just waiting for modern electronics to catch up with them.

The Middle Ages – starting to feel the strain

1936: The birth of the strain gauge

The strain gauge was born into existence by an aerospace engineer, Charles M. Kearns. He just got hired at the Hamilton Standard Propeller Company and was assigned to solve the problem of inflight propeller failures. Existing methods to detect mechanical strain used carbon-based paints, which were painted onto propeller blades. As a keen amateur radio enthusiast, Kearns had an idea to grind down a carbon resistor, stick it to the blade and measure its changing resistance to calculate how much the propeller had bent. Within four years, in-flight propeller failures were eliminated as a problem and the strain gauge was born.

1938: The strain gauge grows up

By the end of the 1930s, the race was on to improve the strain gauge force sensor. Bonded wire strain gauges were developed by two different researchers simultaneously – E.E. Simmons at Caltech and Arthur Ruge at MIT. Simmons was faster to apply for a patent. Wire strain gauges became the new standard.

1952 – 1953: Foiled again

The development of the force sensor had only begun. In early 1952, an engineer called Peter Jackson grew frustrated by various flaws made the wire strain gauges useless for his helicopter application. He decided they were a waste of time and invented the foil strain gauge in 1953.

1967 – 1969: Silicon strain gauges

With the rise of semiconductor fabrication, it was inevitable that strain gauges would eventually be made out of silicon. By 1967, Zias and Egan of the Honeywell Research Center had patented the technology. In 1969, Hans W. Keller patented the first batch-fabrication process. By the 1970s, silicon strain gauges were ready for robotics to enter the scene.

The early modern period – robots arrive!

1949 – 1951: Teleoperation kicks it all off

Before the first industrial robots had even been dreamt up by George Devol, Raymond Goertz at the Argonne National Laboratory was integrating force feedback into robotic teleoperations, which he then used to handle radioactive materials. His first teleoperations were purely mechanical linkages, that gave them inherent force feedback. In 1951, he patented an electrical manipulator. This was a major milestone for teleoperation and haptic research and has been the forefront of force control research ever since.

1961: Industrial robots arrive

The first industrial robot, Unimate, came on the scene in 1961. It did not have force control. Neither did most industrial robots, until quite recently. They used motion control techniques (e.g. position or velocity controllers) which meant they had to follow strictly defined paths and could not monitor the force applied by the robot on the task.

Although force control wouldn’t reach factory robots for some time, researchers started to integrate force sensors into their robots.

Modern day – The force awakens

1970s or thereabouts: The 6-Axis force torque sensor arrives

In the 1970s, the 6-axis force torque sensor was invented. Unlike the previous developments is not clear who first invented it. Patents were filed around 1980 for multi-axis load cells and there were research papers around the same time. Strain gauge based mechanical load cells had been around for a while; it is likely the first 6-axis force torque sensor was the result of steady research conducted by several research groups, like most modern robotic developments.

1970s – 21st Century: Research abounds

As early as 1973, force control was integrated into robotic manipulators. At some point, somebody stuck a 6-axis force torque sensor onto a robot. It’s not clear who did this first. From the late 1970s right up to present day and beyond, force control research has increased in strength.

2016 and beyond: What’s next in the world of force sensors

Now, there are so many new force sensing technologies it’s hard to keep up. The 6-axis force torque sensor has only started to take off in industry, but in the world of researchers there are new sensing technologies coming out every day. After such a long history, the force sensor has definitely become an integral part of robotics.

Over the years, there has been quite a lot of talk within the robotic research community
about a lack of standard experimental procedures. A EURON (European Robotics Network) workshop back in 2010 tried to tackle this issue head on. They said that although there are many good experimental scientists in the robotic community, there has not been uniformly good experimental work and reporting within the community as a whole.

One of the reasons for this might be because experimental design is not integrated into robotic research the way it is in, for example, medicine or psychology. This could partially be due to a lack of education. Experimental practices are usually learned from existing research papers, which can be of varying quality. This means that robotic researchers sometimes don’t even learn the basics of experimental design. This post will give you a starting point from where you can learn how to implement good robotic research experiments.

Of course, experimental design will be different depending on the type of robotics you are researching. An experiment with a UAV (Unmanned Aerial Vehicle) will be very different from an experiment with surgical robotics. However, there are still common aspects between the two. For now, we will focus on the aspects of experimental design that are similar across all experiments. In future posts, we will look further into specific aspects of experimentation.

Heilmeier’s Catechism: What Are You Trying To Do?

George H. Heilmeier was a pioneering engineer and inventor of the Liquid Crystal Display (LCD). He also introduced a useful set of questions for anyone proposing a research project or new product development. These are:

1. What are you trying to do? Articulate your objectives using absolutely no jargon.
2. How is it done today, and what are the limits of current practice?
3. What’s new in your approach and why do you think it will be successful?
4. Who cares? If you’re successful, what difference will it make? What are the risks and the payoffs?
5. How much will it cost? How long will it take? What are the midterm and final “exams” to check for success?

You can use these questions to structure a good introduction for a research paper. They are also a good way to get started with experimental design. If you can’t answer these questions about your experiment before you start then you will just be fishing around in the dark.

Is Robotics Suited to Experimental Design?

You might wonder whether the lack of strict experimental design procedures might be because they are not applicable to robotics. Other fields have stricter experimental design practices, including the pure sciences, medicine, psychology and the social sciences. You might wonder: Is robotics simply different?

This is quite a reasonable question to raise, especially as there is very little information about experimental design in robotics. However, the simple answer is that although, yes, robotics is quite a different field from many others, it could still be improved with stricter experimental design. Many people within the field agree and are trying to improve the experimental quality of robotic research by introducing benchmarks, standards, competitions and common software and hardware platforms.

I would recommend you read this paper from Willow Garage’s Leila Takayama, as a good introduction to why robotics needs more rigorous experiments and how we can achieve this.

