Development of Inertia Sensor Motion Analysis System and Market Available Systems Accuracy Evaluation

• Main Authors: Ching-Wei Chang, 張清瑋
• Other Authors: Sai-Wei Yang
• Format: Others
• Language: zh-TW
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/34597081284253578775

碩士 === 國立陽明大學 === 生物醫學工程學系 === 104 === The phenomenon of applying the inertial sensor devices to motion analysis is a trend in recent years. Therefore, by understanding the algorithm and characteristics of inertial sensor devices applied to motion analysis, the study aims to develop an inertial-sensor-device system of motion analysis through using inertial sensor devices’ module boards in the market to develop into our own wireless inertial sensor’s hardware and an algorithm calculating various human joint angles. An algorithm developed in the study provides the analyzing function to kinematics. The verifying process is as follows: by wearing inertial sensor devices on the body, we could collect each body segment movement’s data to infer to the human’s joint angles. Then, we use the optical motion capture systems and industrial robot to testify the accuracy both of the algorithm developed in the study and of that in the market. Besides, we try to thoroughly compare our own inertial sensor hardware and the algorithm with those in the market. This study evaluates nine inertial sensor devices among six commercial brands, including Xsens, APDM, STT, NORAXON, Delsys, and YOST Labs. The results show that the accuracy of verification method proposed by the study is more reliable than that in the previous studies. Moreover, the algorithm to kinematic analysis developed by the study can be used in different inertial sensor devices and without any sensor placements’ restrictions. Although there is still room for improvement in verifying the accuracy of ankle joints, the algorithm developed in the study is better than that of any of commercial products mentioned in the study. The value of accuracy of the internal / external rotation of knee joint in the inertial sensor device developed in the study is 3.08±0.4, and it is the most accurate data among all the inertial sensor devices verified in the study. And the algorithm developed in this study apply to other brands of inertial sensor devices can improve their values of accuracy.