PhDs receiving financial support through Engage. Initial contact with PhDs can be made through Engage.
The final reports provide a summary of the research – the published PhD theses will be available from the hosting universities.
Thematic challenge: N/A
The general objective of DiSpAtCH (Decision Support System for Airline Operation Control Hub Centre) is to elaborate on artificial intelligence technologies and how these technologies could efficiently support decision making in an Airline Operation Control Hub Centre (OCC) in unexpected or very complex situations.
The daily operation of airlines is often disrupted by unplanned events. As an airline it is therefore essential to operate an OCC to be able to react and mitigate any consequences from the initial disruption. The most challenging task is the information management task. This task includes monitoring, recognition and projection of relevant information out of all information available including current and future situations.
Today the decision making process mainly relies on the experience of the staff working in the OCC. Like in other industries, the desire of using Decision Support Tools (DST) based on machine learning (ML) algorithms is also increasing in the aviation industry. ML algorithms, like neural networks, need a large amount of data to be trained with. The focus of DiSpAtCH is to develop a DST which aims to help the staff in an OCC during disrupted situations. Therefore, three ML modules have been defined of which one aims to propose a suitable action/solution in a disrupted situation. To train the algorithm a database including information about disruptions as well as the implemented solutions from past disrupted situations is needed. Since these kinds of data are not available to researchers and often not recorded by airlines themselves, an approach was needed to get some data to start training algorithms and to validate that certain DST can be developed and support the disruption management process within an OCC. With a decision support system like DiSpAtCH the decisions within an OCC can be optimized which will result in fewer overall disruption cost.
DiSpAtCH provides an approach of using an airline simulation to generate generic operational data of an airline and its daily operations. Synthetic data are generated and ML algorithms are trained to predict actions/solutions for disrupted situations. A first validation shows that a four step classification process including two neural networks can be used to predict actions/solutions in disrupted situations with an accuracy of around 95% and therefore reduce the overall disruption cost by 61% compared to randomly selected actions/solutions.
Thematic challenge: 2 – Data-driven trajectory prediction
The objective of this Engage KTN PhD study is to explore and present state of the art AI/ML algorithms towards planning conflict-free trajectories in computationally efficient ways, for a large number of trajectories in airspaces comprising multiple FIRs, following a methodology combining data-driven and agent-based approaches.
In the context of this study the conflicts-free trajectory planning task is defined to incorporate trajectory prediction and conflicts detection and resolution. While trajectory prediction concerns predicting the spatiotemporal evolution of the aircraft state along a trajectory (also called, trajectory evolution), conflicts detection and resolution concerns the detection of conflicts that breach separation minima (loss of separation) between flights and their resolution by appropriate actions. Therefore, the objective of the conflicts-free trajectory planning task is to predict the evolution of trajectories, and regulating flights to avoid loss of separation.
While trajectory planning may take place at the pre-tactical phase of operations, we expect the methods developed in this study to have a large impact in the tactical phase of operations.
Aiming to model stakeholders’ decisions to planning conflict-free trajectories, the major emphasis of this study is to imitate flights’ trajectories and air traffic controller’s behavior according to demonstrations provided by historical data.
The challenges that this study addressed are as follows:
1. Plan trajectories, considering complex ATM phenomena and operational constraints regarding traffic and conflicts among trajectories.
2. Follow a data-driven approach to learn stakeholders’ preferences on the evolution of trajectories and on resolving conflicts: stakeholders include airspace users (for trajectory prediction) and air traffic controllers (for conflicts’ detection and resolution actions).
3. Address optimization in trajectory planning w.r.t. multiple objectives, preferences and constraints of stakeholders involved, as these are demonstrated by historical data.
4. Address scalability: demonstrate the efficiency of the methods to be applied in settings with a large number of flights.
Contributions that this study makes are as follows:
1. The problem of modelling air traffic controllers’ behavior has been split into two well-defined problems: modelling air traffic controllers’ reactions on whether and when conflicts’ resolution actions should be applied, and modelling air traffic controllers’ reactions on how conflicts should be resolved, i.e. what resolution actions should be applied.
