Research On Targets Detecting And Tracing Based On Wireless Sensor Network

Wireless sensor network (WSN) consists of several small nodes with capability of sense, compute and communicate which are random distributed in physical environment. The network is very suitable for moving target detection and tracking tasks for it's self-organized easily arranged and small nodes as well as close observation of the environment. The WSN targets tracing system has a wide prospect and space of application. It will provided a prefect monitor platform for many kinds of application, such as defence area invasion, enviorment monitoring, forest fire prevention even to seabed investigation and planet probing.The paper, in the first place, studied and analyzed the target detecting and tracking mechanisms using dynamic cluster by deviding the tracing progress into four steps: single target locating, route estimating, nodes awaking and recovery mechanism. Proposed a algorithm for target location which based on least squares two-step optimum strategy for the problem of improving the positioning accuracy and reducing the cost of network computing in practical applications of targeting position. Introduced Piecewise quadratic curve fitting method to calculate target trajectory and estimate the target position, which is a path selection mechanism to overcome the low curve fitting discontinuous, to ensure the accuracy of low trajectory curve fitting and reducing the computation costs. Furthermore, in this paper, a recover strategy for possible target lost was proposed. The simulation results indicated that the ameliorated target tracking algorithm had improved the effects of target positioning and prediction, obtained high tracing accuracy and low network computing costs.On that basis, the paper expended the tracking algorithm to the multiple target tracking area. Proposed the combined tracking strategy for the undesirable effects which was arised from sensor nodes collection to multiple target signals aliasing, that is to agglomerate more than one track into a dynamic tracking cluster when the targets were closed to each other and the locate and maintain strategy of the dynamic cluster was introduced in the paper. At the targets separating stage, a decomposition algorithm of dynamic cluster was proposed according to boundary searching to determine the scale of target separation effectively and to lay the foundation for the new dynamic cluster of follow-up. The simulation results demonstrated that the algorithm can merge or separate dynamic tracking cluster and can realize multi-targets dynamic tracking effectively.