Distributed particle swarm optimizer for wireless sensor networks.

摘要

In this work, a diffusion particle swarm optimization (DPSO) algorithm is proposed to cooperatively estimate a monitored parameter by the sensor nodes in an ad-hoc wireless sensor network (WSN). Here, every sensor node of a wireless sensor network is equipped with a PSO algorithm to estimate a parameter of interest. A novel diffusion scheme is used to cooperatively estimate the parameter by sharing the local best particle and the corresponding particle error value to the neighboring nodes. The performance of the DPSO algorithm is improved by applying different enhancements to the PSO algorithm. Therefore, different types of the DPSO algorithm proposed are: the DPSO algorithm with variable inertia weight (DPSO-VIW), the DPSO algorithm with variable constriction factor (DPSO-VCF), the diffusion modified PSO (DMPSO) algorithm, and a hybrid DPSO-LMS (DPSO-LMS) algorithm. The simulation results reports a great improvement brought about by the DPSO algorithms over the non-cooperative PSOLMS (NCPSO-LMS) algorithm, the diffusion least-mean-squares (DLMS) algorithm and the diffusion recursive-least-squares (DRLS) algorithm.;Keywords : wireless sensor network (WSN), particle swarm optimization (PSO) algorithm, least mean squares (LMS) algorithm, recursive least squares (RLS) algorithm, diffusion protocol, inertia weight, constriction factor.