Current students

Ruben van Hilten (BSc) – The practical possibilities of a Deep Turnaround based alarm system

The Deep Turnaround (DT) system introduces the use of multiple cameras at aircraft VOPs to capture visual data of turnaround operations. This data is processed by a machine learning algorithm to interpret and predict ground handling events using historical datasets. While DT is primarily intended to improve turnaround predictability and efficiency, the visual data it generates can also support additional operational applications. In this project, improved situational awareness during pushback operations could enhance both safety and coordination between Schiphol, LVNL, and the airlines.
This thesis investigates the practical possibilities of using Deep Turnaround data to develop an alarm system for Air Traffic Control/Departure Manager (DMAN) that detects and reports pushbacks initiated without clearance. The research focuses on whether such an application is technically and operationally feasible as a future operational concept, and what implications it would have for implementation in the airport environment. The study explores how DT data can be translated into reliable alerts, assesses potential benefits and limitations for ATC and other stakeholders, and outlines a conceptual system framework. The intended outcome is an evaluation of feasibility and a structured design of the DT-based alert system.

Jaden Mauricia (BSc) – Analysis of Meteorological Conditions and ATC vectoring effects on Atypical Approaches

During arrival operations, aircraft are sequenced and vectored by air traffic control (ATC) to intercept the final approach and establish stabilized approach conditions prior to landing. Flight crews are required to control indicated airspeed, vertical flight path, lateral track and landing configuration in accordance with stabilized approach criteria and operational procedures. An approach that does not meet these criteria or operational requirements is considered atypical.

Atypical approaches are categorized as non-stabilized approaches (NSA), where aircraft parameters such as airspeed, descent rate, or landing configuration fall outside prescribed limits, and non-compliant approaches (NCA), where operational procedures or requirements are not met. These approaches are operationally significant, as they may reduce approach stability and pose a risk to safety.

Previous studies identified and quantified atypical approaches at Amsterdam Schiphol Airport. However, the influence of meteorological conditions and ATC vectoring on the occurrence of atypical approaches has not yet been fully assessed. This thesis analyses historical surveillance, meteorological, and vectoring data for arrivals to runways 18R and 18C at Amsterdam Schiphol Airport, with the objective of identifying the underlying causes of atypical approaches.

Sten den Hartog (MSc) – A Retrieval-Augmented NLP Framework for KDC Research Knowledge

Over the past decades, the Knowledge and Development Centre Mainport Schiphol has built an extensive archive of research reports and innovation studies in Air Traffic Management (ATM). While this represents significant intellectual capital, the scale and diversity of documents make systematic knowledge reuse increasingly challenging. This project develops a Retrieval-Augmented Generation (RAG) framework tailored to KDC’s research environment. Documents are segmented and stored in a dedicated vector embedding database, enabling semantic similarity search instead of keyword-based retrieval. Retrieved fragments are supplied to a Large Language Model (LLM), allowing it to generate structured, context-aware responses grounded in KDC’s own research outputs.
The challenge is not implementing RAG itself, but ensuring reliable reasoning and controlled generation in a research-critical context. The system must handle domain-specific terminology, maintain traceability to source material, and minimise hallucinations.
Beyond improving knowledge access within KDC, this work lays the foundation for applying RAG-based, context-aware generation within LVNL, supporting future knowledge management and research alignment across the Dutch ATM ecosystem.

Jort Simons (MSc) – Uncertainty Quantification for Inbound Air Traffic Demand in LVNL’s Decision Support Tool

Efficiently managing inbound air traffic flows requires precise demand prediction to prevent congestion and ensure safety. Currently, LVNL’s Decision Support Tool (DST) supports this by providing deterministic point predictions. However, these predictions do not convey the range of possible outcomes. In a dynamic operational environment like Schiphol, relying solely on a single predicted number can lead to overconfidence and inefficient resource allocation when actual demand deviates from the forecast.

This research aims to enhance the DST by explicitly quantifying the uncertainty associated with inbound air traffic demand. The project focuses on developing a wrapper model that functions alongside the existing prediction system to generate statistically robust prediction intervals. Crucially, this approach accounts for the heteroscedastic nature of air traffic data—recognizing that uncertainty is not constant, but varies significantly depending on the situation. By utilizing adaptive techniques, the prediction intervals will dynamically expand or contract to reflect the changing reliability of the forecasts.

By moving beyond point estimates and providing a clear, quantified range of uncertainty, this project intends to support more risk-informed decision-making. This will allow controllers and planners to better anticipate potential fluctuations in arriving traffic and optimize capacity management strategies accordingly.

