Zero-Shot Terrain Generalization for Visual Locomotion Policies
This addresses the problem of terrain generalization for legged robots, which is crucial for real-world deployment, though it is incremental as it builds on multi-task reinforcement learning with a novel sensor approach.
The paper tackles the challenge of enabling legged robots to navigate diverse real-world terrains by learning locomotion controllers that generalize across 13 different environments, including stairs and cluttered offices, using raw LiDAR inputs without preprocessing.
Legged robots have unparalleled mobility on unstructured terrains. However, it remains an open challenge to design locomotion controllers that can operate in a large variety of environments. In this paper, we address this challenge of automatically learning locomotion controllers that can generalize to a diverse collection of terrains often encountered in the real world. We frame this challenge as a multi-task reinforcement learning problem and define each task as a type of terrain that the robot needs to traverse. We propose an end-to-end learning approach that makes direct use of the raw exteroceptive inputs gathered from a simulated 3D LiDAR sensor, thus circumventing the need for ground-truth heightmaps or preprocessing of perception information. As a result, the learned controller demonstrates excellent zero-shot generalization capabilities and can navigate 13 different environments, including stairs, rugged land, cluttered offices, and indoor spaces with humans.