The Five Components of a Well-Designed Experiment

Now that you can answer questions about the purpose of your experiment, you should define all of the key components. One way to look at experimental design is to split it into five components:

1. The Hypothesis or Hypotheses

The hypothesis is the big question that your experiment aims to answer. It’s so big, in fact, that we will devote another post to it later. Some examples of hypotheses are:

• Our inverse kinematic algorithm calculates the result of a 6-R robot in less time than three popular existing methods.
• Our robot gripper uses less electrical power than another similar gripper doing the same task.
• Human tests subjects respond more positively to Robot Design A than they do to Robot Design B.
• etc…

Quite simply, the hypothesis answers the question: What do you want to learn from this work?

It can be tempting to pick any old hypothesis and then start experimenting. However, it is best to sit down and formulate a good hypothesis. Here is one step-by-step guide on creating a hypothesis and here is another.

2. Variables

The variables of your study are those factors you are going to test in the experiment. The classic example is from medical drug research, where you have two groups of test subjects: a “drug group” and a “placebo group”. Apart from these variables, all other factors should either remain constant in the experiment or be removed some other way.

The variables that you will change in the experiment are known as independent variables. A classic example in robotics is to change the control algorithm and keep the rest of the experimental setup the same.

The variables which you will not change (but are affected by the independent variables) are known as dependent variables. One dependent variable for an algorithm might be, for example, computation time.

It can be difficult to decide which variables to choose, especially when you are comparing robotic platforms with many differences between them. However, as Takayama discusses, you can compare different systems if (and only if) you can control or eliminate any other variables of the system that might affect your variables. Is the speed of the robot important? Is the energy consumption important? Only compare the variables that make sense.

It is important to be objective about which factors are relevant to the hypothesis.

3. Measures or Metrics

When you have decided on your variables, you need to decide how you are actually going to measure those variables. For some variables, such as the computation speed of an algorithm, this might be a simple case of “time to solution”. However, for other experiments there may be several potential options for each variable.

For example, if you were comparing “user comfort” for various robot designs, you could ask the human subjects how the design made them feel. Alternatively, you could watch their actions when placed in the room with the robot. Both metrics might be valid but could produce very different results and present various challenges. Often, good research experiments will measure a variable in several different ways.

It is important to choose metrics that are not affected by confounding variables. This means variables which are neither those being tested, nor those which are controlled.

Confounding variables can come from many sources, because robots are such complex systems. Always try to think of other explanations for any result and don’t jump to conclusions. For example, imagine you set up an experiment to test the energy consumption of a robot to compare the same algorithm on two different electronic boards. Imagine that Board-A gave you a huge energy consumption; while Board-B used hardly any energy. A confounding variable might be the fact that Board-A turned the WiFi chip on by default; whereas Board-B turned the chip off by default. The energy-hungry WiFi chip would have nothing to do with your algorithm, but you could wrongly conclude that Board-B was more energy efficient.

Always question your results and seek out confounding variables.

4. Experimental Units

Now that we’ve picked our independent variable(s) and our dependent variable(s), and worked out how we are going to measure them, the next thing is to decide your “experimental units”, which can simply be thought of as: the number of things/subjects with which you’re going to test your system. In the case of algorithm research, this often means “data sets”. With human-based research, one experimental unit is usually comparable to one human subject, such as in human-robot interaction studies. For manipulation research, a robot setup may be tested on several different test setups.

In general, the more experimental units you test with, the better chance you have at producing a publishable research study. However, just because you have many results doesn’t necessarily mean that your experiment is well designed. Your results must be well aligned with the hypothesis, free of confounding variables and repeatable.

5. Assignment Method

The final step is to decide how to apply the variables to the experimental units. For example, imagine you have 1,000 data sets and you want to compare them with 10 different robotic grasping algorithms. You might ask questions like:

• Do you run all 10 algorithms on all 1,000 data sets?
• If not, how do you decide which data sets to try with each algorithm?
• How many times do you want to run each algorithm on each data set?
• etc…

The decisions you make for each of these five steps will make up the Experimental Design section of any research paper or report you write.

It is important to be able to defend the choices you make. Even if you don’t write everything in the paper (due to lack of space usually), you may be asked questions about it by others. Thinking through each stage of the experimental design before actually doing the experiments will make your research much stronger.

June 26, 2015 – The Tethered Robotic Explorer (TReX) robot being described by Patrick McGarey to a visiting group at the University of Toronto Institute for Aerospace Studies (UTIAS) at the 2015 Field and Service Robotics conference. Photo credit: François Pomerleau – University of Toronto

What Type of Person Works in Robotics?

Roboticists are a combination of opposites. As specialists, we are skilled in the fine details of our respective specialisms. As generalists, we are able to see “the bigger picture” — something our broad knowledge base allows us to do.

Robotic engineers fall largely into the thinking (investigative) and doing (realistic) categories. This means that roboticists need to be a good mix between two opposing working styles. “Investigative” people generally like to solve problems by thinking, reading and studying. On the other hand, “realistic” people are practical – they like to solve problems by “getting their hands dirty,” so to speak.

Robotics is a delicate balance between hard study and “fiddling about” (as I like to call it), i.e. working on physical things.

10 essential skills for roboticists

In this list we’ve taken 25 career skills and grouped them into 10 essential skills for roboticists.

1. Systems Thinking

A project manager once told me that many people with robotics degrees turn out to be project managers or systems engineers. This makes a lot of sense, as robots are very complicated systems. We have to be good at mechanics, electronics, electrics, programming, sensing and even psychology and cognition.

A good roboticist is able to understand how all of these different systems work together and is comfortable with the theory behind all of them. Whereas, a mechanical engineer could reasonably say: “that’s a programming or an electrical problem, it’s not my job”, a roboticist must be well versed in all of the different specialisms.

2. The Programming Mindset

Programming is a pretty essential skill for robotics. It doesn’t matter if you’re involved in low-level control systems (only using MATLAB to design controllers) or if you’re a computer scientist designing high-level cognitive systems. Robotic engineers can be involved at any stage of the programming abstraction. The main difference between robotics and other programming disciplines is that robotic programming interacts with hardware, electronics and the (messy) real world.

