A wireless evaluation system for in-home physical rehabilitation

This paper describes an evaluation system, which can visually display and evaluate the user's performance in his physical rehabilitation. The system is composed of a body sensor network with low cost MEMS inertial measurement units (IMUs), which is capable of measuring the orientations of all body joints. Every joints orientation is represented relative to its parent joint according to human skeleton and they are used as features to evaluate. The system uses OGRE (Object-Oriented Graphics Rendering Engine) to reconstruct and display the three-dimensional motion of a human body in virtual scenes on computers. To evaluate the user's performance of rehabilitation, the body motion data are analyzed with Dynamic Time Warping (DTW) algorithm, and are compared with the standard motion sequences from health experts. The DTW algorithm can measure the distance between sequences of varying speed, and avoid the problem of different speed of doing a same motion.