基于AGA-BPNN的微小型弹药气动辨识  

Aerodynamic identification of miniature ammunition based on AGA-BPNN

作　　者：豆登辉 范军芳[1] 王海森 Dou Denghui;Fan Junfang;Wang Haisen(Beijing Key Laboratory of High Dynamic Navigation Technology,Beijing Information Science&Technology University,Beijing 100101,China)

机构地区：[1]北京信息科技大学高动态导航技术北京市重点实验室,北京100101

出　　处：《战术导弹技术》2022年第3期1-8,共8页Tactical Missile Technology

基　　金：国家重点研发计划(2020YFC1511705);国家自然科学基金(61801032)。

摘　　要：将自适应遗传算法和神经网络结合应用于微小型制导弹药气动辨识问题中,提出了一种自适应遗传算法-反向传播神经网络(Adaptive Genetic Algorithm-Back Propagation Neural Network,AGA-BPNN)气动辨识模型。用交叉、变异概率自适应调整的遗传算法对神经网络的初始权值和阈值进行优化,建立AGA-BP神经网络模型;分别用BP神经网络、GA-BP神经网络、AGA-BP神经网络算法模型对弹药气动参数以及压心位置进行辨识,并对辨识结果进行弹道仿真飞行。结果表明,相比于BPNN和GA-BPNN,AGA-BPNN模型在微小型弹药气动参数辨识方面具有更小的均方误差和更快的收敛速度;在弹道仿真中辨识速度误差为0.37 m/s,位置误差为22.97 m,辨识模型达到了较好的辨识效果。The adaptive genetic algorithm and neural network are combined and applied to the aerodynamic identification problem of miniature guided ammunition,and an adaptive genetic algorithm-back propagation neural network(AGA-BPNN)aerodynamic identification model is proposed.The genetic algorithm with adaptive adjustment of crossover and mutation probability is used to optimize the initial weight and threshold of neural network,and establish AGA-BP neural network model.The BP neural network,GA-BP neural network,and AGA-BP neural network algorithm models are used to identify the aerodynamic parameters of the ammunition and the position of the pressure center,and the results of identification are carried out for trajectory simulation flight.The results show that compared with BPNN and GA-BPNN,the AGA-BPNN model has a smaller mean square error and faster convergence speed in the identification of aerodynamic parameters of miniature ammunition.In the trajectory simulation,the identification speed error is0.37m/s,and the position error is 22.97m.The identification model achieves a good identification effect.

关 键 词：微小型制导弹药 参数辨识 神经网络 自适应遗传算法 静稳定性 气动布局 弹道仿真 

分 类 号：TJ760.2[兵器科学与技术—武器系统与运用工程]