Abstract: In wireless sensor networks,thewireless sensornodes have to cover the area to be monitored effectively.In order to reduce the coverage holes and improve the coverage rate in wireless sensor networks,this paperproposed a new deployment strategy of wireless sensor network nodes based on improved particle swarm optimization.Taken network coverage as the fitness function,thedeployment of sensor nodes would be formalized as an objective optimization problem.By employing the k-means clustering algorithm,the population was divided into several sub-populations.In addition,the population was re-dividedinto new sub-populations dynamically,which could weaken particles on the pursuit of local optima,realize the improvement of basic PSO algorithm,effectively solve the “premature” problem of basic PSO algorithm,and accelerate the convergence ofthe algorithm.Experimental results show that this deployment strategy can reduce the coverage holes in wireless sensor networks as much as possible and effectively improve the network coverage rate.Compared with the results of elementary particle swarm optimization,the conventional genetic algorithm and swarm optimization algorithm,its coverage rate was increased by 4.11%,9.75% and 5.25%.

Key words: wireless sensor networks, PSO, k-means clustering, sub-population