Distributed detection and tracking in sensor networks

Our work develops a framework for using a wireless ad-hoc network of distributed sensors for detection and tracking of moving objects. Possible applications might be vehicular traffic monitoring in highways or in battlefields, detecting and tracking people movement in secure areas, clots flowing through the bloodstream or even nuclear devices moving around cities. The central idea is to utilize decentralized processing within the network to build a better picture of what is going on (locally), before reporting the results (centrally). This notion, which relies on collaboration among sensors, reduces false alarms and avoids sending all the sensor measurements back to a centralized processing unit. Simple assumptions are made about the structure of the network and the capabilities of the individual nodes. Uncertainty is introduced in the form of uncertain node positions, uncertain position estimates, false events and an unknown number of objects.