There are over 1500 programming languages in the world. Although you clearly don’t need to learn all of them, a good roboticist will be have ‘the programming mindset’. They will be comfortable learning any new language if and when it is required. Which leads us nicely into the next one …

3. Active Learning

There are so many topics within robotics that it is impossible to learn all of them before you need them for a project. Even after a 5 year undergraduate degree in robotics and a 3 year PhD, I had only scratched the surface of the topics in robotics.

Being good at active learning is an essential skill throughout your whole career. Therefore, having a good level of reading comprehension and a grasp of the learning strategies that work for you personally will help you to learn new things quickly and easily when the need arises.

4. Mathematics

There are not many “core” skills in robotics (i.e. topics that can’t be learned as you go along). One of these core skills is mathematics. You would probably find it challenging to succeed in robotics without a good grasp of at least algebra, calculus and geometry. This is because, at a basic level, robotics relies on being able to understand and manipulate abstract concepts, often representing those concepts as functions or equations. Geometry is particularly important for understanding topics like kinematics and technical drawing (which you’re likely to see a lot of in your career, even if it’s only on the back of a napkin).

5. Science or other Applied Mathematics

There are some people (pure mathematicians for example) who only need to handle mathematics without applying the concepts to the real world. Roboticists are not this type of person. Skills in science and other applied mathematics are important for robotics, because the real world is never as exact as mathematics. Being able to decide when the result of a calculation is “good enough to actually work” is a key skill for a robotics engineer. Which leads us neatly to…

6. Judgement and Decision Making

Being a good roboticist means continually making engineering decisions. Should you program with ROS or another system? How many fingers should you give your robot? Which sensors should you use? Robotics is full of choices and there is almost never one correct solution.

Thanks to the wide knowledge base of roboticists, you might find yourself in a better position to weigh up certain problems than engineers from more specialized disciplines. Judgement and decision making are essential to make the most of your position. Skills in analytical thinking will allow you to analyze the problem from various angles while critical thinking skills will help you to use logic and reasoning to balance the strengths and weaknesses of each solution.

Further reading:

How many axes does my robot need?

New collaborative robots — and product lines — at iREX 2015

How to calculate a robot’s forward kinematics in 5 easy steps

Where to place fuses and how to protect your robot and motors

Dry vs wet, oil vs grease: Choosing the right lubrication for your robot

Selecting a camera and lens for your robot

Things to think about when transporting your robot

Sign up for our newsletter.

7. Good Communication

As a roboticist, your generalist knowledge will mean that you often have to explain concepts to non-specialists. For example, you might have to explain a high-level programming issue to a mechanical engineer, or a structural mechanics problem to a computer scientist. Good roboticists are a channel of communication between the different disciplines. Therefore, communication skills are vital. Being able to use your speaking and writing skills effectively is important. Also, if you have good instructing skills this is a big bonus.

8. Technology Design

Being proficient at technology design means being able to design things that actually work. It also means being able to figure out why something isn’t working properly and come up with possible solutions, meaning skills in repairing. Robotics involves a wide range of technologies, so skills in technology design mean you can effectively isolate the source of problems and propose effective solutions. An almost magical ability to “get it working” (whatever it is and however it is broken) is one of the marks of a truly talented roboticist.

9. Complex Problem Solving

As we’ve seen from the previous skills, a lot of robotics is about using your complex problem solving skills. This includes foreseeing problems, to fix the problems before they’ve even arisen, and troubleshooting them if they do arise.

10. Persistence

Finally, given the complex nature of robotics, persistence is a pretty essential skill. It might be persistence in trying to find the solution to a particularly difficult problem, or persistence in trying to explain a complex topic to others. Good roboticists will also support their persistence with dependability, proving themselves to be as knowledgeable and adaptable as robotics requires them to be.

As everyone starts to look for “the big technology trends of 2016”, it seems that one technology is overshadowing all others – the Internet of Things. We take a look at why robotics is actually ahead of the game with the IoT trend, and show that some industrial robotic companies have been involved with it for 10 years already. 

It’s around this time of year that people start making a lot of predictions about the future (ourselves included, here were our predictions for 2015). Technology journalists get all excited about what they think will be the big “Tech Trends” of the New Year.

For 2016, popular upcoming robotic trends include affordable social robotics (e.g Pepper and Buddy), an increasing use of robotics in mining and agriculture, connected autonomous cars, drones, drones and more drones, a proliferation of robotics startups, among others. But one tech trend is overshadowing them all for the upcoming year: The Internet of Things (IoT).

Predictions about the IoT being “the next big thing” are coming from all technology sectors, from information technology and computer security to investment. It’s not all hype: there are real devices being produced. At the end of 2015, IoT devices had begun to show up on the consumer market, including a smart cup that tells you how much water to drink and a stealth gadget that ensures the security of other connected IoT devices.

Interestingly, in the field of industrial robotics, the IoT has been well underway in the field of industrial robotics for a quite few years now. Back in 2013, one of our top takeaways from RoboBusiness 2013 was how IoT is an important concept in robotics. Then, in our recent takeaways from RoboBusiness 2015, we heard about numerous examples of companies using IoT concepts.

Why is the Internet of Things Important for Industrial Robotics?

Let’s look at how two big robotics companies have already implemented IoT concepts into their systems:

ABB

In 2015, ABB placed itself at the heart of IoT technologies in industrial robotics by becoming part of the Industrial Internet Consortium. Although IoT might seem new, ABB announced earlier this year that their IoTSP (Internet of Things, Services and People) strategy has been underway for the last 10 years. They have been finding ways to integrate all of the sensors and devices on a manufacturing shop floor to improve all areas of their operations, including: reducing the downtime of robots, improving the reliability of systems and optimizing processes.

