ABSTRACT
Event detection allows for a wireless sensor network to reliably detect application-specific events based on data sampled from the environment. Events, in this context, may range from the comparatively easy to detect outbreak of a fire to the more complex stumbling and falling of a patient, or even to an intrusion into a restricted area. Event detection is a special form of in-network data processing for wireless sensor networks (WSNs) that pushes the logic for application-level decision making deeply into the network. Raw data samples from the sensors are evaluated directly on the sensor nodes in order to extract information that is semantically meaningful for the given task of the WSN. The centralized evaluation architecture is possibly the most widely used: All nodes in a WSN communicate exclusively with the central base station, which has much more computing power and energy resources than the sensors nodes.
