Deep Whole-Body Control: Learning a Unified Policy for Manipulation and Locomotion

TL;DR

This paper introduces a reinforcement learning-based unified control policy for legged robots with arms, enabling dynamic manipulation and locomotion without hierarchical decoupling, and demonstrates its effectiveness in various tasks.

Contribution

It presents a novel reinforcement learning approach with Regularized Online Adaptation and Advantage Mixing for whole-body control of legged manipulators, addressing the Sim2Real gap and coordination challenges.

Findings

Unified policy enables dynamic, agile behaviors.

Effective transfer from simulation to real-world robots.

Improved coordination over hierarchical control methods.

Abstract

An attached arm can significantly increase the applicability of legged robots to several mobile manipulation tasks that are not possible for the wheeled or tracked counterparts. The standard hierarchical control pipeline for such legged manipulators is to decouple the controller into that of manipulation and locomotion. However, this is ineffective. It requires immense engineering to support coordination between the arm and legs, and error can propagate across modules causing non-smooth unnatural motions. It is also biological implausible given evidence for strong motor synergies across limbs. In this work, we propose to learn a unified policy for whole-body control of a legged manipulator using reinforcement learning. We propose Regularized Online Adaptation to bridge the Sim2Real gap for high-DoF control, and Advantage Mixing exploiting the causal dependency in the action space to overcome local minima during training the whole-body system. We also present a simple design for a low-cost legged manipulator, and find that our unified policy can demonstrate dynamic and agile behaviors across several task setups. Videos are at https://maniploco.github.io