Impedance Matching: Enabling an RL-Based Running Jump in a Quadruped Robot

TL;DR

This paper introduces a frequency-domain impedance matching framework to improve sim-to-real transfer of RL-based control policies, enabling a quadruped robot to perform high jumps and stable walking with enhanced agility.

Contribution

The proposed impedance matching framework reduces the sim-to-real gap, allowing RL policies to achieve dynamic movements like high jumps and fast walking in real quadruped robots.

Findings

Achieved jumps of 55 cm distance and 38 cm height in real robot.

Enabled stable walking at 2 m/s forward/backward and 1 m/s sideways.

Demonstrated one of the highest RL-based jumps in quadruped robots.

Abstract

Replicating the remarkable athleticism seen in animals has long been a challenge in robotics control. Although Reinforcement Learning (RL) has demonstrated significant progress in dynamic legged locomotion control, the substantial sim-to-real gap often hinders the real-world demonstration of truly dynamic movements. We propose a new framework to mitigate this gap through frequency-domain analysis-based impedance matching between simulated and real robots. Our framework offers a structured guideline for parameter selection and the range for dynamics randomization in simulation, thus facilitating a safe sim-to-real transfer. The learned policy using our framework enabled jumps across distances of 55 cm and heights of 38 cm. The results are, to the best of our knowledge, one of the highest and longest running jumps demonstrated by an RL-based control policy in a real quadruped robot. Note that the achieved jumping height is approximately 85% of that obtained from a state-of-the-art trajectory optimization method, which can be seen as the physical limit for the given robot hardware. In addition, our control policy accomplished stable walking at speeds up to 2 m/s in the forward and backward directions, and 1 m/s in the sideway direction.