摘要
针对装有侵彻战斗部的导弹飞行末端攻角归零的需求,提出了过载收攻角自动驾驶仪和姿态收攻角自动驾驶仪两种攻角收敛方案。将导弹末端攻角收敛问题,转化为在某期望时间后攻角绝对值收敛于某小量的优化问题,待优化量为攻角收敛驾驶仪的外回路控制参数,采用最速梯度下降方法对该问题进行求解。使用导弹线性化状态空间作为优化算法的数学模型,设计了2种攻角收敛自动驾驶仪的控制参数;以战术导弹的某典型动力系数为例,分别优化了过载收攻角自动驾驶仪和姿态收攻角自动驾驶仪的控制参数。使用数学仿真对参数优化设计结果的有效性进行了验证,并使用根轨迹法对两种驾驶仪进行优劣对比,结果表明姿态收攻角驾驶仪对舵机性能要求更高。
In order to meet the requirement of terminal angle-of-attack convergence for missiles with penetration warhead,the structures of two classical angle-of-attack convergence control loops were proposed,one is acceleration autopilot and the other is pitch attitude autopilot.The convergence problem of the terminal angle-of-attack was innovatively transformed into the optimization problem that the absolute value of the angle-of-attack converges to a small amount after a certain expected time,and the optimization value is the out-loop control parameters of the angle-of-attack convergence autopilot.The steepest-gradient-descent method,one of the popular algorithms in artificial intelligence,was used to solve the optimization problem,and the linearized state space of the missile was used as the mathematical model of the optimization algorithm.Taking a typical dynamic coefficient of a tactical missile as an example,control parameters of the two angle-of-attack autopilots were optimized respectively,with the effectiveness of parameter optimization design verified by mathematical simulation.The advantages and disadvantages of the two kinds of autopilot were compared by root-locus method,and the conclusion that the attitude angle-of-attack autopilot requires higher servo performance was obtained.The mathematical simulation curves under different servo performance were used to verify the conclusion.
作者
管茂桥
崔晓曦
王林平
贾鑫
邹景锋
GUAN Maoqiao;CUI Xiaoxi;WANG Linping;JIA Xin;ZOU Jingfeng(Norinco Group Institute of Navigation and Control Technology, Beijing 100089, China)
出处
《兵器装备工程学报》
CSCD
北大核心
2021年第9期83-88,共6页
Journal of Ordnance Equipment Engineering
关键词
收攻角
优化控制
根轨迹
过载驾驶仪
姿态驾驶仪
最速梯度优化
angle-of-attack covergence
optimal control
root locus analysis
acceleration AOA autopilot
pitch AOA autopilot
gradient descent optimization