The increasing number of Remotely Piloted Aircraft (RPA) in the skies worldwide raises growing concerns over safety and the effective management of air traffic.
Across the globe, the number of drones – also known as Remotely Piloted Aircraft (RPA) or Unmanned Aerial Vehicles (UAV) – is steadily increasing within airspace traditionally used by “general aviation“. This sector of civil aviation encompasses «non-military flights that do not involve the paid transport of passengers or goods, nor patrols, infrastructure inspections, or search and rescue operations».
General aviation typically operates in uncontrolled airspace and is reserved for private flights carrying passengers, using «light jets, helicopters, gliders, hang gliders, paragliders, hot air balloons, and similar aircraft, with purposes primarily recreational or tourist-oriented» [source: Sharing Airspace with Uncrewed Aerial Vehicles (UAVs): Views of the General Aviation (GA) Community – Journal of Air Transport Management, July 2022].
The growing presence of drones in this space, used for tasks such as land mapping, aerial photography, environmental monitoring, surveillance, security, and search-and-rescue missions, is driven by a global market for professional-commercial RPA. Since the post-Covid period, this market has grown at an annual rate of 23.7%, with a projected total value of $21.69 billion by 2030 [source: Allied Market Research].
To grasp the extent of the traffic generated in general aviation airspace, consider the example of the United States, where the Federal Aviation Administration (FAA) «reports approximately 200,000 crewed aircraft daily, with only a few thousand flights occurring simultaneously during peak hours. Meanwhile, over 500,000 recreational drones and more than 300,000 professional-commercial drones are already registered» [source: Federal Aviation Administration].
TAKEAWAYS
Harmonising drone traffic: preparing for the arrival of flying taxis
In “Decentralized Traffic Management of Autonomous Drones” (Swarm Intelligence, 11 July 2024), a research team from Eötvös Loránd University in Budapest highlights how the density of Remotely Piloted Aircraft in shared airspace «is expected to surge even further in future smart cities. These cities will feature drone-like aircraft that not only monitor urban landscapes and enhance city security but will also deliver goods and transport passengers, integrating into our daily lives and making safe, reliable air mobility management a top priority».
Regarding the use of drone-like aircraft for passenger transport, a notable article published on 30 May 2024 in National Geographic (“Would You Travel by Flying Taxi? Here’s Everything You Need to Know“) drew significant attention. It describes flying taxis as no longer a mere future hypothesis but a present-day reality, particularly in China, New York, and the UK, where they have moved beyond the design phase and are now operational.
According to the article, «more than 150 companies worldwide are developing flying taxis, with widespread agreement that these will resemble drones». These vehicles, specifically eVTOL (electric Vertical Take-Off and Landing) aircraft, are electric, piloted for now, and capable of vertical take-off and landing without the need for a runway, much like drones.
«Similar to the rise of electric cars, much of the innovation is coming from China, where, in 2023, drone manufacturer EHang gained approval from China’s Civil Aviation Authority for its flying taxi. The company later received orders for 100 of its aircraft for use in sightseeing and shuttle flights over the city of Hefei, and it is now developing an unmanned version».
In November 2023, the first eVTOL flying taxi took to the skies over New York, «reducing the usual one-hour journey to John F. Kennedy International Airport to just seven minutes». London, meanwhile, is preparing for a series of tests by 2026 involving five-seater eVTOL air taxis. These will take off and land vertically at dedicated structures known as vertiports, which are already being planned by the city’s authorities.
Coordinating drones in autonomous air traffic: past approaches
This global scenario, where drone traffic intersects with general aviation airspace, has become increasingly dense and complex, presenting numerous challenges in the management and safety of aerial mobility.
The aforementioned study by Eötvös Loránd University recently proposed a decentralised solution for managing drone traffic. The approach relies on a combined method that starts with route planning and is followed by detection and avoidance processes to navigate obstacles along the flight path.
Before diving into this specific solution, it’s worth reviewing the body of work over the past 15 years that has contributed to the coordination of multiple drones within autonomous air traffic, which – as the Hungarian researchers point out – differs significantly from drone “swarms” or “flocks“, where the task is entirely different:
«Large drone swarms execute flights where the goal is to synchronise movements and fly ‘together’. In autonomous traffic, however, rather than moving together, aircraft must be able to move independently along their own routes, but in a coordinated manner»
Initially, the approach was based on centralised multi-drone route planning, «where the challenging task of generating non-overlapping 4D trajectories was solved by a central computer either before take-off or during flight».
Notable methods included mathematical models like Mixed Integer Linear Programming (MILP), as explored in “Path Planning for UAVs Under Communication Constraints Using SPLAT! and MILP” (Journal of Intelligent & Robotic Systems, 2011), or Sequential Convex Programming, discussed in “Generation of Collision-Free Trajectories for a Quadrocopter Fleet: A Sequential Convex Programming Approach” (IEEE, 2012).
From 2020 onwards, bio-inspired centralised planning methods emerged, such as those mimicking ant colonies, detailed in “Swarm-Based 4D Path Planning for Drone Operations in Urban Environments” (IEEE, 2021).
However, centralised systems have a critical flaw that has limited their scalability and adaptability over time: they require highly stable communication infrastructures.
Decentralised methods
In recent years, decentralised methods for multi-drone route planning have gained prominence as the most flexible tools for managing complex air traffic. Unlike centralised systems, no single computer controls or guides the aircraft in decentralised systems. Instead, each drone navigates autonomously, avoiding collisions using its onboard computer.
