Researchers at the Harvard Institute for Biological Inspiration Wyss have created a set of fish-like underwater robots that can navigate autonomously and find each other, collaborate on tasks or simply learn together.
Just as air drones are proving useful in post-industrial industries, submarines can revolutionize the environment, shipping and other areas where a permanent underwater presence is desirable but difficult.
In the last few years, there have been interesting new autonomous submarines or AUVs, but the most common type is almost torpedo – effective for sailing in open waters, but not for working through the corners of a coral reef or marina.
To this end, it seems practical to see what Nature itself has deemed appropriate to create, and the Weiss Institute has made a specialty by doing so and creating robots and machines in imitation of the natural world.
In this case, Florian Berlinger, Melvin Gauci and Radhika Nagpal, all co-authors of a new article published in Science Robotics, decided to imitate not only the shape of the fish, but also the way it interacts with its brethren.
Inspired by the sight of learning to fish while diving, Nagpal pursues the question, “How do we create artificial agents to demonstrate this kind of collective coherence when a whole team looks like one agent?”
Their answer, Blueswarm, is a collection of small Bluebots, 3D printed in the shape of a fish, with fins instead of propellers and eye cameras. Although neither you nor I would mistake them for real fish, they are far less scary than an object a normal fish can see than a six-foot metal tube with a propeller spinning hard in the back. Bluebots also mimic nature’s innovations in bioluminescence, illuminating with LEDs the way some fish and insects signal to others. The LED pulses change and adjust depending on the position of each bot and the knowledge of its neighbors.
Using the simple sensors of the cameras and the photosensor from the front, basic swimming movements and LEDs, Blueswarm automatically organizes itself into group swimming behavior, establishing a simple “milling” model that holds new bots when released from any angle.
Robots can also work together on simple tasks, such as searching for something. If the group is given the task of finding a red LED in the tank in which it is located, everyone can look independently, but when one of them finds it, he changes his own flashing LED to alert and call the others.
It is not difficult to imagine how these technologies are used. These robots can safely approach reefs and other natural objects without alerting marine life, monitoring their health or searching for specific objects that their camera eyes can detect. Or they could meander under docks and ships that inspect hulls more effectively than a ship can. Maybe they can even be useful in search and rescue.
The study also improves our understanding of how and why animals swarm together.
With this research we can not only build more advanced robotic teams, but also learn about the collective intelligence in nature. Pisces must follow even simpler patterns of behavior when swimming in schools than our robots. This simplicity is so beautiful, but difficult to find, “said Berlinger. “Other researchers have already approached me to use my Bluebots as fish substitutes for biological research in swimming and fish training. The fact that they welcome Bluebot among their laboratory fish makes me very happy. “