EAVS: Energy aware virtual sensing for wireless sensor networks

In order to extend the operational span of wireless sensor networks, we propose to trade energy consumption for average response time by extending the deactivation periods of sensors with a particularly high energy consumption. To compensate for the temporal inavailability of these sensors, alternative, low-power hybrid sensors generate estimates on the probabilities of the occurrence of interesting events, waking up the corresponding main sensor when detection probability justifies it. We demonstrate the principle on a case study of gas detection and analyze its efficiency formally using probabilistic model checking, which is able to compute probabilistic quantified properties pertaining to energy consumption, lifetime expectancy, and response time. The preliminary results confirm significant savings in energy consumption while retaining an acceptable average response time.