Abstract
Full-duplex (FD) communication is considered to be a promising technology to address the problem of radio spectrum congestion due to the increasing demand for wireless communication services. This thesis investigates the problem of asymmetric data flow between downlink and uplink in a system with symmetric radio resource allocation, i.e. FD. The aim is to devise a model by exploiting the coupling nature of FD downlink and uplink transmission powers to maximise the downlink data rate while ensuring the quality of service (QoS) of uplink transmission. Whilst the aim is to maximise the spectrum efficiency gain that can be achieved with FD enabled cellular networks, this thesis presents D-MAC: a novel distributed MAC protocol. D-MAC allows interfering nodes to cooperate and implement an interuser interference estimation mechanism to gather the interference information with minimum overhead. Furthermore, D-MAC includes interference aware joint user equipment (UE) scheduling and power allocation methods. Computing the optimal resource allocation and transmission rates is non-linear, non-convex combinatorial optimisation and achieving a globally optimal solution is not possible in practice. Therefore, D-MAC uses a low complexity iterative approach for resource allocation and ensures that channel bandwidth is shared fairly across all UEs. Finally, a detailed evaluation of D-MAC performance is presented and compared to the half-duplex (HD) system, FD system with round robin and FD system with exhaustive search. It is shown that the D-MAC achieves as much as 62% throughput improvement compared to an HD system with symmetric traffic model and 49% with asymmetric traffic model in an outdoor multi-cell scenario under the limits of practical self-interference cancellation. Further, D-MAC also reduces the round-trip propagation delay, involved from the point at which the UE sends the request to the base station to the point at which the UE realize the result of contention.