Green hydrogen is emerging as a viable alternative to kerosene for medium-range aviation, with airports playing a critical role in the distribution of liquid hydrogen (LH2). This study uses a simulation-based optimisation model to investigate hydrogen refuelling operations at Rotterdam The Hague Airport. Different analyses reveal that on-time performance and capacity are strongly influenced by hydrogen-flight penetration rate, refuelling-truck fleet size, and turnaround times. Operational efficiency depends on maintaining adequate fleet sizing and minimising LH2 refuelling durations, while factors like truck speed and trailer swap time have limited impact in compact airport layouts. Moderate integration by 2040 requires gradual investment, but high penetration by 2050 demands substantial infrastructure and vehicle expansion. As hydrogen usage grows, a second refuelling truck becomes necessary between 22% to 39% penetration, depending on the specific day and required refuelling time. Also, end-of-day and multiple close LH2 departures risk exceeding allowable departure time slots. Safety zone constraints and trailer capacity trade-offs further highlight the need for coordinated planning. These results underline that a successful transition to hydrogen aviation hinges on strategic fleet sizing, optimised processes, and phased infrastructure upgrades to ensure reliable and sustainable airport operations.

Airport surface movement operations present a promising opportunity for enhancing airport sustainability, as up to 25% of airport emissions originate from taxiing aircraft. This study explores the use of Electric Towing Vehicles (ETVs) to tow aircraft to runways, eliminating the operational use of aircraft engines during taxiing. Given the high cost of ETVs, their effective deployment is essential. This research adopts a Multi-Agent Pickup and Delivery (MAPD) framework to optimize ETV operations, addressing both the task assignment and path planning aspects of this problem. A novel task assignment method based on Ant Colony Optimization (ACO) is proposed, capable of efficiently scheduling towing and charging tasks for ETVs. The proposed method outperforms benchmark Mixed Integer Linear Programming (MILP) methods in terms of runtime and scalability, while delivering a competitive solution quality. The ACO-based task assignment method is integrated with a state-of-the-art motion planning algorithm based on Priority Based Search (PBS) and Safe Interval Motion Planning (SIMP). The resulting MAPD solution method is evaluated through its application in a high-fidelity simulation of Amsterdam Airport Schiphol (AAS). The simulation results indicate that deploying 22 ETVs at AAS can achieve a 57% reduction in the annual addressable CO2 emissions, while ensuring the economic feasibility of the operational concept.

At airports, taxiway maintenance is a complex task affecting airport ground operations, specifically airport surface movement operations managed by ground controllers. It is expected that taxiway maintenance also affects the task load experienced by ground controllers. However, to our best knowledge, no studies were performed so far to quantify and assess this influence. To fill this gap, this research studies the impact of taxiway maintenance on the task load of ground controllers. It is essential to develop this understanding to ensure that taxiway maintenance is planned safely, efficiently, and robustly. This research uses Schiphol Airport as a use case, for which an existing airport surface movement simulator was used to model the ground traffic with a single-agent trajectory planning algorithm. To realistically represent the traffic patterns at Amsterdam Schiphol Airport, the simulated model was calibrated using real trajectory data and expert opinion.
To quantify the task load of ground controllers, the well-known Dynamic Density model and a novel measure, based on controller's mental re-plannings, were used and compared to each other. Based on a maintenance scenario, simulations were performed for two runway configurations, with the aim to study the differences in task load between the nominal and maintenance scenario for different flight schedules. The main outcome of this analysis is that conflicting flows, resulting from taxiway maintenance, have a significant contribution to the task load of ground controllers. It was further found that the effects of these conflicts on the task load are localized and do not propagate further to other areas, and other ground controllers who do not have responsibility over the specific area in question do not experience this impact in task load.

