A new mutton logistics tracking algorithm for Internet of things based on PSO and neural network

Minning Wu¹, Fei You¹, Feng Zhang^{1, 2} 

¹School of Information Engineering, Yulin University, Yulin 719000, China

²School of automation, Northwestern Polytechnical University, Xi’an 710072, China

In order to improve the particle filtering precision and reduce the required number of particles, to solve the neural network training algorithm has slow convergence speed, easily falling into local optimal solution, proposed a target tracking algorithm based on PSO particle filter, using of Bayesian method to sample the prior information and coupled PSO algorithm. For the existence of intelligent wireless sensor network energy constrained sensor nodes, limited communication features, the PSO optimization is introduced into the distributed particle filter algorithm to solve the existing distributed particle filter network traffic load is heavy and node energy consumption of high disadvantage. Then, we propose a new particle filter algorithm based on PSO and neural integration the algorithm makes full use filter tracking historical information, combined with predictions of particle filter, the detection signal of the sensor nodes were isolated, thus achieving the target tracking. Simulation results show that the target tracking algorithm based on particle filter PSO and neural integration can use a smaller computational cost, multi-target tracking problem solving, and practical system to meet the demand.