This study investigates how navigation uncertainty affects conflict detection and resolution (CD&R) for uncrewed aircraft in U-space. Position and velocity errors are modelled as zero-mean Gaussian noise consistent with ADS-L accuracy, and propagated through conflict metrics using Monte Carlo and analytical approximations. Under uncertainty, state-based detection becomes probabilistic. The probability of detection depends on both the level of uncertainty and the encounter geometry, and falls below 50 % when the nominal intrusion time equals the look-ahead. Operationally, detection is re-evaluated over time as the encounter develops, yielding multiple observations with varying probabilities. Two resolution algorithms are compared: Modified Voltage Potential (MVP) and Velocity Obstacle (VO). MVP proves more robust under uncertainty because it explicitly maximises distance at the closest point of approach (CPA). By maximising CPA distance, MVP maintains an outward push and avoids reversal behaviour during the manoeuvre, whereas VO performance degrades at low relative speeds and shallow angles. BlueSky simulations confirm these effects: MVP achieves higher intrusion-prevention rates and larger post-resolution miss distances across conflict scenarios, with its advantage most pronounced at low relative velocity. The findings highlight the importance of maximising CPA distance as a conflict resolution strategy. Moreover, the look-ahead horizon and protected zone can be tuned to achieve a desired target level of safety.

Reinforcement learning (RL) is a method that has been studied extensively for the task of conflict-resolution and separation management within air traffic control, offering advantages over analytical methods. One key challenge associated with RL for this task is the construction of the input vector. Because the number of agents in the airspace varies, methods that can handle dynamic number of agents are required. Various methods exist, for example, selecting a fixed number of aircraft, or using methods such as recurrent neural networks or attention to encode the information. Multiple studies have shown promising results using these encoder methods, however, studies comparing these methods are limited and the results remain inconclusive on which method works better. To address this issue, this paper compares different input encoding methods: three different attention methods – scaled dot-product, additive and context aware attention – and long short-term memory (LSTM) with three different sorting strategies. These methods are used as input encoders for different models trained with the Soft Actor–Critic algorithm for separation management in high traffic density scenarios. It is found that additive attention is the most effective at increasing the total safety and maximizing path efficiency, outperforming the commonly used scaled dot-product attention and LSTM. Additionally, it is shown that the order of the input sequence significantly impacts the performance of the LSTM based input encoder. This is in contrast with the attention methods, which are sequence-independent and therefore do not suffer from biases introduced by the order of the input sequence.

The growing density of civil air traffic is tightening operational safety margins and motivating the search for data-driven conflict-resolution policies. However, the rising compute demand for the training of AI models collides with the need to minimize its environmental impact. In an effort to reduce this climate impact, this paper investigates mixed-fidelity reinforcement learning (MiFi RL) as an alternative to training in high-fidelity (HiFi) simulators only, by first pre-training in a computationally lightweight low-fidelity (LoFi) environment before fine-tuning in HiFi. We analyze this paradigm across five single-agent algorithms – A2C, PPO, DDPG, SAC, and TD3 – using a fixed training budget of 3 million timesteps. Off-policy methods yield a large curriculum benefit: with a 60% LoFi / 40% HiFi split, SAC achieves a 24% increase in evaluated HiFi reward and a 20% reduction in wall-clock training time relative to pure-HiFi training; DDPG attains gains of 37% and 16% at a 40% LoFi share. In contrast, the on-policy algorithms exhibit negligible or negative improvements, possibly underscoring the replay buffer’s role in mitigating the domain shift between simulators. Efficient curriculum setup can alleviate computational load and environmental impact while improving final policy performance.

