Hotspot-Based Cellular Vehicle-to-Everything Networks (C-V2X) for Real-time MEC-Assisted Applications

As more connected vehicles pervade our roads, achieving their original objectives which include safety, efficiency and enjoyable experience becomes more important. Achieving CAV
(Connected and Autonomous Vehicles) objectives is hinged on reliable and efficient communication between participating road entities. Cellular Vehicle-to-Everything (C-V2X) networks have been touted to have great potential in supporting CAV operations. The development of more Mobile Edge Computing (MEC) assisted applications for CAVs means more vehicles will require access to cellular infrastructure. In high vehicular demand scenarios, cellular infrastructure might struggle to satisfy both traditional and vehicular users. A cluster and relay approach have been suggested to create a hotspot scenario, where strategically located users could stream and relay information to vehicles within its cluster. In highly dynamic driving environments, clusters tend to be unstable leading to poor Vehicle-to- Vehicle (V2V) link performance.
This thesis describes how we approached these problems by answering three key questions; could the re-clustering process be remodelled to improve V2V network performance? At what point in time should re-clustering be initiated to minimize overhead while sustaining performance? Is there an optimal number of clusters that satisfies efficient bandwidth resource utilization for both Vehicle-to-Infrastructure (V2I) and V2V communication? We sought answers to these research questions through simulation and explored two distinct driving environments: the urban and highway environment. By limiting our work to use-cases requiring short-time real-time download of traffic data, we were able to focus our efforts towards improving cognate performance indices such as throughput, jitter and reliability. Our re-clustering remodelling effort and scheduling schemes yielded significant improvement in throughput and jitter and stability performance, while our attempt on bandwidth resource utilization succeeded in improving V2I and V2V user bandwidth.