Robots about the house

The robots are coming, as we’ve been hearing for quite a while. Nearby, in The Hague, scientists, engineers and TU Delft students are working on robots we might encounter in the coming years, for instance in an abandoned factory.

Racks, stacked sky-high. Dozens of them, row after row. He entered un­seen, but without proper directions he would never have found the com­partment where the latest iPhones are stored. Just try to make sense of the jungle of barcodes hanging here. Easy navigation for a picking robot of course, but totally useless for a human being. Then again, this hall isn’t meant for humans. He only has to climb up a metre. That’s not too bad. As he holds on to a post with one arm and fumbles around inside the compartment with the other, he glimpses a device approa­ching around the corner. It resembles a moving trash can, apart from the haze of purplish light beneath its bulk. Rolling at considerable speed, the ro­bot navigates between the racks. Just keep hanging quietly and nothing will happen. Suddenly, the robot stops. Its purple light turns red. A camera under its plastic dome turns his way. It makes a sound: “Shunplees”. The robot ap­proaches, the camera swivels up, and a light turns on around the lens. “Identi­fication, please.” Now he understands. Silly thing. He jumps off the racking, kicks the robot and runs away.

Sensors
“We do not make robot bodies”, Lustig clarifies. He estimates that robot SAM is a procured platform for 60% and 40% self-built. What RSS adds to the mix are primarily sensors and sensor data fusion and intelligence. The actual producti­on of robots also happens outside the company. Here, on a floor and in the basement of the Lobeco Fire + Security company near Benoor­denhout in The Hague, forty men and women employed by Robot Ro­bots Company and its subsidiaries, most of them originating from Delft University of Technology, work on robots that can collaborate with people in clever ways to the benefit of security, geriatric care and pas­senger transport.

Cognitive robotics
“We procure a blind system, and  teach it how to see”, explains Jonker, the man who introduced robot soccer in the Netherlands. This professor of Intelligent Vehi­cles & Cognitive Robotics envisions ‘seeing’ – i.e. perception – as the beginning. After that, the questions start: Where am I? What is coming at me? And what should I do about it? That brings us into the field of cognitive robotics. “A security company does not pro­vide a guy with a moustache; it supplies security through cameras, surveillance cars, security guards, an emergency centre and possibly a robot. The robot can take on a num­ber of jobs, including surveillance. A robot will probably do that job bet­ter than a security guard who is ma­king the same round for the ump­teenth time and is mostly focused on getting back to his post as soon as possible. If a security company can deploy 1.5 to 2 fewer people, they will be able to lower their price bids considerably”, says Lustig, who has worked in the security industry for fifteen years. “In those days, I apprehended three intruders, but I fired twenty employees for sleeping on the job.” Robots do not have tho­se problems. Jonker (aged 64) launched a busi­ness alongside his professorship to safeguard his research legacy af­ter he retired, since it was unclear if he would have a successor. The whole adventure has taught him a great deal about putting his know­ledge into practice. “We have had a time when the emphasis was on publications. That is all nice for the rankings, but it does not put the technology on the market in the Netherlands. That only really hap­pens when you bring a bunch of your graduates together in a company. Pieter Kruit (AS) did that with Mapper. He was my role model. I do that with my companies too now. And I’ll have more time for it as from next year.” The irony is that Jonker will in fact have a successor as of 1 April 2016, namely Dr Dariu Gavrilla, originally from Daimler Chrysler, who will focus mainly on data fusion, pattern recognition and deep learning in intelligent vehicles.

 

Foto © Sam Rentmeester . 20160303 . Patrouille-robot in parkeergarage Scheveningen, RRC - Edwin Lustig, Delft Integraal DI // thema

Foto © Sam Rentmeester . 20160303 .
Patrouille-robot in parkeergarage Scheveningen, RRC – Edwin Lustig, Delft Integraal DI // thema Robots

Stay informed about the research

Receive the Delft Outlook newsletter 4 times a year