Abnormal Event Detection and Localization in Wireless Sensor Network Using Compressed Sensing

The main challenge in wireless sensor network is the accuracy and timeliness of event detection and localization. However, for a wireless sensor network, abnormal events are relatively sparse compared with all the events in monitoring area. Compressed sensing theory proposed the idea to recover a sparse signal from a few measurements. In this paper abnormal event detection and localization in wireless sensor network are formulated as a compressed sensing problem. On the assumption that the transmission signal of sensors is binary, the performance of the basis pursuit algorithm combined with alternating direction method of multipliers and other compressed sensing recovery algorithms are analyzed. Simulation results verify the preponderance of the basis pursuit algorithm on accurate detection and localization.