Wireless Backhaul Architectures for 5G Networks

This thesis investigates innovative wireless backhaul deployment strategies for dense small cells. In particular, the work focuses on improving the resource utilisation, reliability and energy efficiency of future wireless backhaul networks by increasing and exploiting the flexibility of the network. The wireless backhaul configurations and topology management schemes proposed in this thesis consider a dense urban area scenario with static users as well as an ultra-dense outdoor small cell scenario with vehicular traffic (pedestrians, bus users and car users). Moreover, a diverse range of traffic types such as file transfer, ultra-high definition (UHD) on-demand and real-time video streaming are used.
In the first part of this thesis, novel dynamic two-tier Software Defined Networking (SDN) architecture is employed in backhaul network to facilitate complex network management tasks including multi-tenancy resource sharing and energy-aware topology management. The results show the proposed architecture can deliver efficient resource utilisation, and QoS guarantee.
The second part of the thesis presents wireless backhaul architectures that serve ultra-dense outdoor small cells installed on street-level fixtures. The characteristics of vehicular communications including diverse mobility patterns and unevenly distributed traffic are investigated. The system-level performance of two key technologies for 5G backhaul are compared: massive MIMO backhaul using sub-6GHz band and millimetre (mm)-wave backhaul in the 71 – 76 GHz band.
Finally, innovative wireless backhaul architectures delivered from street fibre cabinets for ultra-dense outdoor small cells with vehicular traffic is proposed, which can effectively minimise the need for additional sites, power and fibre infrastructure. Multi-hop backhaul configurations are presented in order to bring in an extra level of flexibility, and thus, improve the coverage of a street cabinet mm-wave backhaul network as well as distribute traffic loads.