K-均值聚类与SVM结合的地空通信干扰识别方法  

Ground-to-air communications interference method of K-means combine with SVM

作　　者：张自豪[1] 马方立 裴峥[1] 

机构地区：[1]西华大学无线电管理技术研究中心,四川成都610039 [2]四川省无线电监测站,四川成都610016

出　　处：《济南大学学报（自然科学版）》2015年第6期420-424,共5页Journal of University of Jinan(Science and Technology)

基　　金：国家自然科学基金(61175055);四川省科技支撑项目(2013GXZ0155);西华大学创新基金(ycjj2014038)

摘　　要：通过对设置在不同地形的监测设备所采集的地空通信信号研究发现,其音频信号可作为识别地空通信干扰信号的研究对象,但音频信号特征通常是基于短时间域上的,无法直接应用在识别过程中。故提出利用K-均值聚类算法构建地空通信干扰音频信号的特征集合,并将通过遍历选择参数的支持向量机作为分类器用于地空通信干扰音频信号识别。实验表明,该方法可以很好地表现出音频信号的统计特性,快速、高效地识别出不同的地空通信干扰信号。Through the study of ground-to-air communication interference signals which are collected by different monitoring equipments of different terrain and find that its audio signals can be regarded as the recognized object of ground-to-air communication interference signals. Since the feature of audio signals is based on short time domain so that it is not used directly in recognition. Therefore,we propose that K-means clustering algorithm is used to structure the feature extraction of ground-to-air communication interference audio signals,and the SVM algorithm of traversal selected parameters are regarded as a classifier to recognize the interference of ground-toair communications audio signals. Experimental results prove that this method can well show the statistical characteristics of audio signals and has a high recognition rate for different ground-to-air communication audio interference signals in a shot time.

关 键 词：地空通信音频信号 K-均值聚类算法 支持向量机 

分 类 号：TU7[建筑科学—建筑技术科学]