一种基于模糊聚类的物理小区识别分配方案  

A physical cell identification allocation scheme based on fuzzy hierarchical clustering

作　　者：刘濛 涂山山 肖创柏[1] 林强强 LIU Meng;TU Shanshan;XIAO Chuangbai;LIN Qiangqiang(Faculty of Information,Beijing University of Technology,Beijing 100124,China;Beijing Key Laboratory of Trusted Computing,Beijing University of Technology,Beijing 100124,China)

机构地区：[1]北京工业大学信息学部,北京100124 [2]北京工业大学可信计算北京市重点实验室,北京100124

出　　处：《现代电子技术》2019年第17期14-20,共7页Modern Electronics Technique

基　　金：北京市科技计划项目(Z171100004717001);北京市自然科学基金重点项目(L172049);国家自然科学基金项目(61671030);北京工业大学研究生科技基金(ykj-2017-00850)~~

摘　　要：用户在异构蜂窝网络(HCN)中寻求网络数据服务时,需要用到物理小区识别(PCI)对环境中不同的蜂窝小区进行识别。然而PCI的数量是一定的,传统的解决方案缺少合理利用有限数量的PCI机制,无法满足小蜂窝的大规模随机部署,同时也难以保障用户的网络质量(QoS)。针对以上问题,引入蜂窝小区“活跃度”概念,提出基于模糊分层聚类的PCI分配方案,采用欧氏距离法进行不同蜂窝小区相似度的求解,从而得到不同阈值下的蜂窝小区基站分类集群,并依据统计学中的方差分析法查找最佳阈值,优先对“活跃度”高的蜂窝小区基站群进行PCI的分配和复用。与现有方案比较,有效提高了用户的QoS以及PCI的分配效率。仿真实验结果表明,文中提出的基于欧氏距离法的聚类方法与传统的曼哈顿距离法、夹角余弦值距离法以及切比雪夫距离法相比,蜂窝基站分类结果更为合理,同时具备较低的PCI冲突混淆率。When the users are searching for online service in heterogeneous cellular network (HCN),physical cell identification(PCI) is needed to distinguish different cellular cells. Nevertheless,the number of PCI is limited,and traditional PCI allocation methods are weak in saving the PCIs,and can hardly meet the requirements of massive deployment of small cells and the quality of service(Q oS) of users. In view of these issues,a PCI allocation method based on fuzzy hierarchical clustering is proposed according to "activeness" concept of cellular cells,and the Euclid distance method is used to derive the similarity from different cellular cells. As a result,the clustering of cellular cell base stations is obtained at different threshold values, then the optimal threshold is found by using the variance components method,and the PCIs are preferentially allocated and reused in cellular cell base station groups with higher "activeness". In comparison with the existing schemes,the proposed method can improve the users′ QoS and efficiency of PCI allocation. The simulation results show that the method has a better clustering result and lower possibilities of PCI conflict and confusion when compared with the Manhattan distance method,A ngle Cosine distance method and Chebyshev distance method.

关 键 词：异构蜂窝网络 蜂窝小区 聚类方法 活跃度 模糊聚类 物理小区识别 

分 类 号：TN915.02-34[电子电信—通信与信息系统]