摘要
战场辐射源识别已成为电子侦察和雷达威胁告警的基本要素,其关键技术之一——辐射源特征聚类算法的研究也显得日益重要。在分析常用误差反传(BP)网络算法对辐射源特征聚类的不足后,提出利用基于粒子群优化的神经网络算法对多特征参数进行聚类。通过比较该优化算法和传统BP网络算法在聚类正确率和收敛速度方面的差别,验证了基于粒子群优化的BP算法在辐射源特征聚类中相对于传统BP算法的优越性,仿真结果证明了该方法具有较好的实用价值。
The emitter identification in the battlefield has become a basic element of electronic reconnaissance and radar threat warning,researching into one of its crucial techniques——the emitter characteristic clustering algorithm has become more and more important.This paper analyzes the shortages of normal error back propagation（BP）network algorithm clustering the emitter characteristic,then brings forward the method to cluster the multi-characteristic parameter by means of the neural network algorithm based on particle swarm optimization,validates the superiority of the BP algorithm based on particle swarm optimization than that of traditional BP algorithm through comparing the differences of accuracy and convergence speed of the optimization algorithm and the traditional BP algorithm,the simulation result proves that the method has preferable practical value.
出处
《舰船电子对抗》
2010年第3期66-68,95,共4页
Shipboard Electronic Countermeasure
关键词
粒子群优化
BP网络
神经网络
辐射源识别
particle swarm optimization
BP network
neural network
emitter identification
