Each robotic fish, known as a Bluebot, is equipped with
cameras and blue LED lights that sense the direction and distance of others
inside water tanks.
They swim using flapping fins rather than propellers, which
improves their efficiency and maneuverability compared to standard underwater
drones.
"It's definitely useful for future applications -- for
example a search mission in the open ocean where you want to find people in
distress and rescue them quickly," said Florian Berlinger, the lead author
of a paper about the research that appeared in Science Robotics on Wednesday.
Other applications could include environmental monitoring or
inspecting infrastructure.
Existing underwater multi robot systems rely on individual
robots communicating with each other over radio and transmitting their GPS
positions.
The new system moves closer to mimicking the natural
behavior of fish, which show complex, coordinated behavior without following a
leader.
The robots use their camera "eyes" to detect other
robots in their peripheral vision, then engage in self-organising behaviour,
which include flashing their lights simultaneously, arranging themselves in a
circle, and gathering around a target.
Berlinger described a test in which the robots were spread
out across a water tank to seek out a light source.
When one of the robots found the light, it sent out a signal
to the others to gather around, in a demonstration of a search-and-rescue
mission.
"Other researchers have reached out to me already to
use my Bluebots as fish surrogates for biological studies on fish swimming and
schooling," said Berlinger, explaining that the robot collectives can help
us learn more about collective intelligence in nature.
He hopes to improve the design so that it doesn't require
LEDs and can be used outside laboratory settings such as in coral reefs.