Study On The Localization Alogrithms Of Wireless Sensor Network Based On Control Strategy

The research on the localization algorithms of wireless sensor network (WSN) is focused on the control strategy in this dissertation. First, the analysis framework of localization algorithms based on control strategy is set up, and the analysis tool is also given, according to the modular idea. Second, the existing localization algorithms are classified and researched by using the analysis framework, and the deep undertanding of the algorithm structure is achieved on the basis of preliminary work, and the corresponding relationship between localization algorithms and analysis framework is also set up. Finally, the improved control strategies of localization algorithms designed for three major application environments are proposed based on the analysis framework, and they are proved the effectiveness by simulations. The details of dissertation are desrcibed as below:1. According to the modular idea, the analysis framework of localization algorithms for wireless sensor network is set up in this dissertation. First, the composition principle is achieved on the basis of deep analysis of the influence factors and the evalution index of localization algorithms, and the IPO pattern for modular analysis and control strategy are also proposed. Second, the semantic web is chosen as the analysis tool for analyzing the modules of localization algorithms, and discovering the operation mechanism and constraints, which are the support for the following design of algorithms.2. Seven control frameworks for localization algorithms in this dissertation are proposed on the deep analysis of control strategy which are self-optimization, self-learning, self-identification, self-adaption, self-stability, self-organization, self-coordination. The functions, objectives, and the application patterns for each control framework are also deeply analyzed. The semantic web for analyzing the modules of algorithms is also described in details, including the basic components and the operation mechanism, and the applications of semantic web are also described with examples considering the influnence factors of localization algoritms.3. According to the IPO pattern for localization algorithms, three major application environments are chosen for localization of wireless sensor network, which are distributed localization,3D localization, and the path planning for mobile beacon. The corresponding reference algorithms for the these environments are also proposed, including the DV-Hop, APIT and the S-curve. Based on the analysis of application environments, the coresponding control strategy suites are proposed, and the implementation details are also designed via semantic web. In distributed localization, CRFDV-Hop algorithm based on the hybrid computing model is designed in chapter three. Through the division of the localization system, CRFDV-Hop obtains the estimated location of unknown node in the system initialization phase by using the RSSI and the location feedback mechanism which is on the basis of DV-Hop algorithm. CRFDV-Hop establishes the new localization pattern based on sink in the system run phase, and it significantly reduces the power consumption of localization with the rise of localization accuracy; In chapter four, RTD-APIT based on APIT, proposes the 3D localization algorithm supported by the triangle transformation. By the extension and mirror operation of triangles, RTD-APIT solves the In-To-Out error and the Out-To-In error in APIT, and extends the triangle transformation to the 3D environment. Meanwhile, RTD-APIT solves the strict requirements properly by establishing the location feedback mechanism and the pattern of promoting the virtual beacons; In large scale wireless sensor networks, VFST which using the virtual force model for path plan is proposed in the chapter five. Using the difference between the estimated location and the real location as the judging factor for virtual force, VFST meets the goal of path plan which aiming for the localization requests, and it uses the S-curve algorithm as basic moving pattern which reduces the collinear error effectively. In order to fit the real environment well, VFST proposes the obstacle sensing and avoiding model. The CRFDV-Hop, RTD-APIT, and the VFST which proposed in the paper, are proved having good localization performance and the stability of algorithms by simulations.