3.0

credits

Average Course Rating

Sensorimotor coordination in robotics is a fundamental topic that studies the coordination between a robot's sensory and motor systems, enabling reliable interaction between robots and the physical world. The impressive movement we saw robots perform today is the accumulation of sensorimotor coordination techniques developed over the past three decades, encompassing aspects of perception, planning, control, decision-making, and a recent surge in embodied intelligence. This course will guide students through these aspects, with a deep dive into how advances in machine learning bring robots' sensorimotor capabilities to the next level. Specifically, the topics will cover supervised/unsupervised learning, imitation learning, reinforcement learning, model learning, and adaptation between simulation and real environments (sim2real, real2sim, and real2sim2real). Each course will also engage students in discussions on the necessity of data-driven approaches in different aspects of robotics, as well as what has and has not been solved in embodied AI. Required Course Background: coursework in machine learning or deep learning. Students may receive credit for only one of EN.601.495/EN.601.695.