基于RBF神经网络的大迎角机动空空导弹控制系统设计  

A/A Missile Control System Design for High Angle of Attack Maneuvering Based on RBF Neural Networks

作　　者：张友安[1] 吴进华[1] 杨华东[1] 

机构地区：[1]海军航空工程学院教研室,山东烟台264001

出　　处：《飞行力学》2003年第3期48-51,共4页Flight Dynamics

基　　金：航空科学基金资助项目(99D12001)

摘　　要：以某型空空导弹为对象,基于双时标的假设,将导弹动力学模态分离为快变状态动力学模态和慢变状态动力学模态,并对其分别进行了动态逆设计。考虑到空空导弹在高速、大迎角飞行时,空气动力系数具有强非线性,直接采用了真实的多维非线性气动力模型(未进行小迎角线性化处理),用RBF神经网络逼近了气动力矩系数,在非仿射条件下用反解设计方法求得舵偏角控制命令。最后,给出了控制系统六自由度数学仿真结果。For an A/A missile, its dynamics is separated into two parts, the dynamics of fast state and the dynamics of slow state based on the two timescale separation. Considering the strong nonlinearities of aerodynamic coefficients of the A/A missile under the conditions of high speed and high angle of attack, the real nonlinear multidimensional aerodynamics is not linearized. Using RBF neural networks the moment coefficients are approximated and the commands are obtained under nonaffined conditions. In the end, 6DOF simulation results have shown the rightness and effectiveness of the proposed scheme.

关 键 词：空对空导弹 非线性控制 RBF神经网络 大迎角机动 控制系统 

分 类 号：TJ765.2[兵器科学与技术—武器系统与运用工程]