基于双重权值优化的分布式定位算法  

Research of the Distribute Localization Based-double Optimal Weights

作　　者：李晋[1] 熊炎[1] 陈永生[2] 

机构地区：[1]信阳师范学院,信阳464002 [2]同济大学电子与信息工程学院,上海200092

出　　处：《科学技术与工程》2013年第29期8666-8671,共6页Science Technology and Engineering

摘　　要：针对无线传感网络目标定位精度偏低,利用传感器节点的特性,并结合数据融合理论,提出了应用于传感网络的目标定位算法。该算法通过双重权值优化,并融合多个传感器节点数据,实现目标定位。在权值选取价段,先使用最小均方误差算法LMSE产生系统的权值初始值,然后采用自适应调整权值理念,并采用递归最小二乘算法RLS优化权值,以提高系统的定位精度。对静态和运动目标的定位数据融合算法进行了仿真,结果表明,相比单节点定位,融合算法的定位精度有约一到两个数量级的提高。A deep research is done for the poor precision of wireless sensor network location due to low-accuracy of location in wireless sensor networks.According to the node character of sensor and data fusion theory,location algorithm in wireless sensor network is proposed.The target location is realized by double-weight optimization and fusing node data from multiple sensors.In the process of weight selection,firstly,the initial value of weight is generated by the minimum mean square error.Then the adaptive cumulative value idea is introduced and the recursive least squares algorithm （RLS） is used to optimize the weights in order to improve the accuracy of location.The number simulations of the data fusion algorithm for static and mobile targets positioning are given,and results show that compared with the single node localization,the positioning variance of fusion algorithm presented in this paper decreases on the order of 10-1 to 10-2.

关 键 词：分布式定位 权值 数据融合 最小均方误差 递归最小二乘算法 

分 类 号：TP393.02[自动化与计算机技术—计算机应用技术]