As they describe in the above video, there are great benefits when every robot is able to store and analyze its own usage data, and then is able to communicate that data smartly to other connected devices. For example, it means that ABB doesn’t have to schedule the maintenance of a robot by a simple, old-fashioned “after 10,000 hours of uptime”, which might be overly conservative. Instead, the robot itself can monitor its own actual usage and report on its performance. ABB can schedule maintenance of the robot at the perfect time to avoid interfering with its operations, by looking at when the robot is actually used most. It can also mean that problems are caught before they have even become problems.

KUKA

In 2014, KUKA integrated with Microsoft’s Azure IoT platform to create a “connected factory” of 60,000 devices and 259 robots. This 24-hour automated factory is able to manufacture automobiles very flexibly thanks to the IoT approach. This application shows that IoT concepts can be well and truly integrated into industrial robotics. We can expect to see many more examples like this turning up in the coming years.

Is There a Common Protocol for the Internet of Things?

One of the big issues facing IoT is whether there should be a common protocol, because, in order for everything to be able to communicate, the devices must be “speaking the same language.”

In the consumer world, for the moment at least, there are already four big competing protocols – Qualcomm’s AllJoyn, Intel’s IoTivity, Apple’s HomeKit and Google’s Brillo. There are also a load of smaller platforms. Although these consumer protocols might be used for home robots, industrial robotics is going to need a slightly different protocol. Security will be paramount, not least because a successful hack could do real damage if it gained access to a robot. There are a few potential options for an Industrial Internet of Things (IIoT), including the Connext DDS protocol from RTI and Predix from General Electric.

You might think that it’s going to be difficult to decide on a standard. However, one of the proposed strengths of the IoT approach is that it (supposedly) does not rely on a standard protocol. It looks like the IoT might not end up with those long, heated discussions that we’ve had in the field of robotics research before we finally ended up with ROS as an (almost) universal platform.

Instead, the IoT is all about the communication paradigm between many devices regardless of the specific protocols used. For example, all of the robots in my factory might communicate using one platform, while all the robots in your factory might use a different platform. Even though my robots couldn’t communicate directly with your robots, we would both still be using the IoT approach. Despite this, it is likely that some sort of linking protocol will become necessary in the future. But for now, we shall just have to wait and see.

Calculating kinematics is a cornerstone skill for robotics engineers.  Kinematics can sometimes be a pain, however, being told to “go and calculate the Forward Kinematics” is almost robotics research shorthand for “go and get familiar with this robot”. It’s the vital first step when using any new robot in research, particularly for manipulators.

Even though I had learned the theory of kinematics in university, it wasn’t until I had calculated various kinematic solutions for a few real research robots that the whole process started to feel intuitive. Even then, because I was not calculating kinematics every day I had to go back to my notes to “remind myself” how to do it every time. It would have been really helpful to have a step-by-step guide to remind myself which stages to go through, without having to read through hundreds of pages of academically written equations in text books. A sort of kinematics “cheat sheet” would have been useful.

This post is exactly that sort of cheat sheet. For now, we’ll primarily focus on the Devanit-Hartenberg (DH) approach to forward kinematics, as it’s the most common, and we’ll get to inverse kinematics in a future post.

Step 1: Get a Pencil and Paper

It can be tempting to jump straight for the computer when starting out with a new robot. However, even if the robot looks like a “standard” 6R manipulator, I always sit down with a pencil and paper to draw out the kinematic diagram. This simple task forces you to carefully consider the actual physical configuration of the robot, avoiding false assumptions that can wreak havoc later on during coding.

There are various ways to draw a kinematic chain. Pick whichever style you prefer. I favor simple cylinders for the revolute joints and lines for the links. Do a Google Image Search for “kinematic diagram” and see some of the different styles available.

As you draw, work out which way each joint moves and draw this motion as double-ended arrows onto the diagram.

 
Step 2: Figure Out Your Axes

The next key step is to draw the axes onto each joint. The DH approach assigns a different axis to each movable joint.

If you set up your axes correctly then working with the robot will be easy. Set them up incorrectly and you will suffer countless headaches. These axes will be required by simulators, inverse kinematic solvers, and even other researchers (nobody wants to solve a forward kinematic problem if someone else has already done it).

The z-axis should lie on the axis of rotation, for a revolute joint, or axis of extension, for a prismatic joint. The x-axis should lie along the “common normal”, which is the shortest orthogonal line between the previous z-axis and the current z-axis. Have a look at the video below to see how to set them up:

Personally, I draw the axes using the following coloring: z-axis (blue), x-axis (red) and y-axis (green). This is the coloring scheme that’s used in the RViz visualizer from ROS. Back in my undergraduate days, our lecturer made an axis “sculpture” out of three colored straws stuck into a sphere of blue-tack to explain the theory to us. Though this might seem a bit “playschool,” it can sometimes actually be helpful as you can position the sculpture next to the physical robot to make sure you’ve got the axes pointing in the right direction. For a virtual version of this, check out this interactive tool.

Step 3: Remember Your End Effector

The goal of calculating the forward kinematics is to be able to calculate the end effector pose from the position of the joints.

Most forward kinematic tutorials will generalize the end effector as a single distance from the final joint. This is fine for a simple “open-close” gripper. However, as modern grippers are often more complicated than this, it’s worth considering how the end effector operates. The Robotiq 3-Fingered Adaptive Gripper, for example, has a few different gripping modes. Each mode will correspond to a slightly different desired end effector pose. You should consider the end effector carefully when formulating the kinematic model.

Step 4: Calculate the DH parameters

Denavit-Hartenberg parameters are often required to enter the robot model into a simulator and start performing any sort of analysis on it.

The best way to visualize the DH parameters is to once again watch the video above. Have a look one more time with the DH parameters in mind.

The DH parameters break down each joint of the robot into four parameters, each taken with reference to the previous joint. They are calculated in reference to the “common normal” described above. Note that if the previous z-axis intersects the current z-axis, which is often the case, the common normal has a length of zero.

• d – the distance between the previous x-axis and the current x-axis, along the common normal
• θ – the angle around the z-axis between the previous x-axis and current x-axis.
• a (or r) – the length of the common normal
• α – the angle around the common normal to between the previous z-axis and current z-axis.

