Node Localization Algorithm Based on Hop-count Quantization in Wireless Sensor Networks

A novel algorithm based on hop-count quantization and extended Kalman filter based on multidimensional scaling (MDS-HE) is proposed to improve the localization accuracy of nodes in wireless sensor networks. The integer hop-count can be transformed into a real number hop-count by partitioning a node's one-hop neighbor set into three disjoint subsets and estimating the distance between nodes by the areas of the intersection regions of hop ring segmentation. The transformed real number hop-count is a more accurate representation of distance between nodes. The real number hop-count matrix is applied to the multidimensional scaling (MDS) method, and the extended Kalman filter is applied to refine accurately the coordinates of nodes. The localization performance of MDS-HE algorithm is simulated and analyzed in WSNs which is composed of nodes deploying randomly over a region. Simulated and experimental results show that the performance of the MDS-HE algorithm outperforms the DV-Hop method and the classical MDS method in the case of different number of nodes. The MDS-HE algorithm is exceedingly accurate in case of the enough anchor nodes.