Monitoring Scheme for Event and Danger Detection in Wireless Sensor Networks
Kálmán Tornai, András Oláh and János Levendovszky
Event and danger detection is of primary importance in surveillance systems. In this paper, we propose a novel distributed event detection scheme based on data acquired by a Wireless Sensor Network. The proposed event detection scheme takes into account the spatial and temporal correlation of data. This scheme is appropriate for complex monitoring systems because it requires a relatively low computational power from the sensor nodes but ensures a high rate of correct detection. Distributed operation is ensured by running the outlier detection (being one of the main components of the scheme) on the sensor nodes. The other main component is the clustering which is executed on the base station. Implementing these two components on separate system modules, we have managed to considerably decrease the number of radio transmissions needed for event detection, as a result, the proposed scheme is energy efficient and maximizes the lifespan of the system.
Keywords: Event detection, monitoring systems, sensor networks, Ad-hoc networks, energy efficiency, prediction based detection method, spatio-temporal correlation