Go through each joint on your drawing and write down the DH parameters for each joint. Each joint should have one value, which is a variable representing the actuated joint. For a more detailed explanation and some examples, I recommend this handout by Peter Corke or this chapter from Introduction to Robotics.

Alternatives to the DH parameters — The DH approach is the most common approach to forward kinematics, but it’s not perfect. One of its failings is that it doesn’t handle parallel z-axes very elegantly. There are various alternatives, including Screw Theory representations, Hayati-Roberts and other geometric modelings (see this paper for a comparison). Although these may (or may not) be better approaches, most of the kinematic libraries below do accept the DH parameters, and for that reason it’s a reasonable approach to begin with.

Step 5: Choose Your Library

There are loads of kinematic software libraries, and many of them do far more than just calculate forward kinematics. Most of them include inverse kinematic solvers, dynamics, visualization, motion planning and collision detection, to name just a few features. These libraries will transform your DH parameters into matrices, which are then multiplied together to calculate the relationship between joint positions and end effector pose.

Of course, you can also quite easily “roll your own” forward kinematic library using your favorite programming language, if you’re inclined that way. I’ve taken this approach myself, but apart from being a good learning exercise it didn’t offer many more benefits than using an existing library.

Some good development libraries include: Robotics Library, Orocos Kinematics and Dynamics Library, ROS MoveIt, OpenRave, RoboAnalyzer and the Matlab Robotics Toolbox.

Even though you’ll usually require inverse kinematics to actually control the robot, computing the forward kinematics is a necessary step to get familiar with any new robotic arm. If you found this post useful, make sure to bookmark it so you can find it when you next encounter a new robot!

What is your favorite method of getting familiar with a new robot’s kinematics? Do you prefer another method over the DH parameters? Have you got any questions on implementing Forward Kinematics in your research robot? Tell us in the comments below.

Enjoy this post? Find more tutorials and robotics how-tos here. 

Robots have featured in the movies since as early as 1897, with the first celluloid robot turning up in a French movie called Gugusse and the Automaton. Since then, the history of movie robots has been filled with mechanical maids, cruel cyborgs and angry androids.

Most fictional robots are not realistic. Back in February on the Robotiq blog, we investigated how realistic the robots in Star Wars really are. However, even some of the earliest movie robots are still far beyond what our modern technology can produce. Perhaps we’ll never be able to create robots that live up to our fictional ideal. Neill Blomkamp, director of the recent movie, Chappie, thinks not. Recently he said that his research for the movie brought him to the conclusion that humans are incapable of producing an AI as advanced as science-fiction portrays. It might not be such a bad thing if he’s right.

Even though AI might never catch up, some of the other associated robotic technologies have developed quite significantly; enough that we are now able to create some robotic technologies which were first dreamed up for the silver screen.

Lets take an informed eye to four classic movie robots, to separate fanciful robot fiction from realistic robot fact …

1. Power Loader – Aliens (1986)

Exoskeletons are a perfect example where technology is getting pretty close to the movies. The reason is partly that they don’t require much AI – all of the high-level processing is done by the human brain.

This year, the Center for New American Security released a report saying that the US military is aiming to provide soldiers with the TALOS exoskeleton by 2018. There are also other contenders, such as Lockheed Martin’s HULC.

Some have compared the XOS 2 exoskeleton to Tony Stark’s suit in the recent Iron Man movies. However, given the bulk of modern exoskeletons, they’re really more like the Power Loaders from Aliens. Panasonic also claims that they are developing a power loader, which you can see an early design of in this video:

https://youtu.be/nTRUoanbYjU

Whichever movie robot you choose to compare them to, exoskeletons have well and truly arrived.

2. ED-209 – Robocop (1987)

The ED 209 was a giant, machine-gun wielding robot that malfunctions at the start of Robocop, killing a room full of suited executives — possibly the most catastrophic case of Vorführeffekt ever (a German word meaning “demo effect” that refers to the fact that something will always go wrong in a live demo).

Military robots with guns are making their way onto the battlefield whether we like it or not. The HDT Protector might not have legs like the ED 209, but it does have a machine gun. It’s not automated, but new military robots are capable of automated targeting. The physical design imagined for ED 209 was an imposing two-legged affair, and looked worryingly similar to the real-life US mech-suit Megabot 2.

3. KITT – Knight Rider (1982 TV series and 1991 movie)

If the rumors of a new Knight Rider movie coming out in 2016 were true (it seems they’re not, but it’s hard to tell with internet rumors) it would be perfect timing to correspond with the birth of a real life Knight Rider. Next March, Chinese startup Youxia Motors is set to release its KITT inspired electric car. Modeled on the famous robotic automobile, it comes complete with an Android-based operating system called KITT OS.

The Youxia X might not be a complete KITT in itself, but autonomous cars have really come far in the last few years and several of the technologies featured on-screen are now possible. Self driving and collision avoidance are looking to be standard features on many new cars. The so-called “anamorphic-equalizer” — a flashing line of red lights that acted as the cars “eyes” — were infra-red on TV but nowadays they’d probably be LIDAR (though hopefully not so easily hacked as today’s LIDAR). Also, KITT’s engine was a true hybrid, running on various fuels including hydrogen and electric, like a growing number of modern cars.

4. R2D2 – Star Wars Trilogy (1977 to 1983)

Yes, we have covered Star Wars robots already. Despite this, R2D2 really deserves a mention. This droid has earned its place in the robot hall of fame because it’s actually one of the most realistic robots on this list. Largely, this is because the physical design of the robot is pretty basic. However, many of its more advanced, futuristic technologies have actually become possible since they were first dreamed up in the 70s.

R2D2’s impressive list of equipment included a fusion welder, buzz saw and all-terrain treads; all well-proven technologies. In fact, modern technology has surpassed some of the droid’s more futuristic capabilities. Instead of the universal connectivity it got through a plug-in “scomp link”, we’d now probably use bluetooth. These days, its “power recharge coupler” could be upgraded to wireless charging. Its flight capabilities were similar to that of modern autopilots.

