This project will investigate how machine learning based artificial intelligence strategies can be used to control both the radio resources and network topology in small cell 5G networks, likely to be used in dense urban areas, when traffic is dominated by ultra-high definition video from vehicular users. It is expected that base stations and users will have multiple (smart) antennas, enabling rapid adjustment in the coverage and capacity density. Wireless backhaul/fronthaul is expected in such networks, and the project will examine how multi-hop mm-wave techniques can be exploited. Resource assignment strategies will be developed which deliver the required capacity with low latency. A mixture of simulation and analysis will be used to assess performance. Stochastic Geometry, Game theory and Markov analysis will be particularly important analytical tools.