基于神经网络剩余时间模型的协同制导律设计  

Impact Time Cooperation Guidance Law Design Based on Time-to-go Estimating Model Using Neural Network

作　　者：金泽宇 刘凯 尹中杰 王龙[2] Jin Zeyu;Liu Kai;Yin Zhongjie;Wang Long(School of Aeronautics and Astronautics,Dalian University of Technology,Dalian 116024,China;China Airborne Missile Academy,Luoyang 471009,China)

机构地区：[1]大连理工大学航空航天学院,大连116024 [2]中国空空导弹研究院,洛阳471009

出　　处：《战术导弹技术》2021年第4期103-109,116,共8页Tactical Missile Technology

摘　　要：控制多枚导弹同时协同拦截运动目标时,需要估算拦截导弹的剩余飞行时间。针对此问题,分析了用于剩余飞行时间估算的神经网络模型输入变量对模型估算精度的影响,确定了神经网络模型的结构及输入变量类型。结合时间可控制导律,完成具有协调变量的双层协同制导律设计,并将训练所得剩余飞行时间计算模型与协同制导律相结合。对所设计的神经网络模型以及协同制导律进行了仿真分析。结果表明,神经网络模型在经过样本的训练后,可以精确估算拦截导弹的剩余飞行时间,同时以该估算结果作为依据确定协调变量值,可以实现多枚导弹的时间协同拦截。The time-to-go of intercepting missile should be estimated when controlling multiple missiles to intercept moving target simultaneously.To solve this problem,the influence of the input variables of the neural network model for the estimation of time-to-go on the estimation accuracy of the model is analyzed,and the structure of the neural network model and the type of input variables are determined.Combined with the time-controllable guidance law,the design of double-layer cooperative guidance law with coordinated variables is completed,and the calculation model of time-to-go obtained from training is combined with the cooperative guidance law.A simulation analysis is carried out to evaluate the performance of neural network model and cooperative guidance law,the result shows that the neural network model after the sample training can accurately estimate the time-to-go of the interceptor missile,and determine the value of the coordination variable based on the estimation result,so as to realize the time cooperative intercept of multiple missiles.

关 键 词：神经网络 剩余飞行时间 时间控制 协同制导律 智能算法 时间协同 导弹拦截 

分 类 号：TJ761.3[兵器科学与技术—武器系统与运用工程]