Interval data fusion with preference aggregation in wireless sensor network: Energy-accuracy trade-off in presence of outliers

For balancing measurement accuracy and energy consumption in a wireless sensor network in presence of outliers it is proposed sensor accuracy enhancement algorithm SensAcc and active node selection algorithm ActiveNode based on the interval data fusion method IF&PA. The results of numerical experimental investigation of the developed algorithms are presented. It is shown that the SensAcc provides the reduction of the uncertainty of measurement result at least tenfold comparing with the uncertainty of multisensor readings under possible existence of failed nodes. Simulation results have shown the ActiveNode allows to reduce the cluster nodes energy consumption approximately threefold.