To accommodate the rising demand for airport resources and airspace capacity, a more efficient and robust orchestration of the components that facilitate air travel is required than is the case today. Over recent years, camera systems for automatic identification of ground-handling events were developed, providing the opportunity to automatically detect ground-handling delays as they manifest. Detecting ground-handling delays early allows the Air Navigation Service Providers (ANSPs), ground handlers and aircraft operators to reallocate their resources, facilitating more stable planning and robustness of the air traffic network. This study proposes a novel interpretable machine learning approach to predict initial ground-handling end time adherence. This approach is based on a discrete Bayesian network composed of a central network structure and modular sub-networks, representing distinct turnaround processes often found on the critical path. The Bayesian network structure and discretization were found to be adequate to model most ground-handling operations for narrow-body aircraft. The model proved to be more accurate than the ground handler in predicting initial Target Off-Block Time (TOBT) adherence at all prediction moments. At the initial TOBT-0 [minutes], a maximum in model accuracy was obtained amounting to 65.76%. In addition, the model demonstrated the capability to detect ground-handling delays earlier. At the initial TOBT-20 [minutes], an average recall of 50.4% was obtained for delays within 5 and 15 minutes. In contrast, the ground handler identified only 4.3% of small ground-handling delays at this prediction moment. The discrete Bayesian network approach is also inherently interpretable, providing transparency into the decision-making process of the model.

A novel security checkpoint concept has been proposed to address the continuing growth in passenger traffic that airports are experiencing globally. The new security checkpoint intends to employ Automated Guided Vehicles (AGVs) instead of conveyor belts for baggage tray transportation. However, the new concept poses new challenges to the classic Multi-Agent Path Finding (MAPF) problem due to its small environment and high agent density. The AGV fleet will thus have to effectively coordinate their paths to avoid gridlocks within the crowded security checkpoint environment. Firstly, this paper presents a detailed Agent-Based Model (ABM) to accurately represent the new automated security checkpoint and estimate its passenger throughput. Secondly, the paper presents a new MAPF algorithm for the AGV fleet: the Windowed Hierarchical Idle-Push (WHIP) algorithm. WHIP is a prioritised MAPF solver that uses Push-operations to efficiently resolve the bottlenecks of the automated security checkpoint, avoiding gridlock even at very high agent densities. Experimentally, WHIP achieves path efficiencies 5 percentage points higher than leading prioritised MAPF algorithms, while requiring only half the computational runtime. Using the WHIP algorithm, the automated security checkpoint is capable of achieving an 18% increase in passenger throughput with respect to existing security checkpoints using conveyor belts.

Urban Air Mobility (UAM) has seen remarkable growth over the past decade, especially considering the delivery of goods. Notably, the delivery of medical supplies via Unmanned Aerial Vehicles (UAVs) has emerged as a promising application, which is driven by its societal benefits and has been demonstrated successfully in rural environments. However, expanding these operations to urban areas presents a unique set of challenges, such as high traffic density and the need for online path planning of UAVs. Effective path planning and coordination are of utmost importance to allow for safe and efficient delivery operations in urban environments. Currently, path planning of UAVs is relatively unexplored, and for drone delivery models the trajectories of UAVs are often oversimplified. This research aims to fill the research gap by developing and evaluating an online path planning and coordination mechanism in 3D for a cooperative fleet of autonomous UAVs delivering medical supplies. The Rotterdam Area is selected as a use case, but the model can be adapted to other cities. Utilising an agent-based model formulation, this paper formulates a novel approach for multi-agent pathfinding in 3D and real-world environments, where the travelled paths of UAVs in drone delivery models are computed by taking into account the urban environment and kinematics of UAVs. Windowed Cooperative Safe Interval Path Planning (WC-SIPP) is used as a path planning and coordination mechanism, where kinematic constraints of UAVs and safety separation distances between UAVs are incorporated. Experimental studies show the capability of the algorithm to plan UAVs in different multi-layer urban environments. We establish that including more than one vertical layer is beneficial for UAVs to avoid conflicts by allowing them to change altitudes during flight. Additionally, we show that an increase in the number of layers does not lead to an increase in the delivery time of packages. However, an increase in available layers does come with a decrease in computational performance. Besides, the proposed method is able to plan the trajectories of UAVs and ensure safety separation between UAVs within a dynamic environment, where up to 20 unexpected obstacles are introduced into the environment. Furthermore, we demonstrated that for our operation model, assigning prioritisation based on either mission type or package request time did not influence the delivery time of packages.

