基于深度学习的突防控制博弈对象匹配方法CSCD

针对基于博弈理论设计应对多枚拦截弹的协同突防控制方案时需要确定博弈对象的问题,提出了一种基于长短时记忆(LSTM)网络的拦截弹攻击对象匹配方法。基于传统防空导弹飞行时序与流程构建拦截弹飞行轨迹库,以轨迹库为训练样本对LSTM网络进行训练,并以此为基础构建航迹预测模型与对象匹配模型,实现对拦截弹攻击对象的识别。仿真结果表明,该方法能够有效识别拦截弹拦截目标,为后续的巡航弹突防研究提供支撑。

Competitor Matching Method for Penetration Control Based on Deep Learning

For the problem that the competitor needs to be identified when designing a cooperative penetration control scheme against multiple interceptors based on game theory, an attack object matching method of interceptor missile based on long short-term memory (LSTM) network is proposed. Based on the flight sequence and process of the traditional interceptor missile, the interceptor missile flight trajectory library is constructed, and the LSTM network is trained with the trajectory database as the training sample. At the same time, based on this, a track prediction model and an object matching model are constructed to realize the target recognition of the interceptor. The simulation results show that the method can effectively identify the interceptor target and provide support for the follow-up cruise missile penetration research.