Wearable IoT System for Hand Function Assessment Based on EMG Signals

Abstract

Evaluating hand function presents a significant challenge in the realm of remote rehabilitation, particularly when highlighting the need for comfort and practicality in wearable devices. This research introduces an innovative wearable device-based Internet of Things (IoT) system, specifically designed for the assessment of hand function, with a focus on a wearable wristband. The system, enhanced by cloud technology, offers comprehensive solutions for remote health management and therapeutic services. Firstly, it uses electromyography (EMG) signals from the arm to assess hand function. By employing sophisticated classification and regression models, this system can automatically identify user gestures and accurately measure grip strength. Additionally, the integration of additional sensor data ensures that the system fulfills essential criteria for hand function assessment. Leaving conventional grip strength classification methods, this study explored four distinct regression models to accurately represent the grip strength curve. The findings reveal that the Random Forest Regression (RFR) model is the most effective, achieving an R2 score of 0.9563 on the test data. This significant outcome not only confirms the practicality of the wearable wristband, which relies on EMG signals, but also underscores the potential of the IoT system in assessing hand function.