Detection of Information Flows

    T. He and L. Tong

    Submitted to IEEE Transactions on Information Theory, 2007.

    Abstract
    The detection of information flows in a wireless network is considered.
    Given transmission epochs of a set of monitored nodes, the problem is
    to decide whether there is an information flow through these nodes by
    analyzing the transmission timing. Under bounded delay or bounded
    memory constraint, we show that when an information flow is embedded
    in chaff noise, there exists a threshold on the level of chaff below
    which the flow can be detected with vanishing error probabilities. When
    the level of chaff noise is above this threshold, on the other hand,
    there exist transmission schedules that make the information flows
    completely undetectable. The threshold on chaff noise for two-hop flows
    using Poisson transmission schedules is obtained in closed form. It
    is shown that while information flows with a small number of hops can
    be hidden in chaff noise, the rate of undetectable information flow
    vanishes as the number of hops increases.