Not Only Rewards But Also Constraints: Applications on Legged Robot Locomotion

TL;DR

This paper introduces a reinforcement learning framework that incorporates both rewards and constraints to efficiently train legged robot controllers, reducing reward engineering effort and improving interpretability.

Contribution

The work proposes a novel RL framework with two constraint types and an efficient optimization algorithm, enabling simpler and more generalizable controller training for complex robots.

Findings

Controllers trained with fewer reward terms perform well on challenging terrains.

The framework reduces reward engineering complexity and tuning effort.

Both simulation and real-world experiments validate the approach.

Abstract

Several earlier studies have shown impressive control performance in complex robotic systems by designing the controller using a neural network and training it with model-free reinforcement learning. However, these outstanding controllers with natural motion style and high task performance are developed through extensive reward engineering, which is a highly laborious and time-consuming process of designing numerous reward terms and determining suitable reward coefficients. In this work, we propose a novel reinforcement learning framework for training neural network controllers for complex robotic systems consisting of both rewards and constraints. To let the engineers appropriately reflect their intent to constraints and handle them with minimal computation overhead, two constraint types and an efficient policy optimization algorithm are suggested. The learning framework is applied to train locomotion controllers for several legged robots with different morphology and physical attributes to traverse challenging terrains. Extensive simulation and real-world experiments demonstrate that performant controllers can be trained with significantly less reward engineering, by tuning only a single reward coefficient. Furthermore, a more straightforward and intuitive engineering process can be utilized, thanks to the interpretability and generalizability of constraints. The summary video is available at https://youtu.be/KAlm3yskhvM.