Very-low-level (VLL) urban air operations have been extensively investigated as a solution for mitigating congestion in cities. However, the manner in which the management of such traffic should be performed is still actively investigated. One important component of such a system is the conflict detection and resolution (CD&R), mainly composed of the strategic and tactical CD&R module. While many approaches towards these have been studied, insufficient analysis has been conducted on their compatibility when functioning within a unified, hybrid system. Additionally, their robustness to operational uncertainties such as wind and departure delays is often overlooked. In this work, we investigate the performance of strategic planing methods when combined with tactical CD&R and subjected to a wide range of traffic demand levels and uncertainty conditions. Simulations indicate that the performance of the strategic deconfliction module is highly sensitive to the presence of wind and delay. This decline in performance is partially mitigated by the tactical deconfliction module. Thus, the results suggest that increased use of tactical CD&R could lessen the required level of detail of strategic deconfliction methods, leading to improved compatibility between the two modules.

The concept of urban air mobility is rapidly advancing, with much research being dedicated towards the development of the air traffic management services required for such operations. An important component of unmanned air traffic management (U-space/UTM) is conflict detection and resolution (CD&R), tasked with ensuring the operational safety of such systems. Strategic flight plan optimisation and tactical CD&R methods have generally been studied independently, leading to suboptimal performance when deployed simultaneously in simulated high-density very-low-level constrained urban airspace environments. Furthermore, the limited flexibility of pre-departure 4D trajectory planning methods towards dynamic and uncertain environmental and operational conditions (i.e., wind and delay) produces a degradation in safety that is difficult to mitigate using tactical manoeuvring. In this work, we design a traffic-flow capacity strategic optimisation method that aims to achieve robustness against flight plan deviations and to better complement tactical CD&R manoeuvring. The performance of the proposed strategic and tactical deconfliction module is tested within constrained urban airspace traffic scenarios simulated using the BlueSky Open Air Traffic Simulator. The results are compared with other methods, such as 4D trajectory planning and state-based CD&R.

The increasing integration of Uncrewed Aerial Systems (UAS) into controlled airspace presents significant operational and safety challenges, particularly in conflict detection and resolution (CD&R) for Beyond Visual Line of Sight (BVLOS) operations. Ensuring reliable separation management in U-Space requires robust e-conspicuity solutions that address uncertainties in Communication, Navigation, and Surveillance (CNS) systems. This study evaluates the CERTIFLIGHT UTM Box, an e-conspicuity device designed for General Aviation (GA) and UAS operations, incorporating authenticated GNSS tracking, blockchain-based data integrity, and conflict resolution advisory services. Flight validation tests were conducted using a GA aircraft and a UAS to assess the system's effectiveness in detecting and resolving conflicts under realistic operational conditions. The results indicate that while the UTM Box successfully provided conflict advisories, navigation uncertainties and communication delays exceeding five seconds affected its performance. The study highlights the importance of incorporating CNS system uncertainties into CD&R algorithms to ensure safe separation. Future work will focus on refining conflict resolution strategies, integrating advanced filtering techniques to mitigate sensor noise, and enhancing pilot interface design for improved situational awareness and decision support.

With the rapid increase in the use of Unmanned Aerial Systems (UAS) for commercial applications such as medical and parcel delivery, the need for safe airborne separation in airspace has become critical. This paper examines the impact of position uncertainty on autonomous separation methods within U-Space, a European Union initiative for managing drone traffic. The study focuses on evaluating various conflict resolution algorithms—specifically, Modified Voltage Potential (MVP) and Velocity Obstacle (VO) variations—under conditions of navigational uncertainty. Through Monte Carlo simulations using the BlueSky ATM simulator, position uncertainty stemming from Global Navigation Satellite Systems (GNSS) errors is modelled and analysed. The research compares the effectiveness of different conflict resolution strategies in preventing conflicts between UAS, measuring intrusion prevention rates and the closest point of approach during encounters. The results indicate that MVP provides superior performance in handling positional uncertainty, offering more robust conflict resolution capabilities than VO-based methods especially at shallow angles conflict situation. These findings are critical for ensuring the safe integration of UAS into increasingly congested airspace environments, guiding future developments in U-Space operations.