Stijn Koelemaij (MSc) – Developing an improved metric to better predict aircraft noise annoyance based on number of flights and complaints data

Predicting the impact that aircraft noise will have on the community around an airport is a challenging problem. A human being experiences sound based on more than just the sound pressure (dB), and annoyance can increase even though measured dB levels do not. With the currently used metric Lden, a 3 dB reduction in engine noise during testing allows for 2x the number of flights to be flown in the same time, without the metric being affected. This 3 dB reduction is barely noticeable by the human ear on the ground, so the annoyance increases, which is reflected in complaints and survey data. In my project I will develop a new metric that incorporates the number of flights more explicitly, aiming to better predict the annoyance caused than the current Lden metric. This will be achieved by analyzing flight tracks, number of flights and peak levels and comparing them to the complaint data obtained from Bewoners Aanspreekpunt Schiphol. Together with this data analysis, listening experiments will be conducted at the TU Delft that test the effect of a higher number of marginally quieter flyovers on experienced annoyance, which will help to substantiate or reject the developed metric.

Javier Crespo Núñez (MSc) – Development of a data driven realistic agent based model for current ground surface operations at Schiphol

The development of this tool aims to provide a modular and flexible platform for users to easily develop and simulate Agent Based Models (ABM) for current Airport Surface Operations. Because of the future air traffic growth expectations, many researchers and companies around the globe are focusing on new developments and the optimization of current Air Traffic Management (ATM) processes and procedures. All of these proposed concepts are being tested and validated with different simplistic self-made simulations or expensive and rigid private commercial simulators.

The architecture to be implemented aims to provide an open source simulator framework that could bridge the gap between custom self-made simulations and private commercial simulators. By having a common open-source framework to easily implement and test new research, students, researchers and companies could accelerate their development and better contribute to the challenges faced today in a more standardized and collaborative way.

Furthermore, a realistic and well calibrated model of the current operations is critical to serve as the baseline and first step to test, evaluate, verify and analyze any future or current implementation. The availability of such a model would be incredibly useful to quickly and flexibly test new procedures, human machine interfaces, interactions between the stakeholders, infrastructure sizing analysis, workload, etc.

Amine Nari (MSc) – Modeling Delay Propagation for LVNL’s Decision Support Tool (DST)

Air traffic flow management depends on accurate demand prediciton, yet a major source of error comes from knock-on delays, which are primary delays that propagate to subsequent flights. These cascading effects create significant uncertainty in the European network, affecting both planning and operational decisions. This research focuses on understanding how delays develop and spread over ti

me across flights and connections, using Schiphol Airport as the primary case study. By analysing large-scale delay data, the study aims to identify key patterns and mechanisms driving delay propagation.

The goal is to develop a model that represents how delays evolve throughout the network and can simulate their impact on traffic demand. The model will be integrated into LVNL’s Decision Support Tool (DST) to improve demand prediction and support operational decision making.

Albert Chou (MSc) – Analysis Noise Abatement Procedures Schiphol

Noise Abatement Procedures (NAPs) are specific procedures designed to reduce noise impact in the airport surroundings. At Schiphol Airport, three main NAPs are considered; Noise Abatement Departure Procedures (NADP), Continuous Descent Approaches (CDA), and reduced flap setting descents. These NAPs reduce noise through a reduction in aircraft thrust and drag. However, the implementation of these NAPs is sensitive to aircraft type, aircraft mass, weather, air traffic, and other factors. Therefore, analysing compliance of flight tracks with NAPs is challenging. This research aims to incorporate a data-driven approach to analyse the compliance of flight tracks with NAPs such as NADP, CDA, and reduced flap setting descents through Automatic Dependent Surveillance–Broadcast (ADS-B) and Automatic Dependent Surveillance–Contract (ADS-C) data. By improving NAP detection, this research supports better evaluation of NAP compliance and aircraft noise emissions at Schiphol.

Coen Schinkel (BSc) – Artificial Intelligence Applications for Air Traffic Management

The integration of artificial intelligence (AI) in air traffic management (ATM) is expected to play a key role in improving efficiency, prediction, and safety. International programs such as SESAR and NextGen outline recurring domains where AI could add value, including trajectory prediction, conflict detection, speech recognition, anomaly detection, and human–machine teaming. For LVNL, however, it is not yet clear which of these applications are both relevant and feasible in daily operations.

My thesis focuses on identifying and evaluating AI applications that can realistically be adopted within the Dutch ATM context in the short term. The research begins by reviewing international frameworks and academic studies to determine which applications are sufficiently mature to be considered within the coming five years. These applications are then linked to LVNL’s organizational functions and assessed against three feasibility dimensions: operational fit, compliance with safety and cybersecurity requirements, and acceptance by air traffic controllers. The outcome of the study will be a prioritized set of AI opportunities that combine potential benefits with practical feasibility. In this way, the research provides LVNL with a concrete basis for deciding which AI applications to explore further, supporting both operational efficiency and the continued modernization of Dutch air traffic management.