But what about the famous holographic projector? This August we all learned how to make holographic projectors out of our smartphones, using the same technique that full sized holographic projectors have been using for a few years.

There is even a real-life R2D2 Builders Club, whose members often meet to show off their hundreds of R2D2 replicas. Most of these replicas probably don’t have the droid’s more advanced features, but it would only take one enterprising robot-builder to turn R2D2 into a reality.

***
As more technologies catch up with fiction, we might continue to see some of the classic movie robots become realistic possibilities. Right now, it looks like AI will always be the limiting factor from making truly realistic replicas, but you never know.

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

• The real soft robots that inspired Baymax, with Chris Atkeson
• The robots in Ex Machina: VFX Q&A with Andrew Whitehurst
• Why do people love robots? Find out in this Botscar-winning Robot Film Festival short
• From sideshow to main event, drones are making their mark in entertainment

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

Could robot chefs really have arrived? Moley’s approach, the robotic kitchen, is basically a flexible workcell, and in manufacturing that’s nothing new. But providing the essential ingredients – feeding the robot chef – needs further consideration.

Is 2015 the pivotal year for robot chefs, the year when it’s started to look realistically possible? When Moley announced its robotic kitchen back in April, the media jumped on the story as a promising glimpse into the future. Could robot chefs really have arrived? As robotics’ professionals, we’re understandably skeptical about such news stories – since the 1960s we’ve heard that robot servants are coming. But if Moley manages to overcome one major issue, their approach could have real potential. Why? Because their kitchen is basically a flexible robotic workcell, and in manufacturing that’s nothing new.

When you tell people that you work in robotics, one of the most popular responses is: “So, when are you going to make a robot that cooks and does the washing up for me?”

Robot cooks and maids have been promised to us since personal robotics were first predicted in the 1960s: In 1966, Shakey was introduced as “the first electronic person”, in the 1980s Omnibot could carry a tray of drinks and in 2009, several “Robot Chefs” were center-stage at Japan’s Food Machinery and Technology Expo, flipping pancakes and arranging plastic sushi.

Finally, this year in April a London company called Moley robotics gave the first live demo of their motion-capture based robotic kitchen.

Could robotic chefs really have arrived?

Why robot cooks are so challenging

The traditional challenges for creating a cooking robot have been the challenges of operating within a human environment, including navigating in cluttered environments, operating safely around humans and animals, as well as classifying unknown objects.

We are probably still a long way off from producing safe, multi-purpose robotic servants which can operate reliably within a cluttered human environment.

Moley has removed many of these challenges by changing the requirements of a robotic cook.

Their approach seems to be: Why make things difficult by trying to design a robot which can work in any kitchen?

Instead, they have chosen to supply the whole robotic kitchen. Safety is provided by a protective window, in the same way that we ensure safety for industrial robots.

The Moley robotic kitchen is more like a flexible manufacturing workcell than a “robot servant”. Instead of manufacturing mechanical parts, it flexibly manufactures food. The biggest difference is that it is programmed by a motion-capture of a real chef preparing the dish, and is designed to copy the organic motions of the chef as closely as possible.

This approach to programming could be a source of complications. However, it is far less problematic than the question of how to reliably chop and prepare the ingredients before cooking.

Moley would do well to learn some lessons from the existing cooking robots that are already used in the world today. Though several robots exist, none of them have reliably solved the Feeder Problem (i.e. providing the essential ingredients for your robotic cell).

Food manufacturing: robots already make your food

It should come as no surprise – we eat food prepared by robots all the time. There are many robots already used in the food manufacturing industry. Chances are you have probably eaten at least something recently that has been touched by a robot. For example, robots are often used to butcher meat.

But, using a robot to make a home-cooked meal is not the same as processing raw ingredients in a factory. Even so, people have been eating robot-cooked meals for several years.

It turns out that the Moley robot is just the latest in a whole line of robotic cooks, with well-established machines like the Thermomix being a starting point.

Although it doesn’t look much like a robot, a Thermomix is more than a robotic food processor. It follows simple programs, but lacks advanced sensors. It chops, cooks, steams and blends, allowing it to cook a lot of varied meals, making it as versatile as many of the existing cooking robots in China.

China: home of the robot chef

Right now, China seems to be the home of the robot chef. During the last five years, various restaurants have been using wok-tossing robots, which cook stir-fried food using pre-measured and pre-chopped ingredients. The same cooking design has even been turned into vending machines and robot restaurants.

Measuring and chopping are supposedly complex tasks and so are still carried out by humans.

These simple machines are really no more complicated than the Thermomix, and are far more involved than the very successful Ultraman noodle-shaving robot, which has basically removed a menial job from the hands of human chefs entirely. Its humanoid appearance is little more than a marketing gimmick.

Such robots demonstrate that profitable markets for highly specific cooking robots exist, if those robots are not overly complex.

The brand-new Sereneti kitchen robot, announced at the end of September, has taken the same approach. It basically consists of a pot-stirring robot. Feeder trays, very much like those of the wok-tossing robots, pour pre-measured and pre-chopped ingredients into the pot.

However, these are obviously far less complex than the robotic kitchen that Moley is proposing.

The Feeder Problem: the challenge for true robot chefs

Truly versatile robot chefs are a bit more complex than a wok-tosser, as is clear when you watch this demonstration of a Motoman dual-arm with Roboiq hands cooking an egg sandwich.

However, just as with the simpler systems, there is still no easy solution to the feeder problem.

Every ingredient must be measured out and placed in the right location for the cooking operation to be a success. This is the core challenge that Moley will have to deal with. It currently does not chop or prepare any of the ingredients (they are placed in position by human operators).

As the CRC Handbook of Mechanical Engineering says: “Experience so far has shown that in most instances, it is the feeder that fails in the workcell, not the robot. Feeding methods must be carefully considered when designing a workcell.”

