反导系统对弹道群目标分离识别仿真

Simulation of Ballistic Target Group Separation and Identification in Anti-Missile System

反导中段假目标数量众多,多个目标可能会位于同一雷达波束而无法分辨和识别。针对上述问题,首先使用独立成分分析算法进行弹道群目标混波的盲源分离,并从分离前后的各目标回波中提取若干种物理意义明确、提取难度较低、类别可分性强的特征,构成模糊支持向量机的训练及测试样本向量。然后基于改进的样本模糊隶属度函数和训练样本精简算法,利用一对多模糊支持向量机分类思想对弹道群目标进行真假判决。仿真结果表明,将独立成分分析与模糊支持向量机相结合可以成功解决反导中段群目标混波的分离难题,同时以较高识别率实现了真实弹头的有效识别。

The number of false target in the middle stage of anti - missile is large, multiple targets may be located on the same radar beam and cannot be separated and identified. Aiming at the problem, Firstly, the blind source separation of ballistic target group mixed echo is carried out by using the independent component analysis algorithm, and several kinds of features with clear physical meaning and strong separability are extracted as sample vector of FS- VM. Then, based on improved fuzzy membership function and training sample reduction algorithm, the rest FSVM is used to make true and false judgments of ballistic group target. Simulation results show that the combination method of ICA and FSVM successfully solved the problem of the separation of ballistic target group mixed echo, and the effective identification of the real warhead is achieved with high recognition rate.