As an alternative to kerosene (Jet-A1), green hydrogen is considered a potential solution to reduce the environmental impact of the medium-range aviation industry. While the technical feasibility of liquid hydrogen (LH2)-powered aircraft is already being researched and demonstrated, the required infrastructure, as well as its impact on the current ground operations has received little attention in literature. This research paper aims to fill this research gap with the objective of exploring the potential impact of commercial LH2-powered aircraft on the ground operations, focusing on (the transition towards) 2050. In contribution to this main research objective, five important turnaround-related factors for ground handling LH2 were considered to explore their respective impact on the ground operations. These five
parameters include the number of LH2 fuel trucks, LH2 flow rate, LH2 aircraft penetration rate, LH2 refueling safety zone diameter and the level of restricted parallel turnaround processes during LH2 refueling. Through varying these parameters in experiments using an agent-based simulation model, meaningful results could be obtained which are useful to consult airports and airlines on what operational and infrastructural measures need to be accommodated to ensure competitive ground operations in terms of efficiency involving LH2-powered aircraft. The impact of the different experiments on the efficiency was expressed in terms of the turnaround time, number of aircraft delays and on-time performance of both Jet-A1 and LH2-powered aircraft. As a case-study, the ZEROe turbofan was selected to operate at the regional airport Rotterdam The Hague Airport, featuring remote aircraft stands. The results show that the LH2 penetration rate has the greatest influence on the efficiency of the ground operations,
and increasing this rate must be carefully coordinated by the involved stakeholders in accordance with technological developments of the remaining four parameters. For a LH2 penetration rate of up to 25%, at least 50% of the turnaround processes must be allowed to be executed during refueling, the safety zone diameter must not be greater than 30 meter and a single LH2 fuel truck with a flow rate of at least 5 kg/s is required to ensure efficient ground operations. To accommodate LH2 penetration rates up to 50%, there must be no restrictions on parallel turnaround processes, the safety zone diameter should be 15 meter and finally at least two LH2 fuel trucks are required with a flow rate of at least 10 kg/s each.

The Dutch healthcare sector wrestles with rising costs, staff shortages, and increased demand due to healthcare centralization. This, coupled with worsening traffic congestion, underscores the need for efficient solutions like drone-based medical transport. This paper addresses the need for effective fleet management strategies tailored to the unique demands of the healthcare environment. Specifically, it seeks to develop an adaptive strategy that accommodates the network's expansion and the inherently stochastic and urgent nature of pickup and delivery orders associated with medical transport. Utilizing an agent-based model formulation, the paper introduces a novel approach combining adapted temporal sequential single-item auctions for allocation and scheduling with reinforcement learning for drone repositioning to optimize fleet management. Our findings highlight the strategy's consistent efficiency across various demand scenarios, maintaining performance within predefined limits. Notably, the repositioning module significantly enhances the fleet utility and the fraction of served orders, albeit at the expense of increased cost per delivery. Conversely, the reallocation module causes minimal performance improvement. Under heightened stochasticity introduced by urgent orders, the strategy maintains stable costs per delivery while fleet utility and order fulfillment rates decline. Additionally, our investigation underscores the increasing benefits of repositioning in more stochastic scenarios. Moreover, exploring hybrid fleets reveals that while short-range high-payload drones can reduce cost per delivery, they compromise overall fleet utility and order fulfillment rates. Furthermore, we identify the under-utilization of payload capacity in scenarios with orders weighing up to 2 kilograms for a drone with a payload of 10 kilograms.

