Towards Fault-tolerant Quadruped Locomotion with Reinforcement Learning

Modern quadrupedal robots are skilled in navigating through challenging terrains in remote uncontrolled environments with recent advances in reinforcement learning (RL). However, survival in the wild requires not only maneuverability, but also the ability to handle potential critical hardware failur...

Full description

Saved in:
Bibliographic Details
Published in2024 IEEE Conference on Artificial Intelligence (CAI) pp. 1438 - 1441
Main Authors Liu, Dikai, Yin, Jianxiong, See, Simon
Format Conference Proceeding
LanguageEnglish
Published IEEE 25.06.2024
Subjects
Fault tolerance
Fault tolerant systems
Hardware
Lead
n/a
Navigation
Reinforcement learning
Training
Online AccessGet full text

Cover

More Information
Summary:Modern quadrupedal robots are skilled in navigating through challenging terrains in remote uncontrolled environments with recent advances in reinforcement learning (RL). However, survival in the wild requires not only maneuverability, but also the ability to handle potential critical hardware failures. How to grant such ability to quadrupeds with RL is rarely investigated. In this paper, we propose a novel methodology to enable fault tolerance for RL-based quadruped locomotion controller with joint teacher-student framework for fast zero-shot knowledge transfer that can be deployed to a physical robot without any fine-tuning. With no dedicated reward design for gait guidance, the designed simulation and training strategy can be easily added on top of existing RL-based controllers and generalized to unseen situations. Extensive experiments show that our fault-tolerant controller can efficiently lead a quadruped stably when it faces joint failures during locomotion.
DOI:10.1109/CAI59869.2024.00257