인체 동작 분석을 위한 관성센서 비교 검증에 관한 연구

This study was conducted to evaluate a depth sensor based Kinect camera which can be utilized to evaluate a subject’s kinetic and kinematic data during motion. To fulfil our purpose, a manufactured dummy, which is an adapted shoulder to make abduction and adduction motion, were used. And seven healthy male adults with no musculoskeletal or nervous system disorders selected to participate in experiments. Dynamic movement of the dummy and humans were induced by shoulder motion of abduction and adduction up to maximum angle of 60˚. The dummy and human’s motion of shoulder were measured simultaneously using ten infrared camera based three-dimensional (3D) motion capture system, two depth sensor based Kinect camera and IMU sensor. In the test results, phase difference between VICON and IMU had no significantly differences for all of the standing anges and showed maximum angle differences of 3.37~6.35˚ in dummy, and 2.27~5.20˚ in human. In the case of VICON and KINECT, the phase differences were shown that no significantly differences and its maximum angle difference were 2.89~4.14˚ in dummy, and 2.94~5.65˚ in human at the standing angle of 0˚, 30˚, 60˚ and 90˚. However, the phase difference of the dual Kinect camera had significantly diffrences of 8.10±0.59˚ in dummy, and 6.46±0.96˚ in human at the standing angle of 45˚. These results showed that the dual Kinect camera could provide measureing accuracy about maximum 9% of error at 0˚, 30 ˚, 60˚ and 90˚, and maximum 14.5% error at 45˚.