Research On Node Localization Algorithm In Anisotropic Wireless Sensor Networks

Wireless Sensor Networks(WSN)have been widely used in many fields such as ecological monitoring,industrial security,military warfare,agricultural production and target tracking due to their high reliability,low cost and easy deployment.Node location technology is the foundation of wireless sensor network applications.However,most positioning studies at home and abroad focus on the standardized network environment,namely,isotropic WSN(nodes are located in the regular monitoring area)and homogeneous WSN(nodes have the same communication radius).But in some practical applications,anisotropic WSN and heterogeneous WSN are ubiquitous.Therefore,the research of node localization algorithms for anisotropic WSN and heterogeneous WSN has more theoretical value and practical significance.The localization algorithms of anisotropic WSN and heterogeneous WSN are analyzed in detail.In this paper,we have proposed an optimized localization algorithm for anisotropic WSN or heterogeneous WSN based on artificial intelligence algorithm.The specific research contents and main innovations are as follows:1)For anisotropic wireless sensor networks,a novel expected hop progress(EHP)localization algorithm based on particle swarm optimization(PSO)is proposed.Firstly,we select the effective anchor nodes by introducing the control parameter MaxHop;Secondly,the average hop distance in anisotropic WSN is calculated by the EHP method of isomorphic networks;Finally,particle swarm optimization algorithm is used for localization optimization.The experimental results demonstrate that our algorithm has higher positioning accuracy and faster convergence speed in anisotropic WSN.2)For heterogeneous wireless sensor networks,a localization algorithm based on support vector regression machine is proposed.On the basis of EHP algorithm in heterogeneous WSN,we use support vector regression machine(SVR)to model and analyze the location problem.Firstly,the algorithm reduces the computational complexity and achieves faster convergence speed by improving the classical EHP algorithm.Secondly,the anchor nodes are divided into target anchor nodes and training anchor nodes.The distance vector between target anchor nodes and training anchor nodes and their specific coordinates are used as training samples of support vector regression machine to obtain the learning model.Then,the SVR model is used to predict the coordinates of unknown nodes.The experimental results demonstrate that this localization method can achieve better localization effect in heterogeneous wireless sensor networks.Even for anisotropic heterogeneous WSN,our algorithm always outperforms its counterparts.