Compression Techniques for Minimum Energy Consumption

Low power sensor nodes are integral parts of largescale wireless sensor networks, which find extensive applications in domains such as military surveillance, transportation control, and environmental monitoring. This paper explores energyconsumption trade-offs associated with lossless data compression in low power sensor nodes. We focus specifically on compression of acoustic signals. The processing of acoustic signals in digital form and subsequent transmission of the signals across a network involve significant energy consumption by the transceiver of a sensor node. To reduce the volume of data, and the associated energy consumption of transmission, it is useful to apply data compression. The process of data compression will generally reduce the transmission energy at the expense of energy expended due to additional computation in the embedded processor. This paper focuses on this trade-off between computational and communication-oriented costs associated with using data compression in sensor nodes. Understanding this trade-off requires an integrated approach that considers the strength of compression, the energy expended by the embedded software in performing the compression, and optimization of the embedded software to further improve overall impact of compression. This paper develops a framework for design and experimentation to facilitate these goals, and reports preliminary experimental results based on this framework.