On the interference channel with causal cognition

This paper considers the causal cognitive interference channel that consists of two full-duplex transmitter-receiver pairs sharing the same channel, where one transmitter can causally learn the message of the other transmitter through a noisy link. This channel models unilateral source cooperation. The work focuses on the generalized degrees-of-freedom of the symmetric, i.e. the two interfering links and the two direct links have the same strength, sum-capacity for the Gaussian noise channel. It is shown through evaluation of various achievable schemes that known sum-rate upper-bounds are achievable to within a constant gap regardless of the strength of the channel parameters. The achievable schemes are quite simple in the sense that only superposition coding is used, while it is shown that more complex schemes using binning can achieve a smaller gap.