Engineers, scientists and other researchers will often look to the natural world to find solutions to problems. That is what a team at the University at Buffalo, New York recently did when it created a beaver-inspired robot which, believe it or not, could turn out to be a lifesaver in a disaster zone.
The robot built by graduate students Maira Saboia Da Silva, Vivek Thangavelu, and others doesn’t physically resemble a beaver. Instead, it’s a mini-rover vehicle which uses a camera, custom software, and a robotic arm to lift and deposit nearby objects — modeled after the way the beaver builds dams from whatever it can find in its location. In this case, however, it uses the nearby objects it finds to build ramps which allow it to overcome obstacles to reach its destination. In the real world, the robot could be used for rescue missions in disaster zones — for instance, utilizing whatever rubble and other objects it can find to help it reach trapped people.
“Animals don’t make a complete plan before they start,” Nils Napp, assistant professor in the Department of Computer Science and Engineering, told Digital Trends. “Instead, they keep evaluating cues as they build and respond to them. For many animals, we don’t exactly know which cues they respond to, and how those cues map to the final function of the emerging structure. In the case of beavers, we know they respond to the sound of moving water, and that stopping water from rushing results in a working dam. In our robots, we were able to map geometric cues to a final function, which is to enable mobility. That’s where the inspiration came from. In the future, we would also like to expand the types of functional structures our robots can build including dams and levees.”
At present, the autonomous robot has only worked with beanbags of different sizes to simulate various objects. In 10 tests, it moved anywhere from 33 to 170 bags, each time creating a ramp to reach its target location.
“It’s really difficult for robots to work in messy, real-world, outside environments,” Napp continued. “In factories and inside homes the world is structured, and that allows robots to reason about what they should do next. We were looking at how animals solve problems in the wild, and one common approach seems to be that they continuously analyze and modify partially built structures until they fulfill some specific function. We applied this idea to robots, enabling them to assess structures and add material where necessary until their task is complete.”