Font Size: a A A

Research On Range-Free Node Localization In Anisotropic Wireless Sensor Networks

Posted on:2016-02-02Degree:DoctorType:Dissertation
Country:ChinaCandidate:W X HeFull Text:PDF
GTID:1108330464469540Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
Node localization is one of the important issues in WSNs(Wireless Sensor Networks). Current range-based localizations usually need complicated hardware, which are cost ineffective and are difficult in the deployment aspects as compared to range-free ones. To overcome the limitations(i.e., hardware cost, energy consumption, scalability, limited range, etc.) of the range-based localization schemes, range-free solutions have been proposed. Range-free solutions always assume the WSNs are isotropic and ignore the irregular deployment and non-ideal communication model(anisotropic). They calculate the coordinates of sensor nodes, either, based on the radio connectivity information among neighboring nodes, or based on the sensing capabilities that each sensor node possesses but less precise than range-based methods and easily lead to the localization error. Nowadays, the range-free algorithms are more popular than the range-based methods due to the cost factor.In this dissertation, the mobile-beacon aided and kernel-based range-free localization algorithm is proposed, in order to achieve better accuracy. Simulations are performed to show the advantage of the proposed algorithms. The main research results of the dissertations are as follows:(1) Three mobile-beacons aided range-free localization algorithm is proposed in two dimensional spaces. The algorithm is based on the idea of equal triple coverage and can generate multiple virtual beacons at the same time. By trilaterat the simultaneous localization can be achieved and thus high localization accuracy is obtained at the expense of small energy consumption. The simulation results show the less average traffic and shorter beacon moving distance as compared to the earlier algorithm with DV-Hop and MB DV-Hop.(2) The concept of coplanarity degree is introduced and a layered 3D localization algorithm with multiple mobile beacons is proposed to solve the problem of coplanarity of node localization in three dimensional spaces. The algorithm extend the algorithm proposed in chapter 3 and screen the virtual beacons for localization by setting coplanarity threshold. Finally, the quadrilateration is used to localize the unknown node. The idea of equal triple coverage in two dimensional spaces is extended to three dimensional spaces by layering path of mobile beacons. The simulation results show that the algorithm can achieve higher localization accuracy and less average traffic compared with 3D DV-Hop.(3) A range-free node localization algorithm based on kernel function is proposed in this dissertation. Many traditional localization algorithms performed better by utilizing adjacent node information in WSNs provided that the WSNs are isotropic. When the WSNs are anisotropic, the proposed algorithm utilizes the global information of WSN by kernel function. The relationship between the estimation distance and the Euclidean distance of any two nodes is described by a nonlinear model. Simulation results show that the algorithm can achieve higher accuracy as compared to the DV-Hop algorithm, especially in the anisotropic case.(4) Using the single kernel function to deal with the case of large scale network, unisomorphic samples and the unevenly distributed data is unreasonable in high dimensional eigenspace. Whereas the combined kernel function which consists of two or more kernels is taken into account performs better. A range-free node localization algorithm is proposed based on the combined kernel which combined RBF and polynomial kernel, which combines the advantages of the RBF and polynomial kernel. The simulation results show that the algorithm can achieve higher localization accuracy as compared to the localization based on a single kernel.
Keywords/Search Tags:WSNs, node localization, anisotropic network, mobile beacon, support vector regression, kernel function
PDF Full Text Request
Related items