Research On Dynamic Deployment In Wireless Sensor Networks

Wireless sensor networks (WSNs) which covers sensing, communications, electronics and other fronts is a hot area of academic research in recent years. With the advantages of self-organization and robustness is WSNs extremely suitable for monitoring information in harsh environments by tossing nodes randomly. Then the problems about monitoring hole and connectivity caused by random tossing will seriously affect the normal operation of the network.This paper studies the dynamic deployment technology of wireless sensor networks. Two different deployment strategies are presented for different scenarios. The two strategies all make the hole eliminating and coverage enhancing as the main target.The structure and characteristics of WSNs are firstly introduced in this paper then the evaluation of dynamic deployment is also presented pointedly. After analyse the existing deployment strategies combining with specific application scenarios, two deployment strategies are presented as follow:(1) Presented a clustering deployment strategy based on virtual force (VF) which was divided into two phases. One was inter-cluster deployment and the other one was nodes deployment in same cluster. First a clustering was done based on the location information of the nodes. Then the clusters would be overall moved to make a effective dispersion in intensive nodes. To enhance the dispersion effect a cluster splitting mechanism was introduced. Based on the characteristics and purpose of the two-stage mobile units two different VF algorithm were designed. The deployment strategy significantly improved the area coverage comparing with other strategy based on VF.(2) Presented a deployment strategy based on virtual force directed particle swarm optimization (PSO) which aimed at mixed deployment containing both static nodes and mobile nodes. Taking into account both the lack of VF algorithm in a mixed deployment in which mobile nodes could not move because of the shackles from static nodes and the lack of PSO algorithm which had a slow convergence speed and a poor global search capability, the VF and PSO algorithm were combined in the this strategy. The deployment strategies was significantly better then the strategies which only relied on VF or PSO algorithm in optimization effect and convergence rate. Based on binary perceptual model, the simulations about on the two strategies presented were done. The results showed that the improving effect of two strategies was obvious and the coverage could be more than90%after optimized.