Fundamental Limits of Communication over Fading Interference Channels

Absence of perfect network state information and presence of output feedback links in wireless networks have both attracted significant amount of interest in recent years. In this paper, we study the impact of lack of perfect network state information on the capacity region of the binary fading interference channel where noiseless output links are available from the receivers to the transmitters. Our results include a new achievability strategy that systematically utilizes the stale network state information and previously received signals at the receivers. Our achievability scheme reveals the importance of output feedback links in the sense that they provide cooperation opportunity among the transmitters, as well as the usefulness of the stale network state information that helps us categorize the previously transmitted bits and elevate them into a subproblem that can be solved optimally. By deriving new outer-bounds on the capacity region of such networks, we show that the delay in learning network state information results in a loss in the capacity region.