A swarm of bees or a colony of ants works together to complete tasks that would be impossible for individuals to carry out. Researchers in Belgium have now shown that a group of robots can pull off the same feat—including figuring out the correct sequence in which the tasks must be performed.
The researchers deployed a swarm of 20 two-wheeled robots, each 50 millimeters high, in a hexagonal playing field. Their job was to enter three booths in the correct order. The key to the robots’ success? Forming chains, the researchers report this month in Science Robotics. The chains extended the robots’ limited perception of their surroundings and allowed them to know not only where to go, but also the correct sequence.
Over the course of the experiment, the robots took on one or more of four roles—runner, guardian, tail, or link. After entering a booth, the runner would turn into a guardian that would keep other runners from entering. The guardian would start to build a chain, with tails and links used to help the chain negotiate its environment. Based on the positive or negative feedback from infrared signals in the booth, guardians would direct runners to the next booth in the sequence.
The study overturns conventional wisdom that robots either use deductive reasoning to plan ahead or that they react in programmed ways to a given situation. The new approach could come in handy in search and rescue missions, say the researchers, in which the sequence of steps to achieve a goal isn’t known ahead of time. Perfecting that, however, would require testing larger robot swarms outside.