The proliferation of pervasive computing applications relying on battery-powered devices and wireless connectivity is posing great emphasis on the issue of power optimization. While node-level models and approaches have been widely discussed, a problem requiring even greater attention is that of power associated with the communication protocols. We propose a high-level modeling methodology based on Finite State Machines useful to predict the energy consumption of given communication tasks with very low computational cost, which can be applied to any protocol. We use this methodology to create a power model of Bluetooth that we characterize and validate experimentally on a real implementation.
ISLPED ''04: Proceedings of the 2004 international symposium on Low power electronics and design