Wireless Sensor Network Localization Algorithm With Grouping Particle Swarm Optimization

Wireless sensor network(WSN)as an emerging technology has gotten widely attention since it appeared.WSN are widely used in military,civilian and other scenes.In WSN,Node localization is one of the key technologies.In most cases,only with location information,the data collected has the research value.Usually,it is unrealistic to install GPS equipment for each node Since Limited resources of node and cost constraints.So self localization algorithms are important and many researches are focus on it.Particle swarm optimization(PSO)is one of Swarm intelligence algorithms.As it is simple and usually can obtain accurate solution so it is widely used in searching the best solution.The standard PSO is easily trapped in the local best but the global,in other words the particle swarm is precocious.When premature convergence occurs,the result may be not able to meet requirement.How to avoid premature convergence and accelerate convergence is a huge challenge.Many researchers tread node location as a problem of searching best value and get the location by PSO after optimization.In these algorithms,every node has a random particle swarm.Generally,the location result is good.This paper proposes grouping particle swarm algorithm with composite optimization(CGPSO)for node location.First,divide the particle swarm of every node into two parts.Then every sub particle swarm use different algorithm to search the best solution.During that,the two swarms communicate with each other by predetermined strategy.This paper studies the details of grouping firstly and then develops the communication strategies for increasing the energy utilization rate and reducing the probability of premature.This paper develops two different algorithms for each sub particle swarm.They both optimize the location information for each dimension.The two algorithms are favorable to improve the performance of communication and finally result.The experimental results show that the CGPSO perform better than standard PSO and grouping PSO without optimization.