Improving preamble sampling performance in wireless sensor networks with state information

Abstract

Wireless sensor networks (WSN) are characterized by stringent power constraints. Researchers focus on improving their power efficiency in order to allow building up networks that survive for long periods of time on limited battery resources without any human intervention. Improvements have been proposed for every hardware and software component, including every networking layer of the sensor nodes. Concerning the MAC layer, different strategies have been proposed, among which preamble sampling is promising. This technique limits the energy spent by nodes that are not recipients of a frame, and transfer a part of this cost to the transmitter. However, the basic technique leaves room for enhancements. We show that such optimization is possible by adapting the duty cycles of the nodes taking into consideration their depth in the routing tree. We experimentally demonstrate that a significant reduction in energy consumption may be obtained when using different preamble lengths throughout the network. Our results show that, with such inter-layer cooperation, the energy efficiency of nodes closer to the sink can be improved by up to 50%.