관성센서를 이용한 앉은 자세 분류에 최적 센서 위치

초록

ABSTRACT Incorrect sitting posture causes many diseases. For example, spine repletion, cyphosis and disc are caused by incorrect sitting posture. It is not easy to maintain correct sitting posture for a long time. So, various automatic system which help maintain correct sitting posture have suggested. Specially, classifying sitting posture with inertial sensor has a lot of benefits. It consists of simple hardware system and it is free for restriction of measurement environment. In this paper, we proposed an optimal sensor position which can distinguish all sitting postures by using just 1 inertial sensor. Data classification was processed by Support Vector Machine (SVM) which is machine learning model. The result of classification was used for finding optimal sensor position. It was found that back of neck is optimal position for classifying sitting posture.