基于线性回归的无线传感器网络加权质心定位算法

Linear-regression-based Weighted Centroid Localization Algorithm in Wireless Sensor Network

在建立定位算法求解数学模型和定位性能的基础上,提出一种基于线性回归的加权质心定位算法。该算法首先利用节点之间的跳数对质心定位算法进行加权处理、中心化处理,然后利用线性回归的数学模型对节点进行校正,最后对校正位置进行修正。通过Matlab程序仿真,与原始质心算法相比,改进算法的平均定位误差下降了50%以上,并且在信标节点密度较低的情况下,可使节点定位比率提高至接近1。

On the basis of the establishment of mathematical models, this paper proposed a new localization algorithm for wireless sensor network linear-regression-based weighted cen- troid localization algorithm. This algorithm has improved weighted centroid localization algorithm by using the hops between nodes, centering and calibrating nodes＇ position by using linear re- gression mathematical model. Finally, the algorithm corrected the position by using certain pa- rameters. After using matlab for simulation and comparing with the original centroid localization algorithm, the new algorithm＇s average localization error was found to drop by more than half and the localization ratio was increased to close to 1 in the case of anchor node in low density.