Intelligent System for Upper Limb Motor Assessment Using Inertial Sensors and Machine Learning for Telerehabilitation Therapies

Neurological injuries can lead to a significant disability, particularly in upper limb motor function, impairing patients' ability to perform Activities of Daily Living (ADLs). Telerehabilitation has emerged as a promising solution for remote patient monitoring and rehabilitation. Hence, there is a need to develop technologies to facilitate the patient assessment process remotely. This study presents an intelligent evaluation system for assessing upper limb motor function using three Inertial Measurement Units (IMUs). The upper limb joint trajectories are used as inputs to a machine learning model to recognize twelve activities based on ADLs and functional movements. Furthermore, a trajectory assessment is conducted comparing the trajectories performed by patients and non-disabled users based on similarity indexes calculated using the Dynamic Time Warping (DTW) distance. Finally, the similarity indexes are leveraged to estimate the patients ' degree of impairment, classifying their motor disabilities into mild or moderate. The system has been evaluated by 31 survivors of neurological injury with motor impairment and 21 individuals from a control group. The results demonstrate the feasibility of using the system to recognize different activities and to assess upper limb motor function. This approach has the potential to enhance the monitoring and rehabilitation of stroke patients in a home setting.