The current ATM system is complex and dynamic. It ensures the efficient movement of aircraft while keeping millions of people traveling safely in the sky. With the rapid growth of air traffic and the emergence of new types of vehicles, such as uncrewed aerial vehicles (UAVs) or drones, the ATM system faces unprecedented challenges and opportunities. AI is a key technology that can help address these challenges and unlock new possibilities. AI can enhance the capabilities of human operators, automate repetitive or low-value tasks, improve decision-making and prediction, and enable new modes of operation. In this post, we will explore how AI can transform the ATM system as UAVs begin flying in the congested airspace of crewed aviation.
One of the main improvements AI brings to ATM is the automation of various aspects of airspace management, such as flight planning, route optimization, conflict detection and resolution, and demand and capacity balancing. AI-enabled platforms leverage data from multiple sources, such as sensors, radars, satellites, weather forecasts, and flight plans, to generate optimal solutions for airspace users and service providers. In addition, these platforms adapt to dynamic conditions and learn from past experiences to improve performance. As a result, automation can dramatically improve efficiency and reduce operating costs by safely optimizing aircraft spacing requirements, efficient weather and capacity-based routing, and reducing human workloads for pilots, air traffic controllers, and ground operations crews. Many companies already see the benefits of automation as they introduce AI functionality into their product lines. One such company is Thales, a leader in ATM solutions, which uses AI to predict traffic flow, optimal routings, and estimated take-off and arrival times. This results in improved operations and reduced operating costs for all the stakeholders in air safety. Thales’ automation complements human capabilities by reducing repetitive workloads thus keeping humans in the loop to focus on more critical tasks. “In this way, AI will enable controllers to cope with the expected growth in air traffic, as well as the complexity of integrating new vehicles,” says Beatrice Pesquet-Popescu, Research and Business Innovation Director for ATM at Thales. “In addition to the growth expected in traditional aircraft, we will have to cope with new vehicles such as drones and stratospheric balloons, circulating in low- or high-altitude airspace.” Thales AI-powered automation provides services in over 85 locations worldwide and roughly two-thirds of all aircraft globally [1].
Another area where AI is improving ATM is in the communications and coordination between different airspace users, especially for beyond-visual-line-of-sight (BVLOS) drone operations. BVLOS operations are those where the drone operator cannot directly see the drone or its surroundings or where the drone is operating entirely without human control. This method of operation requires a reliable and secure way of exchanging information with other airspace users and authorities to ensure safety and compliance. AI helps by providing a distributed network of highly automated systems that communicate via application programming interfaces (APIs) rather than voice. These systems provide real-time constraints and guidance to drone operators, air traffic controllers, commercial crewed aviation providers, and ground crew responsible for safely managing their operations. For example, the Federal Aviation Administration (FAA), National Air and Space Administration (NASA), and other partners are collaborating to develop an Unmanned Aircraft System Traffic Management (UTM) system that enables multiple BVLOS drone operations at low altitudes (under 400 feet AGL) in airspace where FAA air traffic services are not provided. Both organizations have jointly developed a UTM Research Plan to focus on program objectives and roadmap capabilities incrementally implemented over the next several years [2]. One such system that most commercial Part-107 drone operators are familiar with is the Low Altitude Authorization and Notification Capability (LAANC). Using LAANC, operators can file flight plans and receive authorizations in near real-time. LAANC processes most authorization and airspace deconfliction automatically; however, depending on the proposed flight location, requests can be routed to the cognizant air traffic control (ATC) facility for approval from a human airspace controller.
