基于再励学习的歼击机安全着陆横侧向协调控制  

Fighter safe landing lateral-directional control based on reinforcement learning

作　　者：文博武[1] 胡寿松[1] 

机构地区：[1]南京航空航天大学自动化学院,南京210016

出　　处：《东南大学学报（自然科学版）》2005年第A02期203-206,共4页Journal of Southeast University：Natural Science Edition

基　　金：国家自然科学基金重点资助项目(60234010);航空科学基金资助项目(05E52031);国防基础科研资助项目(K1603060318)

摘　　要：基于大系统递阶控制思想,提出了一种运用再励学习算法设计歼击机自动着陆横侧向协调控制系统的方法.将再励学习算法引入歼击机着陆阶段的横侧向子系统的协调控制中,对滚转通道和航向通道的控制器参数进行在线寻优,提高了在线补偿神经网络消除非线性系统逆误差的能力,有效地改善了歼击机着陆阶段横侧向控制子系统的鲁棒性能.通过将这种方法应用于某型歼击机自动着陆控制的实时仿真研究,仿真结果表明了本文方法的有效性和可行性.A fighter safe landing lateral-directional control method is presented based on reinforcement learning algorithm （RL）, using hierarchical control of dynamical large-scale systems theory. The method, searching optimized controllers＇ parameters of rolling channel and yawing channel by applying RL algorithm in the lateraldirectional coordination control of fighter, is used to regulate the error in the plant inversion due to modeling uncertainties and disturbances, and effectively improves the robustness of fighter lateral-directional control subsystem during the landing phase. The application of the method in safe landing control of fighter is studied by the real-time simulation. Results show that the method is effective and feasible.

关 键 词：递阶控制 再励学习 侧向控制 安全着陆 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]