Moley would do well to focus heavily on their feeder method. Perhaps it would make more sense to employ modified industrial food slicers instead of trying to keep every aspect of the operation anthropomorphic, as they are trying to do now. Or perhaps they’ll come up with an even more ingenious method.

The expected release date of the robot kitchen is 2017. So, we will have to wait until then to see if Moley manages to overcome the feeder problem. I look forward to seeing how they solve this core manufacturing challenge.

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

• Shadow Robot’s Dexterous Hand can do more than cook up a storm
• Robot as marketing strategy: Bosch takes a bet on Rotimatic’s consumer flatbread maker
• Robotic flatbread maker gets $11.5 million from two venture funds
• Yaskawa robot learns the art of the samurai
• Blueberry picking robot could win $250,000 prize

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

image: Bill Wesner/flickr

Robots have been getting their names in the record books. Whether it’s taking the first pictures on the moon or observing the deepest volcanic eruption, robots are setting Guinness World Records alongside the most impressive humans. We take a look at 15 official world records which have been set, and beaten, by robots.

History was made recently, on the 27th of October, when a robot walked the equivalent of more than 3 ultramarathons in 54 hours, completely untethered and without recharging. Robot Xingzhe 1 (Walker 1) walked continuously for 134.04 km using only 0.8 kWh of power, totally smashing the previous record of 65 km set by Cornell University’s Ranger robot in 2011.

Though world records can often raise a laugh – particularly when you consider records such as: “fastest marathon dressed as a toilet (male)” and “longest eyelash on a dog” – they are an entertaining way of demonstrating a very serious point. They show us the very limits of our current abilities as a human race and the limits of the modern world.

Although Xingzhe 1’s walk seems to be an unofficial record (i.e. not adjudicated by the Guinness World Records), it demonstrates just how energy efficient we can make robots if we really try.

So, on this monumental occasion, what better excuse do we need to take a look back at some of the ‘official’ world records that have been set by robots?

Robots: A Whole New Category of World Record

World Records are set all the time, with more robots getting involved each year, although they’re not always the ones setting the records. This September a flailing robot hit a Guinness World Record adjudicator on the head as she judged the World’s Largest Meccano Structure – a temporary bridge built in Belfast, Ireland.

Most of the records listed are specifically set for robots. For Guinness World Records, it would seem, humans and robots are just separate categories of record makers.

Some robot records are “firsts” such as First Robot Trumpeter and First Robot Olympics. They mark a change in our modern world but can never again be broken. Other records will be broken again and again as our technology improves (meaning that some of this post is bound to go out of date).

Guinness World Records That Were Set by Robots

Strongest Robot Arm

The Fanuc M-2000iA/1200 is a pretty massive robot with an impressive payload of 1,200kg. In October 2008, it beat the previous record of 1000kg, which had been held by the KUKA Titan for the previous 16 months.

No doubt the record will be broken at some point in the future. However, the fact that seven years have passed since it was last broken suggests that industrial robot manufacturers have other priorities than breaking world records and making bigger robots. With both KUKA and Fanuc now moving into smaller collaborative robotics, perhaps this record will stand a bit longer.

Longest Journey by an Unmanned Autonomous Surface Vehicle

Although Xingzhe 1’s 134 km walk is impressive, it’s nothing when you compare it to the 14,703 km which were covered by the Wave Glider robot (nicknamed Benjamin Franklin) in February 2013. This impressive water vehicle used wave power to generate propulsion and solar panels to power its sensors. Setting off from San Francisco in November 2011, the craft arrived in Queensland, Australia in February 2013 after having crossed the Pacific Ocean.

First Pictures from the Moon

Back in July 1969, Neil Armstrong made “one small step for man” by becoming the first human to stand on the surface of the moon. We all know this. However, it’s less known that 3 years before, in February 1966, a robot broke another world record by sending back the first photos from the moon. The Luna 9 was sent up by the Soviet Union and transmitted photos back to earth for a total of four days.

Robot records will probably continue to be set in space due to the hostile environment. In May 2008, the First Successful Mars Polar Lander first set down its wheels onto the surface of the red planet.

Double Record: First Companion Robot in Space and Highest Altitude for a Robot to Have a Conversation

Speaking of robots in space, last March a Japanese humanoid called Kirobo set two Guinness World Records simultaneously as it boarded the International Space Station to join another robot record breaker, the Canadarm 2 – the Largest Robotic Arm in Space.

Designed in collaboration with the University of Tokyo, Kirobo is part of a project which aims to see if robots can become good companions for people in space. Could we be one step closer towards a real HAL 9000?

Along with other Guinness World Records like First Robot Babysitter and First Restaurant with Robot Waiting Staff, Kirobo’s world records could mark a change towards our acceptance of social robots as a real possibility for the future.

Deepest Observed Volcanic Eruption

One of the few robots that have set a non-robot-specific world record was a submersible robot named Jason. In December 2009, US scientists used Jason to video the West Mata Volcano erupting at more than 1.2 km below the Pacific Ocean near Samoa. It was the very first time that scientists had seen molten lava flowing on the sea floor.

First Human to be Killed by a Robot

Safety is a key concern with industrial robots and fatal accidents do happen though they are few and far between. Often accidents can be traced back to human error. The first human to be killed by a robot was Robert Williams, who worked in a Ford Motor Company factory and died in January 1979 when accessing a storage compartment shared with a robotic arm.

Since then, robots have moved from the factories and onto the battlefield. The record for First Use of Military Robots in Combat went to the Hermes robot which was used in 2002 in Afghanistan.

Fastest Robot to Solve a Rubik’s Cube

Although other robot record-breakers beyond those on this list do exist, it makes sense to finish on the Rubik’s Cube solving robots.

Guinness World Records for solving various types of Rubik’s Cube are much fought for by both robot makers and human solvers alike. As a result, this particular record is likely to change rather quickly.

The human Rubik’s Cube record was set in April this year by Collin Burns, who can be seen in this video solving the cube in just 5.25 seconds.

