Robots unchained

It’s now more than 50 years since the world’s first industrial robot clocked on for work at GM’s Ternstedt car plant in Trenton, New Jersey in the United States. Fast forward to 2015 and robots have all but conquered the global automotive industry.

Sales of industrial robots are now at record levels. According to the International Federation of Robotics (IFR), 2014 saw unit sales of robots exceed 200,000 for the first time, up more than 27 per cent on the previous year. Demand, which is greatest in China and South Korea, is being driven primarily by the automotive and electronics industries.

Large industrial robots are normally bolted to the factory floor and caged to protect the workforce. These tethered giants, prized for their precision, speed and ability to repeat the same task endlessly, require a huge level of capital investment. As a consequence, they are largely confined to high-volume assembly lines. But that could all be about to change.

“There are several factors that are really drawing our attention at the moment,” says Jean-Philippe Jahier, director of Innovation and Industrialisation of New Technologies with Thales Alenia Space-France. “The first is that off-the-shelf robotic components are becoming very affordable – you can acquire a robotic arm for around €60,000. The second factor is the development of cobotics: robotic systems that can work safely alongside humans.”

Cobots – collaborative robots – are a recent development.  Unlike the majority of industrial robots, cobots are adaptable, mobile and designed to work with people. They’re attracting the attention of industry because they’re easy to deploy and put the power of robotics within reach of low-volume manufacturers for the first time.
 

“Off-the-shelf robotic components are becoming  very affordable – you can acquire a robotic arm  for around €60,000”
Jean-Philippe Jahier, director of Innovation and Industrialisation of New Technologies  with Thales Alenia Space-France

“Flexibility is one of the cornerstones of cobotics,” says Jahier. “Conventional industrial robots are often dedicated to a single function. But with cobotics, it’s more like having a Swiss Army knife: by changing tools, you can adapt a cobot for different tasks quickly. The capability to move within the factory between different applications is very exciting and opens up new doors for us.”

One of the breakthroughs with cobots is that they learn by imitation: “Instead of needing specialised programmers, the end user is able to teach the cobot how to perform a given gesture,” explains Jahier. “This is important, because it empowers operators on the shop floor. It’s also socially important, because the person using the tool is able to shape it to his own usage.”
 [[asset:signpost:134831 {"mode":"full","align":"","field_admin_bool_image_filter":[0],"field_admin_link":["https:\/\/www.linkedin.com\/pulse\/when-robots-offer-helping-hand-even-space-patrice-caine"]}]]

As one of the world’s leading satellite and payload manufacturers, Thales Alenia Space invests continuously to optimise production. Launched earlier this year, the company’s “Tomorrow’s Factory” initiative underlines this commitment, with an emphasis on combining state-of-the-art technologies, such as robotics and cobotics, with human expertise.

The need for innovation in satellite production is being brought into focus by major changes in the space market. One of these is the growing momentum behind what is being described as the “internet in space”, the idea of delivering broadband connectivity on earth using satellite networks. This will require the deployment of huge constellations of small satellites and will trigger demand for satellite production on a scale not previously seen in the space industry.

“Shifting from building a single communications satellite in 30 months to rolling out a satellite every two days is a big change,” says Jahier. “If you want to address the large constellation market, with satellites in the hundreds, you have no choice but to deploy robotic or cobotic assistance.”

Jahier envisages a number of shop floor applications for cobotic assistance. One of these is “kitting”: the painstaking job of assembling all the components needed to carry out a given production task.
“Efficient kitting is important and it’s linked to the concept of lean operations. The idea is to minimise disorder near to the place where you work,” explains Jahier. “The cobot selects the components required for the job and these are given to the operator a few minutes before work is due to start. This is a key issue as regards the efficiency of the production flow.”

Getting cobots to select components from the parts bin has other advantages. Using its built-in camera, a cobot can check components for damage and even measure them. “This eliminates a lot of errors – it’s difficult for a human operator to tell if a screw is 5mm too long or too short, for example, but for a cobot, it’s not a problem,” says Jahier.

