If you’ve ever been to an aquarium, you probably spent a moment watching a school of fish swim. Now, scientists have taught 3D-printed fish how to swim in a group, a feat that could improve decentralized autonomous robotics in the future.
The seven robot fish, dubbed Bluebots, have three LED lights and two wide-angle cameras for eyes and use a 3D vision-based coordination system and 3D locomotion. Each fish navigates its own path in the water tank by tracking the LED lights on the other fish and reacting accordingly using their onboard Raspberry Pi and custom algorithm that helps them accurately gauge direction and distance.
Just like real fish, the robot fish watch the other fish in the swarm, dubbed the Blueswarm, for cues and synchronize their movement with each other and without a single leader. They have multiple fins, just like real fish have, which helps them precisely navigate movement forward and backward, side to side, and up and down. They’re also pretty adorable looking.
Between this hardware and the algorithm, the Bluebots are able to perform complex behaviors, like swarming, dispersion, and milling. Previously, scientists were only able to achieve this type of movement in a 2D simulation. This is the first time scientists have been able to demonstrate several complex behaviors in an 3D environment underwater, without human assistance or external base stations, which was accomplished by a team at the Harvard John A. Paulson School of Engineering and Applied Sciences and the Wyss Institute for Biologically Inspired Engineering.
Teaching robots to more efficiently synchronize movement and coordinate with each other has striking potential. It could improve autonomous vehicles while lessening collisions, allow for robots to work alongside humans in warehouses, and even make it easier to deploy robot teams to Mars to set up shelters and perform other tasks that humans couldn’t (though it also puts us one step closer to Skynet, which is terrifying).