Research And Design Of Motion Capture System Based On Inertial Sensors

Biomechanics seeks to understand the mechanics of living systems by the combination of biology, kinematics and kinetics method. With the development of science and technology, measuring instruments and analysis systems for Biomechanics are increasing, such as surface electromyography instrument, foot pressure distribution measure system, optical motion capture system, force measuring platform,3D image measurement and analysis system. Generally speaking,3D image measurement and optical motion capture system are most commonly used in the full body motion capture, but data analysis and user operation are very complex. In addition, high-precision optical motion capture system is quite expensive, the measurement accuracy of low-cost system is difficult to meet customer needs. Because of the problems mentioned above, we designed a kind of motion capture system to capture arm motion based on a9-axis inertial measurement unit (3axis accelerometer,3axis gyroscope,3axis magnetometer). The main research works are as follows:(1) Design inertial measurement unit hardware and software platform based on WIFI communication.After analyzing the different characteristics of the inertial sensors, we designed9-axis inertial measurement unit which integrated accelerometer, gyroscope and magnetometer. By using the wireless WIFI communication technology, the problem of complex installation was solved. In the end,we established the hardware and software platform which realized inertial unit asynchronous data transmission and real time motion capture(2) To achieve a inertial measurement unit real time attitude by studying the multi-source data fusion method.A two stage EKF (Extended Kalman Filter) algorithm aimed at the attitude fusion was proposed to simplify the computational complexity. Then, an adaptive acceleration error covariance operator was adopted to reduce the influence of harmful acceleration in translational motion. (3) To achieve the arm motion capture by establishing the modelBased on anatomy principles and characteristics of arm movement, a simple rigid hinge arm model with four degrees of freedom was established. Then we studied the measurement methods of single joint and two joints movement. In addition, we designed a fast and efficient calibration method to compensate the initial error between the inertial unit and joint. At last,we calculated the elbow rotation center position and achieved the calculation method of elbow angel.Compared with the commercial motion capture system and clinical commonly used joint angle measuring instrument, the experiment results showed that the attitude solution is reasonable and acceptable. At last, we rebuilt the arm motion by visualization technology that verified the feasibility of the system. The work in this paper can lay a solid foundation for building and researching the whole body motion capture system in the future.