Huggable robots with 'feelings'

A soft robot has been given a ‘nervous system’ so it can sense when you touch it.
04 December 2018


Our world is filled with robots. From large-scale manufacturing machines to Furbys, robots are present at almost every level of human life. Traditionally, robots have been made out of hard materials like metal or plastic, but there’s a new type of robot that scientists are working hard to make: soft robots. These are robots made out of inherently soft, stretchable materials like rubber or silicon, and they have many advantages over traditional hard robots.

Hard robots can be too dangerous for human contact when functioning, whereas soft robots might allow humans to work closely alongside one with a lower risk of accidents. Another advantage of soft robots comes when robots are trying to mimic parts of the human body - a soft robotic prosthetic could be more comfortable and allow the wearer to experience something much closer to human touch and agility than hard prosthetics currently allow.

“I guess the dream soft robot would be a companion robot, right?” says Ilse Van Meerbeek, a PhD student at Cornell University who is working on making soft robots a reality. As technology moves even further into the home, achieving a safe, functional companion or medical robot (think Baymax in ‘Big Hero Six’) is like the holy grail.

However, “one of the challenges in soft robotics is being able to reliably control them” explained Ilse. For a robot to manoeuvre itself around efficiently it must be able to sense where its different parts are, just as we need to know what position our legs are in to be able to walk. With a traditional hard robot this is relatively simple if you have enough sensors on the joints, as a hard robot can only really move in ways it has been programmed to move. Soft robots, however, have the capacity to change shape in almost infinite directions (“like an octopus”) thanks to the stretchable and flexible materials they are made of. This means that for a soft robot to understand where its different parts are and what is happening to them, they need something more akin to a nervous system that could report whatever type of deformation, or change in shape, it is experiencing.

To overcome this problem, Ilse embedded a network of sensors - optical fibres sensing light intensity - in a piece of silicone foam. She then changed the shape of the foam by bending it or twisting it, which changed the way light was moving through the foam. This change in light intensity was sensed by the optical fibres. Ilse collected the information reported by the optical fibres and used this to generate a computer model of shape changes. By combining this data with artificial intelligence she showed that the silicone foam was able to use this network to accurately detect what was happening to it - whether it was being bent or twisted, and to what degree. She recently published her work in the journal Science Robotics.

"I think what the research I’ve done tells us is that we may be able to make soft robots that can sense really complex types of deformation," said Ilse, describing how a soft robot that can feel when it is touched in different ways is a big leap forward in the field. She concluded that the goal is to "be able to use that information to interact with the world and with the environment in more meaningful ways".


Add a comment