Design of Weapon Detection Device Control System
|School||Harbin Institute of Technology|
|Course||Control Science and Engineering|
|Keywords||servo system neural network control RBF inverse models|
Servo System detection device is a system that test indicators of servo system. How to improve loading accuracy in the device is key issue of Servo system detection device.The loading motor follows the position changes of the rudder passively and so the influence of redundancy torque becomes inevitably which is the mutual problem of passive systems. As a kind of intense disturbance, the redundancy torque effects the loading precision and the dynamic performance badly and can’t be controlled by conventional control strategy based on accurate model. Aiming at the problem of redundancy torque, this paper presents a new kind of inverse model control strategy based on RBF neural networks. Using the excellent nonlinear function approximation of RBF neural network, the identification neural network identify the inverse model of the object on real-time, and it’s copy is put into the forward channel of the system as a forward compensation controller. Ideally, the transfer function of forward channel close to 1,the output would follow the order accurately.Firstly, the mathematic model of the system is obtained and the effect of the redundancy torque is analyzed via simulation.Then a control method based on traditional forward feedback method is used in designing the control method of system, and we point out inadequate of this method Subsequently, an improved RBF arithmetic is proposed too, which is generated offline and updated online. By fully using the existed knowledge of the object and optimizing the network parameter locally, the calculating burden and the size of network are observably lessened.The simulation analyzing is carried out in the Matlab/Simulink environment, the simulation result show that the proposed control strategy can restrain the redundancy torque effectively, improve the dynamic performance and loading precision under different loading frequency and different loading grads.