Intelligent Traffic Flow Management Techniques for Dense Low Altitude Airspace and Urban Air Mobility

Urban Air Mobility (UAM) promises to revolutionize urban transportation, offering an attractive solution to the challenges of traffic congestion and air pollution in rapidly growing cities. Central to this vision is the deployment of Unmanned Aircraft Systems (UAS), which, while promising reduced travel times and lower carbon emissions, face significant hurdles due to the current safety-related limitations associated to low-altitude airspace operations, especially over densely populated areas. Traditional Air Traffic Management (ATM) systems, primarily designed for manned aircraft, are under considerable strain from the increasing demand for UAS services and lack adequate provisions for low-altitude operations. These factors highlight the critical role of UAS Traffic Management Operators (UTMOs) in maintaining a balance between demand and capacity in congested airspaces. However, legacy ATM systems are ill-equipped to handle the unique challenges of UAS operations, necessitating a paradigm shift in traffic management approaches. In response to the above challenges, this thesis explores the opportunities offered by the integration of Artificial Intelligence (AI) into Traffic Flow Management (TFM) systems. A thorough evaluation of AI algorithms was conducted to enhance Demand and Capacity Balancing (DCB) services, aiming to alleviate the workload on UTMOs and elevate operational efficiency and safety. AI algorithms promise a transformative impact on UAM by enabling more dynamic, adaptive traffic management strategies. Furthermore, the study delves into the critical aspect of AI explainability and trustworthiness, emphasizing the need for transparent and fair algorithms to foster user confidence in these automated systems. This aspect is crucial in ensuring equitable and unbiased decision-making processes in UAM operations. Additionally, the interconnected nature of ATM Communications, Navigation, Surveillance, and Avionics (CNS+A) systems, while offering several opportunities for AI integration, also presented the range and impact of potential challenges. These include vulnerabilities to both cyber and physical security threats, with an expanded attack surface and additional propagation mechanisms in CNS+A systems, requiring new defensive measures both at system and component level. This study identifies potential vulnerabilities in the proposes CNS+A architecture and sets foundations for future research on possible AI-driven mitigation strategies to enhance ATM and UAS Traffic Management (UTM) systems. The uncertainty inherent in UAM services, primarily due to fluctuating demand and capacity in urban low-altitude airspace, is the one of the key challenges tackled in this thesis. Technical solutions to enhance system safety and operational efficiency are explored, highlighting the importance of robust and flexible management systems in adapting to these uncertainties. The culmination of this research is the proposal and development of an adaptable TFM system framework, utilising hybrid AI algorithms tailored explicitly for UTM applications. The practical application of this research is demonstrated in the design, implementation and testing of a prototype TFM system. This prototype underwent rigorous functional testing and performance assessments within a simulated UTM environment, offering valuable insights into the feasibility and effectiveness of the proposed solutions. In summary, this paper provides a comprehensive analysis of integrating AI with Urban Air Mobility (UAM), addressing critical issues such as airspace resource scarcity, air traffic management system limitations, cybersecurity threats, and operational uncertainty. This research significantly contributes to the development of UAM by creating an adaptable AI-enhanced Traffic Flow Management (TFM) system, which paves the way for a more efficient, safer, and sustainable UAM framework. The findings and methods presented in this study have the potential to shape the future of UAM, facilitating machine-assisted automation in airspace operations and laying the groundwork for fully automated airspace management. The findings and methodologies outlined in this study have the potential to inform future developments in UAM, offering a blueprint for the challenges and opportunities of integrating highly automated aircraft and UASs into urban environments.