Interference networks with cognition and cooperation

Interference is a central issue that needs to be addressed in multi-user wireless networks. This dissertation examines the potential of mitigating interference through the cognitive capability of transceivers and user cooperation. From an information theoretic perspective, we study the performance limits of such networks in the context of the classical interference channel. The cognitive radio network is modeled as an interference channel with one cognitive transmitter (ICCOT) from both the noncausal and causal perspective. For noncausal ICOCT, we propose a novel coding scheme, which yields a new achievable rate region that generalizes previously existing rate regions in the literature. In the absence of the primary transmitter, this proposed achievable rate region recovers Marton\u27s region for the general broadcast channels. The proposed coding scheme gives the capacity region for a class of semi-deterministic IC-DMS. The dissertation also examines the causal ICOCT which is in practice more relevant as it does not require the non-causal knowledge of the primary transmitter\u27s messages at the cognitive transmitter. Instead, noisy channel feedback is utilized in a causal fashion to harvest the benefit of the cognitive capability. Capacity bounds for this causal cognitive radio channel model are proposed for various interference regimes. Many proposed cognitive radio technologies allow spectrum sharing as long as the interference imposed on the primary receiver from the cognitive transmitter, the so-called interference temperature is negligible. Motivated by this real world constraint as well as the intractability of the general cognitive radio channel with causal feedback, we study a simplified model, the so-called cognitive Z channel where the link between the cognitive transmitter and the primary receiver is absent. Capacity bound for the cognitive Z channel is obtained for various interference regimes and the impact of cognitive capability is carefully calibrated to better understand the performance cost trade-off. In addition to the cognitive radios, we also study the role of user cooperation in the framework of two-hop interference networks. The study is motivated by the emerging wireless mesh network and we characterize achievable symmetric rates for different parameter regimes under decode-and-forward (DF) relaying and amplify-and-forward (AF) relaying. The results show that one-hop interference channel capacity is not necessarily the bottleneck for the transmission in the two-hop networks, for both the DF and AF schemes