基于K-均值聚类的动态频谱接入技术  

Dynamic Spectrum Access Technical Based on K-means Clustering

作　　者：赵陆文[1] 缪志敏[1] 周志杰[1] 张皓[1] 

机构地区：[1]解放军理工大学通信工程学院,南京210007

出　　处：《信号处理》2009年第12期1825-1829,共5页Journal of Signal Processing

基　　金：中国博士后科学基金资助(No.20080441320);江苏省博士后科研资助计划(No.0802075C)

摘　　要：为了有效降低次用户接入授权信道时对主用户造成的干扰以及次用户被阻塞的概率,提出了一种基于K-均值聚类的动态频谱接入方法。该方法首先将关注的频谱划分为若干子信道,尔后对每个子信道进行特征提取,借助频谱特征,用K-均值算法对子信道进行聚类,将子信道分别纳入白色、灰色和黑色频谱池。次用户在接入信道时,重点选择白色频谱池中的信道接入。理论分析和仿真结果表明,利用该方法能够有效降低次用户对主用户的干扰和次用户被阻塞的概率,提高信道接入效率。In order to reduce the jamming between SUs and PUs, and the blocked probability of the SUs, a novel dynamic spectrum access technical based on K-means clustering is presented. Firstly, the interested spectrum is divided into many subchannels. Then, the characteristics of these subchannels are extracted, and these subchannels are clustered by K-means clustering algorithm. These subchannels are classed into white, grey and black spectrum pools. The subchannels in white pools are priority to be accessed for SU. Analysis and simulation results show that, the jamming and the blocked probability can be reduced, and the access efficiency of subchannels can be improved by this method.

关 键 词：认知无线电 K-均值聚类 频谱池 动态频谱接入 

分 类 号：TN92[电子电信—通信与信息系统]