Adaptive Energy Regularization for Autonomous Gait Transition and Energy-Efficient Quadruped Locomotion

2024-03-29

arXiv.org

AI Summary

Quadrupedal gait learning suffers from low energy efficiency, reliance on predefined gaits, and complex reward engineering. Method: We propose a velocity-adaptive, simplified center-of-energy reward mechanism—first directly embedding biologically inspired energy minimization into the reinforcement learning reward function. Built upon the PPO algorithm, the method is trained in IsaacGym and deployed on physical ANYmal-C and Unitree Go1 robots. Without prior gait knowledge, it autonomously selects optimal gaits across the full speed spectrum (e.g., tetrapod walking at low speeds, diagonal trotting at high speeds) and enables smooth gait transitions. Results: Experiments demonstrate stable velocity tracking in both simulation and real-world deployment, with significantly lower energy consumption than conventional multi-stage reward approaches. The method achieves superior energy efficiency and generalizability, validating its effectiveness for adaptive, efficient quadrupedal locomotion.

Technology Category

• Intelligent Robots: Motion and Path Planning
• Machine Learning: Imitation Learning & Inverse Reinforcement Learning
• Search and Optimization: Learning to Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Energy management for devices in mobile Web and WoT environments
• Search and Retrieval-Augmented AI: Agentic search
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

In reinforcement learning for legged robot locomotion, crafting effective reward strategies is crucial. Pre-defined gait patterns and complex reward systems are widely used to stabilize policy training. Drawing from the natural locomotion behaviors of humans and animals, which adapt their gaits to minimize energy consumption, we propose a simplified, energy-centric reward strategy to foster the development of energy-efficient locomotion across various speeds in quadruped robots. By implementing an adaptive energy reward function and adjusting the weights based on velocity, we demonstrate that our approach enables ANYmal-C and Unitree Go1 robots to autonomously select appropriate gaits, such as four-beat walking at lower speeds and trotting at higher speeds, resulting in improved energy efficiency and stable velocity tracking compared to previous methods using complex reward designs and prior gait knowledge. The effectiveness of our policy is validated through simulations in the IsaacGym simulation environment and on real robots, demonstrating its potential to facilitate stable and adaptive locomotion.

Problem

• Develop energy-efficient locomotion for quadruped robots.
• Autonomous gait transition based on speed adaptation.
• Simplify reward strategies in reinforcement learning for legged robots.

Innovation

• Adaptive energy reward function for gait transition
• Velocity-based weight adjustment for energy efficiency
• Simplified reward strategy for quadruped locomotion