2. The problem of trajectory planning (either with or without considering conflicts) has been formulated as an imitation learning problem, based on historical flown trajectories.
3. AI/ML methods have been developed and tested on learning models regarding the evolution of 4D trajectories, using data-driven approaches, i.e. based on historical real-world data.
4. AI/ML methods have been developed and tested on learning models regarding air traffic controllers’ reactions and policy using data-driven approaches, i.e. based on historical real-world data.
5. This study has proposed an elaborated evaluation method for data-driven imitation learning techniques predicting air traffic controllers’ reactions, considering the uncertainties involved in the evolution of trajectories, in the assessment of conflicts, and in the reactions of ATCO.
6. Challenging issues due to inherent data limitations have been addressed and thoroughly discussed.
7. The study provides an integrated trajectory planning approach, where data-driven trajectory predictions are intertwined with data-driven conflicts detection and resolution.
Thematic challenge: N/A
Among the navigation means, Global Navigation Satellites Systems (GNSS), and namely the Global Positioning System (GPS), have become essential and the availability of a GNSS navigation solution on board seems completely natural. However, the quality of the position calculated by the on-board equipment may be reduced when the received signal is degraded. This degradation can find its origin in a defect of the signal generation system, carried by the satellite, or in the receiving conditions, typically when interferences or multipaths are in addition to the desired signal.
The objectives of the thesis were to detect, classify, identify and finally reduce the impairments of the GNSS signals seen by the on-board receiver, by means of Machine Learning techniques.
More specifically, the performance of Machine Learning methods has been assessed on the signal at the correlator output, the correlator output in short. Indeed, the correlator output is a key element in the calculation of the aircraft’s position by the receiver, and, consequently, it is the link in the signal processing chain where the degradations have the most significant impact.
Correlations of the received signal with a local replica over a (Doppler shift, propagation delay)-grid are mapped into grayscale 2D images. They depict the received information possibly contaminated by multipath propagation. The images feed a Convolutional Neural Network (CNN) for automatic feature construction and multipath pattern detection.
The issue of unavailability of a large amount of supervised data required for CNN training has been overcome by the development of a synthetic data generator. It implements a well-established and documented theoretical model. A comparison of synthetic data with real samples is proposed.
The complete framework is tested for various signal characteristics and algorithm parameters. The prediction accuracy does not fall below 93% for Carrier-to-Noise ratio (C/N0) as low as 36 dBHz, corresponding to poor receiving conditions. In addition, the model turns out to be robust to the reduction of image resolution.
Thematic challenge: 2 – Data-driven trajectory prediction
The goal of air traffic flow and capacity management (ATFCM) is to ensure that airport and airspace capacity meet traffic demand while optimising traffic flows to avoid exceeding the available capacity when it cannot be further increased. In Europe, ATFCM is handled by EUROCONTROL, in its role of Network Manager (NM), and comprises three phases: strategic, pre-tactical, and tactical. This thesis is focused on the pre-tactical phase, which covers the six days prior to the day of operations.
During the pre-tactical phase, few or no flight plans (FPLs) have been filed by airspace users (AUs) and the only flight information available to the NM are the so-called flight intentions (FIs), consisting mainly of flight schedules. Trajectory information becomes available only when the AUs send their FPLs. This information is required to ensure a correct allocation of resources in coordination with air navigation service providers (ANSPs). To forecast FPLs before they are filed by the AUs, the NM relies on the PREDICT tool, which generates traffic forecasts for the whole European Civil Aviation Conference (ECAC) area according to the trajectories chosen by the same or similar flights in the recent past, without taking advantage of the information on AU choices encoded in historical data.
The goal of the present PhD thesis is to develop a solution for pre-tactical traffic forecast that improves the predictive performance of the PREDICT tool while being able to cope with the entire set of flights in the ECAC network in a computationally efficient manner. To this end, trajectory forecasting approaches based on machine learning models trained on historical data have been explored, evaluating their predictive performance.
