Personalized Lower-limb Exoskeleton Assistance via Preference-based Bayesian Optimization

2026-08-09

AI Summary

This study addresses the challenge of dynamically adapting lower-limb exoskeletons to individual user preferences to simultaneously optimize assistance efficiency and comfort. Existing preference-driven parameter tuning methods suffer from low sample efficiency and high user burden due to extensive human-in-the-loop interactions. To overcome this, the authors propose a preference-based Bayesian optimization framework (PbBO) that incorporates a prior over sampling distributions, dramatically improving sample efficiency. The approach converges to personalized control parameters within only 20 iterations and integrates with a hierarchical controller for real-time torque delivery. Experimental results demonstrate that the optimized parameters reduce metabolic cost by 14.5%–15.4%, lower heart rate by 6.3%–7.6%, and decrease muscle activation by 6.7%–31.5% across both treadmill and outdoor walking conditions, with a user preference prediction accuracy of 90.7%.

Abstract

A significant challenge in exoskeleton robotics is the need to dynamically adapt control profiles to individual motion preferences, thereby ensuring both efficient and comfortable assistance. Currently, since user experience can serve as a comprehensive metric for evaluating the effectiveness of assistance, user preference-based optimization methods have been widely studied for parameter tuning. However, the existing methods rely heavily on extensive human-robot online interactions and suffer from slow optimization speed, which not only induces user fatigue but also compromises optimization effectiveness. Therefore, this paper aims to explore an efficient preference-based optimization framework for personalized exoskeleton assistance that can learn optimal parameters with minimal interaction. We propose a preference-based Bayesian optimization (PbBO) approach that can improve sample efficiency by leveraging knowledge about the sampling distribution of candidate sets. For optimizing six control parameters, PbBO can converge to user-preferred parameters with 90.7% validation accuracy via 20 iterations. Moreover, the hierarchical controller is designed to generate personalized torque for different tasks and achieve interaction torque tracking in real time. The results of treadmill and outdoor experiments demonstrate that the optimized parameters can reduce metabolic rate by 14.5%-15.4%, heart rate by 6.3%-7.6%, and muscle activation by 6.7%-31.5% compared to unassisted walking.

Problem

• personalized exoskeleton assistance
• preference-based optimization
• sample efficiency
• human-robot interaction
• control parameter tuning

Innovation

• Preference-based Bayesian Optimization
• Personalized Exoskeleton Assistance
• Sample-efficient Optimization
• Hierarchical Control
• Human-in-the-loop Learning