Research And Application On Localization Algorithms Of Wireless Sensor Networks

Wireless sensor networks (WSN), a newly emerged technology, is increasingly utilized in modern industry, daily life and so on. As the Internet of Things is vigorously promoted by government, WSN which is a core technique of Internet of Things will have a great boom. During this process, localization which is a critical technique in WSN will lay a solid foundation for the other applications.The research on WSN localization in this dissertation is separately carried out from three phases of a localization system:1) Distance estimation. The four models of Received Signal Strength Indicator are analyzed, and a data processing algorithm which is divided into two steps including abnormal value elimination is designed according to the easily interfered feature of RSSI.2) Position computation. A novel side deployment of anchors is presented and the corresponding characters are deeply investigated. Consequently, the definition of ill-conditioned unknown node is provided. Through numeral simulation between two types of anchor deployment, the position computation methods are analyzed and contracted.3) Refinement algorithm. The refinement algorithms based on optimization algorithm are discussed, which advances the boosting-speed refinement algorithm. Furthermore, compare and contrast between it and the speedest refinement algorithm are studied by simulation. Finally, the selection principles of several distributed and centralized localization algorithms are presented.The localization algorithms are applied and tested in the real WSN localization system after they are researched and designed by numeral simulation:1) Exposition of the independently developed localization platform is regarding design of communication protocol, further development of Z-Stack, construction of man-machine interface and so on.2) 3D visual simulation is introduced and applied in the man-machine interface, which changes the status quo that the majority of available localization platforms are 2D.3) After the algorithm is utilized by MATLAB and Visual C++ hybrid programming, the experimental environment and procedure are described, and the experiment results are obtained by comparation with CC2431.