This Low-Cost Quadruped Robot Can Traverse Almost Any Obstacle
Developed by Carnegie Mellon and UC Berkeley engineers, this system enables small legged robots to maneuver in challenging environments.
Engineers from Carnegie Mellon University's School of Computer Science and the University of California, Berkeley have developed a small, low-cost quadruped robot that can navigate nearly any obstacle and across multiple types of terrain. The robot can easily walk up and down stairs, traverse rocky, slippery and uneven ground, and even walk over gaps.
"Empowering small robots to climb stairs and handle a variety of environments is crucial to developing robots that will be useful in people's homes as well as search-and-rescue operations," explains Deepak Pathak, an assistant professor at the Robotics Institute. "This system creates a robust and adaptable robot that could perform many everyday tasks."
The researchers tested the robot on uneven stairs and hillsides at public parks, along with its ability to walk across stepping stones and over slippery surfaces. The robot gained its impressive ability to quickly adapt and master the terrain using its vision and a small onboard computer. Most robots with the same navigation abilities typically use cameras to create a map of their surroundings, which it uses to plan movements before executing them.
The team, on the other hand, chose to use 4,000 clones of the robot in a simulator, where it practiced walking and climbing on challenging terrain. The simulator's speed allowed the robot to gain six years of experience in a single day of testing. The simulator also stored the robot's motor skills learned during training in a neural network, which the engineers ported over to the real robot without any additional programming of its movements. To that end, what the robot sees determines its movements, even without any specifications of how its legs should move. This breakthrough could help improve the limitations of legged robot navigation in the near future.