The Algorithm Research Of Humanoid Robot Arm Control System Based On The Inertial Motion Capture

With the deepening of the humanoid robot research, humanoid robot arm as important institution to execute function, become the forefront of current research. Based on multi-DOF humanoid robot arm, how to guarantee the operator to control robot arm simply and flexibly under the control accuracy is a problem worthy of study. To solve this issue, a control method using human arm attitude to control humanoid robot arm was proposed in this paper. By studying the inertial sensor attitude algorithm, humanoid robot arm configuration and control algorithm, humanoid robot arm control system based on inertial motion capture was realized. The main contents are as follows:Firstly, through reading a large number of related literatures at home and abroad, analyzed the researched status of humanoid robot arm and the motion capture technology both at home and abroad, and designed and demonstrated system overall architecture and implementation plan. Secondly, inertial attitude capture algorithm was researched. The sensor raw data was preprocessed and the system reference coordinate system and the definition of attitude Angle were determined. The direct attitude algorithm was realized by quaternion method. Studying the attitude fusion algorithm, introduced the index of dynamic degree, put forward a kind of adaptive complementary filter algorithm based on dynamic degree. Analyzing human arm motion constraints, proposed posture calibration algorithm, improve the attitude algorithm’s robustness. Thirdly, humanoid robot arm configuration and control algorithm were researched. By analyzed Arm motion anatomy structure, a kind of 5 DOF humanoid robot arm configuration was proposed and the kinematics was analyzed by simulation. The motion relation between upper arm and fore arm was analyzed, and a kind of humanoid robot arm control algorithm based on human arm attitude was proposed. Human arm posture was transformed into humanoid robot arm control variable parameters by the algorithm of attitude mapping, decompose and speed estimation. Finally, the algorithm and the whole system were tested and validated by the software simulation and hardware circuit experiment. Experiments show that this algorithm has high static and dynamic accuracy, control system realized the function of using arm attitude to control humanoid robot arm, achieve the desired research goal.