Tiny ping pong ball shaped robots developed by researchers at the University of Boulder may be the first of a new breed of robots designed to do work in remote locations. The researchers there note that virtually all life forms rely on natural swarms of living cells or even organisms to do their work. This new effort underscores their importance in how life works, and perhaps opens the door to creating new types of robots altogether. In a similar vein, researchers at the University of Pennsylvania have been developing swarms of quadrocopters that are capable of flying in formation so precise, it’s difficult to tell just looking at them that they are not connected somehow.
In a profile on the UoB News site, the researchers working on the ping pong ball shaped bots say that programming many small elements to work together to perform a certain task allows for the creation of not just new products in remote locations, but machines that make them. Consider, they say, the case of very small balls that can be programmed to induce a certain amount of magnetism – including turning it off altogether. If the balls are programmed just right, and if enough of them are used, then they can congregate in ways that mimic other structures – a chair for example, or a house. Taking the idea even further, they could be programmed to mimic movement in the form of a liquid such as water, allowing for the transport of materials say, from one point on Mars, to another. This is what they are doing, the team says, adding more and more smarts to the little balls, all the while working to make them ever smaller. The ideal they say, would be balls as small as molecules – building blocks for artificial organic material such as plants or robot animals.
In a similar vein, the researchers at UoP have figured out how to program their quadrocopters to be controllable with a single joystick, by having them work as a swarm, which means they are all constantly aware of one another and their own position in the swarm. That means directions from a controller need only be given to one of the members of the swarm regarding where to move to next, and the others follow in lockstep automatically. This they say, allows for easy control of large masses of swarmed bots.
It’s not difficult to see how the work at the two locations, while different in nature, is all leading to the same eventual destination – intelligent swarms that do our bidding with a minimal amount of input.