Research On Model Identification And Control Method For Stewart Robot

The relative position and attitude of the primary mirror and the secondary mirror have important influence on the high quality imaging in the spatial multiple function optical facility.Because the primary mirror's quality and volume is relatively large,so we often design the secondary mirror as a multiple degree of freedom adjustment mechanism.Stewart robot has the advantages of compact structure,high precision,high rigidity,high stability and bearing capacity.It can correct the position of the secondary mirror to improve the quality of the image.Stewart robot is a kind of high precision six degree of freedom adjustment mechanism.In order to achieve good effect of position and attitude adjustment,the research of the control system is particularly important.In this paper,the key problems in the control such as the establishment of the control model,model parameter identification and control method are studied and the experiment is carried out.The main contents are as follows.Based on the kinematic analysis of the Stewart robot with known structure parameters,the inverse kinematics model can be used to convert the six rod length which needed to control system.The forward kinematic model can be used to convert the length of the six rod into the upper platform position in real time.Based on the inverse kinematics model,the method of Stewart robot parameter identification error model is proposed,which avoids the difficulty of solving the forward kinematics and the low efficiency and existence error of the numericalsolution.Secondly,the solution problem of identification model was analyzed and the Jacobian matrix of the model was solved.For the problem of the coefficient matrix irreversible,an improved Gauss Newton iteration method and a specific experimental procedure are presented.Finally,the phenomenon of measurement error of least squares principle leading to a greater impact on the accuracy was analyzed.In the MATLAB,the identification experiment of Stewart robot is completed.When the measurement error is 0.1um,the identification accuracy is increased by 56 times,and the average precision is only 7.8 times when the measurement error is below 1um.After the identification of the prototype,the average position precision is increased by 30.77 times,and the attitude accuracy is increased by 20.73 times.The results show that the parameter identification method based on inverse kinematics of Stewart robot,can get more accurate parameters of the real structure,so as to effectively improve the position accuracy of Stewart when the robot is working.Based on DSP and FPGA architecture,the design and implementation of Stewart robot motion control system was studied.Including DSP and FPGA system,DC brushless motor driver's hardware circuit design and production,software design and implementation.The control system of the host computer software was complete by using the VC++6.0.Finally,the control experiment of single pole motor in the production of motion controller using PI position regulator and integral separation PI speed regulator was completed and the control precision can reach 9.17''.