In the application of machine learning techniques to demand trajectory prediction, three fundamental methodological choices have to be made: (i) approach to trajectory clustering, which is used to group similar trajectories in order to simplify the trajectory prediction problem; (ii) model formulation; and (iii) model training approach. The contribution of this PhD thesis to the state of the-art lies in the first two areas. First, we have developed a novel route clustering technique based on the area comprised between two routes that reduces the required computational time and increases the scalability with respect to other clustering techniques described in the literature. Second, we have developed, tested and evaluated two new modelling approaches for route prediction. The first approach consists in building and training an independent machine learning model for each origin-destination (OD) pair in the network, taking as inputs different variables available from FIs plus other variables related to weather and to the number of regulations. This approach improves the performance of the PREDICT model, but it also has an important limitation: it does not consider changes in the route availability, thus being unable to predict routes not available in the training data and sometimes predicting routes that are not compatible with the airspace structure. The second approach is an airline-based approach, which consists in building and training a model for each airline. The limitations of the first model are overcome by considering as input variables not only the variables available from the FIs and the weather, but also route availability and route characteristics (e.g., route cost, length, etc.).
The airline-based approach yields a significant improvement with respect to PREDICT and to the OD pair-based model, achieving a route prediction accuracy of 0.896 (versus PREDICT’s accuracy of 0.828), while being able to deal with the full ECAC network within reasonable computational time. These promising results encourage us to be optimistic about the future implementation of the proposed system.
Thematic challenge: 2 – Data-driven trajectory prediction
Air Traffic Management’s (ATM) aim is to ensure separation management of aircraft in an efficient way, minimizing possible delays and costs. The expected increase in air traffic demand across manned and unmanned traffic requires a higher level of automation to support the decision making. Adaptive self-Governed aerial Ecosystem by Negotiated Traffic (AGENT) was an exploratory research project supported by the H2020 Research and Innovation Programme, which proposed a system where the avoidance of potential loss of separations is done in a distributed and collaborative way while the controllers monitor the process. This PhD project is built on AGENT’s future work proposals and seeks possible improvement of several critical aspects of the system through the application of Machine Learning (ML) techniques. There were two clear goals in this project: define airspace complexity in a way that challenges current definitions and overcomes their limitations and investigate how ML can be applied to safety in aviation. We investigate these problems in en-route traffic at the tactical level, as well as UAV systems.
The first major contribution of this thesis has been modelling air traffic as a graph in the context of airspace complexity and conflict resolution. We define a graph with aircraft as nodes and interdependencies between them as edges of the graph. This definition allows for problem specific definitions of interdependencies. We further extend the definition of air traffic as a graph by including the time domain, which creates dynamic graphs. We define airspace complexity as graph connectivity and propose four indicators that combine different topological information and the severity of interdependencies to give a complete and nuanced picture of complexity. These indicators are able to provide a dynamic evolution of complexity by leveraging the modelling choice of air traffic as a dynamic graph. Simulation results indicated that the indicators we propose give detailed information and overcome drawbacks of existing metrics. We evaluated our approach using real and synthetic traffic and demonstrated that the indicators express different facets of complexity, confirming that all indicators are needed. The way we define complexity also provides a new framework in the design of conflict resolution algorithms which considers the reduction of airspace complexity in addition to safety preservations. Conflict Resolution (CR) algorithms could be discouraged from providing solutions that increase the overall complexity of the airspace.
