一种DBSCAN聚类点密度的加权质心定位算法  

A Weighted Centroid Localization Algorithm Based on DBSCAN Clustering Point Density

作　　者：李轶[1,2] 张亮 张然[1,2] 张申 

机构地区：[1]中国矿业大学物联网(感知矿山)研究中心,江苏徐州221008 [2]中国矿业大学信息与控制工程学院,江苏徐州221008

出　　处：《河南科技大学学报（自然科学版）》2018年第2期36-39,47,共5页Journal of Henan University of Science And Technology:Natural Science

基　　金：国家科技支撑计划基金项目(2013BAK06B05)

摘　　要：加权质心定位算法是无线传感器网络中最常用的定位算法,为了进一步提高其定位精度,本文将聚类算法引入到无线传感器网络定位中,提出了一种具有噪声的基于密度的聚类算法(DBSCAN)聚类点密度的加权质心定位算法。该算法根据共线度理论选择参数,构建定位三角形的集合。选择集合中定位效果较好的一部分三角形对未知节点进行定位,并对所得的初始定位结果进行DBSCAN聚类。在排除误差较大的定位坐标后,将聚类后每个簇的核心点个数视为权值,采用加权质心定位算法得到未知节点的最终定位位置。仿真结果表明:该算法与传统加权质心定位算法相比,平均定位误差减小80%以上,有效提高了无线传感器网络定位精度。The weighted centroid localization algorithm was the most commonly used localization algorithm in wireless sensor network.In order to improve its positioning accuracy,the clustering algorithm was applied to the localization in wireless sensor networks. A weighted centroid localization algorithm based on densitly-based spatial clustering of application with noise( DBSCAN) clustering point density was proposed.The parameter was chosen according to the degree of collinearity,and the set of localization triangle was established.Then a part of triangle which had better localization effectiveness in this set was chosen to locate unknown node,and the obtained result of initial positioning was clustered by DBSCAN. After the large error location coordinate was eliminated,the number of core points in each cluster was regarded as the weight,and the weighted centroid localization algorithm was used to obtain the final location of the unknown node. The simulation results show that compared with the traditional weighted centroid localization algorithm,the average localization error of the proposed algorithm is reduced by more than 80%. The localization accuracy of wireless sensor networks is improved effectively.

关 键 词：无线传感器网络 聚类 DBSCAN 加权质心定位算法 共线度 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]