Guided Reinforcement Learning for Omnidirectional 3D Jumping in Quadruped Robots

TL;DR

This paper presents a guided reinforcement learning method for quadruped robots to perform omnidirectional 3D jumping efficiently and safely by integrating physical models with learning.

Contribution

It introduces a novel guided RL approach combining Bezier curves and UARM model for explainable, efficient jumping control in quadruped robots.

Findings

The approach outperforms existing methods in simulation and real-world tests.

It reduces training time and improves motion predictability.

The method enhances safety certification of robot jumps.

Abstract

Jumping poses a significant challenge for quadruped robots, despite being crucial for many operational scenarios. While optimisation methods exist for controlling such motions, they are often time-consuming and demand extensive knowledge of robot and terrain parameters, making them less robust in real-world scenarios. Reinforcement learning (RL) is emerging as a viable alternative, yet conventional end-to-end approaches lack efficiency in terms of sample complexity, requiring extensive training in simulations, and predictability of the final motion, which makes it difficult to certify the safety of the final motion. To overcome these limitations, this paper introduces a novel guided reinforcement learning approach that leverages physical intuition for efficient and explainable jumping, by combining B\'ezier curves with a Uniformly Accelerated Rectilinear Motion (UARM) model. Extensive simulation and experimental results clearly demonstrate the advantages of our approach over existing alternatives.