| Summary: | 碩士 === 國立虎尾科技大學 === 資訊工程系碩士班 === 104 === This paper mainly studies the intruder detection in the wireless sensing network. The proposed methods track the intruder by estimating the possible region of its location. And predict its moving path. There are many kinds of positioning technologies designed for indoors and outdoors. Current research issues focus on how to arrange sensors, how to detect the target, and how to position the location. This thesis proposes new intruder detection methods in wireless sensor network by using the concept of repellent force and attractive force. A virtual force model transferring the distance between sensors to moving distance is proposed to make the sensors organize themselves while sensors are deployed. The virtual force model will make the sensors fully cover the deployed area. Next the Delaunay Triangulations are sketched and used to mark the possible positions of intruder. The connected Delaunay Triangulations which mark the intruder will point out the tracking of the intruder. By dynamically changing the position of sensors and reorganizing Delaunay Triangulation, the position accuracy can be effectively improved.
This thesis proposes three tracking methods. The first one is Virtual force Delaunay triangulation (VDT), which uses the Delaunay Triangulation and track the intruder. To improve the accuracy, the method using VDT with the ability to predict the position of intruder named Prediction of Virtual Force and Delaunay Triangulation(VDT) is proposed. To enhance the VDT to track the high moving speed intruder, the Level virtual force Delaunay Triangulation(L-VDT) is proposed. The intruder with straight moving path and curve moving path are discussed. Simulation results show that, in the case of high deploying density the proposed. VDTP method generates small error on average position area than the triangulation positioning method. The VDTP positioning area on the case that intruder moving path is straight can be reduced 34.3% if sensor are deployed in high density, and reduce 24.1% for the case of curve moving path. When the scenario changes to a large deploying area, the VDTP can increase 41.2% positioning accuracy for the straight moving path and increase 20.7% for the curve moving path. The overall average position error of VDTP is less than VDT about 8.4%.
|
|---|