Yoari Karelsz (MSc) – Managing Idle Descent Trajectory Uncertainties at Schiphol

Idle descents are continuous descent operations flown at almost zero thrust. These descents are more fuel efficient and cause less noise compared to normal powered descents. However, air traffic control cannot intervene during such descents, or the additional efficiency will be lost. Change in wind, pilot inputs and other factors, normally compensated with thrust, cause these descents to be less predictable. Therefore, an additional spacing buffer of two till four minutes is required to safely fly idle descents. This has a large impact on the landing capacity of Schiphol, hence only nighttime operations incorporate idle descents when possible. This research aims to provide insight into the trajectory uncertainties of idle descents by analysing flight recorder data. Quantifying and reducing the trajectory uncertainties of idle descents can potentially reduce the required spacing buffer and help with flying more idle descents.

Niels Prins (MSc) – Trajectory-Based Operations in Dutch medium-altitude airspace under mixed ADS-C equipage conditions

Trajectory-Based Operations (TBO) aim to improve the predictability, efficiency, and safety of air traffic management through the use of intent-based trajectory information. With the growing availability of Automatic Dependent Surveillance–Contract (ADS-C) technology, TBO is becoming increasingly feasible. This thesis focuses on developing a trajectory management framework for the Dutch Flight Information Region (FIR) within the FL100–FL260 range, managed by Amsterdam Area Control Centre. Designed for mixed equipage conditions, it will accommodate both ADS-C-equipped and conventional aircraft, managing inbound and outbound traffic while integrating a conflict detection and resolution module. Operating on a time horizon of minutes to tens of minutes, the framework enables continuous trajectory planning and strategic conflict resolution, while retaining tactical intervention capability. Tested with real-world Dutch traffic scenarios, and potentially extended to lower altitude sectors, the framework represents a pragmatic, evolutionary step towards TBO, enhancing planning capabilities while preserving the flexibility required in current operations.

Julia Huigen (MSc) – Environmental performance of TBO implementation evolutions at Mainport Schiphol

The aviation industry contributes significantly to global greenhouse gas emissions, and its environmental impact is expected to grow as air traffic continues to increase. As a result, both Airlines and Air Navigation Service Providers (ANSP) are required to reduce emissions. Numerous studies have shown the benefits of Continuous Descent Operations (CDO) in reducing fuel consumption and emissions. However, these studies often focus on simplified situations and do not fully consider the real challenges of busy and complex airspace using real-world data. This research aims to provide practical insights into the potential of TBO in the Dutch airspace. This thesis will investigate how vertical flight performance can be improved by addressing inefficiencies through the implementation of TBO, particularly during descent. Airline data will be utilised as a necessary input to understand real-world operations and identify areas where performance can be improved.

Jens Bremer (MSc) – Improving the departure manager (DMAN) at Schiphol through PEGT Integration and optimization of the 10-minute bin mechanism

Schiphol’s current Departure Manager (DMAN) system, built around Target Off-Block Times (TOBT) and fixed 10-minute bins, offers operational stability but lacks the granularity needed to efficiently sequence departures in a high-density environment. This research investigates the integration of the Predicted End of Ground Handling Time (PEGT) into the DMAN. This new PEGT is a machine learning-based estimate developed by Schiphol Aviation Solutions’ Deep Turnaround project. PEGT leverages image-based machine learning algorithms to monitor over 70 unique turnaround events in real-time, generating predictive insights from over 150,000 historical turnarounds. These predictions enable early delay detection (up to 40 minutes in advance), supporting more precise TOBT estimates and improved gate and runway slot usage. Building on previous work and aligned with SESAR’s integrated DMAN-AMAN vision, this study also questions the effectiveness of rigid 10-minute departure bins. Using historical PEGT & TOBT data together with simulation models inspired by traffic flow constraints, this project evaluates dynamic bin sizing and delay damping mechanisms. The goal is to find a balance between operational stability and responsiveness in pre-departure sequencing. By combining turnaround predictions with more flexible pre-departure scheduling, this research aims to improve DMAN’s ability to manage departure flows efficiently, reduce last-minute gate conflicts, and enhance overall airport throughput at complex hub airports like Schiphol.

Jorn van Beek (MSc) – Evaluation of Arrival Manager Horizon Extension in a Trajectory Management context

The Arrival Manager (AMAN) system is used to provide regulation on aircraft entering the Terminal Maneuvering Area (TMA). New regulations require the AMAN to freeze the sequence further ahead than the current 14 minutes with the aim of reducing fuel burn and decreasing controller workload. The increased horizon combined with uncertainty on arriving aircraft cause these aircraft to “pop-up”. This causes sequence errors which increase workload and fuel burn. These effects are more pronounced when using the AMAN in a Trajectory Based Operation (TBO) environment. Previously, research on the horizon extension and pop-up impact and solutions has been performed, although in idealized scenarios without a realistic solution. This research aims to model the Arrival Manager in high fidelity to support decisions on the Arrival Manager design, and to develop and test realistic solutions to decrease the impact of the pop-up traffic when extending the AMAN horizon towards TBO operations.