Research On Control Of Sliding Mode Variable Structure Artillery Servo System Based On Fuzzy Neural Network

With the improvement of modern military capabilities,artillery has gotten great development in the aspect of high precision,long range,fast response and digital direction.Artillery servo system is an important factor to improve its accuracy and shooting range.For artillery servo system is nonlinear,time-varying,coupling,the traditional control strategy cannot meet the requirements of precision and stability for artillery servo system control.This article done research in advanced modeling methods and control strategy,in order to improve the control performance of artillery systems.The main work of this paper is showed as the following aspects:(1)Firstly,this paper set the hardware-in-the-loop simulation test bench,introduced the principle and structure of the system,used the vector control method to decouple the AC servo motor,then a simplified treatment was carried out on the curent loop and speed loop in order to derive the whole mathematical model of artillery servo system,finally provided the foundation for advanced control strategy and experimental analysis.(2)For nonlinear and time-varying in artillery servo system,it is difficult to establish a accurate mathematical model.Meanwhile,RBF neural network has good ability of approaching a model.This paper adopted the RBF neural network to identify system.As a result of the poor self-learning ability and fixed parameters of RBF neural network,differential evolution algorithm was introduced to optimize the system.Finally the results show that the identification of RBF neural network based on differential evolution algorithm has high precision through offline training and validation.(3)In view of the sliding mode variable structure control is not sensitive to external disturbance characteristics,this paper adopted sliding mode variable structure control for servo system.Considering problems of chattering on the sliding mode variable structure,this paper adjusted switch gain of sliding mode controller in real time according to introduce RBF neural network.In view of the shortcoming of RBF neural network such as poor self-learning ability and fixed parameters,the system cannot switch gain in time and adjust the switch gain accurately.So a method of fuzzy control to modify the structure parameters was promoted in order to improve its adjustment ability of the switch gain.Simulation results show that the design of the fuzzy RBF neural network sliding mode variable structure control strategy has the advantages of high control precision and good robustness,finally meets the system control requirements.(4)Baesd on the system simulation platform,the experiment is carried out to verify the algorithm above.According to analysising the result of the experiment,it is concluded that the fuzzy neural network sliding mode variable structure control has good dynamic and static index and meets the performance index very well.