The Airbus Shopfloor Challenge with Curtis Carson

Airbus assembly floor. Source: Airbus Group

Transcript below

In this episode, Andrew Vaziri speaks with Curtis Carson, Head of Research and Technology in Industrial Strategy and Systems at Airbus. They speak about the Airbus Shopfloor Challenge to be held at ICRA 2016, and discuss the need for a new generation of industrial robots for aircraft manufacturing.

More information can be found on the competition website.

Curtis Carson

Curtis Carson is the Head of Research and Technology in the Industrial Strategy and Systems Centre of Competence at Airbus, a leading aircraft manufacturer headquartered in Toulouse, France. Holder of an engineering degree from Ryerson Polytechnic University in Canada, Curtis has extensive international experience in both the aerospace and automotive industries. Curtis’ current role includes developing new and innovative aircraft production processes through the application of emerging technologies. This role spans a host of technologies in the framework of the future factory, from augmented reality to collaborative robotics to digitalization to name a few examples.

Links:

• Download mp3 (12.9 MB)
• Subscribe to Robots using iTunes
• Subscribe to Robots using RSS
• ICRA Shopfloor Challenge

Transcript

This transcript has been edited for clarity.

Andrew:  Hello, welcome to Robots Podcast! Can you introduce yourself and describe what you do?

Curtis:  I’m Curtis Carson and I’m responsible for the research and technology on the industrial system side with Airbus in France. My role encompasses everything you can imagine around the production of a commercial aircraft from jigs machinery equipment to robotics automation to new innovative techniques, enhancing and optimizing through IOT and connected objects.

Andrew:  What is IOT?

Curtis:  Internet of things, so connectedness and bridging multiple sensors and systems around our worker or automation on our shop floor.

Andrew: For people who aren’t familiar with Airbus, could you describe your products?

Curtis: There’s a high likelihood if you’ve travelled anywhere, it was one of our Airbus products. We develop commercial aircrafts from a smaller range, which we call single aisle aircraft A320, to the world’s largest passenger aircraft, the A380.

Andrew: What are the challenges in manufacturing aircraft?

Curtis: It might be hard to imagine from an outside perspective, but some of crafts have been developed years ago and we’ve been producing them for many, many years. Our new and innovative products, like our A350, have just gone to our first customers and the industrial system that we built behind to produce all this is quite large. But also expansive in terms of footprint and age and the diversity, is quite significant. We’re not building a factory every year and we’re not designing new aircraft each year. This makes it challenging to find ways to innovate.

Andrew: You said that the lifetime can be significant. What is the average time one aircraft design might be in production?

Curtis:  It is quite typical for an aircraft design to be in production for 20, 30 years or more. The time frames are large to very long, the more successful the program the longer the production.

Andrew:  How many aircraft does Airbus produce in a year?

Curtis:  It was over 630+ aircraft produced last year. This might seem like a small number versus an automotive domain – where you’re producing thousands and thousands of cars – but this is some of the largest quantities of commercial aircrafts ever produced in a given period, or a given year.

Andrew:  How many aircraft are currently in the backlog to be purchased from Airbus?

Curtis: I think it’s close to 7,000 aircrafts in backlog for production time-frame, don’t quote me on that. If you take the numbers per year, the backlog is over quite a number of years.

Andrew: To address the large number of aircraft currently in the backlog, what are some initiatives Airbus is undertaking for its manufacturing processes?

Curtis: We’re looking to innovate assembly process in the building of aircraft that are part of that building process and it might be difficult to see this without looking at really what’s inside a production system. As well as looking into automation to help people, or robotics and automation that we can use to assemble and fabricate parts of the aircraft.

We’re looking from across different facets of innovation: from smart tools, that know where they are in the aircraft; to autonomous robotics systems, that can manage and move themselves around in our production environment; to innovative technologies, like 3D printing and what it can bring to production efficiency and performance in the building of our aircraft. We’re looking across a large spectrum of technologies and see how we can effectively boost our productivity and our performance in building our aircraft.

Andrew:  What robotics initiatives has Airbus undertaken in the past?

Curtis:  In the area of robotics we’re using industrial robots and systems you may see if you were in an automotive environment for specific tasks – like drilling holes on a panel or a skin structure to assemble frames on the aircraft – we have big machines that we use today for building large scale composite parts up to the size of wings. We have a number of different automated systems we use. These are the typical industrial automations that we’re applying. With those typical industrial automations we can target a small percentage of the production activities to the large quantity of parts you can have in an aircraft, and the large diversity of the types of parts.

That’s what exists today. What we’re moving towards is where we can have more light-weight robotic systems that are flexible and agile. Systems that can move in and around our structures versus designing from the ground up a fixed and rigid system… instead you have a robot from within. Once you realize the size and the scale of the aircraft, you’ll see it’s not something that’s readily set up to bring the product to the robot or for the robot to do the work, and then move the product along. We have to be more agile and flexible in our building process.

Andrew:  Airbus is sponsoring this year’s challenge at the International Conference on Robotics and Automation. Can you tell us about that?

