Iteratively Learning Muscle Memory for Legged Robots to Master Adaptive and High Precision Locomotion

2025-07-18

AI Summary

To address low trajectory tracking accuracy, slow adaptation, and high online computational burden in legged robots under unmodeled dynamics, external disturbances, and varying environments (e.g., speed, terrain, gravity), this paper proposes a scalable hierarchical control framework integrating iterative learning control (ILC) with a bio-inspired generalized torque library. The torque library stores optimal joint torque profiles—both periodic and aperiodic—offline, enabling zero-shot, cross-condition rapid transfer. ILC operates online to compensate for modeling errors and disturbances, enhanced by hybrid-system trajectory optimization and real-time learning for improved robustness. Experimental validation on Cassie and A1 platforms demonstrates an 85% reduction in joint tracking error within seconds, a control update frequency 30× higher than state-of-the-art methods, and successful execution of ramp walking and dynamic terrain adaptation.

Technology Category

• Intelligent Robots: Behavior Learning & Control
• Machine Learning: Imitation Learning & Inverse Reinforcement Learning
• Search and Optimization: Learning to Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

This paper presents a scalable and adaptive control framework for legged robots that integrates Iterative Learning Control (ILC) with a biologically inspired torque library (TL), analogous to muscle memory. The proposed method addresses key challenges in robotic locomotion, including accurate trajectory tracking under unmodeled dynamics and external disturbances. By leveraging the repetitive nature of periodic gaits and extending ILC to nonperiodic tasks, the framework enhances accuracy and generalization across diverse locomotion scenarios. The control architecture is data-enabled, combining a physics-based model derived from hybrid-system trajectory optimization with real-time learning to compensate for model uncertainties and external disturbances. A central contribution is the development of a generalized TL that stores learned control profiles and enables rapid adaptation to changes in speed, terrain, and gravitational conditions-eliminating the need for repeated learning and significantly reducing online computation. The approach is validated on the bipedal robot Cassie and the quadrupedal robot A1 through extensive simulations and hardware experiments. Results demonstrate that the proposed framework reduces joint tracking errors by up to 85% within a few seconds and enables reliable execution of both periodic and nonperiodic gaits, including slope traversal and terrain adaptation. Compared to state-of-the-art whole-body controllers, the learned skills eliminate the need for online computation during execution and achieve control update rates exceeding 30x those of existing methods. These findings highlight the effectiveness of integrating ILC with torque memory as a highly data-efficient and practical solution for legged locomotion in unstructured and dynamic environments.

Problem

• Enhances trajectory tracking for legged robots under disturbances
• Extends learning control to nonperiodic tasks for better generalization
• Reduces online computation by storing learned control profiles

Innovation

• Integrates Iterative Learning Control with torque library
• Combines physics-based model with real-time learning
• Generalized torque library enables rapid adaptation