Two-Stage Learning of Highly Dynamic Motions with Rigid and Articulated Soft Quadrupeds

TL;DR

This paper introduces a two-stage learning framework combining evolution strategies and deep reinforcement learning to enable quadrupedal robots, especially soft articulated ones, to perform complex dynamic motions efficiently from scratch.

Contribution

It presents a novel two-stage learning approach that simplifies reward shaping and effectively handles the control challenges of soft articulated quadrupeds.

Findings

Successfully learned complex motions like pronking and back-flipping.

Reduced need for predefined reference motions and extensive reward shaping.

Enhanced control of soft quadrupeds with high adaptability.

Abstract

Controlled execution of dynamic motions in quadrupedal robots, especially those with articulated soft bodies, presents a unique set of challenges that traditional methods struggle to address efficiently. In this study, we tackle these issues by relying on a simple yet effective two-stage learning framework to generate dynamic motions for quadrupedal robots. First, a gradient-free evolution strategy is employed to discover simply represented control policies, eliminating the need for a predefined reference motion. Then, we refine these policies using deep reinforcement learning. Our approach enables the acquisition of complex motions like pronking and back-flipping, effectively from scratch. Additionally, our method simplifies the traditionally labour-intensive task of reward shaping, boosting the efficiency of the learning process. Importantly, our framework proves particularly effective for articulated soft quadrupeds, whose inherent compliance and adaptability make them ideal for dynamic tasks but also introduce unique control challenges.