Researchers from Harvard University created a small army of robots capable of aligning themselves and forming 2D shapes. They are called the Kilobots, because the number of robots is 1024. All the robots are identical and each of them is the size of a hockey puck. Radhika Nagpal, Michael Rubenstein, and Alex Cornejo from Harvard University form the team behind the project.
How a thousand robots assemble themselves is the result of a widely encountered characteristic, cooperation. The novelty is that this is the largest group of robots to communicate with each other without a centralized command center. Scientists use infrared to transmit a certain task, but the Kilobots have to decide on their own how exactly the objective will be achieved.
“This is a ‘collective’ of robots – a group of robots that work together to complete a common goal,” said Harvard computer scientist Michael Rubenstein, according to Reuters. “If you call collective artificial intelligence the ability of a ‘collective’ to start to behave as a single entity, you could call this collective artificial intelligence.”
There is a new wave of robots entering the market. Scientists say that we need to overcome the need to anthropomorphize robots and focus instead on functional design, like the first family robot to be mass produced starting with 2015.
Thousand robots assemble in 2D shapes without constant centralized directions
Each robot has a diameter of about 1.2 inches and its fabrication materials cost only $14. The tiny robots have limited communication systems. They can communicate with each other through infrared light, sense ambient light and a LED indicates their status by changing the color. Charging such a large number of devices was managed by the researchers by using an adapted system. The machines were placed one on top of the other between two conductive surfaces, Arstechnica notes.
The team of scientists devised the swarm of robots with a single purpose in mind for now. They wanted to see how the robots would cooperate to arrange themselves into a specific shape without continuous central directions.
Four of the robots have some special roles. They receive the objective and are the first to communicate to the others what they should do. Basically, the four robots roughly mark the desired shape. Further, they communicate with the other robots, who learn how far they are and how much they have to move.
The most exciting fact is how the robots cooperate in order to avoid jamming and achieving efficient positioning. According to some strategies, the thousand robots assemble themselves by monitoring each other to overcome faulty positioning, while not relying too much on precise locations.
The article “Programmable self-assembly in a thousand-robot swarm” was published in Science journal.