基于神经网络优化直/气复合控制导弹迅捷转弯  

Control of Quick Turning of Missile with Lateral Thrust and Aerodynamics Based on Neural Network

作　　者：裴新月 于勇[1] 李政 李佳讯 于剑桥[1] PEI Xinyue;YU Yong;LI Zheng;LI Jiaxun;YU Jianqiao(School of Aerospace Engineering,Beijing Institute of Technology,Beijing 100081,China)

机构地区：[1]北京理工大学宇航学院,北京100081

出　　处：《兵工学报》2024年第10期3564-3576,共13页Acta Armamentarii

摘　　要：空空导弹迅捷转弯控制是在保证导弹稳定飞行的前提下,在尽可能短的时间内完成整个机动过程。为实现上述战术指标,考虑到导弹迅捷转弯的快时变性、强非线性和大干扰性,设计基于滑模变结构控制和模糊逻辑的直/气复合控制器。针对实际飞行中可能出现的初值不确定和外界干扰过大等现象,引入固定时间收敛理论和扩张状态观测器,有效避免干扰带来的控制输入抖振并保证导弹姿态角能在固定时间内快速稳定收敛。利用神经网络对滑模变结构控制的设计参数进行优化,在保证导弹迅捷转弯的同时减轻导弹侧向发动机的质量。通过仿真验证新提出的控制和分配方案的有效性。研究结果表明,所设计的控制器在存在模型不确定性和外部干扰等复杂情况下具一定的鲁棒性、稳定平滑的控制效果、较强的抗干扰能力以及优异的姿态跟踪性能,同时参数优化后的控制算法跟踪效果更优,在降低了原收敛时间的12.6%的同时,降低了原最大控制力的9.6%。The control of the quick turning of air-to-air missile is to complete the entire maneuvering process in the shortest possible time under the premise of ensuring the stable flight of the missile.In order to achieve the above tactical indicators,a lateral thrust/aerodynamics composite controller based on sliding mode variable structure control and fuzzy logic is designed considering the fast time denaturation,strong nonlinearity and large interference of the missile quick turning.In view of the uncertain initial value and excessive external interference that may occur in actual flight,the fixed-time convergence theory and the extended state observer are introduced to effectively avoid the control input jitter caused by interference and ensure that the attitude angle of missile can converge quickly and stably within a fixed time.The neural network is used to optimize the design parameters of sliding mode variable structure control,which not only ensures the quick turning of missile,but also reduces the mass of missile's lateral engine.The effectiveness of the proposed control and distribution scheme is verified by simulation.The designed controller has certain robustness,stable and smooth control effect,strong anti-interference ability and excellent attitude tracking performance in complex situations such as model uncertainty and external interference,and the tracking effect of the control algorithm after parameter optimization is better,which reduces the original convergence time by 12.6%and the original maximum control force by 9.6%.

关 键 词：迅捷转弯 复合控制 滑模变结构控制 扩张状态观测器 模糊控制 神经网络 

分 类 号：V249[航空宇航科学与技术—飞行器设计] TJ765.2[兵器科学与技术—武器系统与运用工程]