基于偏最小二乘的无线传感网多跳定位算法  

Multihop Localization Algorithm Based on Partial Least Squares for Wireless Sensor Network

作　　者：窦如林[1] 阎浩[1] 严筱永[2] 

机构地区：[1]金陵科技学院软件工程学院,江苏南京211169 [2]金陵科技学院智能科学与控制工程学院,江苏南京211169

出　　处：《金陵科技学院学报》2014年第3期19-25,共7页Journal of Jinling Institute of Technology

基　　金：金陵科技学院博士启动基金(jit-b-201411)

摘　　要：提出了一种基于偏最小二乘的多跳定位算法(ML-PLS)。算法借助偏最小二乘对已知节点间的跳数与真实距离的建模,通过跳数、距离两者间的最大协方差共同推动节点位置的估计。ML-PLS算法对复杂部署环境具有较强的适应性,克服了传统算法只适用于各向同性网络的不足。仿真结果表明,ML-PLS算法定位精确度高且性能较稳定,并能适应不同网络环境。An improved multihop localization algorithm （ML-PLS ） based on partial least squares is proposed.This algorithm uses the partial least squares to build the localization model of hop-count and the real distances between beacons,and employs the maximun covariance of hop-count matrix and distance matrix to jointly promote the estimation of node location.ML-PLS has strong adaptability for complicated deployment environment,and overcomes the short-age of traditional multihop algorithm which is only suitable for isotropic networks.Simulation results show that ML-PLS algorithm has high estimate precision and stable performance,and can adapt to different network environments.

关 键 词：多跳定位 无线传感器网络 偏最小二乘 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]