应用于水下防御的小目标动态多属性威胁评估  

Application of Dynamic Multi-Attribute Threat Assessment of Small Target in Underwater Defense

作　　者：陈玄真 李诺 段江涛 张学磊[1] 石建飞[1] CHEN Xuanzhen;LI Nuo;DUAN Jiangtao;ZHANG Xueei;SHI Jianfei(The 3rd Research Institute of China Electronics Technology Group Corporation,Beijing 100015,China)

机构地区：[1]中国电子科技集团公司第三研究所,北京100015

出　　处：《电子信息对抗技术》2024年第2期10-19,共10页Electronic Information Warfare Technology

基　　金：陆军预研项目(30102100102)。

摘　　要：面向近海水下防御的信息融合,提出一种危险小目标的动态多属性威胁评估方法。通过对典型目标的特性分析,构建了威胁评估指标体系。利用K均值聚类为各规范化指标添加威胁度标签,从而生成指标值-威胁度样本。基于此,将指标威胁隶属度函数的确定和计算转化为径向基函数神经网络的训练和预测,连接权值及偏置等网络参数利用量子粒子群优化算法求解,目标属性及时间序列的权重分别由熵权法和正态分布累积函数计算,应用逼近理想解排序法得到目标威胁度。算例分析中,通过专家评分和对比实验验证了所提方法的有效性和可靠性。Towards the problem of information fusion in inshore underwater defense,a dynamic multi-attribute threat assessment method of small dangerous targets is proposed.Firstly,an index system of threat assessment is established by analyzing the characteristics of typical targets.Then,K-means is utilized to add threat labels to each normalized index values.Thus,index-threat samples are generated.On this basis,the determination and calculation of the threat mem-bership functions of indice are transformed into the training and prediction of radial basis func-tion neural networks(RBFNN).Connection weights and offset of network parameters are solved by quantum particle swarm optimization(QPSO).Weights of target index and time series are cal-culated based on information entropy and cumulative function of normal distribution,respective-ly.Finally,threat degrees of the targets are obtained using technique for order preference by sim-ilarity to ideal solution(TOPSIS).In example analysis,the effectiveness and reliability of the proposed method are verified through expert evaluation and comparative experiments.

关 键 词：威胁评估 水下防御 直觉模糊集 神经网络 群体智能 

分 类 号：TN97[电子电信—信号与信息处理]