The Research Of Localization Algorithm In Wireless Sensor Networks

Wireless Sensor Networks,which integrates wireless communication, embedded computing and information processing,is a new distributed and organized networks of data acquisition.It can be used to various areas,ranging from military application to people tracking and environment monitoring.Localization is an fundermental issue of many wireless sensor networks applications,and has extremely vital significance in research.First of all,this paper introduced the reseach background of WSN and its application in domestic and abroad,summarized related research work in positioning technology and algorithm,the problems faced as well as the means of classification,discussed several the most representative localization algorithms and systems on the basic principles and its application limitations in detail.Secondly,around the MDS-based localization algorithm for wireless sensor networks,the MDS technology was explained in detail.The application of MDS technology,which is a multi-variable method of statistical analysis,was analyzed,and several MDS-based localization algorithms were discussed.Then,the method of coordinate conversion for MDS which is a mapping algorithm for relative coordinates to absolute coordinates was analyzed.Householder transformation and its application in MDS-based localization algorithm were intensively discussed.An improved algorithm based on Householder transformation was proposed.The improved algorithm was applied to MDS-MAP(P) algorithm and was intensively compared with primary algorithm by simulation.The result of simulation show that the improved algorithm which use householder transformation can reduce the impact of anchor nodes' distribution and achieves better localization accuracy and lower time costs compared with other algorithms that adopt Householder method too.At last,an absolute localization algorithm which based on anchor node dissemination was developed in order to reduce the large localization error of fringe nodes when the distribution of anchor nodes is heterogeneous.The proposed algorithm adopted clustering and anchor node disseminating scheme to achieve the aim.Simulation demonstrate that the range of its application is very specific.When the distribution of anchor nodes is undesirable the algorithm can effectively reduce the localization error of fringe nodes by making use of anchor nodes' position.