Research On Optimization Localization Algorithm Based On Received Signal Strength Indication In WSN

The deployment environment of nodes in wireless sensor networks (WSN) is usuallycomplex，which causes the communication model between nodes immeasurable. Thedynamic changes of WSN in large-scale environment, as well as seasonal and temperatureinfluence on the nodes, make a big difference between transmitted signals of nodes atdifferent times. Therefore, an adaptive algorithm for the model parameters is needed topropose. Only getting the model parameters through experimental data environment orempirical models are not accurate, this paper aims to combine measured data and thecurrent environment characterized by empirical models to design a parameter adaptivealgorithm. Meanwhile, NLOS errors can not be ignored in complex environments, whichmay solve by the filter for filtering the received signal strength value. Finally, this paperprovides a high-precision localization algorithm suitable for complex environment basedon the adaptive parameters and NLOS error suppressed model.Firstly, the paper sets up some experiments based on Mica2platform, including sightpropagation experiment; single propagating experiment and so on. We research on thepropagation characteristics of received signal strength value by analysis of experimentaldata, and the impact on radio propagation model parameters. In contrast to existingmethods for solving the propagation model parameters, we study an adaptive algorithmthat ensures the model parameters adapt to the dynamic changes of the environment.Meanwhile, as existence of NLOS path in complex environment, the measurement valueincludes large NLOS errors. By comparing the existing judgment for NLOS error, wepropose an improved algorithm that adapted to the actual environment. And we designsbiased Kalman Filter to suppress NLOS error. Finally, based on the weighted centroidlocalization algorithm, we proposed an improved weighted centroid localization algorithmthat based on estimation distance and PSO. The algorithm is validated by simulation andpractical tests. The experiments show that the algorithm can reach better localizationaccuracy in a complex environment.