Furthermore, we model CR as Multiagent Reinforcement Learning Problem (MARL). We initially investigate CR only in a pairwise setting using Multiagent Deep Deterministic Policy Gradient (MADDPG) as a learning algorithm. We propose a novel state representation that combines positional information with speed and heading of the aircraft. Additionally, we propose a reward function that not only guides agents towards solving the conflict but also to consider factors such as fuel consumption, airspace complexity and delays. Our results indicate that the agents are capable of solving the conflicts and further learning desired behaviours such as solving them as soon as possible with minimal manoeuvres. However, this method suffers from issues of scalability and nonstationarity. In order to overcome these issues, we utilize Graph Neural Networks (GNNs). GNNs inherently allow communication between agents which facilitates cooperation between them. We apply Graph Convolutional Reinforcement Learning (DGN) in CR for Unmanned Aerial Vehicles (UAV) to solve conflicts with 3 and 4 present aircraft which we assume to be cooperative. We achieve impressive performance with the agents being able to always solve the conflicts. Furthermore, they learn a strategy that increases the distance between them, without previous knowledge of the environment. Currently, we are using this application domain to investigate some fundamental questions in MARL such as agent coordination, heterogeneity and transparency in environments where agents have individual and common goals.
Thematic challenge: 2 – Data-driven trajectory prediction, and 3 – Efficient provision and use of meteorological information in ATM
Weather has a strong impact on ATM. Inefficient weather avoidance procedures and inaccurate prognosis lead to longer aircraft routes and, as a result, to fuel waste and increased negative environmental impact. A better integration of weather information into the operational ATM-system will ultimately improve the overall air traffic safety and efficiency.
Covid-19 pandemics affected aviation severely, resulting in an unprecedented reduction of air traffic, and gave the opportunity to study the flight performance in non-congested scenarios. We discovered noticeable inefficiencies and environmental performance degradation, which persisted despite significant reduction of traffic intensity. The PhD thesis proposes a methodology that allows us to distinguish which factors have the highest impact on which aspects of arrival performance in horizontal and vertical dimensions.
Academic Excellence in ATM and UTM Research (AEAR) group operating within the Communications and Transport Systems (KTS) division at Linköping University (LIU), together with the Research and Development at Luftfartsverket (LFV, Swedish ANSP) develops optimization techniques to support efficient decision-making for aviation authorities.
In this thesis, we apply probabilistic weather modelling techniques, taking into account the influence of bad weather conditions on the solutions developed in our related projects and integrate them into the corresponding optimization frameworks. First, the PhD student enhanced the optimization framework for arrival route planning in TMA, with the convective weather avoidance technique. Next, the probabilistic weather products were used to obtain an ensemble of staffing solutions, from which the probability distributions of the number of necessary ATCOs were derived. The modelling is based on the techniques recently developed within several SESAR projects addressing weather uncertainty challenges. The proposed solutions were successfully tested using the historical flight data from Stockholm Arlanda airport and five airports in Sweden planned for remote operation in the future.
Thematic challenge: 2 – Data-driven trajectory prediction
Accurate and reliable trajectory prediction (TP) is required in several air traffic management (ATM) systems, for instance, to design air and ground-based decision support tools and safety nets. Estimating the aircraft trajectory in the vertical plane typically requires the knowledge of a pair of aircraft intents (e.g., constant Mach and minimum throttle), information which is seldom available, besides for the ownship (i.e., one’s own aircraft) trajectory planning system. In the flight execution phase, the aircraft is directed by the (auto) pilot through a series of sequential guidance modes that might override some of the planning phase aircraft intents. Thus, guidance mode is defined as a combination of constraints/commands that specify how the aircraft should behave to perform a desired trajectory.
Reliable guidance mode information is fundamental for next generation of air- or ground-based TP, especially in the context of trajectory-based operations (TBO) and advanced decision support tools for aircraft crew and/or air traffic control e.g., to improve conflict detection (and resolution) algorithms, conformance monitoring, departure/arrival managers, separation assurance systems, etc. These new tools might result in increased safety, capacity, predictability and cost-efficiency for the future European ATM system.
This research is concentrated on identifying aircraft guidance modes in the vertical plane. The final goal of this study is to indicate that acquiring the knowledge of aircraft guidance mode significantly affects the TP problem, and subsequently, the new ATM systems. In this PhD i) we provided a new probabilistic perspective of the trajectory prediction problem using signal processing mathematical tools, ii) we review state-of-the-art and the main challenges for the design of novel or enhanced TP systems that should enable future ATM paradigms, iii) we develop an optimal guidance mode identification using a Kalman filtering approach, iv) we analyse the impact of model mismatch on the interacting multiple model (IMM) filtering technique, v) we propose a robust linear-constrained IMM filtering under model mismatch, vi) we also propose a new methodology based on Bayesian inference to identify the aircraft guidance modes, and finally, vii) we evaluate the methodology to indicate the effect of known guidance modes on the TP accuracy.
