Exploiting a mmWave Large Antenna Array from a High-Altitude Platform (HAP) for Rural Vehicular Communications

Intelligent Transport Systems (ITS) are envisioned to revolutionize the quality of experience for drivers and passengers by facilitating a broad array of vehicular services and applications. Several network architectures have been proposed in the literature which exploits cellular infrastructure and Roadside Units (RSUs). However, these approaches suffer from various challenges, particularly on rural roads. First, because of the rural topology, coverage and connectivity issues arise due to the need to deploy a sufficient number of RSUs. Second is that current technologies, i.e., Dedicated Short-Range Communications (DSRC), can only support a maximum data rate of 27 Mbps. This is grossly inadequate considering the growing requirements of next-generation vehicular applications.

In this thesis, we address some of these challenges by proposing the exploitation of a High-Altitude Platform (HAP), operating in the mmWave band, and equipped with a large antenna array. Firstly, a framework for characterizing the HAP-assisted vehicular network with a representative traffic demand is presented. Our approach is efficient, scalable, incorporates real-world maps and publicly available demographic data. To improve the network management, we propose a clustering scheme for grouping vehicles into clusters and show the effectiveness of our algorithm in achieving stable clusters. Secondly, by exploiting the HAP model, we investigate the coverage and capacity performance of the clustered group of vehicular users. To the best of our knowledge, this is the first such attempt that characterizes the performance of a HAP-assisted vehicular network for a rural scenario. Finally, we propose an optimal relay selection scheme for extending the performance of vehicles with poor radio conditions. In lieu of an exhaustive search, the relay selection is formulated as an optimization problem and solved using the Kuhn-Munkres algorithm. Simulation results show our scheme can provide an 82.4% improvement in throughput when compared to without relaying.