Research Of Neural Network Controller For Certain Artillery Servo System

With the need of modern warfare and the development of technology,automation and precision of weaponry is more and more demanding, andservo system plays an important role in it. This paper proposes a method toimprove the stability, rapidity and accuracy of certain artillery servo systemwhen it works under heavy inertia, variable load and mighty impact condition.Neural network is a hot field of intelligent control, and an effectivesolution to the problem of nonlinear system identification and control. Manyspecialists and scholars pay attention to the research of neural network, andmake progress both at home and abroad. This paper applies BP neuralnetwork in modeling and control of servo system, and carries out simulationbased on MATLAB.The content of this paper is summarized as follow:(1) Application of digital AC servo system can improve servo system’sperformance a lot. This paper introduces hardware and principle ofexperiment table firstly, and then chooses the AC servo system according tothe load and operation requirements.(2) DSP can calculate fast with high precision, and is easy for programmingand test, so we use DSP as microprocessor in the servo control system.(3) We can make a transfer function model for the servo system in traditionalway. BP neural network is used as the model in this paper so as to get theexact characteristic of the controlled plant. The authority and threshold aretrained offline with servo system’s input and output sample data.(4) We design a controller for servo system with the hope of prefect dynamicperformance and robustness to variable parameters, and it can’t be accomplished simply by a PID controller. In adaptive PID controller based onBP neural network, three parameters of PID controller can be adjusted by theBP neural network online to make optimized control effect. I combineBang-Bang control method and adaptive PID controller based on BP neuralnetwork to design an intelligent controller.(5) Finally, I test the responses of servo system to step signal, velocity signaland sin signal on the experiment table.Results of simulations and experiments indicate that, the intelligentcontroller has the high quality of static and dynamic performance withrobustness to environment changes.