Study On Node Self-localization Algorithm For Wireless Sensor Networks

The wireless sensor network node localization technology is one of the supportingtechnologies and it not only provides localization information of the monitoring events,but also it is the basis of network topology control and routing assist management. Inthe real applications, the random distribution of sensor nodes and the complexity ofenvironment have a big influence on the localization performance. Therefore, how toimprove the localization accuracy in wireless sensor networks has become a hot anddifficult research field. As a core part of localization technology, the research for nodeself-localization algorithm in wireless sensor network has important academicsignificance and application value.This paper introduces the architecture, basic features, key technologies andapplications of wireless sensor network. After making an analysis of the basic conceptof node self-localization, the basic principles and classification of localization algorithm,this paper discusses the criteria of localization performance evaluation in detail, focuseson analyzing some typical existing range-free and range-based node self-localizationalgorithms. Based on an in-depth study on centroid algorithm and DV-Hop algorithm,this paper proposes the corresponding improved localization algorithm respectively.The main tasks are as follows:①In order to solve the centroid algorithm’s large localization error and lowpositioning coverage in the wireless sensor network with sparse beacon nodes and smallcommunication radius, this paper puts forward a weighted centroid algorithm based ondistance estimation. This algorithm uses distance vector routing principle to makeunknown nodes can get more location information of anchor nodes, moreover,introduces the distance estimation in the weight calculation of anchor nodes to describethe degree of influence on unknown nodes localization. When calculating the distances,this algorithm defines a new measure to estimate the neighborhood distances by relatingthe proximity of two neighbors to their connectivity difference and uses a roughestimation method for non-neighborhood distances. Simulation results show that thisalgorithm can effectively improve the localization accuracy and positioning coveragewithout additional hardware.②Aiming at the large localization error of DV-Hop algorithm in the anisotropicwireless sensor network with irregular network topologies, this paper proposes a calibration based DV-Hop localization algorithm, which corrects the average hopdistance of unknown nodes and not only considers the whole network characteristics ofaverage hop distance, but also the local characteristics. When calculating the distancebetween neighbors, this algorithm optimizes the calculation method, which solves thehop-distance ambiguity problem effectively. In order to further improve the localizationaccuracy, this algorithm adds the calibration step, which takes the optimizedneighborhood distance as the calibration standard to move the estimated location ofunknown node to a more reasonable one. Simulation results show that this algorithmcan improve the localization accuracy in the irregular network topologies obviously.