A rehabilitation exercise assessment system based on wearable sensors for knee osteoarthritis

Abstract

In order to enable the knee OsteoArthritis (OA) patients to manage their own rehabilitation progress, this study develops a rehabilitation exercise assessment system for knee OA using three wearable sensors mounted on the chest, thigh and shank of working leg. The system derives 51 features from the calculated angles, spectrum, and means of the acceleration signals to judge the exercise type and determine whether their postures are correct or not. After ten subjects performed three kinds of rehabilitation activities, we got 99.29% accuracy for exercise type classification, and 90.14% accuracy for wrong exercise recognition. The experimental results demonstrate that the proposed system can help the physician and patients to monitor the rehabilitation progress efficiently. © 2013 Springer-Verlag Berlin Heidelberg.