- 機械力学・計測制御講演論文集 (ISSN:13480235)
- vol.2014, pp."130-1"-"130-7", 2014-08-25
The aim of this study is to improve of stick picture using wearable motion sensor, in order to estimate human joint center and stride. Human joint center was estimated by least square method of sphere and marker points on the human arm were observed by motion analysis system (MAC3D System). First, as a basic experiment, instead of measuring human joint, we made use of a stick which has 490 mm length and two markers were affixed at the end of the stick. Next, it is hard to measure the movement of lower body, we performed an experiment on the movement of arm. As a result, estimated value by sphere method has a higher accuracy than estimated by circle. Finally, gait acceleration value was observed by acceleration sensor. Then, gait acceleration value had coordinate transform toward direction of movement and filtered by low pass filter. After, Gait velocity and acceleration were calibrated to 0 and integrated. In a result, stride was calculated with high accuracy. In this paper, it is shown that estimated stride and value of human joint center were calculated with high accuracy. To summary, these experiments, it is feasible that the accuracy of stick picture using wearable motion sensor is to be improved.