Distributed Detection of Information Flows in Chaff

Distributed detection of information flows is considered. The detector detects the presence of information flows by collecting timing information from nodes of interest through channels of finite capacity. The information flows are assumed to be perturbed up to a bounded delay and interleaved with chaff. Joint compression and detection schemes are proposed to achieve reliable detection with inaccurate measurements. Detection performance is analytically evaluated by robustness against chaff as functions of the capacity constraints in the data collection. The proposed detectors are proved to be optimal for their corresponding quantizers. A comparison of their performance gives guidelines on quantizer design.