The increasing demand and complexity of air traffic management (ATM) systems necessitate significant advancements in automation to ensure safety and efficiency. Artificial intelligence (AI) and machine learning (ML) are emerging as promising solutions to manage this growing complexity, offering enhanced decision-making and predictive capabilities. However, the effectiveness of ML models in ATM heavily relies on the availability of extensive, high-quality data. In many cases, such data is scarce or incomplete, which presents a major challenge for training robust models. Synthetic data generation (SDG) represents a viable solution to address this data scarcity, enabling the creation of realistic, diverse datasets that can effectively fill data gaps and unlock the ML value proposition. The Terminal Maneuvering Area (TMA) is a crucial segment of airspace where aircraft perform complex maneuvers including rare go-arounds, and holding patterns. The TMA is characterized by high traffic density and diverse trajectory types, necessitating precise, granular data to model these dynamic scenarios accurately. The main research objective of this thesis project was to investigate the applicability of Generative Adversarial Networks (GANs) in generating synthetic 4-dimensional aircraft landing trajectories capable of capturing traffic patterns in this airspace. An unconditional recurrent-based network (TimeGAN) was applied and complemented with an investigation into the merit of conditioning to reap the rewards of transfer learning between classes. The resulting synthetic trajectories were evaluated in terms of data diversity, fidelity and usefulness. The main challenge identified during this research was the imbalance in data classes, which affected the models' ability to accurately capture data patterns, particularly in less frequent scenarios. Generating synthetic data based on separate groupings showed promise in addressing these imbalances, although this approach was sensitive to the number of clusters chosen. Smoothing filters frequently proved necessary to ensure realistic landings were devoid of unwanted noise. Overall, this work proves the capability of TimeGAN in generating diverse, realistic trajectories that are difficult to differentiate from real historical data.

Through several contractions, stiff competition, and increasing passenger expectations, airports must evolve continually. One of the main avenues for this has been improving the efficiency of the security check- points, which are airports’ primary bottlenecks. Operational optimisation methods, such as resource and task scheduling are relatively mature fields of research, however, they require accurate forecasts. Current forecasting approaches seldom use useful information such as the flight schedule, nor are they able to re- present uncertainty or integrate real-time information. Therefore this paper aims to develop and evaluate a real-time probabilistic security checkpoint arrival rate forecasting model by utilising a Bayesian frame- work. This is achieved using a probabilistic programming language to create a bottom-up model, where per-flight arrivals are predicted. The passenger arrival rate for each flight is determined by estimating the total number of passengers and their temporal arrival distributions probabilistically. The combination of arrival rates from all flights in the flight schedule then provides the full checkpoint forecast. Furthermore, an updating scheme is proposed, that updates the expected number of passengers for each flight through Bayesian inference. Results show that the static forecasting model has promising performance, while suc- cessfully capturing uncertainty. However, the proposed real-time updating approach does not function as intended, due to a consistent negative bias. This has been attributed to a fundamental asymmetry present in the problem. Finally, this study includes an application of lane requirement estimation, which yielded highly favourable results. Allowing decision-makers to minimise costs while keeping the probability of poor checkpoint performance to acceptable levels.

In response to the increasing challenges of Cyber Electromagnetic Activities (CEMA) in urban settings, characterized by dense electromagnetic (EM) signals and rising data traffic, this research introduces an Agent-Based Model (ABM) aimed at prioritizing critical signals. The primary goal of this research is to deploy a Unmanned Aerial Vehicle (UAV) swarm operating to selectively interfere with communication lines in a CEMA environment. The research goal is segmented into a multi-stage approach, focusing on the following mission tasks for the UAV swarm: i) Transmitter Search, ii) Communication Line Search, iii) Communication Line Mapping, and iv) Interference Point Finding. This research proposes and evaluates various methodologies for these task. A methodological contribution is the development of the Heuristicdriven Utility by Regression-based Metrics Synthesis (HURMS) framework. This framework addresses the subswarming coalition formation problem in Multi Transmitter Search. The HURMS framework utilizes the benchmarking Monte Carlo Tree Search (MCTS), a heuristic search method, to enhance the Maximum Utility Configuration (MUC), a transparent utility-based method, overcoming heuristic search limitations and complexities in creating utility functions. While the HURMS-enhanced MUC method effectively located all transmitters in the task, comparative analysis showed MCTS to be about 45% faster and 56-71% more successful in transmitter detection. This highlights potential areas for enhancing the MUC algorithm further under the guidance of the HURMS framework. Furthermore, the Particle Swarm Optimization (PSO) Search was utilized as a velocity controller in the context of Multi Transmitter Search methods, guiding the speed and direction of UAVs. Regarding Single Transmitter Search, a significant observation was that the PSO Search was approximately 38% faster than the novel Pair-gradient Search method in locating a single transmitter. The study also examines Communication Line Mapping by comparing two frontier-based methods with the Multi-agent Tabular Q-Learning method. The frontier-based methods provided better coverage, while the Tabular Q-Learning excelled in precision and adaptability for multi-agent mapping. Data from these methods were applied to find the interference point. Additionally, a Reinforcement Learning (RL)-trained agent was used for Interference Point Finding, proving to be faster but less accurate.

