基于RBF网络的BTT导弹逆控制研究  

Design of Inverse Control of BTT Missile System Based on RBF Neural Network

作　　者：杨志峰[1] 雷虎民[1] 李小兵[1] 王强[2] 

机构地区：[1]空军工程大学导弹学院,三原713800 [2]中国航天部二院空军代表室,北京100854

出　　处：《系统仿真学报》2009年第12期3694-3697,共4页Journal of System Simulation

基　　金：总装预研基金(9140A04050407JB3201);航空科学基金(2006ZC12004)

摘　　要：运用RBF神经网络结构和最近邻聚类算法,对导弹系统逆动力学系统进行动态模型辨识,并以辨识模型为控制器与BTT导弹控制系统串联构成一个动态伪线性系统,进而应用逆系统方法设计了一种用于解决BTT导弹非线性控制问题的经典控制与神经网络在线自学习相结合的控制方案,实现了导弹三通道的线性化控制和输出的渐近无差跟踪。仿真结果表明,该方案可依据设计指标的要求,实现对BTT导弹的非线性控制,且具有较强的鲁棒性。A RBF neural network applying nearest neighbor clustering lgorithm was used to realize the identification of the inverse dynamic system model of missile. And the system which makes the identification model and the BTT missile control system are in series was dynamic pseudolinear. Then a method of on-line self-learning control strategy which combines inverse control based on classical control with neural network, was proposed in the system. Thet method can realize linearization control of missile and make the output＇s indistinctive tracting gradually. With the help of simulations, the design of BTT missile control system can not only realize nonlinear control, but also possess excellent robustness and meet the designing requirement.

关 键 词：BTT导弹 RBF神经网络 逆控制 在线自学习 最近邻聚类算法 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]