Learning Low-Frequency Motion Control for Robust and Dynamic Robot Locomotion

TL;DR

This paper demonstrates that a quadruped robot can achieve robust, dynamic locomotion at low control frequencies (as low as 8 Hz) using deep reinforcement learning, challenging the assumption that higher control frequencies are necessary for robustness.

Contribution

The study shows that low-frequency control policies can be effective for robust locomotion and are less sensitive to actuation delays, enabling successful sim-to-real transfer without dynamics randomization.

Findings

Low-frequency policies maintain robustness and agility.

Effective sim-to-real transfer without dynamics randomization.

Low-frequency control reduces sensitivity to actuation delays.

Abstract

Robotic locomotion is often approached with the goal of maximizing robustness and reactivity by increasing motion control frequency. We challenge this intuitive notion by demonstrating robust and dynamic locomotion with a learned motion controller executing at as low as 8 Hz on a real ANYmal C quadruped. The robot is able to robustly and repeatably achieve a high heading velocity of 1.5 m/s, traverse uneven terrain, and resist unexpected external perturbations. We further present a comparative analysis of deep reinforcement learning (RL) based motion control policies trained and executed at frequencies ranging from 5 Hz to 200 Hz. We show that low-frequency policies are less sensitive to actuation latencies and variations in system dynamics. This is to the extent that a successful sim-to-real transfer can be performed even without any dynamics randomization or actuation modeling. We support this claim through a set of rigorous empirical evaluations. Moreover, to assist reproducibility, we provide the training and deployment code along with an extended analysis at https://ori-drs.github.io/lfmc/.