Composite Hypothesis Testing with Intermittent Observations

In this paper, we consider the detection problem with intermittent observations, due to the unreliable shared communication link between local sensors and the fusion center. Detection performance is analyzed using Neyman-Pearson criterion of maximizing the probability of detection, for a given probability of false alarm. The detector performance is compared, with and without intermittent observations, and a formal approach is presented to restore the original detector performance.