基于区间数聚类的无线传感器网络定位方法

WSN Localization Method Using Interval Data Clustering

在基于接收信号强度指示(Received signal strength indicator,RSSI)测距的无线传感器网络(Wireless sensor network,WSN)定位方法应用过程中,信号强度与对应通信距离的对数成线性关系的假设在实际无线通信环境下几乎不能满足,从而导致定位误差较大.针对此问题,本文首先利用区间数表示方法结合实际定位环境中RSSI数据的统计信息表示RSSI的分布区域,并采用区间数聚类方法实现距离估计,以减小由于RSSI值不确定性引起的距离估计误差,然后利用这些距离估计值实现基于测距的WSN定位方法.采用三种实际通信环境下RSSI测量数据完成的定位实验结果表明,本文提出的基于区间数聚类RSSI-通信距离(RSSI-D)估计的定位方法可有效地提高定位精度.

When the wireless sensor network （WSN） range-based localization method is applied to communication distance estimation, RSSI （received signal strength indicator） is assumed linear with the logarithm of corresponding communication distance. But it is always in contradiction with the real communication environment and leads to big localization error. So in this paper, interval data combined with statistic information of RSSI data are used to express the distribution region first, then soft and hard interval data cluster algorithm are used to estimate the communication distance for different uncertainty levels of RSSI data. Next, the RSSI-D estimation results are used in range-based localization methods. Finally, real RSSI data in three typical communication environments are used to evaluate this method. Experiment results show that the proposed localization method using interval data clustering RSSI-D （distance estimation based on RSSI） estimation can get better precision in different environments.