Random access with imperfect state information: An information theoretic approach

This paper studies the throughput of a slotted random access system with cognitive users. Transmission success is determined by the information theoretic capacity of the channel and the rates of the active users. Users have the ability to estimate the other user's state through sensing, and to adapt their respective transmission. We derive an outer bound on the capacity region for an arbitrary channel distribution. A transmission strategy based on single user detection and time division is optimal for the binary deterministic channel, and for the Gaussian channel the capacity is within a constant gap of the outer bound. The throughput region is parameterized by the traffic arrival rate and sensing capability. Rate adaptation improves the throughput only for some arrival rates and sensing errors.