Quantitative and qualitative assessment of assisted strength exercising

Abstract

Human motion detection and tracking systems, which are not video camera based, are becoming more and more popular in rehabilitation, sportsmen tracking or elderly monitoring. The most commonly used motion sensors in these systems are accelerometers and gyroscopes, which are used along with other sensors as a part of wearable wireless body area network (WBAN). In this work we present an algorithm for closed-loop assisted exercising based on Finite State Machine (FSM). The algorithm can be used either for rehabilitation purposes or physical exercise training. As inputs, the algorithm uses real time signals acquired from an accelerometer and a gyroscope of a sensor node in a WBAN. For testing purposes, signals were recorded from 13 healthy subjects during three different strength training exercises (lateral raise, inner-bicep curl, and seated shoulder press) while the sensor node was attached on the wrist of a subject's dominant arm. Qualitative and quantitative assessment of the accelerometer and gyroscope signals was made. Results of tests confirm that assisted exercising using on-line feedback enable more precise movement control closer to prescribed pattern in time and intensity. The proposed FSM based algorithm is suitable for implementation on embedded systems. One of the potential applications of the proposed algorithm is in e-health systems such as HeartWays system providing advanced solutions for supporting cardiac patients in rehabilitation.