Censoring for improved performance of distributed detection in wireless sensor networks

Abstract

In this paper, we consider the problem of binary hypothesis testing for distributed detection in wireless sensor networks in which a transmission censoring scheme is employed. The sensor nodes transmit binary decisions to the fusion center (FC) for final decision making. Sensor nodes with unreliable observation samples censor transmission to FC. By having two thresholds at each sensor node, a sensor node censors transmission if its log-likelihood ratio (LLR) falls between the two thresholds, whereas the more informative sensor nodes transmit their decisions to the FC. In this case of censoring some of the less-informative sensor nodes, and under our system model assumptions, we demonstrate that censoring can lower the probability of error at FC even if there is enough power and rate to support the transmissions of the less-informative sensor nodes.