A third application of AI in ATM is providing sense and avoid capabilities for collision avoidance. Using advanced sensors, AI algorithms, and decision-support tools, large amounts of data can be processed on the aircraft, providing timely and accurate alerts and recommendations to pilots, drone operators, and other ATM/UTM users. These systems – called Edge-AI because the processing, analysis, and decisions are made on the aircraft as opposed to a connected ground processing facility – use cameras, lidars, radars, and machine learning to detect and classify objects, estimate their positions and velocities, and generate collision-free trajectories [3]. Technology in use today can even coordinate planning and avoidance strategies in real time with other aircraft sharing the same airspace. In addition to real-time information and guidance, technology is being developed to provide airspace deconfliction in pre-flight using strategic deconfliction. With this technique, the system can coordinate flight routes and spacing minimums during flight planning before aircraft leave the ground. In a joint research project between Airbus (a leading aircraft manufacturer) and The Johns Hopkins Applied Physics Laboratory (APL), results show that cooperative strategic deconfliction dramatically reduces the risk of mid-air collisions in complex high-density traffic situations [4]. As recent as last year, Unmanned Systems Inc. (UMEX) a world-leading provider of subject matter expertise in UTM, and a nationwide partnership of companies and government agencies showcased capabilities in a successful live flight test of emerging Advanced Air Mobility (AAM) technologies. The live demonstration used a Bell 407 helicopter as a surrogate for an electric Vertical Take-Off and Landing (eVTOL) aircraft and an air route between Hillwood’s Alliance Texas Flight Test Center in Justin, TX and the University of North Texas Discovery Park. The demonstration highlighted two critical aspects of AAM, including pre-flight cooperative strategic deconfliction and in-flight real-time airspace deconfliction. During the pre-flight phase of the flight test, the aircraft filed a flight plan and received automated authorization. Once in flight, the system detected a simulated airspace conflict and automatically issued a new flight plan authorization instructing the aircraft to modify its flight path in real time. In addition to simulating autonomous operations in a high-demand, high-density air traffic route for uncrewed, autonomous cargo and passenger-carrying air transports, the demonstration also provided vital data for industry standards in airspace management, vehicle-to-vehicle-to-infrastructure (V2V2I) communications, and autonomous flight operations [5].
Current regulations require UAV operations to have human pilots fully in the middle of the control loop. Autonomous flights are either performed within the line of sight of a human pilot or monitored by a human pilot during BVLOS operations. Entirely autonomous UAV operations are those where UAVs can perform their tasks without human intervention or supervision. These types of operations have many potential benefits, including [6]:
Enhancing safety – Fully autonomous UAVs can execute dangerous or difficult tasks safely and efficiently, saving time, money, and lives. Application areas include disaster response, search and rescue, firefighting, inspection, surveillance, critical infrastructure protection, asset protection, and delivery of goods and services in hazardous or remote areas [7].
Increasing efficiency – Fully autonomous UAVs can optimize flight paths, avoid conflicts, and adapt to dynamic environments using AI and sensors. This results in improved performance, reduced fuel consumption and emissions, and increased payload capacity. Application areas include traffic management, agriculture, mapping, surveying, and logistics [7].
Expanding capabilities – Fully autonomous UAVs can access areas difficult or impossible for crewed aircraft to reach, such as low-altitude or high-altitude airspace. In addition, they can operate in swarms or formations to achieve complex missions requiring coordination and collaboration. Application areas include scientific research in remote environments, natural resource monitoring, exploration, entertainment, and defense [7].
AI has great potential to transform ATM and UTM systems as we move towards integrating fully autonomous UAV operations with crewed aviation. AI can provide increased automation, communication and coordination, and sense and avoid capabilities to enhance air traffic safety, efficiency, and sustainability. However, fully autonomous UAV operations also pose challenges and risks that must be addressed carefully. These include ensuring the reliability, security, ethics, and accountability of the UAV operations; maintaining human oversight and control; balancing the trade-offs between autonomy and regulation; and harmonizing the standards and norms across different application domains, geo-political boundaries, and geographic regions. Therefore, it is essential to foster a collaborative approach among all the stakeholders involved in ATM, including air navigation service providers (ANSPs), civil aviation authorities (CAAs), airlines, airports, manufacturers, researchers, regulators, and users.