Research And Design Of Space Motion Tracking System Based On MEMS Sensor

Teaching playback is the major way of robot path tracking. Its working principle decide the complex degree of robot working. With the technology of human motion tracking,we can simplified robot teaching program and improve the safety of working environment.This article focus on the human motion tracking based on the technology of MEMS. With the 9-axis inertial sensor, which include an accelerometer, a gyroscope and a magnetometer,and the method of complementary data fusion, the data accuracy of hands will improved.The results of linear path motion and space continuous path motion experiment show that the date calculated by this method has the high accuracy, can be used in robot teaching playback.Firstly, this paper introduced the research background and status of Teaching/Playback robot, then discussed the five common methods of human body motion capture. After analyzing and comparing the characteristics of different methods, it explains the advantages and challenges of the human capture method in selecting MEMS Technology. Then focuses on the research status of MEMS technology in the motion capture of human body.Secondly, before the attitude calculation, this paper analyze the principle of human arm's motion and established the kinematic model of human arm. Then based at the different working characteristics of MEMS accelerometer, gyroscope and magnetic sensor,established three error models of three kinds sensor. After this, this paper calibrated of three kinds of sensors in the methods of six attitudes correction, zero correction and ellipse correction. In this way, the data accuracy of sensors would be improved.Thirdly, this paper described the basic principle of using MEMS sensor to calculate the attitude and three representation method of attitude calculation. With the compared the advantages and disadvantages of various methods, the representation method of Euler angle and four element number is adopted. In dealing with the problem of attitude solving, this paper uses the method based on complementary filtering, through the fusion of different sensor data, to improve the accuracy of attitude calculation. The final hand trajectory is obtained by using the spatial data in the 3D spherical coordinates.Finally, through the experiment of linear motion and space continuous movement of the sensor, the final trajectory data is obtained. In the analysis of the trajectory, we can find that the method of this paper can get a more accurate trajectory, and the attitude of this paper can be used in combination with the traditional robot teaching.