Research On The Dynamic Modeling And Control Method Of Cleaning Mechanism For Navigation Lamps

Aim at the problems of Delta parallel mechanism based on kinematics modeling and control methods that tracking accuracy is low,control energy is large,robustness is not strong,and difficult to realize the complete decoupling control,a set of dynamics control based on dynamics modeling and computing torque control system is designed.In this paper,the research object is the cleaning device of navigation lamps,the dynamic control of its end is realized,which has ensured that the cleaning operation is fast,accurate and anti-interference.In this paper,error compensation,dynamic modeling,parameter identification of the inverse kinematic solution model of Delta cleaning mechanism,the research on the computational torque control strategy based on fractional order PID,and the design of the whole control system are mainly carried out.Firstly,a method based on RBF neural network is proposed to improve the kinematic control precision of Delta cleaning mechanism.Based on the existing kinematic control system of parallel cleaning mechanism,the factors affecting the control precision and the limitations of the existing methods to improve the control precision are discussed.Then a workspace analysis was performed on the Delta cleaning mechanism,and training samples is collected experimentally in the workspace.Taking the actual position of the terminal as the input sample and the error between the expected position and the actual position as the output sample,the RBF neural network model has been trained to obtain the nonlinear mapping relationship between the actual position of the terminal and the position deviation,and the position compensation controller is designed based on this.Trajectory tracking experiments of the end is conducted on the Delta cleaning mechanism platform to verify the method,the control error of the end is reduced from±30mm to±5mm,effectively reducing the error of the end position,providing a simple and feasible method for accurate control of the Delta cleaning mechanism,and laying a foundation for the accuracy of the dynamic control system.Secondly,a parameter identification method based on maximum likelihood estimation is proposed to solve the inaccuracy of the dynamic model due to simplified model,inaccurate parameter measurement and friction between joints.Based on the simplification of Delta cleaning mechanism model,the mathematical expression of Delta cleaning mechanism dynamics model is deduced by Euler-Lagrange equation.Considering the actual cleaning requirements,three excitation trajectories is designed and the trajectory tracking experiment has been carried out to obtain the motor's angle,velocity,acceleration current and end position information.A Min Batch Gradient Descent method is used to evaluate the likelihood function iteratively,so as to obtain the parameters to be identified of the dynamic model.Then,some samples are randomly selected as the test set for verification.The adjusted determination coefficient R2_adjusted is used as an index to evaluate the identification results.The results showed that the fitting degree of the kinetic model obtained by this method is not less than 92.6%.Thirdly,the integral order PID?IOPID?controller and fractional order PID?FOPID?controller are derived theoretically,and the numerical realization method of fractional order PID controller is given.The parameter tuning of fractional order PID is carried out by using quantum particle swarm optimization?QPSO?algorithm.The simulation experiment proves that FOPID is better than IOPID in response speed,response precision,anti-interference performance.Based on this,in order to further explore the five parameters of FOPID influence on control system,the simulation experiments of the single variable method is carried out,the results show that the Kp,Ki,Kd changes of FOPID impact on the system is the same as the traditional PID,the section points order?affect system settle state precision,the differential order times?affect system overshoot.Finally,a dynamic control system of the cleaning mechanism of navigation lamps is designed based on the above three parts,and the interface design,programming and debugging of the control system are completed.Combined with the actual cleaning requirements,the experimental verification trajectory is planned.The response of motor angle,3D space error and torque of motor output are used as the indexes of response speed,control precision and anti-interference performance.The results show that the FOPID based on the calculated torque control strategy has faster response and higher anti-interference performance for the tracking task of the parallel robot.The speed,accuracy and stability of the control system are verified.