基于连通度的分布式加权多维尺度节点定位算法

A Connectivity-Based Distributed Weighted-Multidimensional Scaling Algorithm for Nodes Location in Wireless Sensor Network

研究了迭代优化方法在无线传感器网络节点定位中的应用,针对多维尺度分析定位技术和传统的梯度迭代优化方法,根据数值实验确定了迭代步长和网络连通度之间的函数关系,提出了一种基于连通度的分布式多维尺度分析节点定位算法(a connectivity-based distributed weighted multidimensional scaling algorithm,简称dwMDS(C))。该算法首先根据网络的平均连通度确定迭代步长,然后对每个未知节点的局部代价函数进行优化求解。实验表明该迭代算法收敛快速且稳定,比基于SMACOF算法的dwMDS(G)算法在定位精度上有明显的提高。

This paper focuses on the methods of localization with iterative optimization in wireless sensor networks. After studying the Multi-dimensional scaling algorithms and traditional gradient optimization methods,we determine the function relation between iteration step size and network connectivity based on numerical experiments and introduce a connectivity-based distributed weighted multi-dimensional scaling algorithm. First, this method calculates the iteration step size with the average value of connectivity, then it optimizes the local cost functions. Experiments show that this method performances a faster and more sta- ble convergence than dwMDS（G）algorithm which is based on SMACOF algorithm.