Climate change poses a significant global challenge and the aviation industry, a major contributor to greenhouse gas emissions, must address its environmental impact. As the demand for air travel continues to rise, transitioning to sustainable forms of aviation becomes crucial. This paper focuses on the challenges and opportunities associated with implementing full electric aircraft at regional airports, using Rotterdam the Hague Airport (RTHA) asa case study. The research aims to determine the optimal infrastructure sizing of RTHA and provide insights into the integration of electric aviation. A novel mixed integer linear programming (MILP) model is proposed to optimize the charging schedule and determine the required number of charging stations and electricity grid capacity. The study considers different scenarios of electric aviation uptake, aircraft types, charger types, and spatial aspects of charging locations. The findings offer comprehensive recommendations for regional airport infrastructure planning and address uncertainties through sensitivity analysis.

Low-altitude, high-density air traffic is expected to grow in the coming decades with several companies being certified to initiate urban operations for both freight and passenger transport. However, traditional human-centered Air Traffic Control operations (ATCos) are not scalable to handle the increased demand to maintain safe separation between aircraft. Thus, new autonomous solutions to resolve potential conflicts between aircraft must be developed, allowing humans to supervise and understand machine actions. We propose a centralized Deep Reinforcement Learning (DRL)-based framework to provide speed advisories to vehicles, ensuring safe and efficient operations. We used Proximal Policy Optimization (PPO) to train the intelligent controller and show that our framework is capable of handling challenging merging point settings. We evaluated our model extensively using a custom OpenAI Gym environment and proved that it can achieve a 99% success rate in conflict resolution across multiple merging point configurations.

This research addresses the process of task allocation in a heterogeneous multi-agent fleet through the introduction of a novel mechanism in existing decentralised consensus algorithms: bid intercession. Bid intercession refers to the principle of agents biding on behalf of other agents in decision-making architectures leveraging market-based decision strategies. The method exploits and extends existing consensus-based allocation processes through the redistribution of responsibilities in the auction process to achieve various degrees of centralisation in the task allocation process. It is demonstrated that the extension proposed allows for hybridising multiple allocation methods together and structuring the auction process (notably through embedding hierarchies and decision trees directly in the decision-making process) all the while retaining the convergence robustness and performance guarantees provided by the underlying algorithms. The Search and Rescue case study is investigated to assist in framing the research and provide a reference scenario for the application of such concepts. This concept, unexplored so far in consensus-based approaches, not only opens up a sway of coordination architectures and optimisations but also paves the way to novel ethically compliant autonomous systems while retaining essential performance and robustness properties crucial in high-stake applications.

Many large airports aim to have complete autonomous airside operations in the future. Amsterdam Airport Schiphol (AAS) for example, launched the Autonomous Airside Operations program to achieve this goal. Our main contribution is to present a Multi-agent Pickup-and-Delivery (MAPD) model that uses a centralized task allocation mechanism to improve the performance of integrated task allocation and path planning for autonomous ground handling operations compared to previous research. This study models a global multi-vehicle Pickup and Delivery Problem with Time Windows (PDPTW) for the scheduling of autonomous ground handling tasks. A warm start multi-objective mixed integer linear programming model is proposed to solve the scheduling problem where the initial feasible solution is obtained by an insertion heuristic. This multi-agent task allocation model, when combined with multi-agent path planning, forms a MAPD model for modeling autonomous ground handling operations. Multi-agent path planning is solved using prioritized Safe Interval Path Planning (SIPP). A replanning model is developed to assess the resilience of our model to disruptions of operations. Also, a mixed integer nonlinear programming model, which includes an additional non-linear objective, is proposed to generate more realistic task assignments by minimizing the waiting time of vehicles on the aircraft stands. In this study, a four-hour planning window with three aircraft stands at AAS is used for the experiments. The results show that the proposed approach improves the computational time of the task allocation model with 48% for the normal traffic scenario, compared to the previously published results. The conflict-free routes of all ground support equipment (GSE) vehicles are all successful and close to the shortest path results, with an average increase of 0.04% and 10% for the path length and the duration of the path, respectively. Our model is therefore able to generate complete, high quality solutions in less than three minutes.

