Bipedalism for Quadrupedal Robots: Versatile Loco-Manipulation through Risk-Adaptive Reinforcement Learning

TL;DR

This paper introduces a risk-adaptive reinforcement learning framework enabling quadrupedal robots to walk on their hind legs, thereby enhancing their interaction capabilities and robustness in complex environments.

Contribution

It presents a novel risk-adaptive RL approach for bipedal locomotion in quadrupedal robots, balancing safety and performance during unstable tasks.

Findings

Superior performance over baselines in simulation

Successful real-world deployment on Unitree Go2 robot

Enabled versatile tasks like cart pushing and obstacle probing

Abstract

Loco-manipulation of quadrupedal robots has broadened robotic applications, but using legs as manipulators often compromises locomotion, while mounting arms complicates the system. To mitigate this issue, we introduce bipedalism for quadrupedal robots, thus freeing the front legs for versatile interactions with the environment. We propose a risk-adaptive distributional Reinforcement Learning (RL) framework designed for quadrupedal robots walking on their hind legs, balancing worst-case conservativeness with optimal performance in this inherently unstable task. During training, the adaptive risk preference is dynamically adjusted based on the uncertainty of the return, measured by the coefficient of variation of the estimated return distribution. Extensive experiments in simulation show our method's superior performance over baselines. Real-world deployment on a Unitree Go2 robot further demonstrates the versatility of our policy, enabling tasks like cart pushing, obstacle probing, and payload transport, while showcasing robustness against challenging dynamics and external disturbances.