Two key studies illustrate these solutions: “EGO-Swarm: A Fully Autonomous and Decentralized Quadrotor Swarm System in Cluttered Environments” and “Safe Tightly-Constrained UAV Swarming in GNSS-denied Environments”, both published by the IEEE (Institute of Electrical and Electronics Engineers) in 2021. These studies focus on Remotely Piloted Aircraft (RPA) navigating environments with high obstacle density, introducing a decentralised algorithm for autonomous navigation and testing it with up to three real drones.
Despite originating from different regions – one study conducted in China, the other in the Czech Republic – both research efforts share the commonality of validating their techniques through numerous simulations and real-world experiments, albeit with a limited number of aircraft.
An especially intriguing study is “Sky Highway Design for Dense Traffic” (Science Direct, 2021), led by Beihang University in Beijing. This research focuses on designing geometric air highways, allowing drones to organise themselves into sky lanes, akin to “roads in the sky“, where each aircraft follows its designated path. Although this project has also been tested through practical demonstrations, it revealed significant limitations on the free movement of aircraft.
Subsequent studies mostly presented simulation results without conducting real-world experiments with drones. That was until July 2024, when Eötvös Loránd University in Budapest introduced its decentralised traffic management approach in “Decentralized Traffic Management of Autonomous Drones”, as previously mentioned. The Hungarian team tested their system on a fleet of hundreds of unmanned aerial vehicles. Let’s explore this innovative approach in more detail.
Real-time aerial route planning
The decentralised flight planning algorithm developed by researchers at Eötvös Loránd University in Hungary is, in many ways, bio-inspired. As the researchers explain:
«We designed a drone speed controller that incorporates interaction terms based on models describing the collective movement of animal herds, characterised by ‘repulsion‘ and ‘alignment’ dynamics derived from mammalian reflexes. This allows us to adopt a fully decentralised approach, governed by simple pairwise drone interactions, resulting in scalable systems»
This method, called “sense-and-avoid“, relies on «repulsive and speed-alignment interactions between drone pairs», which manages the hierarchical priorities among the drones.
Drones operating in autonomous air traffic, without a central control system, use this algorithm to remain within the repulsion-alignment framework. Rather than exiting the system, «they continuously combine the two dynamics in an adaptive, symbiotic manner».
Each drone, several times per second, transmits its position, flight speed, and destination to nearby drones, while simultaneously receiving the same information from them. «Our control algorithm uses these inputs to determine the desired speed for each aircraft» the team explains.
This model was successfully tested through simulations, demonstrating self-organised high-speed traffic involving up to 5,000 drones. Following the virtual simulations, live demonstrations were conducted using a fleet of 100 autonomous drones. «Their task was to independently reach their assigned destinations while avoiding collisions in open space, coordinated by the decentralised system».
Lastly, the researchers explored layered flight scenarios, mapping solutions for particularly dense drone traffic, anticipating the challenges of future smart cities.
Glimpses of Futures
The described strategy, involving the decentralised control of multiple drones with independent tasks, demonstrates how real-time coordination and conflict resolution as aerial vehicles move through shared airspace is a key element in the safe management of drone traffic.
Using the STEPS matrix, we will now attempt to anticipate potential future scenarios, analysing the impacts that the evolution of decentralised flight planning algorithms for drones could have from social, technological, economic, political, and sustainability perspectives.
S – SOCIAL: as the EASA – European Union Aviation Safety Agency states, «…the potential for mid-air collisions between drones and other aircraft is a safety concern due to the increasing accessibility of unmanned aerial systems», decentralised approach to coordinating drones in autonomous air traffic – like the one presented by the researchers at Eötvös Loránd University – emerges as an efficient and flexible method, adaptable to different types of drones. This, in a future scenario, could make such a method «part of a global standard for decentralised management of unmanned aerial system traffic».
T – TECHNOLOGICAL: from a technological perspective, future research may not only focus on the challenges posed by outdoor drone traffic, where the aircraft avoid each other based on absolute position data actively transmitted by each vehicle, but also on the safety of Remotely Piloted Aircraft in smart cities. In the coming years, unmanned systems will need the ability to avoid so-called “passive obstacles” (such as static objects like skyscrapers) and tackle the issue of traffic in “confined spaces”, which characterise urban environments.
E – ECONOMIC: although the drone sector is expanding globally (in the EU alone, it is expected to generate over 10 billion euros annually by 2035 [source: “Drones: Foreseeing a ‘risky’ business? Policing the challenge that flies above” – Science Direct, November 2022]), the economic impact of poor management of drone traffic is equally significant. Such mismanagement can lead to prolonged disruption of operations, affecting critical infrastructure. A prime example is the 2018 incident at London Gatwick Airport, where a drone sighting on the runway led to the airport being closed for two days, resulting in an estimated loss of 64 million euros for airlines, with an overall cost of 75 million euros due to delays, diversions, and expenses incurred from mobilising the military and police forces.
P – POLITICAL: a real-time drone route planning solution, such as the one introduced by the research team at the Hungarian university, aligns with the European PODIUM project for the safe and secure management of drone traffic. This project aims to develop a system capable of interacting with all types of aircraft, including piloted ones. Currently, the project is being tested at five real sites (both rural and urban) across three EU countries, with trials involving «operations at airports, in controlled airspace, and mixed environments with piloted aviation».
S – SUSTAINABILITY: the decentralised drone traffic management system described, with its focus on safety and reliability, would support the future direction of “advanced air mobility”, which prioritises sustainability in all aspects, including environmental. This would involve reducing noise and gas emissions from aircraft. In future smart cities, the flexible and decentralised coordination of drone traffic could indirectly reduce road congestion, while also improving access to more remote areas, which are often less well-served by conventional transport compared to central locations, thereby supporting urban sustainability.
