Simulation Research Of Fuzzy PID Servo Motor Control System Based On Neural Network

Based on the background of micro-UAV laser wireless power transmission,the working principle and control method of the existing two-axis rotary table servo motor are researched.Six different controllers were designed theoretically,including PID controller,fuzzy controller,fuzzy PID controller,neural network PID controller and fuzzy neural network PID controller.The corresponding servo motor control system was built and analyzed by simulation.The control accuracy,speed,advantages and disadvantages of the six control systems.Firstly,the servo motor of the existing two-axis turntable in the laboratory is taken as the research object.By analyzing its working principle and the control method of the traditional current loop,position loop and speed loop(three-loop),the dynamic model of the servo motor is established.The Bode diagram and step response curve of the servo motor are simulated by using Simulink module in Matlab,and its dynamic performance and static performance are explored.On this basis,the different control methods are used to improve the working performance of the two-axis rotary table servo motor.Thereby improving the feasibility of the laser beam tracking accuracy and speed of the moving target.Firstly,the traditional PID control method is used to control the servo motor model.The simulation results show that the method obviously has the disadvantages of inconvenient parameter setting,poor robustness,slow tracking speed when moving target motion.In order to overcome the above disadvantages,an attempt is made to add another fuzzy controller to the above conventional PID controller,which improves the adaptive tuning capability and robust performance of the PID controller.However,fuzzy PID control also has its own shortcomings.For example,the inference rule is an approximate behavior;Selecting a fixed value of the quantization factor will cause the control system to generate an unstable response.Considering the actual tracking process,the fuzzy PID controller cannot adapt to the flexibility of the UAV flight.At the same time,on the theoretical level,the fuzzy PID control method has great limitations in the nonlinear system.Neural networks have the ability to find the best nonlinear function by learning.Therefore,in order to overcome the shortcomings of the above fuzzy PID control system,it is proposed to add a neural network method to the above fuzzy PID control system.A BP neural network is added to the fuzzy PID controller.The desired servo motor rotation angle is used as the input of the neural network.Construct a three-layer neural network system with input layer,hidden layer and output layer,and the output is the quantization factor of the fuzzy controller.The input-output training sample is the expected servo motor rotation angle and the artificial tuning factor.The fitting training of the constructed three-layer neural network realizes the online adjustment of the quantization factor of the fuzzy controller.The fuzzy neural network PID control system not only has the characteristics of good robustness and adaptive parameter-adjusting ability of the fuzzy PID controller,but also has the advantage of good approximation ability to the nonlinear model.Finally,the control system model is built in Simulink,and the program is added in the fuzzy control module and the neural network module.The programming and modeling are combined to simulate the several control systems.The results show that the dynamic performance and static performance of the fuzzy PID servo motor control system are better than the single fuzzy servo motor control system and PID servo motor control system.Through the BP neural network learning,the quantization factor of the fuzzy controller is adjusted online,and the performance of the fuzzy neural network PID servo motor control system is further optimized,which greatly improves the performance of the original PID servo motor control system.