低信标节点密度传感器网络的启发式定位算法

Heuristic localization algorithm for low density of anchor nodes in wireless sensor networks

针对传统无线传感器网络定位算法在信标节点密度较低时定位误差较大的问题,提出了一种基于加权最小二乘的迭代定位算法。首先,该算法将最小二乘估计器进行线性化处理,并根据节点之间的距离量测的方差进行加权,获得加权最小二乘估计器。然后,通过分析定位过程中影响未知节点定位精度的因素,基于空间点分布的椭圆拟合,对信标节点的线性化程度、离散程度和偏置程度进行量化。以该量化特征作为未知节点定位估计的置信评分,结合信标节点集扩展和信标节点剪裁控制迭代过程中的误差传递。仿真和实验结果表明,与传统定位算法相比,该算法可以在极小信标节点占比情况下获得较高的定位精度,且在各向异性网络中表现出较好稳定性。

For the traditional localization method in wireless sensor networks,a small number of anchor nodes usually result in larger localization error.To decrease the error,a novel weighted least square estimator based iterative localization algorithm is proposed in this study.Firstly,the algorithm utilizes the linear least linear least square estimator with the distance variance as a weight which can provide better accuracy and is robust to the distance noise.Then,by analyzing the influence of the space distribution of the anchor nodes on localization accuracy,the anchor position features based on eclipse fitting are provided.Based on these features,the algorithm makes a decision on localization and anchor set cutting which can reduce the localization error from propagating in the network.Compared with the traditional method,the proposed algorithm can achieve better localization accuracy under the condition of minimal anchor nodes ratio.Besides,the algorithm is robust in the anisotropic nodes deployment.