基于交替修正牛顿法的分布式传感器定位算法

为了克服锚节点位置误差影响定位精度这一问题,提出了一种基于交替修正牛顿法的分布式定位算法。首先,将无线传感器网络表示的无向图划分成多个部分重叠的子图,建立可独立求解的子图内定位问题,子图内未知节点根据不准确的锚节点位置和测距信息采用修正牛顿法得到初步估计位置,再融合求平均得到估计位置；其次,根据第一步结果和测距信息采用修正牛顿法更新锚节点位置,使其位置更为精准；最后,未知节点再根据相对准确的锚节点位置更新估计位置。实验结果表明,与现有的分布式算法相比,所提算法具有更好的定位性能和扩展性,能够应用于较大规模的无线传感器网络。

Distributed sensor localization algorithm based on the alternatemodified newton method

In order to overcome the problem that the localization accuracy is affected by the anchor node location error, a distributed localization algorithm based on the alternate modified newton method is proposed. First, the undirected graph represented by the wireless sensor network is divided into several overlapping subgraphs, and a localization problem in the subgraph that can be solved independently is established. Based on the inaccurate anchor node location and distance measurements, the modified newton method is used to obtain the preliminary estimated location, and then the estimated location is obtained by merging and averaging. Second, based on the results of the first step and distance measurements, the modified newton method is used to update the anchor node location to make the location more accurate. Finally, according to the relatively accurate anchor node location, the estimated location of the unknown node is updated. The results show that compared with the existing distributed algorithms, the proposed algorithm has better localization performance and scalability, and can be applied to large-scale wireless sensor network.