DreamWaQ++: Obstacle-aware quadrupedal locomotion with resilient multi-modal reinforcement learning

1Urban Robotics Lab., KAIST, 2SPARK Lab., MIT

Abstract

Quadrupedal robots hold promising potential for applications in navigating cluttered environments with resilience akin to their animal counterparts. However, their floating base configuration makes them vulnerable to real-world uncertainties, yielding substantial challenges in their locomotion control.

Deep reinforcement learning has become one of the plausible alternatives for realizing a robust locomotion controller. However, the approaches that rely solely on proprioception sacrifice collision-free locomotion because they require front-feet contact to detect the presence of stairs to adapt the locomotion gait. Meanwhile, incorporating exteroception necessitates a precisely modeled map observed by exteroceptive sensors over a period of time.

Therefore, this work proposes a novel method to fuse proprioception and exteroception featuring a resilient multi-modal reinforcement learning. The proposed method yields a controller that showcases agile locomotion performance on a quadrupedal robot over a myriad of real-world courses, including rough terrains, steep slopes, and high-rise stairs, while retaining its robustness against out-of-distribution situations.

Movie S2

Stair-climbing race between DreamWaQ++, DreamWaQ, and the built-in controller.

Movie S3

Resilient stair-climbing on various stairs with different rise and run propoerties.

Movie S4

Emergent probing behavior on uncertain terrains.

Movie S5

Adaptation in out-of distribution situations.

Movie S6

Locomotion performance comparison on extreme slopes.

Movie S7

Scalability of DreamWaQ++ on various robots.

Movie S8

Demonstration of DreamWaQ++ on extreme locomotion tasks such as overcoming high obstacles.

Movie S9

Comparison of the baseline DreamWaQ on different robot platforms.

Movie S8=10

In-depth experiment using of Movie S2 with asynchronous race.

BibTeX

@article{nahrendra2024obstacle,
      title={Obstacle-Aware Quadrupedal Locomotion With Resilient Multi-Modal Reinforcement Learning},
      author={Nahrendra, I and Yu, Byeongho and Oh, Minho and Lee, Dongkyu and Lee, Seunghyun and Lee, Hyeonwoo and Lim, Hyungtae and Myung, Hyun},
      journal={arXiv preprint arXiv:2409.19709},
      year={2024}
    }