Cooperative Interference Mitigation in Wireless Cellular Networks

In this thesis, we explore potentials of cooperative communication at receivers' side in the downlink of wireless cellular networks for interference mitigation and signal enhancement. We investigate two regimes of latency and reliability of communication: one for providing high long-term average user rates for supporting a quality of service that is suitable for human visual and auditory perception, and the other for providing an ultrareliable low-latency machine-type wireless communication for streaming information in control applications. We consider three different models of cooperative communication. First, we study deploying a multi-antenna relay node that provides a nearby cell-edge user with extra dimensions over an out-of-band relaying link. We model this scenario by a Gaussian multiple-input multiple-output (MIMO) relay channel with correlated noise across relay and destination antennas, and analyze the capacity of this channel. This type of relay deployment is most effective when the number of receive antennas is small and the number of relay antennas is large enough. Second, we study deploying a multi-antenna half-duplex amplify-and-forward relay node that simultaneously provides multiple cell-edge users with extra dimensions. We show that the optimized relaying significantly improves the long-term average rates of cell-edge users, even after accounting for the extra bandwidth required for half-duplex relaying, provided that the relay is equipped with sufficiently many antennas. Third, we study cooperation among receivers for combating fading and mitigating interference for ultrareliable low-latency wireless communication. We consider multiple interfering broadcasts from controllers to their corresponding actuators. The recently-proposed Occupy CoW protocol efficiently exploits the spatial diversity of distributed receivers for combating deep fading. It consists of two consecutive phases: the broadcast phase and the cooperation phase. However, it avoids interference by orthogonalization; hence, its required bandwidth for achieving ultrareliability is not scalable in network size. We observe that full frequency reuse in the broadcast phase with successive interference cancellation notably improves the scalability of this protocol. We propose two schemes depending on whether interference cancellation or interference avoidance is implemented in the cooperation phase, and show that both outperform Occupy CoW, each in its own preference region.Ph.D