Robots have been associated with something “hard”, most likely because their roots are in the world of mechanics, being built with cogs and bars. That was something that was clearly setting them apart from living things that are all made by some soft substance (we call it flesh….). Yes, most living creatures have had to develop some sort of scaffolding (internally, like our bones, or externally like the insects’ exoskeleton). Only those floating in water, like jellyfish and octopuses had been able to make do without hard stuff.
Researchers have worked in the last twenty years to find some “soft” materials that could be used as a wrapping for robots, a sort of robotic skin that could be used as a sensing interface or just to provide a more pleasing interaction for those robots that are designed as human companions. That has required finding materials that can be resistant to continuous deformation as well as able to accommodate electronic components (plastic electronics).
The FTI’s report dedicates some thoughts to soft robotics because of its potential to extend the area of application of robots in the coming years. Soft robots would be able to change their shape and adapt to tricky environment much better than their hard siblings. Also, several types of biomedical applications (like prosthetics) would benefit from soft robotics.
In the figure the photo of Octopus, a soft robot developed with European research funding by the SSSA in Pisa. It can float and move in the water and like the real octopus can use its tentacles to grab stuff and move them around. An interesting twist of soft robotics is the focus on developing robots that could self repair. Of course the very first step in self-repair is to be aware that there is something you need to fix. Artificial intelligence and sensors that provides the equivalent of our sense of pain are being used. The next step is to have some material with self-healing properties and this is a whole new area of material science that is joining forces with soft robotics. Here again there is a significant role to be played by artificial intelligence since repairing something requires an understanding of the goal (how is the repaired stuff supposed to be? This goes both for the form and the function). It also involves design approach, robots have to be designed with self-repair in mind, hence the interest in soft robotics since “soft” its easier to handle (from a repair point of view) than hard.
The need to design self repairing structures, hence understand the function required and how to implement them set the stage for systems that support the creation of robots, the so called robot compilers, since a good portion -more and more- is about designing the software (if I may, software for a robot is what a soul is to us…it provides the operating framework, and its execution, i.e. the functionality).
There is quite a bit of (research) work going on motivated by the expected increase in robots adoption and the need to enable users to customise their robots to their specific needs (give them the soul). The ultimate goal is to be able to develop mass market programmable robots, as today we buy PCs and then customise them (with software and data) to fit our needs.