AllGaits: Learning All Quadruped Gaits and Transitions

TL;DR

This paper introduces a unified deep reinforcement learning framework that enables quadruped robots to learn, switch between, and optimize all common gaits and transitions for energy-efficient locomotion across different speeds.

Contribution

It presents a novel DRL-based policy controlling a CPG system to produce and transition between all quadruped gaits, including energy optimization and robustness to leg failures.

Findings

The framework can generate all 9 typical quadruped gaits.

Gait transitions can be instantaneously controlled at any velocity.

The approach demonstrates robustness to leg failures and generalizes to unseen gaits.

Abstract

We present a framework for learning a single policy capable of producing all quadruped gaits and transitions. The framework consists of a policy trained with deep reinforcement learning (DRL) to modulate the parameters of a system of abstract oscillators (i.e. Central Pattern Generator), whose output is mapped to joint commands through a pattern formation layer that sets the gait style, i.e. body height, swing foot ground clearance height, and foot offset. Different gaits are formed by changing the coupling between different oscillators, which can be instantaneously selected at any velocity by a user. With this framework, we systematically investigate which gait should be used at which velocity, and when gait transitions should occur from a Cost of Transport (COT), i.e. energy-efficiency, point of view. Additionally, we note how gait style changes as a function of locomotion speed for each gait to keep the most energy-efficient locomotion. While the currently most popular gait (trot) does not result in the lowest COT, we find that considering different co-dependent metrics such as mean base velocity and joint acceleration result in different `optimal' gaits than those that minimize COT. We deploy our controller in various hardware experiments, showing all 9 typical quadruped animal gaits, and demonstrate generalizability to unseen gaits during training, and robustness to leg failures. Video results can be found at https://youtu.be/OLoWSX_R868.