Abstract: In the case of large area and low distribution density of nodes, the accuracy of localization algorithm was too low because of the excessive measurement error between nodes. The author proposed a node localization algorithm based on dividing the information of each neighbor node into several local network blocks. The algorithm first combined wireless sensor network node localization technology with the dimensionality reduction method in machine learning domain, and then according to the rule of the closer the distance between the nodes, the measurement accuracy would be the higher, relying on the information related to a certain communication radius of each neighbor node to build network blocks. Finally, the network block was built into a global coordinate system, and the coordinates of each node were mapped by using the global as well as the detailed information of anchor nodes. The simulation results show that the proposed algorithm is better than the other node localization algorithms in the localization accuracy of nodes.

Key words: localization algorithm, wireless sensor network, manifold learning, local network block, node, measurement