Fundamental performance analysis of event detection in wireless sensor networks

A wireless sensor network (WSN) is designed to perform various information processing tasks such as event detection, target tracking and data classification. Comparing with traditional centralized networks, networked sensing offers unique advantage in improved robustness and scalability. Measures of performance for these tasks are well defined, including detection of false alarms or misses, classification errors, and track quality. In this paper, we present a fundamental performance analysis of event detection in wireless sensor networks. Our performance analysis is based on a new detection scheme - double sliding window (DSW) even detection. We compare it theoretically against the fixed threshold event detection approach. Results show that the DSW detection scheme performs better than the fixed threshold detection under certain circumstances