The Research On Localization Of Mobile Nodes In WSN

At present, WSN, as an emerging information obtaining technology,can be widely applied in many areas due to its unique advantages. Node localization technology is a basic technology which is based on the application of WSN and its localization accuracy directly determines the quality of applications in WSN. With the expansion of people’s needs, the traditional static WSN localization technology can not meet the needs of people. The mobile WSN localization technology draws more and more attention. The support vector machine is a machine learning method based on statistical learning theory and with the aim of structural risk minimization,it can solve the problems of small sample size and the local minimum.Based on full analysis of previous technologies, the following researches are conducted in this paper:In order to solve the large energy consumption from self-localization and computation of MCL algorithm-based WSN mobile nodes, the small sample SVM model is built. And the nodes to be located can be located by the SVM regression model. The localization way by this model saves the ranging energy consumption of the traditional node localization algorithms.This paper presents a WSN mobile node localization algorithm based on SVM regression training. This algorithm fully inherits the concept of sampling box from the traditional mobile node localization algorithm and introduces the prediction of the node movement direction, which improves the localization accuracy of the node. And as the algorithm omits the large energy consuming process in ranging and location computing, the energy consumption is largely reduced in localization.This paper presents a positioning system solution of vehicle networks based on the MSVML algorithm according to the structured analysis of vehicle networks, command center and mobile unit were analysised and designed datailly.