How Legged Robots Learn to Walk

Legged robots take inspiration from animals with the goal of traversing the world in a way that wheeled robots can’t, seamlessly moving around homes, factories, and outdoor terrain while overcoming obstacles. While legs are more versatile than wheels, they also make for systems that are complex to control. MIT researcher Gabriel Margolis will join us to discuss the uses of walking robots and will explain how reinforcement learning is used to develop locomotion skills through trial and error. After the talk, Gabriel will take questions from the audience.