Bipedalism for Quadrupedal Robots: Versatile Loco-Manipulation through Risk-Adaptive Reinforcement Learning
This work addresses the challenge of enabling quadrupedal robots to perform manipulation tasks without compromising locomotion, which is incremental as it adapts existing RL methods to a new robotic configuration.
The paper tackled the problem of loco-manipulation in quadrupedal robots by introducing bipedalism to free front legs for versatile interactions, using a risk-adaptive reinforcement learning framework that achieved superior performance in simulation and real-world tasks like cart pushing and payload transport on a Unitree Go2 robot.
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.