What happens when you take a swarm consisting of different types of robot all cooperating to achieve a single goal? The result might be described as strangely disturbing ...
The notion of distributed processing isn't new and its application to robotics leads naturally to the idea of a swarm of robots. However most swarms are build up of a lot of very usually very simple and usually identical robots. In this case however things are very different. What we have is a swarm consisting of different types of robot all cooperating to achieve a single goal.
The Swarmanoid project is funded by the European Commission and is the successor to the Swarmbot project. The video below was created to show off what they have achieved and won the video competition at the AAAI conference on artificial intelligence last week.
Each of the robots in the swarm have special capabilities - I almost said powers because they do remind the viewer of a team of super heros tackling some task. The team consists of Hand-bots that can grip and climb, Foot-bots that move around and can join with other modules to move them and Eye-bots that fly over head and survey the terrain.
There is also something quite spooky about they way that they interact in a purposeful way and perhaps this is another aspect of the uncanny valley effect. It could be that swarms of identical robots are more like insects or lower lifeforms and we don't get too scared of them. On the other hand the different forms of the Swarmanoid robots seems to give them different personalities and this makes the swarm look much more social and they seem to operate as a single higher entity, i.e. a Swarmanoid. It's an effect that might be worth exploring.
Watch the video and be amazed:
Ok so they are only retrieving a book from a shelf but imagine the drama that a humanoid robot would make of just walking across the room!
Eventually the Swarmanoid is intended to have 60 autonomous robots of the three types: eye-bots, hand-bots and foot-bots.
It's a very different approach to the problem of building a robot and one that might prove more practical than building a single humanoid robot. But for me, and I'm a robot enthusiast, there is something really disturbing about Swarmanoid...
Benchmarking is both essential and a distortion. The trouble is that once there is a benchmark that anyone pays any attention to, the tendency is to tweak for a high score - even if this results in a [ ... ]