Research On Modeling And Control Strategy Of Weapon Servo System

With the continuous improvement of modern military capabilities,put forward higher requirements on combat performance of Rocket Destroying Obstacles Weapon. The shooting accuracy and stability of Rocket Destroying Obstacles Weapon mainly depend on the performance of the positional servosystem,so,The research on the control of position servo system has important scientific research significance.In this paper,with the Rocket Destroying Obstacles Weapon as the engineering background,with the weapon AC positioning servo-system as controlled object,studying the system modeling and control strategy.Firstly, combining the actual control system of Rocket Destroying Obstacles Weapon, set up the hardware-in-loop simulation platform to reflect the actual system.,introduces components of platform, designs and explains the hardware circuit of the control system.Secondly, Modeling for the actuator of Rocket Destroying Obstacles Weapon-PMSM through the vector control method. And base on this, establish the mathematical model.of Rocket Destroying Obstacles Weapon.Then, becauseof the uncertainty and nonlinear of Rocket Destroying Obstacles Weapon AC servo system, it make difficult to establish a precise mathematical model.so,system identification is introduced to establish the more precise mathematical model. Taking identification of system by using RBF neural network and genetic algorithm optimize of RBF neural network. By comparing simulation results and analysis, we can draw the conclusion that genetic algorithm optimize of RBF neural network has higher accuracy, can satisfy the system requirements.Furthermore, on the basis of the identification model and research on control strategy. Using RBF neural network controller and GA-RBF neural network on-line identification of fuzzy RBF neural network adaptive controller to control, Through the emulation of MATLAB, it is concluded that the fuzzy RBF neural network have obvious advantages.Finally,to testify two kinds of control strategies on hardware-in-the-loop platform, the test results show that the fuzzy RBF neural network adaptive control can satisfy the requirement of technical indicators.