Development of a Multi-Point Foot Pressure Sensing System for In-Home Rehabilitation Monitoring

Abstract

Home-based rehabilitation post orthopedic lower-extremity interventions is difficult to monitor objectively, as clinic-based gait labs are expensive, inaccessible for frequent use, and impractical for everyday settings. Patients also often have no clear feedback on whether they are performing prescribed exercises correctly, which may slow recovery and increase the risk of re-injury. This study aims to develop and evaluate a low-cost, wearable insole system capable of measuring plantar pressure and automatically identifying common rehabilitation exercises in home environments. The proposed system integrates six force-sensitive resistors placed at key plantar regions with an ESP32-based acquisition unit that samples at 50 Hz, filters the signals using a second-order Butterworth low-pass filter, and transmits data via Bluetooth to a custom Android application for real-time visualization and storage. Twenty participants completed split squats, heel raises, and half squats, each performed for three sets of ten repetitions while plantar pressure data were continuously recorded. Features including mean and average peak regional pressure, center-of-pressure progression, symmetry indices, and repetition duration were extracted for analysis. Statistical tests revealed significant differences between exercise phases and across conditions, with inter-cycle variability below 7%, indicating high repeatability. A machine-learning classifier trained on these features achieved 93% accuracy in distinguishing Standing, Heel Raise, Half Squat, and Split Squat. These results show that the insole can reliably quantify dynamic plantar loading and exercise technique, enabling patients to check whether they are performing exercises correctly and supporting clinicians with objective remote monitoring.

Video