ON RANDOM EVENT DETECTION WITH WIRELESS SENSOR NETWORKS

Wireless sensor networks hold great promise as an enabling technology for a variety of applications. Data collection and event detection are two such classes of applications that are broadly representative and which have received considerable attention in the literature. While wireless multi-hop data collection has achieved operational lifetimes on the order of a year, we are unaware of lifetimes exceeding a few days or weeks for wireless multi-hop event detection sensor networks. Our thesis is that sensor networks for event detection are constrained by two factors which do not similarly affect data collection sensor networks. The first factor is that no appropriate sensing, signal conditioning, and signal processing architecture has been broadly implemented to support event detection in distributed systems that are simultaneously energy, space, time, and message complexity-constrained. The second factor is that middleware for services such as time synchronization, localization, and routing are predominantly and unnecessarily proactive. A comparison of data collection and event detection will serve to illustrate the subtle but important differences between these applications. Fundamentally, data collection is a signal reconstruction problem in which the objective is to centrally reconstruct observations of distributed phenomena with high spatial and temporal fidelity. Performance metrics for such applications include the