Development of Rehabilitation Applications by Using Wearable Sensors

Abstract

As aging population becomes a major issue in a number of countries, more medical services are increased. Wearable sensor will substitute for the role of healthcare providers to accommodate increasing requirements of rehabilitation which has characteristics of labor- intensive and time-consuming. We chose two wearable sensors such as 6 degree of freedom inertial measurement unit (6-DOF IMU) and surface electromyography (SEMG) sensor, and proposed rehabilitation applications related to early detection of disorders and home rehabilitation. First, we proposed a novel system for monitoring in-sleep stroke by detecting abnormal activity ratio of the left and right arms from wearable the 6-DOF IMU sensor which consists of an accelerometer and gyroscope sensor. We extracted multiple features for consists of an accelerometer and gyroscope sensor. We extracted multiple features for and detected stroke by sliding window method with stroke thresholds according to the each feature. The system discriminated stroke 75.48% by the accelerometer sensor and 97.12% by the gyroscope sensor in sleep data of the stroke patients with hemiparesis. Second, we tested a feasibility of the SEMG pattern recognition for training of activity daily life. We experimented from simple motions to complicated motions considering variables such as time, electrode position and person change. The results showed that the SEMG pattern recognition is largely influenced by the three variables because of structural problems in the muscle and the SEMG sensor. We concluded that the SEMG is appropriate in simple application such as co-contraction EMG detecting whether a muscle is activated. ⓒ 2014 DGIST