Efficient decision fusion for cooperative wireless sensor networks

This work considers the problem of decision fusion in cooperative wireless sensor networks, where the decisions are transmitted to the fusion center via a number of relays. The likelihood ratio test is incorporated to derive the optimum fusion rule. Moreover, a suboptimal rule is derived to reduce the complexity of the optimum rule while providing reliable probability of error performance. A lower bound is derived to evaluate the error performance of the optimal rule while a closed-form analytical expression is derived for the error probability of the suboptimal rule. The derived analytical results are verified using Monte Carlo simulation under various operating conditions. The obtained results show that the reliability of the used sensors, number of sensors, and the number of relays are the main factors that determine the error performance at the fusion center. Moreover, the probability of error suffers from an error floor that appears at moderate and high signal-to-noise ratios, which is mostly caused by the number and the error performance of the sensors.