Rotterdam the Hague Airport (RTHA) is planning to implement a new self-service luggage check-in system which has a different concept of operations compared to the current self-service system for luggage check-in. To efficiently develop, evaluate and implement this new system, a simulation optimization model is required to test different system configurations. The simulation optimization model is required to have a high level of detail to accurately represent the airport terminal process. An agent-based model will be used as the basis as that approach provides the necessary level of detail. This model is then combined with the metaheuristic tabu search algorithm to optimize the new self-service luggage check-in system. To address challenges related to the large required computational resources to perform simulation optimization with a high level of detail, the Simheuristic framework will be used. Combining an agent-based model with the Simheuristic framework is a novel approach to simulation optimization. This approach requires the development of a deterministic version of the airport terminal model. It was found that implementing an optimally configured self-service luggage check-in system can improve the performance of the check-in process while not compromising the performance of the security checkpoint. Additionally, the Simheuristic framework is capable of reducing the computational resources required.

This study presents the development of a multimodal System of Systems (SoS) framework, to assess the impact of top-level aircraft requirements in providing advanced air door-to-door mobility. The field of Advanced Air Mobility (AAM) is rapidly evolving, and while current research assesses these vehicles from
both aircraft design and operational perspectives, a system of systems perspective combining both domains is required to find the best overall vehicle design and operational concepts. In this work, aircraft design is assessed by its effectiveness in the operating environment, considering the perspective of the stakeholders of passengers, operators and European policy makers. This is done through combining aircraft design in an agent-based multimodal simulation framework, covering both surface and air transport. Doing so allows us
to analyse the impact of top-level aircraft requirements on door-to-door travel time, AAM mode share and energy consumption per passenger. Using this framework, it was found that the most impactful parameter on AAM utilization is the passenger’s mean value of time. Additionally, several SoS effects could be traced to changes at the agent level, where giving passengers more travel options leads to 16% higher AAM adoption, but at shorter (-8.2 %) average trip distances due to preferences for shorter and cheaper flights. Lastly, the
relevance of considering operations for aircraft design was seen, where the best theoretical design for energy efficiency performed 10% worse than the best design considering operations, as the average mission profile in operation is different to the designed profile. In conclusion, the developed framework demonstrates the need for combining operations and aircraft design, and can be used to explore the best concepts in both fields.

Airport terminals are complex sociotechnical systems, in which humans interact with diverse technical systems. A natural way to represent them is through agent-based modeling. However, this method has two drawbacks: it entails a heavy computational burden and the emergent properties are often difficult to analyze. The purpose of our research is therefore to accurately abstract and explain the dynamics of airport terminal operations by means of computationally efficient and interpretable surrogate models, based on an existing agent-based simulation model. We propose a methodology consisting of two stages. Stage I involves the development of faithful surrogates. A sample is collected according to an active learning strategy, upon which Gaussian process regression, higher-order polynomials, gradient boosting, and random forests are fitted. Stage II then applies state-of-the-art techniques from the emerging field of explainable artificial intelligence to these models. Both model-agnostic and model-specific methods are considered, and their results are synthesized in order to explain the emergent properties. We prove the efficacy of this approach by conducting two case studies on AATOM, an existing Agent-based Airport Terminal Operations Model. The first case study examines the total expenditure on discretionary activities, such as shopping and dining. A combination of poor staffing strategies and high occupancy rates on certain flights was found to disrupt the terminal journey of passengers on subsequent flights. As a result of these knock-on phenomena, less free time is left for discretionary activities, which has a negative effect on the total expenditure. The second case study examines the throughput of security checkpoints. While throughput increases with passenger numbers, a clear point was observed where the checkpoint reaches its maximum capacity. This leads to longer queues and therefore higher waiting times. It even goes so far as to put passengers at risk of missing their flight, especially with poor staffing strategies. Altogether, we clearly observed the preservation of emergent phenomena in surrogate models, and conclude that their combination with interpretable machine learning is an effective way to explain the dynamics of complex sociotechnical systems.

