A new adaptive sampling method for energy-efficient measurement of environmental parameters

Abstract

A new method involving adaptive sampling of environmental parameters at sensor nodes has been proposed and developed. The method involves determination of stochastic models of the environmental parameters so that forward and backward predictions could be performed accurately with minimal energy. When difference between measured values and corresponding model-predicted values falls outside predefined threshold interval, measured values are upheld and used to update the model by computing for new model parameters while keeping stochastic order of the model constant. The proposed method has been applied and validated using environmental field data. Favorable results were obtained. The method was also used to determine, with sufficient accuracy, numerical values of missed measurements during a low frequency sampling.