A Miniature Sensor System for Precise Hand Position Monitoring

This paper presents a sensor system to estimate wrist and elbow positions developing two wearable miniature motion sensors. The developed sensors consist of a processor, a motion sensor, which comprise a 3-D accelerometer, a 3-D gyroscope, a 3-D magnetometer, as well as a transceiver operating at industrial, scientific, and medical band frequency of 2.4 GHz. By proposing a modified complementary filter alongside employing hand kinematic problem, wrist and elbow positions for three different exercises are estimated. In order to determine the accuracy and precision of the developed system, the estimated positions are compared to an optical tracking system. The obtained results illustrate root-mean-square error (RMSE) of these two exercises for positions of x-, y-, and z-axes less than 0.4, 0.59, and 0.29 cm, respectively. Also, the RMSEs of joint angles are obtained less than 1.64 degrees in these exercises. These results show a 11%-23% improvement for estimating these positions than Madgwick filter and prove that the proposed system can be used as an accurate position feedback for monitoring hand joint movement in rehabilitation exercises.