Dynamic Voltage Scaling Techniques for Energy Efficient Synchronized Sensor Network Design

Building energy-efficient systems is one of the principal challenges in wireless sensor networks. Dynamic voltage scaling (DVS), a technique to reduce energy consumption by varying the CPU frequency on the fly, has been widely used in other settings to accomplish this goal. In this paper, we show that changing the CPU frequency can affect timekeeping functionality of some sensor platforms. This phenomenon can cause an unacceptable loss of time synchronization in networks that require tight synchrony over extended periods, thus preventing all existing DVS techniques from being applied. We present a method for reducing energy consumption in sensor networks via DVS, while minimizing the impact of CPU frequency switching on time synchronization. The system is implemented and evaluated on a network of 11 Imote2 sensors mounted on a truss bridge and running a high-fidelity continuous structural health monitoring application. Experimental measurements confirm that the algorithm significantly reduces network energy consumption over the same network that does not use DVS, while requiring significantly fewer re-synchronization actions than a classic DVS algorithm.