The push for improved driving experience, including the ultimate goal of driverless vehicles, has fueled the development of Connected and Autonomous Vehicles (CAVs). Connectivity is an important enabler for CAVs, as it allows for autonomous vehicles to directly participate in the intelligent transportation system (ITS) and make collective intelligent decisions. 5G mobile networks play an important role in providing vehicular connectivity as mobile networks support mobility by default. However, the density of vehicles is high in urban areas that poses challenges for the support from mobile networks. Deploying ultra-dense small cells using millimetre wave (mmWave) communication is one promising solution since ultra-dense small cells deployment addresses the network capacity and mmWave addresses the interference arisen from the dense deployment. However, due to the relatively narrow footprint of a mmWave beam, using mmWave communication for fast moving vehicles requires careful allocation to utilise its potential and avoid selfishness in radio resource usage. This project sets an ambitious goal of designing smart and dynamic algorithms to manage mmWave beam allocations in CAV environments. The project first investigates solely on the mmWave beam allocations and proposes smart algorithm using both traditional optimisation technique for benchmarking and modern machine learning algorithms for practical operation. It then studies the radio resource allocation under multi radio access technologies (multiRATs) that is the likely deployment setup before the full mmWave small cell networks become practical in the future. Finally, our study is extended to vehicle-to-vehicle communication using mmWave that can support delay sensitive message exchanges and range extension. As experimentation needs expensive infrastructure, the foreseen secondments offer partners without infrastructure on their premises to test their design on other partners that possess the infrastructure.
