Analytical and experimental sensor node energy modeling in ambient monitoring

Wireless sensor networks are currently being deployed over a vast array of applications to collect, process and reliably transmit data from a source event to a data sink. This in turn requires reliable power supply, mainly in the form of primary or secondary batteries storing energy to supply the various functional discrete components along with accurate prediction of sensor node life time, based on estimated state of charge of the energy source. The paper presents two approaches for energy modeling of sensor node life time in ambient monitoring scenarios concerned with periodical sampling of measurable parameters from the indoor environment. The analytical approach involves modeling the individual components of a sensor node and carrying out simulations in different duty-cycling regimes to obtain an expected life time. Experimental modeling has also been carried out over short and medium term repeatable deployments in order to gather meaningful sets of data suitable for statistical interpretation. Battery discharge characteristics are determined in order to extract high level information used in system design for indoor ambient monitoring. final goal is to incorporate validated experimental results back into simulation approaches and enable a prediction framework for the Memsic IRIS XM2110 motes in order to approach the direct and inverse problem: accurately determine node life time for various sampling and communication regimes or choose optimal configuration parameters for an imposed node life time.