As urban ground transportation congestion increases, there is growing interest in urban air transportation, such as delivery drones and air taxis. However, managing air traffic in densely populated urban areas poses significant challenges, which require effective flight planning, separation management, and airspace design. This paper investigates dynamic capacity balancing methods to manage air traffic in constrained urban airspace, where drones must fly above the existing road network. Specifically, it compares the effectiveness of labelling high-complexity zones using historical data versus real-time aggregate flow data. The results indicate that while both approaches reduce airspace intrusions and improve safety, the best approach depends on traffic demand levels. At lower demand levels, using historical data yields better safety outcomes, whereas using real-time data is more effective at higher demand levels due to its flexibility. At their best, both methods increase the travel distance by less than 6% while reducing airspace intrusions by 30% compared to a case without dynamic capacity balancing.

Conventional Air Traffic Control is still predominantly being done by human Air Traffic Controllers, however, as the traffic density increases, the workload of the controllers increases as well. Especially for the area of unmanned aviation, driven by the rise in drones, having human controllers might become unfeasible. One of the methods that is currently being investigated for replacing the conflict resolution task of Air Traffic Control is Reinforcement Learning. As violation of the required separation margins, also called an intrusion, is an event of relatively low frequency, using Reinforcement Learning for this task comes with difficulties that can potentially be attributed to data imbalance. This paper artificially increased the traffic density during the training phase of the Reinforcement Learning method to investigate what the importance is of a balanced data set on the performance of the Reinforcement Learning method. It was found that as the traffic density increased, the Reinforcement Learning methods started to outperform the analytical methods. Beyond this it was found that methods trained at higher traffic densities, but tested at lower traffic densities, outperformed the methods trained at that specific density. This indicates that it might be better to always ensure that the training scenarios are more complex than anticipated during the execution phase, even if that results in unrealistic scenarios.

U-space/UTM operations are considered an integral part of the future development of cities, with applications ranging from package delivery to urban air mobility. However, this new complex environment also poses challenges for the conflict detection and resolution (CD&R) process, especially if aircraft will have to navigate above the existing street network due to privacy and obstacle constraints. The research at hand aims to investigate how information about the environment and other aircraft can be used to improve the performance of CD&R methods in constrained urban airspace. For this, three algorithms are developed and tested, each with different levels of information availability: the first solely uses current state information for conflict solving, the second includes additional information about the urban environment within the CD&R process, and the last also incorporates trajectory intent data to solve conflicts. These methods are tested within simulations of urban air traffic scenarios at various demand and wind levels to determine their safety and efficiency performance. Results show the use of street geometry information benefits the resolution process greatly, increasing the safety level while minimally affecting efficiency. Intent information is shown to not be critical for achieving this.

U-Space drone operations are expected to be a driver for further urban development, especially through use cases such as medical and commercial parcel delivery. In particular, package delivery using small drones shows great promise, with e-commerce giants such as Amazon deploying limited-scale drone delivery trials in rural areas. As the technology matures, large-scale operations will take place in constrained urban areas, leading to high airborne traffic densities. It is necessary to develop a robust automated separation management system that actively ensures safe separation of drones both in the air and on the ground. This paper focuses on the Strategic element, more specifically on pre-departure planning. The aim of this is to reduce the chance of conflicts around vertiports, where spatial and environmental constraints make tactical resolutions difficult. This work focuses on two scenarios: a single pad for both takeoffs and landings (in a spatially constrained urban area) and 4 takeoff-landing pad pairs (for a distribution center). Several methods are compared for this takeoff sequencing task, coupled with a conflict detection algorithm: A First-Come First-Served method that applied delay to conflicting flights, a Mixed-Integer Programming approach, a Genetic Algorithm, Particle Swarm Algorithm and Simulated Annealing were used. For a single-pad approach, first-come first-served works best in terms of computation time and total deployment time (or makespan). For the multi-pad approach however, changing the flight sequence through metaheurisitic methods and mixed-integer linear programming show a reduction in total deployment time.