Curtis: The competition we are proposing this year is drilling. Why would we target something like drilling? The simple answer is by looking at the production in Airbus in assembling the aircraft I mentioned earlier. We’re drilling over 120 million holes per year. Of those 120 million holes I would say we’re only able to automate a small percentage in terms of production tasks. There’s a high level of manual drilling we still do in the production of our aircraft today. Drilling is an interesting aspect. If we can combine what I said earlier – a lightweight and mobile robotic system – and combine with lightweight tools, then we have an opportunity to move into a new arena, and potential. We could implement an automated system without changing our infrastructure or our production lines to accommodate such a system.

A lot of these holes are drilled today by people. It’s not that a large industrial robot wouldn’t work doing that job, but it is not very mobile. And if you make the robot mobile, well it’s still heavy! If you’re drilling holes inside an aircraft structure barrel, it becomes difficult to put the robot inside without rebuilding your industrial infrastructure. This is what we’re interested to see: how can we push the boundaries of robotics capabilities today in this lightweight mobile module and solve these activities? Drilling is just the start. If we can do it with drilling, we can look to other aspects in production operations.

Andrew:  In the competition itself, what are you asking teams to do?

Curtis:  What we setup might look very simple, but the technical complexities come if we have a metal template. We’re asking them to produce a specific drilling pattern. In this drilling pattern we are asking to produce an automated system with bounding frame or restrictions with regards to size, mass, points for modularity. Then, looking at how quickly they can manipulate the system towards the task. If I have a pattern of holes to drill, how fast can they drill? What’s the level of quality? We’re very, very interested not only to have high speed, but quality.

We need precision. If we adjust the environment and the positions of these templates how can they master their robotic system to accommodate these different positions? How would they simulate the complexities of the geometry and the environment that we’re drilling on? I think for us and during the activities, we’re looking for a robotics system not doing just one task in a repeated mode, over and over, in one position. The types of automation we need, especially when it comes to drilling, are having high volume of repeated drilling tasks over a wide range of different positions.

Andrew:  I understand that the robots are to be modular in their end effectors and weigh less than 100kg and you’ve mentioned reasons for both. I did want to ask: why are you asking people to provide open source software to actually control these robots?

Curtis:  I would say the simple approach is there in the production context we have today and you see it in a number of industries. With this approach we can look at getting the best of the best. You mentioned end effector. Instead of thinking about a closed complete system that we would go buy and maybe the company we’re buying from is advanced in the modular robotic arm system. However, they don’t develop end effectors. The standard approach in most industries would mean we go buy from a supplier and package. It’s not that it doesn’t work, it can work.  But what happens when better end effectors comes along? Or we have another company that’s able to do end effectors better than the company that is integrating the robot and end effectors of the 1^st system?

Well, it becomes difficult and complicated to figure out how to benefit from such an innovative change on the current system. We see the same thing in terms of the software, the control, the algorithms, whatever goes inside. Imagine also we have a huge IT infrastructure on our side across many different aspects of the production lifecycle and this is another level of complexity when it’s closed. We’re then in a mode where we develop each time we have a new system or solution we have to develop a very specific interface for such a solution.

There are a number of different reasons, but I would say 1^st and foremost we want to capture the best capability from the best in the world, wherever they are, for what they’re good at. It’s not common to find somebody that’s good at every single aspect being on all those aspects.

Andrew:  It sounds like a large part of this is foreign partnerships. Are you looking more so towards academia or industry, of a mix of both?

Curtis:  I would say if we’re really good at it, it will be both. Sometimes you can discover innovative things coming out of industry but you can discover interesting things coming out of academia. I don’t think that’s by surprise; academia and industry are linked in a number of different ways and means. A lot of the new businesses you see in today’s start-up world are from academia. What we’re looking for out there is how can we help stimulate towards academia and/or the outside world?

Where are our challenges, what are we looking to resolve and help that makes it feasible? This is also a good reason for the open approach. The more we are open in accessibility to the outside world, whether academia or businesses, the more opportunities and people that could potentially propose ideas that, perhaps, we didn’t think of in resolving. It creates a level playing field for small start-ups to be just as powerful and effective at proposing something valuable versus the big business and industry we’re used to working with today. It creates opportunity.

Andrew:  Beyond just the drilling task that’s been selected for this competition, what are some places where you might see innovation being needed? Or a new approach beyond just making something that’s practical and useful in an industrial setting, but a whole new capability entirely?

Curtis:  I guess it depends on what you mean by a whole new capability. What we look for today is predominantly driven by what we’ve already designed and producing. The aspect that we’re not deeply looking from this type of competition is okay. We’re talking about optimizing the current state of our product based upon how it designed. The design may have been a number of years ago, or 5 years ago. How do you combine those new capabilities and products, materials, or techniques? You use the capabilities in the design of the product that you’re doing and optimize both at the same time.

Andrew:  As you think about your current production process, what are some examples of other low hanging fruit processes that could be adapted to use these technologies we’ve discussed?

