AN ANKLE SPRAIN RECOGNITION SYSTEM FOR IDENTIFYING ANKLE SPRAIN MOTION FROM OTHER NORMAL MOTION USING MOTION SENSOR

Abstract

The purpose of this study was to develop an ankle sprain recognition system which identifies ankle sprain motions from other normal motions. Six healthy male subjects performed a total of 600 simulated ankle sprain motions and normal sports motions. Eight motion sensors were attached to cover the whole foot segment to monitor the linear velocity and angular accelerations of the segment. The data obtained from the motion sensor at the medial calcaneus selected to train up the Support Vector Machine (SVM). The trained SVM model was then verified by another 600 trials from other six healthy male subjects. Among the 300 sprain trials, 291 (97.0%) of them were identified correctly. However, there was still a 14.3% false alarm which normal trials being identified as sprain trails. In general, a good accuracy of 91.3% was achieved.