Polygon-Based Tracking Framework in Surveillance Wireless Sensor Networks

We propose a new tracking framework by organizing nodes into a polygonal spatial neighborhood in order to detect and track unauthorized traversals in surveillance wireless sensor networks. During a tracking, the neighborhood is further constructed ahead of the target traversal, and this features a timely forwarding with guaranteed delivery property. Instead of estimating future movement and position separately in a polygon, we find an original target tracking path in a graph, and create a brink on the graph called "critical region\´\´ by introducing a brink detection algorithm to know a target\´s route, and to also achieve reliable inter-node communications. In addition to the basic design, an optimal sensor selection algorithm was developed to select which sensors to query, and dynamically guide the target information to a sink. Simulation results validated that the proposed approach has better tracking accuracy, reduced localization error, and is robust to strong environment noise, while using a minimum number of sensors.