Distributed detection with fuzzy censoring sensors in the presence of noise uncertainty

Abstract Distributed detection is an essential task of wireless sensor networks (WSNs) with stringent limitations of energy and processing resources. In most of detection scenarios, the optimal threshold depends on the noise power which itself consists of some uncertainty in practice. In this paper, we use fuzzy hypothesis test (FHT) for modeling the noise power uncertainty. Moreover, we propose an optimal censoring scheme by applying the Neyman–Pearson lemma on the FHT. It is proved that the optimal censoring method is obtained by simply comparing the energy of the observed data with a threshold. Interestingly, we show that the threshold would depend on the local communication constraint as well as the noise uncertainty bound. We evaluate the performance of the proposed censoring algorithm by resorting to the equal gain combining (EGC) as the relevant fusion rule, using common receiver operating curves. Simulation results show the effectiveness of the proposed method in counteracting the uncertainty.