Remotely piloted drones, also known as unmanned aerial vehicles (UAVs), have become increasingly important in recent years, making a notable impact on different applications, such as surveillance, precision agriculture, and parcel delivery. As interest in UAVs operating within urban environments continues to grow, the use of UAVs in populated areas presents a significant challenge. Aerial highways (AHs) are emerging as a promising solution for enabling optimal management of UAVs traffic and secure urban operations. Like terrestrial highways, AHs are predefined lanes in the sky that UAVs must adhere to when navigating in the urban sky.
Having identified cellular networks as the key technology for providing ubiquitous and high-performance connectivity for UAVs operations. Owing to their ability to navigate 3D space, due to the typical line of sight (LoS) condition, UAVs experience favourable channel gains across multiple cells within the networks. Although this permits UAVs to perceive sufficient reference signal received powers (RSRPs), it also results in comparable signal power from multiple neighbouring cells, leading to high interference and reduced signal quality. Moreover, the inherited high channel correlation due to LoS and the proximity of UAVs along AHs pose significant challenges when adopting massive multiple-input multiple-output (mMIMO) systems.
In this thesis, we investigate, from a system-level network perspective, how the prior information of the planned AHs can be leveraged by network operators to optimize their deployed terrestrial cellular network to provide connectivity along AHs while maintaining unchanged ground service.
Specifically, we first focus on legacy 4G long term evolution (LTE) networks can be optimized to maintain minimum connectivity along the AHs. To achieve this, we propose a gradient-based optimization framework that leverages AH information and adjust the vertical tilt of terrestrial LTE sectors, ensuring optimal power distribution and meeting coverage requirements along the AH.
Then, moving towards 5G new radio (NR), we investigate how multi-antenna systems can be leveraged, together with the information of the planned AHs, to provide enhanced coverage. Therefore, we propose an optimization framework that utilizes gradient computations to design optimal coverage beams from surrounding network cells, identifying the most critical beams while configuring the remaining according to legacy configuration designed primarily for ground service.
Acknowledging that optimizing solely coverage is insufficient to enhance the UAVs data rate effectively, we propose then a solution capable of merging coverage and capacity optimization. Specifically, we envision that UAVs performances along the AHs can be optimized by solely controlling UAVs serving cells. Therefore, we propose a novel solution to optimally plan 5G new radio (NR) coverage beams across the network to strategically control the UAVs cell association and, in turn, maximize UAVs data rate without affecting ground users. This solution can increase the final UAVs data rate along the AH by adjusting the network solely at its planning stage, allowing the network to remain unchanged regardless of instantaneous conditions. To solve the problem, we propose a heuristic solution based on a novel metric that captures the multiplexing capability, average channel quality gain, and interference. Leveraging this metric, we introduce a two-stage framework in which, first, we optimally split the AH into multiple segments and identify the corresponding set of optimal serving cells. Secondly, we define the optimal set of transmitted new radio (NR) coverage beams from the identified cell set to maximize their coverage.