Study On Localization Algorithm Of Wireless Sensor Network

Location information has played an increasingly important role in many applicationsof wireless sensor networks, such as monitoring activities and so on. The simple, quickand precise way to obtain location information is either to set up manually or to installGPS, which will waste a vast amount of time and human resources. A better way to obtainlocation information is to use the localization algorithm.In this paper, we mainly focuse onthe research of wireless sensor network location algorithm based on multi-dimensionalscaling.Firstly, the research background, significance and status for WSN are summarized inthis paper based on large amount of related literatures. And the framework, characteristicsand typical localization algorithms of wireless sensor networks are introduced.Secondly, in this paper, we give details of classical multidimensional scaling and itsapplication in wireless sensor network location algorithm. Based on the analysis of theclassic MDS-MAP of the location algorithm, this paper presents Hop-Euclidean basedMDS-MAP(D) location algorithm, which uses the clustering method to divide large-scalenetwork into several local networks with cluster head, and in the local positioning, usingthe Hop-Euclidean algorithm in place of the shortest path distance to calculate Euclideandistance between the nodes neighbor. Obviously, this algorithm not only increases thepositioning accuracy, but also is beneficial to the expansion of the network.Thirdly, in this paper, we propose an improved algorithm, which makes up for theshortage of the distributed weighted multi-dimensional scaling localization algorithm thatcan not adapt to the network connectivity change, the irregular network topology and theconvergence speed is slow. We use the weighted mechanism and neighbor choicemechanism that comprehensively considerate the number of one hop neighbors, the nodepositioning precision and the location error. We also introduce the steepest descentmethod to ptimize the goal cost function.Lastly, the two improved algorithms are tested by simulations via Matlab accordingto localization error, network topology and so on. Simulation results show that these algorithms proposed in this paper improve the localization accuracy and are moreadaptable to the irregular topology, large-scale network.