Learning Stable and Energetically Economical Walking with RAMone

TL;DR

This paper employs reinforcement learning to optimize control parameters of the RAMone robot, achieving stable and energy-efficient walking across different speeds by tuning hybrid zero dynamics and motor controller gains.

Contribution

It introduces a reinforcement learning framework for optimizing control parameters of a bipedal robot to enhance stability and energy efficiency during walking.

Findings

Successful optimization of control parameters for stable walking

Demonstrated energy savings across various walking speeds

Validated the approach on the RAMone robot platform

Abstract

In this paper, we optimize over the control parameter space of our planar-bipedal robot, RAMone, for stable and energetically economical walking at various speeds. We formulate this task as an episodic reinforcement learning problem and use Covariance Matrix Adaptation. The parameters we are interested in modifying include gains from our Hybrid Zero Dynamics style controller and from RAMone's low-level motor controllers.