Technology developments has enabled Unmanned Aircraft System (UAS) to be adopted for various applications, including Urban Air Mobility (UAM) – an air transportation system for passengers and cargo in and around urban environments. The operations of UAS in urban environment inevitably raises concerns about the safety impact of UAS.
The operational characteristic of UAS is largely different from the conventional commercial aviation. which brings novel safety issues for which the safety learning process has just started. To address these novel safety issues of UAS operations, it is essential to systematically study them within a formal setting of safety risk assessment.
Safety risk assessment involves a process that comprises risk indicators, risk analysis and risk evaluation. In recent years, regulators and researcher have dedicated significant efforts to developing risk assessment for UAS operations. These approaches are largely adopted from safety risk assessment of commercial aviation. However, it is essential to recognize that UAS operations have large differences with commercial aviation. Therefore there remains shortcomings and improvements to be made to the risk assessment of UAS operations.
This thesis addresses the further development of risk assessment methods for UAS operations for Urban Air Mobility (UAM). The main risk posed by UAM is third party risk (TPR) posed to people on the ground. Therefore, the focus of this thesis is on improving risk assessment methods for ground TPR.
The first study focuses on the TPR indicators for UAS operations. Based on these TPR indicators of commercial aviation, novel TPR indicators and nine separate third party fatality terms are identified. Subsequently, current UAS regulations are evaluated regarding their coverage of these nine third party fatality terms. By doing so, the research provides a more comprehensive understanding of the overall third party risk posed by UAS operations.
The second study aims to develop a safety risk assessment method for the novel ground TPR indicators proposed in the first study. To achieve this, a Monte Carlo simulation based risk assessment approach is proposed and applied to a hypothetical UAS urban parcel delivery case. The results show that the proposed annual ground TPR model and indicators provide an accumulated understanding of the risk posed to people on the ground. The non-negligible level of uncertainty in the models adopted highlights the need for further development of more accurate sub models for UAS ground TPR assessment.
The third study aims to improve the accuracy of the common ground TPR model, where a key limitation lays in the assumption that the product of impact PoF and size of impact area are independent of each other. To address this, an improved characterization is developed and evaluated using dynamical simulation of MBS model of a UAS impacting a human body. The comparison of the novel approach to existing approaches shows significant advantages of the novel developed approach.
The fourth study applies the novel approach developed in the third study to an urban parcel delivery UAS, weighting 15kg, equipped with airbag and parachute. A key motivation is that existing models do not address the risk mitigating effects of equipping a UAS with a combination of airbag and parachute. For the UAS equipped with an airbag Multi Body System (MBS) and Finite Element (FE) models are developed. Subsequently, these models are used to assess ground TPR for different cases with and without airbag and parachute. This analysis show that the method developed in the third study is able to quantify the risk reducing effects of the combination parachute and airbag.
The four interrelated series of studies have developed novel insights and methods in Third Party Risk assessment of UAS operations. These novel insights and methods can provide enhanced safety feedback to a UAS design process, and can stimulate further development of UAS regulation.

Inspired by the vision of fully autonomous airside operations at Schiphol airport, this study aims to contribute to the short-term goal of automated aircraft ground handling. In this research, we design and evaluate a multi-agent system for planning of automated ground handling. There are two main components in the system, task allocation optimization, and multi-agent path planning. To allocate tasks to ground support equipment (GSE) vehicles, an auction mechanism inspired by temporal sequential single item (TeSSI) auction is proposed. Ground handling tasks scheduling for GSE vehicles is modeled as several single-vehicle pickup and delivery optimization problems (SPDP), and the values of the objective functions are applied to generate bids for GSE vehicle agents in the auction. Moreover, Prioritized Safe Interval Path Planning for large agents (LA-SIPP) is used to plan collision-free paths for GSE vehicle agents in the model to execute tasks. Experimental studies have shown that the system is able to perform task allocation and path planning of ground handling tasks for flights in 3 aircraft stands within a 4-hour time in a reasonable computational time. Moreover, the model is capable to replan the tasks for agents when disruption happens. Applying the lowest possible numbers of vehicles used in the current operation, the model can always reach success allocation and path planning rates higher than 81% and 98%, respectively.