多目标击水声信号特征关联方法

Correlation Method of Acoustic Characteristics of Multiple Targets during Water Entry

目标入水产生的击水声信号形式复杂,在海洋尖刺干扰的影响下易造成漏报和虚警,当多目标不等间隔陆续入水,不同观测节点间同一目标的击水声信号难以正确关联。针对上述问题,提出一种多目标击水声信号关联方法。通过经验模态分解将检测到的瞬态信号分解为固有模态函数(Intrinsic Mode Function,IMF),并提取各阶IMF的频谱等特征,利用灰色关联分析,模糊C均值聚类两种算法实现多击水声信号的关联,为后续的分布式定位提供必要条件。仿真结果表明:当特征参数个数为6时,两种方法的目标关联正确率分别可达98%和87%;当测量节点个数为7时,两种方法的目标关联正确率分别可达95%和87%。仿真结果证明了本文方法的可行性和有效性。

The underwater acoustic signal generated by a target entering the water is complex which can easily cause false alarm and miss due to the impulsive interference in ocean.In addition,it is difficult to correctly correlate the underwater acoustic signals of the same target between different observation nodes when multiple targets enter the water at different intervals.In order to solve the above problems,a multi-target water-entry acoustic signals correlation method is proposed.The detected transient signal is decomposed into the intrinsic mode function(IMF)through empirical mode decomposition,and the characteristics of the frequency spectrum and so on are extracted.The grey relational analysis method and fuzzy C-means clustering algorithm are used to realize the correlation of multi-target water-entry acoustic signals,providing necessary conditions for the subsequent distributed positioning.The simulated results show that the target correlation accuracy of the two methods can reach 98%and 87%,respectively,when the number of target feature parameters is 6,and the target correlation accuracy of the two methods can reach 95%and 87%,respectively,when the number of observation nodes is 7.The simulated results prove the feasibility and effectiveness of the proposed method.