Quadrupedal robots trot into the wild

摘要

Deep reinforcement learning enables quadruped robots to traverse challenging natural environments using only proprioception.