On InterFerence Channel with Generalized Feedback (IFC-GF)

This work studies cooperative communication strategies for Interference Channels with Generalized Feedback (IFC-GF). IFC-GF models wireless peer-to-peer networks where several source-destination pairs share the same channel and, because of the broadcast nature of the wireless channel, each transmission can be overheard by the other users. In this model, the interference due to simultaneous communications furnishes the basis for cooperation among otherwise uncoordinated users. For the case of two source-destination pairs, we propose a coding strategy that combines the ideas of (i) information splitting (introduced by Han and Kobayashi for IFC without feedback), (ii) block Markov superposition coding (introduced by Cover and Leung for multiaccess channels with perfect feedback), and (iii) backward decoding (introduced by Willems in the context of multiaccess channels with cribbing encoders). We show that by exploiting the overheard information with the proposed scheme, users achieves collectively higher data rates than the case where the overheard information is neglected. We conclude by showing how our model reduces to well studied multiuser channels.