Learning Linear Policies for Robust Bipedal Locomotion on Terrains with Varying Slopes

TL;DR

This paper presents a simple linear feedback policy learned via a gradient-free algorithm to enable robust bipedal walking on varying slopes and terrains, demonstrating successful simulation and initial hardware transfer.

Contribution

It introduces a lightweight, linear feedback control policy for bipedal robots learned with ARS, capable of handling diverse terrains and behaviors.

Findings

Robust walking on slopes up to 20 degrees in simulation

Ability to walk backwards, step-in-place, and recover from pushes

Initial successful transfer of control policy to hardware

Abstract

In this paper, with a view toward deployment of light-weight control frameworks for bipedal walking robots, we realize end-foot trajectories that are shaped by a single linear feedback policy. We learn this policy via a model-free and a gradient-free learning algorithm, Augmented Random Search (ARS), in the two robot platforms Rabbit and Digit. Our contributions are two-fold: a) By using torso and support plane orientation as inputs, we achieve robust walking on slopes of up to 20 degrees in simulation. b) We demonstrate additional behaviors like walking backwards, stepping-in-place, and recovery from external pushes of up to 120 N. The end result is a robust and a fast feedback control law for bipedal walking on terrains with varying slopes. Towards the end, we also provide preliminary results of hardware transfer to Digit.