BREADCRUMBS: Building up Robust and Efficient routing Algorithms for Drones by integrating Connectivity and Risk awareness in a Suburban air Mobility Bvlos Scenario

In recent years, the use of drones, or Unmanned Aerial Vehicles (UAVs), has expanded across various industries, including agriculture, emergency response, and logistics. However, these operations are generally restricted to within Visual Line of Sight (VLoS), limiting their effectiveness. VLoS refers to drone operations where the pilot can maintain an uninterrupted visual observation of the UAV without any visual aids, such as binoculars or cameras. This type of flight is currently the standard for drone operations in Italy, primarily due to safety and regulatory concerns, ensuring the pilot can immediately react to potential obstacles or hazards. Extended Visual Line of Sight (EVLoS) is also permitted under specific conditions, allowing for slightly longer operational ranges with additional safeguards, such as spotters. Beyond Visual Line of Sight (BVLoS) operations instead open up new possibilities but also introduce complex challenges, particularly in ensuring constant and reliable communication. This thesis aims to develop a novel framework for efficient BVLoS flight planning. This research focuses on a case study in a realistic suburban scenario, also considering a realistic distribution of GSM towers. In such suburban environments, the key challenges revolve around ensuring reliable communication in BVLoS scenarios, adopting the Rural Macrocell (RMA) channel model. Key performance indicators (KPIs) such as communication latency, signal strength, and Quality of Service (QoS) are analyzed to ensure continuous and reliable connectivity. By leveraging existing ground cellular infrastructure, we aim to ensure that UAVs remain under constant control without the need for costly, dedicated networks. One of the central challenges in suburban areas is the variability in communication coverage and quality, especially when using the Rural Macrocell (RMA) channel model. This thesis proposes solutions that propose a graph-based framework, integrating an ad-hoc channel model to provide optimal flight routes. Experimental results demonstrate how drones can navigate through suburban environments while maintaining robust connections with cellular networks, enabling effective communication without the need for new infrastructure. These methods have the potential to be applied in various real-world BVLoS missions, from agricultural monitoring to infrastructure inspections, enhancing both the performance and safety of UAV flights. The results of this work lay the groundwork for future advancements in BVLoS connectivity and their potential deployment in larger-scale applications.