The Research On Localization And Target Tracking In Wireless Sensor Network

As a supporting technology of next-generation Internet-Internet of things, Wireless sensor network (WSN) brings us many challenging researching topics. Broadly speaking, WSN localization technology can be divided into two fields, sensor nodes itself located (node localization) and localization of other nodes or targets (target tracking). Therefore, node localization in wireless sensor network is the basis and prerequisite conditions of target tracking technology.The fault of Beacon nodes, as well as the interference of radio channel environment, will affect the performance of localization. To eliminate these interference factors, a new distributed localization algorithm is proposed based on trust management. Firstly the algorithm splits WSNs into several clusters with a certain cluster protocol, and then initiates the global trust rating (GTR) and local trust rating (LTR) of Beacon nodes. To get more precise location, it needs to localize frequently and update the GTR and LTR of beacon nodes. The method can not only improve the precision of the positioning, but also be used to examine abnormal beacon nodes and infer possible interference factors in wireless channel.Further research work was done based on the study results aboved. The system performance characteristics of a strong non-linear non-Gaussian during WSN target tracking, so particle degradation and particle dilution is the main problem of particle filter which is one of currently hot research in WSN target tracking. This paper proposes two improved program of particle filter algorithm, mainly starting from the selected importance function. Adding measurement information the current time into the probability distribution, we can get the approximation of the true probability distribution. So it can solve the particle degradation problem and the scare of sample problem, and therefore, it enhances the tracking ability while target's state is suddenly changed.Finally, the localization method and the improved particle filter algorithm proposed in this paper are validated and compared with other algorithms in the cluster hierarchy for wireless sensor network. The simulation results show that it can achieve a better positioning accuracy and tracking results by use of these methods.