The Research Of Target Tracking Algorithm In Wireless Sensor Network

The improvement and development of microelectronics technology,wireless communication technology and computer technology have promotedthe development of low-power sensors. Furthermore, wireless sensor network(WSN), a kind of emerging network with unlimited potential has emerged at theright moment. Wireless sensor network is composed of a lot of energy-limitedand low-power sensor nodes in an area. Compared to the traditional network, thewireless sensor network has many advantages such as broad coverage area, lowpower, strong self-organization, high reliability，wireless transmission and so on.Therefore, it is well suited for target tracking and locating such as militarymonitoring, navigation, creature behavior research, traffic management, etc.Meantime, there are some potential problems on its application. For example,the volume of a sensor node is so small that it has limited power. As aconsequence, how to use the energy of nodes in network efficiently is one of themajor challenges for target tracking in wireless sensor network. In targettracking, the reasonable construction of dynamic cluster structure can not only reduce the network energy consumption effectively but also reinforce thelifecycle of the network. And the effective filtering algorithm which can avoidwaking up nodes in the network blindly and reduce the time delay of trackingprocess can further enhance the tracking performance of the network.In this paper, a method of dynamic cluster construction for target trackingin WSNs is proposed. It proves that the new method can reduce the networkenergy consumption effectively in tracking process. And then, the particle filteralgorithm is improved so as to reduce the calculating load. The particle filteralgorithm shows its good performance in solving problems in non-linear andnon-Gaussian motion tracking systems.The main work done in this paper is listed as follow. First, based on thecurrent research on some methods of dynamic cluster construction of targettracking algorithm in wireless sensor network, this paper put forward a methodof dynamic cluster construction. By improving methods of the choose of clusterheads and the selection of the cluster members in the process of dynamic clusterconstruction, this paper is aimed to achieve the further effect of energy saving.The choose of cluster heads is based on the energy of the nodes and the distancefrom node to target. The selection of the cluster members takes the target’smovement speed and the distribution of network nodes into consideration. TheMATLAB simulation showed the effective of the dynamic cluster construction.Second, the paper analyses and compares several kinds of filtering algorithms.Particle Filter algorithm has shown its good performance in processing non-linear motion. However, it must be improved because of its heavycomputation. The simulations show that the algorithm in this paper can reducethe computation burden of particle filter under the tracking precision condition.Thus the algorithm is more suitable for target tracking in WSNs.