Study Of Weighted Centroid Algorithmin Wireless Sensor Networks

As the basis for the implementation of the Internet of Things （IOT）, the WirelessSensor Network （WSN） was confirmed as the critical information transmissiontechnology by national policy.With the development of society, positioning andnavigation services will be more and more needed to people. Such characteristics ofmicrosensor as low power, low cost, multi－function that makes positioning technologybased on WSN has a significant cost advantage and great potential research value.Besides, the positioning accuracy of WSN is not very satisfactory so far that makes thespace of research in this field broadly.Weighted Centroid Algorithm （WCA） was chosen to make the further researchon location technology for its being applied to the localization project of laboratory. Itcombines the advantages of RSSI ranging technology and centroid algorithm as onealgorithm. Implementation of this algorithm is simple, without the installation ofadditional hardware devices. In the process of positioning experiment we found that theinitial weight coefficient values seriously affect the positioning performance of WCAalgorithm, and on the edge of location area the positioning precision is poor in the caseof only a few anchor nodes was laid.In order to truly meet the needs of practicalapplication, two improved algorithms would be proposed to improve the positioningaccuracy of WCA.The main research results in this paper are described as follows.①Through the mathematical derivation and analysis for WCA formula, werealized exactly from the perspective of the derivative how the changing of weightingcoefficients of WCA will affect the location results, which isprovided a theoretical basisfor adjusting parameter for the weighted centroid localization algorithm to makepositioning controllable.②An improved algorithm was proposed by optimizing the initial weightedparameter.A more accurate positioning result will get throughParameter ofWeight Self－Adjusting（PWSA）according to the principle of negative feedback. Simulation resultsshow that the overall positioning accuracy of this algorithm surpass traditional weightedcentroid algorithm.Under the condition that the weight parameter is not less than1.8,the positioning accuracy of this algorithm has nothing to do with the initial weightedparameter, so there is no need to constantly adjust the weighting coefficient during theprocess of algorithm application. ③To further improve the positioning accuracy on the edge of location area, animproved weighted centroid algorithm was put forward based on Virtual Static AnchorNodes（VSAN）.In this paper, VSAN have been proposed to participate in thepositioning, which are fabricated by the combination of location information of theknown anchor nodes and cosines law. This algorithm inherits the favorablecharacteristics from WCA like low－complexity, easy to implement. Without theinstallation of additional anchor nodes, the positioning accuracy can be improvedobviously.