Local-global iterative linear embedding optimization-based positioning algorithm in sensor network

Abstract

ABSTRACT In the distributed iterative positioning algorithm of wireless sensor networks, there will be positioning errors due to the existence of noise and selection of reference nodes; the errors will be further propagated when the nodes selected are taken as the descendant reference nodes. Currently, error propagation is not considered in the distributed iterative algorithm; to reduce the propagation errors, the local linear embedding optimization-based positioning algorithm is proposed in this paper. This algorithm follows the essential criteria of the beacon nodes and takes the mean of the positioned node location and the same node location of the previous generation as the current node location; the particle swarm optimization algorithm of population classification and dynamic learning factor is adopted to accelerate the positioning speed. According to the simulation result, compared with the classical PSO algorithm, the algorithm proposed in this paper has better real-time and effectively improves the positioning accuracy for it has not only saved the positioning time but also effectively reduced the influences of propagation errors on the positioning accuracy.