Learning Locomotion on Complex Terrain for Quadrupedal Robots with Foot Position Maps and Stability Rewards
This addresses the challenge of precise and stable movement for quadrupedal robots on varied terrain, representing an incremental improvement over existing methods.
The paper tackled the problem of quadrupedal robot locomotion on complex terrain by introducing a foot position map and stability reward, achieving improved success rates on both in-domain and out-of-domain terrains.
Quadrupedal locomotion over complex terrain has been a long-standing research topic in robotics. While recent reinforcement learning-based locomotion methods improve generalizability and foot-placement precision, they rely on implicit inference of foot positions from joint angles, lacking the explicit precision and stability guarantees of optimization-based approaches. To address this, we introduce a foot position map integrated into the heightmap, and a dynamic locomotion-stability reward within an attention-based framework to achieve locomotion on complex terrain. We validate our method extensively on terrains seen during training as well as out-of-domain (OOD) terrains. Our results demonstrate that the proposed method enables precise and stable movement, resulting in improved locomotion success rates on both in-domain and OOD terrains.