Abstract
Wireless Sensor Networks have proven their capability to deal with problems where wide and hardly accessible areas need to be monitored. Among the other systems there are also sensor networks in which nodes can (voluntarily) modify their positions to better adapt to changes of monitored phenomenon. One of the major issues arising in these situations is the energy consumption: as all the movements affect the batteries lifetime, the life of a sensor can be extended by equipping the device with power generators that exploit renewable sources, albeit this solution does not always avoid a full battery discharge. In this paper a Fluid Stochastic Petri Net modelling framework is presented to provide a wide evaluation of all the factors that contribute to the energy dissipation in mobile wireless sensor nodes. The framework allows the generation of extensible and composable models capable to evaluate the energy consumption due to sensing, communication and movement functions as well as the impact of power saving mechanisms on the energetical balance of the node. The approach is applied to a marine sensing problem and is validated by comparing the model analysis results with experimental results achieved through an existing off-the-shelf sensor network simulator.