CollECT: Collaborative event detection and tracking in wireless heterogeneous sensor networks

Abstract

Tracking is an important application in wireless sensor networks (WSNs), especially for the urgent event of interest. Recent research has paid much attention to the WSN wherein all sensor nodes are identical in sensing units, but the utilization of different types of sensor nodes has not been widely explored. In the paper, we propose a fully distributed protocol, CollECT, to event detection and tracking in wireless heterogeneous sensor networks (WHSNs), which consists of various types of sensor nodes with different sensing units. The main idea of CollECT is collaboration, by which the same type of sensor nodes construct the attribute region, represented by a convex polygon, whereas the different types of sensor nodes determine whether the event occurs. In CollECT, three procedures, vicinity triangulation, event determination, and border sensor node selection, are proposed to construct the attribute region, to determine the occurrence of the event, and to select border sensor nodes to stand for the event boundary, respectively. Simulation results validate the performance of CollECT in terms of accuracy of event tracking and fitness of the border sensor nodes selected. Approximately 87% sensor nodes within the event region can be correctly identified on average. Additionally, the border sensor nodes selected by CollECT are beneficial to efficiently stand for the event boundary.