9軸センサを利用した身体姿勢推定アルゴリズムの精度検証

抄録

In motion measurement, a measurement method using 9 axis sensor is attracting attention in that they are less expensive, convenient and less restrictive in measurement area than optical motion capture. We focused on the motion information in the relative coordinate system and proposed and verified the Kalman filter algorithm that estimates the 3D body posture with high accuracy by correcting the drift error of the gyrosensor using the acceleration. In this study, we attached each 9 axis sensors to the femoral and shank segments in order to measure the knee joint angle and verified the accuracy of the algorithm to estimate the 3D posture in the relative coordinate system by measuring simultaneously with optical motion capture. As a result, it was found that alignment between sensor and body segment is important. Therefore, we proposed and verified the algorithm of alignment adjustment method. We could find the possibility to correct axis alignment by using the proposed algorithm.