The robotic record is currently held by the ARM-powered Lego robot called Cubestormer 3 which completed the cube in just 3.253 seconds in March 2014.

Who knows which World Records will be set by robots in the near future, but we can be sure that there is still room for a lot of robotic “firsts.”

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

• The robots in Ex Machina: VFX Q&A with Andrew Whitehurst
• A glimpse at our robotic future: 235 start-ups reviewed
• 12 favorite robot movies – new and old
• The real soft robots that inspired Baymax, with Chris Atkeson

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

Robotics is truly an interdisciplinary research field. It draws lessons from other research fields, while also contributing new knowledge to those fields, some more than others. Without a long history of robotics research, we might not have many of the technologies prevalent in the world today.

Of course, it’s hard to say for sure which of the many associated technologies would still exist if robotics didn’t. Robotics “takes” more from some research fields than it “gives back”. Computer vision, for example, has been highly influenced by robotics research. However, as you can see if you look at the history of computer vision, it has largely developed on its own, in parallel to robotics, and would almost certainly exist without robotics.

However, other technologies available to the modern consumer are so strongly influenced by past robotics research that they wouldn’t exist in their current form without it.

1. Inverted Pendulums

Scene from Paul Blart: Mall Cop 2.

These days, it’s not uncommon to see everyone from tourists to security guards riding around on motorized inverted pendulums. Of course, since 2001, the inverted pendulum has been “re-branded” as the Segway, thanks to the battery powered vehicle of that name and its successors.

The theory is simple, the center-of-mass is held above a pivot point that is usually attached to motorized wheels. With Segways, this center of mass is provided by the weight of the human rider. As the system is inherently unstable, the control system for the wheels must compensate to keep the whole system upright. In this way the rider can drive the robot by simply leaning forward.

Cubli, from ETH Zurich.

Inverted pendulums have been a long-standing fixture in robotics and control theory. Since the 1960s, they have been used as a fundamental benchmark to test controller performance, as well as being used to teach control theory. Many different control methods can be used and tested, from simple PIDs to neural networks.

Even though the concept of the inverted pendulum has been around for a long time in robotics, it took until 2001 for anyone to realize its commercial possibilities. When they did, the results were staggering.

How many other, simple robotic concepts are just waiting to be turned into a successful commercial ideas?

2. LiDAR and 3D scanning

Google LIDAR map.

LiDAR (from the words “light” and “RADAR”) is the latest in the line of wave-based range sensors, which began with its predecessors: RADAR (Radio Detection and Ranging) and SONAR (Sound Navigation and Ranging).

It uses light beams — usually lasers — to create high-resolution 3D point maps. Sensors detect the distance of the target by bouncing the beam off the target and measuring how long the light takes to return to the sensor.

Although LiDAR has a long history of development apart from robotics, ranging back to the 1970s, the latest wave of LiDAR 3D scanners has been a game changer for robots. Low cost scanners, such as the Hokuyu Laser Range Finders, have finally given robots a way of reliably detecting their environment. They are now so ubiquitous that it’s rare to see a mobile robot that doesn’t have a laser scanner.

Researchers have started to address many of the challenges that are unique to robotics, such as how to quickly detect objects from point clouds or how to manipulate objects with robotic grippers. These challenges would not have been explored by the original researchers of LiDAR technology. Robots operating in real-time need to perform calculations as quickly as possible, which has pushed the technology further than it would have progressed without robotics.

This is a perfect example of robotic researchers contributing new knowledge to another research field by applying their technology to engineering problems. This knowledge then feeds back into commercial products.

3. Unmanned Aerial Vehicles (Drones)

Drones are another example of a long-standing robotic technology that has recently been commercialized and “re-branded”, this time by the media. Drones (aka, unmanned aerial vehicles or UAVs) are flying robots and incorporate most aspects of robotics, including control systems, sensing, computer vision and navigation. Hobbyist drones are usually quadcopters, while the military drones are more like unmanned airplanes.

Quadcopters, or quadrotors, were the subject of robotic research for quite a few years before their current surge of popularity. Although four-rotor aircraft have existed since as far back as 1907, the modern quadcopter has really taken off in the last decade as electronics have reduced in size and cost. Numerous research groups use them to develop advanced control, such as collaborative mapping and autonomous navigation.

The public has mixed feelings when it comes to drones. On the one hand, they are rightly worried about the use of unmanned aerial vehicles as weapons of war. On the other hand, the public use of hobby quadcopters has skyrocketed (if you’ll excuse the pun). For example, they have revolutionized low-budget filmmaking by allowing anyone to shoot impressive aerial shots.

Drone development, which has long beencarried out by robotic researchers in labs, is now moving quickly into the commercial arena. What area of robotics is next we wonder?

4. Simultaneous Localization and Mapping (SLAM)

Real-time appearance-based map, based on SLAM. Image courtesy: Matthieu Mathieu Labbé

As a research field, SLAM is what we’d call “mature”, meaning it has been going on for a long time (since 1986) and any new developments in the field are relatively advanced. In other words, it is a well proven research field. However, it’s only been recently that we’re starting to see SLAM really take hold in the commercial world. Though it is not exclusively a robotic research field, it has been highly influenced by robotics and is one of the core topics in mobile robotic research.

As it’s name suggests, the idea behind SLAM is to create a map of the robot’s environment while simultaneously calculating where within the environment the robot is located. The classic example that has moved into the consumer world is the robotic vacuum cleaner (like the Roomba), which creates a map of your house while also cleaning up and locating itself in the room.

With the rise of self-driving cars, warehouse robots and location-based mobile applications, SLAM is only going to become more pervasive in the modern world. Without it, robots would find it almost impossible to operate in a human environment.

What’s Next?

The only question now is which key technology will become the next rising star to grow out of robotic research. Just like these four technologies, it will undoubtedly be something that already has a lot of research work behind it and is well proven in real applications. However, hundreds of research fields interact with robotics, while the consumer world can be fickle and unpredictable.

So we will just have to wait and see.