Learning Realistic Joint Space Boundaries for Range of Motion Analysis of Healthy and Impaired Human Arms

2023-11-17

IEEE-RAS International Conference on Humanoid Robots

AI Summary

Existing upper-limb range-of-motion (RoM) modeling approaches neglect anatomical constraints and rely on oversimplified box-shaped boundaries. To address this, we propose a data-driven, anatomy-consistent RoM modeling framework. We introduce the first RoM boundary learning method based on one-class support vector machines (OC-SVM), integrating motion-capture data to explicitly encode inter-individual variability, joint coupling, self-collision avoidance, and neuromuscular constraints. Furthermore, we design a clinically interpretable Injury Index (II) that objectively differentiates healthy from pathological upper limbs—such as hemiparetic or stroke-simulated arms—with statistical significance (p < 0.01). Experimental evaluation demonstrates that our model achieves significantly higher RoM modeling accuracy than conventional methods. Validation via physics-based anatomical constraint simulation confirms its biomechanical plausibility and anatomical fidelity.

Abstract

A realistic human kinematic model that satisfies anatomical constraints is essential for human-robot interaction, biomechanics and robot-assisted rehabilitation. Modeling realistic joint constraints, however, is challenging as human arm motion is constrained by joint limits, inter- and intra-joint dependencies, self-collisions, individual capabilities and muscular or neurological constraints which are difficult to represent. Hence, physicians and researchers have relied on simple box-constraints, ignoring important anatomical factors. In this paper, we propose a data-driven method to learn realistic anatomically constrained upper-limb range of motion (RoM) boundaries from motion capture data. This is achieved by fitting a one-class support vector machine to a dataset of upper-limb joint space exploration motions with an efficient hyper-parameter tuning scheme. Our approach outperforms similar works focused on valid RoM learning. Further, we propose an impairment index (II) metric that offers a quantitative assessment of capability/impairment when comparing healthy and impaired arms. We validate the metric on healthy subjects physically constrained to emulate hemiplegia and different disability levels as stroke patients.

Problem

• Model realistic joint constraints for human arm motion
• Learn anatomically accurate range of motion boundaries
• Quantify impairment levels in healthy vs impaired arms

Innovation

• Data-driven method for realistic RoM boundaries
• One-class SVM with hyper-parameter tuning
• Impairment index metric for quantitative assessment