Researchers at Harvard University have created a school of brightly-lit robotic fish that can swim together in three synchronized patterns, Meagan Cantwell reports for Science magazine.
The seven robotic fish, called Bluebots, can follow each other in a coordinated group called Blueswarm. A study published on January 13 in the journal Science Robotics details how the 3-D-printed fish bots manage the tricky task of navigating together in a tank. The bots rely on their wide-angle cameras to track the bright blue LEDs on their companions, which give them all the information they need to group together, disperse around the tank, or swim in circles together.
“Just by observing how far or close they are in a picture, they know how far or close the robot must be in the real world,” says the study’s lead author Florian Berlinger, a biologist at Harvard, to Matt Simon at Wired. “That's the trick we play here.”
The researchers created seven Bluebots, each about four inches long. The robots were designed with four fins like reef-dwelling surgeonfish so that they could navigate underwater with precision. The goal of the research was to create a group of robots, called a swarm, that could move in sync with each other without constant, individual instructions from human handlers. Previous robot swarms have navigated based on GPS to create impressive aerial shows or navigate on land, in two dimensions. The robot fish had the added challenge of moving in three dimensions through water and working without GPS.
“Robots are often deployed in areas that are inaccessible or dangerous to humans, areas where human intervention might not even be possible,” says Berlinger in a statement. “In these situations, it really benefits you to have a highly autonomous robot swarm that is self-sufficient.”
That’s why the team taught the robots how to interpret the images gathered by their cameras in order to decide where to move next, without relying on WiFi or GPS. The bright blue LEDs on the top and bottom of the Bluebots can tell another bot how close together the two are because if the lights appear large and far apart in an image, then the bot knows its companion is close by. If the lights are small and close to each other, then they know the other bot is far away.
Based on the information a Bluebot gathers about where its companions are, it can decide what direction to swim in. In one test, the researchers put a red light in the tank for the robots to find and group around. First, the bots dispersed from one another around the tank, getting as far away from each other as possible. When one of them approached the red light, it began to blink its own LEDs, and the other Bluebots changed their goal from dispersing to converging on that blinking light.
In addition to dispersing and grouping together, the bots can also imitate a fishy behavior called “milling,” which is when schools of fish swim together in shimmering, mesmerizing circles. To accomplish that behavior with Bluebots, the researchers instructed the bot to turn slightly right if it saw another bot’s lights ahead. If a Bluebot couldn’t see any of its companions, it turned slightly left until it found them, and then joined the clockwise pattern.
“I find it an extremely challenging problem to do these experiments,” says ETH Zürich roboticist Robert Katzschmann tells Wired. Katzschmann wasn’t involved in the new research, but has also developed robotic fish. He adds, “I'm very impressed by them having set this up, because it looks much easier than it actually is.”
The researchers hope that their work could have applications in difficult search and rescue missions and biology research. The bots will need a few modifications before they can get to work in the open sea. For example, the LEDs would not stand out in well-lit water, so the bots would need a different cue, like a pattern on their sides, in order to find each other. Fish schools also don’t navigate by vision alone—they have a unique sensory organ called a lateral line that can detect small changes in water pressure and flow around them.
"Other researchers have reached out to me already to use my Bluebots as fish surrogates for biological studies on fish swimming and schooling," says Berlinger to the AFP. If they confirm that fish behave the same way as Bluebots, the bots could help scientists study collective intelligence.