Visual Whole-Body Control for Legged Loco-Manipulation

TL;DR

This paper introduces Visual Whole-Body Control (VBC), a framework enabling legged robots with arms to perform mobile manipulation using visual inputs, combining low-level and high-level policies trained in simulation for real-world deployment.

Contribution

The paper presents a novel VBC framework that integrates visual perception with autonomous whole-body control for legged robots, enhancing manipulation capabilities.

Findings

Significant improvement in object pickup success across diverse scenarios.

Effective Sim2Real transfer of policies from simulation to real robots.

Enhanced manipulation workspace by controlling legs and arms simultaneously.

Abstract

We study the problem of mobile manipulation using legged robots equipped with an arm, namely legged loco-manipulation. The robot legs, while usually utilized for mobility, offer an opportunity to amplify the manipulation capabilities by conducting whole-body control. That is, the robot can control the legs and the arm at the same time to extend its workspace. We propose a framework that can conduct the whole-body control autonomously with visual observations. Our approach, namely Visual Whole-Body Control(VBC), is composed of a low-level policy using all degrees of freedom to track the body velocities along with the end-effector position, and a high-level policy proposing the velocities and end-effector position based on visual inputs. We train both levels of policies in simulation and perform Sim2Real transfer for real robot deployment. We perform extensive experiments and show significant improvements over baselines in picking up diverse objects in different configurations (heights, locations, orientations) and environments.