Curtis:  The minute we talk about drilling, the next thing is fastening. We have millions of fasteners we’re replacing. We do a lot of surface treatments or ceilings or applying ceilings joints. Painting requires a high demand in terms of manual effort and activity. If you’re painting, then have sanding and abrasion processes. Due to the size and the structure there are confined spaces and restricted access spaces. There are a number of non-ergonomic conditions the workers are in today. This is the whole domain and area of potential when we talk about robotics, especially light weight or new generations of robotics where you see lightweight in modular and adaptable types of systems. We have a huge playing field in terms of opportunity.

Andrew:  How would you see the robots and the people interacting as they assemble an aircraft?

Curtis: The approach we focus on is that we don’t believe, at least today, in having a factory full of robots, and that’s it. We’re looking at worker-robot collaboration. We start from our worker, the person doing the work. How can we enable him to be more effective, better, efficient in his day to day job? It’s like a partner. Any difficult and non-human friendly tasks we can pass to the robot. We can ease the non-ergonomic conditions and reduce the repetitive and straining activities from workers. It’s a combination and collaboration.

We know that as robotics systems evolve they’ll be able to do a little bit more. At the same time, like what I mentioned at the beginning, it’s not just about a robot with a worker, it’s also how we can use the new connectivity of internet of things and connected objects and advanced communication to enable the worker to be more efficient and have the right information in a more dynamic way. Whether it’s through Google Glasses or wearable computing, or sensors in the manual drillers showing the quality of the hole they drilled. These types of automated systems allow a new feedback process for the operator to do the job the best he can.

Andrew:  Could you give us a story or an illustration of what you imagine a worker in this future factory might to on a daily basis that would utilize these technologies, like, internet of things and the general ability to have more information?

Curtis:  I would say the easiest way is just to give you a glimpse of what we’re preparing today. We have tests in preparation for one of our production environments where we will have electronically controlled smart tools. For example, a torqueing wrench for finishing the fastener or a manual drilling pistol for making the hole, with wearable devices. We will equip to three stations or more in one of our production lines and allow them to experience what tomorrow could look like.

The reason I like to focus on the aspect of tomorrow is because it’s not 5 years from now. We have the potential to incrementally put something in place today. What we’re preparing and doing now is the first production trials to see how it would be implemented in day to day for business for our workers.

Andrew:  What do you see as the timeline? If tomorrow is the system you’ve just described what is 5 years from now, 10 years, or 30 years?

Curtis:  It’s difficult for me to project that far ahead. But five years from I don’t think it’s too far-fetched to imagine that connected objects and types of automation will be a prevalent, common thing in any of our production facilities. It could be that tools you might see today like working from paper or drawings, you won’t see. I imagine you’d see a much cleaner environment. You’d see what we consider a high-tech environment in terms of tools, equipment, digital displays, progress tracking interfaces they’re using to prepare job flow of parts to the job that they’re doing.

I would imagine it’s hard to describe it in words unless you visualize and see what a production facility looks like today. I would imagine the proliferation of such devices is not on a linear scale.  You would see it more in an expediential implementation and growth. This is one of the good reasons to consider the open approach. We must have a way, if we want to be fast. It has to be designed in a way that can allow this evolution to happen easily and quickly for robotics systems or connected objects and digital thread that we could use with our workers.

Andrew:  In order to keep pace with that expediential rate of improvement, how serious is Airbus in investing in these opportunities? What would you guess would be the activity in the next few years?

Curtis:  We’re serious about it, and the challenge is just a glimpse into why we’re motivated in this domain. The challenges are real, it’s a real challenge. It’s a real problem that we face today. It’s not something just made up to make a contest for the ICRA event. It’s a real potential application for our shuffler. When it comes to the digitization and the production efficiency and connectivity, we’re very serious about it and we have a number of initiatives running.

Andrew:  On an economic scale people could also be interested, like how much has been spent or planned to be spent. Professors or companies look to the future and will think about possibly partnering with Airbus.

Curtis:  I don’t know if I can put a figure to it, but I would say the best way to see how committed we are to it is to open up a dialogue; ask us! As us in your domain area! It really depends on the value we can potentially get, or bring from it.

We have significant portfolios of research and development activity running on platforms and cooperation with universities, technology centers, businesses. There is a significant amount of activity happening. I can’t really quote to the figures to be honest.

Andrew:  ICRA is the 15 – 21 May in Stockholm, Sweden. When will you be accepting submissions to be part of this competition?

Curtis:  Our submission window for ICRA is open today through March 15^th when we’ll close the submission process.

Andrew:  For these teams, what kind of support is available from the competition?

Curtis:  We’re offering, in certain circumstances, the possibility to any registered team to apply for help with travel, equipment, materials or tools. It depends on the potential and ideas by the teams. We judge it on the case-by-case basis with the teams who register. There is potential for support, but it requires registering and then talking with us.

Andrew:  Where can people find registration forms and more information about the event?

Curtis:  They can go directly to the ICRA 2016 homepage to find the link to the Airbus group webpage. The webpage provides details and process for registering, a general description about the contest, prizes and scoring. The best place is directly on the ICRA webpage.

Andrew:  Thank you for joining us and putting together the Airbus Shop Floor Challenge! We look forward to seeing the results at this year’s ICRA!

Curtis:  Thank you for having me! I’m excited to see the final results from our competitors.

Andrew Vaziri

Andrew Vaziri