基于RBF网络增益自适应调节的滑模制导律  

Sliding Mode Guidance Based on RBFNN Adaptive Gain Adjustment

作　　者：李士勇[1] 章钱[1] 

机构地区：[1]哈尔滨工业大学航天学院控制科学与工程系,黑龙江哈尔滨150001

出　　处：《测试技术学报》2009年第6期471-476,共6页Journal of Test and Measurement Technology

基　　金：国家自然科学基金资助项目(60773065)

摘　　要：针对被动寻的导弹拦截问题,在变结构控制理论的基础上,应用RBF神经网络控制,提出一种新型导引律.首先应用变结构控制理论建立滑模变结构制导律.然后应用RBF神经网络对制导律中变结构项的增益进行调节,克服了变结构制导律中变结构强度项不易确定的缺点,从而减弱变结构控制的抖振,提高了导弹拦截的精度.仿真结果表明,该导引律具有较强的自适应能力,具有实时性好及便于实现的优点;同传统的变结构制导律、比例导引律相比,在脱靶量、拦截时间等方面都有了显著的提高.A new guidance law is proposed by using RBF neural network technology and variable structure control theory for the passive homing guidance. First of all, the sliding mode variable Structure guidance law is derived from variable structure control theory, then we use RBF neural network to adjust the gain of the sliding model guidance. So it overcomes the disadvantage that the strength of variable structure item is not easy to be determined. Therefore it can reduce the chattering and improve the accuracy of missiles interception. This guidance law has strong adaptive capacity and suitable for real-time on-line control, as well as easy to implement. Numerical simulations show that the proposed guidance law yields better performance than the others.

关 键 词：RBF神经网络 变结构 制导律 导弹拦截 

分 类 号：TJ765.3[兵器科学与技术—武器系统与运用工程]