IoT-Based System Using IMU Sensor Fusion for Knee Telerehabilitation Monitoring

Abstract

High costs and clinical limitations restrict access to rehabilitation services, especially in low- and middle-income countries. There is a growing need for affordable, home-based solutions that enable continuous remote monitoring of patient rehabilitation progress. This work proposes an Internet of Thing (IoT)-based system for knee movement monitoring during telerehabilitation. The system comprises wearable inertial measurement units (IMUs) integrated into an IoT architecture. This architecture leverages edge and cloud computing to facilitate remote monitoring and real-time feedback. A sensor fusion algorithm was implemented on the edge to estimate knee joint angle, and a cloud-based application was developed to extract kinematic parameters and assess rehabilitation outcomes. The system was implemented using system-on-chip (SoC) technology, allowing embedded signal processing and wireless communication in a compact and low-power design. Three experimental validation tests were conducted: One hardware test evaluating the performance of the proposed sensor fusion algorithm; goniometer-based static test assessing the impact of environmental interference on system accuracy; dynamic test involving rehabilitation exercises to validate system performance against a gold-standard video-based system in the home context. The results demonstrated that the proposed algorithm achieved an optimal trade-off between accuracy, computational efficiency, and resilience to magnetic distortions. The system showed acceptable accuracy, with an average root mean square error (RMSE) ranging from 3.08° to 6.43° across all exercises. These results are consistent with the current state of the art, highlighting the system’s potential for objective and remote monitoring of knee movement in home-based rehabilitation.