Research On Node Localization And Moving Object Tracking In Wireless Sensor Network

The emergence and development of wireless sensor network has brought a revolution ininformation field, it has gradually affected every aspect of our lives. Node localization is the keyproblem for the widely use of the wireless sensor network, which plays an important role in manyapplication fields. Target tracking is one of the important applications of wireless sensor networks.With the strong tolerance, high-density deployment and accurate tracking ability, wireless sensornetwork obtains the tracking system’s favor.The research work of this paper is about node localization technology and target trackingtechnology. To improve the accuracy of node localization and target tracking, this paper putsforward some theoretical and practical methods and verifies the effectiveness of the proposedalgorithm through the simulation. The main research contents are as follows:Firstly, in order to improve the node localization accuracy, a WSN localization algorithm isproposed based on trilateration, then using UKF algorithm to the nodes precise location with RSSIas the measurement. This method can improve the localization accuracy through the simulation. Byweighing the relationship between the average location error and computational complexity, thenumber of the best-positioned anchor node is given. Finally, through simulating a specificenvironment, the practicality and application significance of this localization algorithm is verified.Secondly, in order to improve the target tracking accuracy, a target tracking method is proposedbased on a variable threshold sensor model in the binary sensor network, then using distanceweighted algorithm for single point position estimation and the polynomial curve fitting method tofit the target trajectory. It confirms the optimal reference threshold and boundary value, which isused as the parameters of target tracking. Compared with the fixed threshold methods, this methodhas a good performance.Finally, this paper further researches on the target tracking in binary sensor network when thereis error in node localization. Tracking precision is affected seriously by the node localization errorwhen the polynomial curve fitting algorithm is used. This paper combines the variable thresholdsensing model and the particle filter tracking algorithm, compared with the polynomial curve fittingalgorithm, tracking using particle filter algorithm has high accuracy.