摘要
为了解决聚类算法需要较多的先验知识,不能自动进行聚类的问题,提出了基因表达式编程和K-Means融合的雷达信号分选算法。从介绍基因表达式编程和K-Means聚类算法的特点出发,针对雷达信号的实际情况,对两种算法进行了优化融合,并通过模拟雷达辐射源数据进行了仿真验证,仿真结果表明该算法在不需要任何雷达辐射源先验知识的情况下即可自动完成聚类分选,具有98.3%的聚类分选精度和较快的收敛速度,其较高的分选精度在电子情报侦察系统上有着广阔的应用前景。
Traditional clustering algorithms need much priori knowledge and can＇t sort radar emitter sig- nals automatically. In order to solve the problem, a novel method based on the combination of gene expres- sion programming and K-means algorithm is proposed. The paper starts from introducing the characteristics of the two algorithms, and then combines the two algorithms based on the characteristics of radar signals. Experiments show that this novel algorithm can not only sort radar emitter signals automatically without pri- or knowledge, but also has the clustering accuracy of 98.3% and high constringency speed. Because of its high clustering accuracy, the algorithm will have a bright future in the field of ELINT.
出处
《雷达科学与技术》
2013年第2期150-154,共5页
Radar Science and Technology
基金
空军装备预研基金(No.KJ2011204)
