Smart Sensor-Driven Gait Rehabilitation Walker Using Machine Learning for Predictive Home-Based Therapy

Area of Science:

• Biomedical Engineering
• Rehabilitation Technology
• Gait Analysis

Background:

• Abnormal gait is a significant clinical challenge, especially in aging populations.
• Existing rehabilitation methods often lack real-time monitoring and adaptive feedback.
• There is a need for low-cost, accessible solutions for gait rehabilitation and monitoring.

Purpose of the Study:

• To develop and evaluate a modular, low-cost Smart Rehabilitation Walker (SRW) with integrated multimodal sensing and haptic feedback.
• To assess the SRW's effectiveness in improving gait biomechanics and neuromuscular activity.
• To validate the SRW's analytical pipeline for gait rehabilitation trajectory modeling.

Main Methods:

• The SRW integrates force-sensing resistors, inertial measurement units, and surface electromyography (sEMG).
• A closed-loop feedback architecture provides real-time corrective intervention.
• A 15-day pilot study with rheumatoid arthritis patients and validation with an amyotrophic lateral sclerosis dataset were conducted.

Main Results:

• Significant reduction in inter-limb load asymmetry (79.26% decrease in Force Symmetry Index).
• Substantial increase in neuromuscular activation (ΔEMG = 4.28, p < 0.001).
• Gaussian Process Regression model achieved high predictive accuracy (R² ≈ 0.9) for rehabilitation trajectories.

Conclusions:

• Sensor-enabled assistive walkers offer a scalable platform for quantitative gait rehabilitation.
• The SRW system enables adaptive feedback and long-term mobility monitoring.
• This technology holds promise for improving gait in individuals with neuromuscular and musculoskeletal disorders.