Cobots also know where components have come from. This information can be collected for quality assurance purposes and fed directly to the back-office system to assist in compliance and inventory control. As well making short work of time-consuming tasks, cobots make light work of awkward lifting operations by offering an intelligent “third hand”. “Even if you’re assembling very fragile objects, the cobotic third hand can help the operator by lifting and holding components in exactly the right place while you go to fetch the missing tool,” says Jahier.

Robots capable of moving around the factory floor on their own are still a relative rarity. But, as with autonomous cars, mobilising robots raises new questions. How do you avoid collisions? And who’s in charge?

Low travel speeds, coupled with the cobot’s ability to learn and respond to its surroundings, means the risk of collision is minimal. “They’re capable of scouting their environment and they can be programmed to stop or move around obstacles,” points out Jahier. And by extending existing ID and authentication techniques, it’s possible to ensure that only authorised users get to issue commands. “The robot will only interact with somebody who’s already logged in. If you want to alter the execution sequence or change the operation, you will have to log in again.”

Despite the productivity gains promised by cobots, Jahier stresses that these robot recruits will be designed to assist, rather than replace, the existing workforce.  “You cannot imagine a factory with 1,000 robots and one person deciding everything,” he emphasises. “Robotics and cobotics are about boosting competitiveness and consolidating growth – not reducing the headcount.”

Spotting the next big thing in robotics is an art as much as a science. What do investors look for? Bruno Bonnell is head of Robolution Capital, a Paris-based private equity firm that focuses on the development of robotics. Picking winners, he says, means focusing on three key criteria.

“The first thing we look for is what we call ‘hypertechnology’ – a proposal has to be disruptive, not simply an improvement of something that’s already on the market,” says Bonnell.

Technology must also meet customer expectations – without exceptions and with no excuses: “It has to be non-deceptive. For example, some of the automatic cars we have analysed depend on video. My first question is: what if there’s fog or it’s dark? If a car can’t work in these conditions, it’s absolutely useless.”
Robotic solutions must be easy to deploy and easy to use. That means delivering disruptive technology with minimum disruption.
“It has to be easy to implement without drilling, cabling or tearing up your living room,” says Bonnell. “The same principle applies in industry: if you have to make huge changes to your infrastructure, it won’t work.”

With a track record in technology stretching back more than 30 years, Bonnell’s career included eight years as CEO of the computer games business Atari – an experience he says was invaluable.
“There’s a natural link between gaming and robotics,” observes Bonnell. “The idea of projecting yourself into a different environment, and then manipulating it, was born in the gaming industry. Today, this principle underpins everything from drone operation to robotic surgery.”

 Stumbling blocks remain.

“Energy is the number one by far,” emphasises Bonnell. “We don’t yet know how to make sophisticated robots work without an external power supply. We are still at the foetal stage of robotics and we need to find a way to cut the umbilical cord.”

There’s also the question of the extent to which robots should be allowed to think for themselves. “I think we have to be careful with this obsessive fight for artificial intelligence,” cautions Bonnell. “Aside from the ethical and practical problems, it’s just much more efficient to have humans controlling robots.”
Where next for robots? Transport will be a major axis of development, thinks Bonnell.
“The self-driving car will be one of the next key evolutions. By 2030, the idea of driving a car yourself will seem ridiculous if not forbidden.”
Aviation is also likely to become increasingly automated. “Leaving control of aircraft exclusively to pilots is going to be a problem. I think we’re going to see much more emphasis on automatic flying, with crews providing emergency backup,” says Bonnell.
The biggest change, though, could be the rise of teleoperation with people controlling machines at a distance.
“You can see the beginnings of this with telepresence,” says Bonnell. “Extending this approach allows us to routinely drive and operate machines remotely in hazardous conditions, such as beneath the sea or in harsh desert or polar environments. The key is combining the performance of robots with the intelligence of people.”