Thematic challenge: 3 – Efficient provision and use of meteorological information in ATM
Uncertainties inherent to convective weather constitute a major challenge for the Air Traffic Management System (ATM), affecting its safety, capacity, and efficiency. Specifically, thunderstorms represent an important threat, as they involve phenomena such as strong turbulence, wind shear or hail. It is essential to avoid them to ensure both passenger comfort and aircraft structural integrity. Thunderstorms’ location and timing are hard to predict with certainty. This stochasticity is an important element that methodologies for aircraft trajectory planning must take into account.
For this purpose, two different methodologies for flight planning in areas of uncertain thunderstorm development are proposed. Both are heuristic approaches that rely on the iterative manipulation of graphs. Moreover, to enhance computational performance and enable real time operation, they are parallelized by means of GPU programming, producing results in less than seconds.
On one hand, the Scenario-Based Rapidly-Exploring Random Trees (Scenario-Based RRTs or SB-RRTs) are introduced, three algorithms for trajectory planning that explore an airspace with a tree structure. This kind of graph grows from the origin and looks for a connection with the destination through a safe sequence of tree branches. On the other hand, the Augmented Random Search (ARS) is proposed for trajectory deformation. This algorithm is applied to a graph, and it looks for the optimal sequence of edges, its relocation, and the best profile of velocities to minimize a combination of time and fuel.
The methodologies are tested with Ensemble Prediction Systems (EPS) that characterize atmospheric uncertainties through a set of possible forecasts. Results reveal that the algorithms are able to ensure safety and minimize objectives, such as time of flight, flight distance or fuel consumption.
Thematic challenge: 4 – Novel and more effective allocation markets in ATM
The main objective of this work is to create a Decision Support Tool to help the Airport Operation Centre with the integration of different approaches at the macroscopic level to make better decisions to minimize airport congestion by mitigating conflicts of critical resources. The main conflicts are related to different processes of the airport management and the capacity, so, the main problems are related to the minimum separation, runway, taxiway, terminal, gates, and ground handling team capacity (overloads) and availability.
We propose a framework as part of the Decision Support Tool to solve the conflicts addressed, we adapted an optimization with simulated annealing heuristic combined with a time decomposition approach (sliding windows).
As part of the solution, we evaluate the performance of the different modules and how the number of conflicts is solved, the final objective is to improve the coordination and efficiency of the operations of an airport. To validate the optimization model and to show the benefits of the macroscopic decomposition approach different computational experiments were performed with real data of one day of operations from Paris Charles de Gaulle airport including the parameters of this airport.
Thematic challenge: 4 – Novel and more effective allocation markets in ATM
Air Traffic Flow Management (ATFM) and airlines use different paradigms for the prioritisation of flights. While ATFM regards each flight as individual entity when it controls sector capacity utilization, airlines evaluate each flight as part of an aircraft rotation, crew pairing and passenger itinerary. As a result, ATFM slot regulations during capacity constraints are poorly coordinated with the resource interdependencies within an airline network, such that the aircraft turnaround – as the connecting element or breaking point between individual flights in an airline schedule – is the major contributor to primary and reactionary delays in Europe.
This dissertation bridges the gap between both paradigms by developing an integrated schedule recovery model that enables airlines to define their optimal flight priorities for schedule disturbances arising from ATFM capacity constraints. These priorities consider constrained airport resources, such as ATFM slots, airport stands or ground handling personnel and different methods are studied how to communicate airline-internal priorities confidentially to external stakeholders for collaborative solutions, such as the assignment of reserve resources or ATFM slot swapping.
The integrated schedule recovery model is an extension of the Resource-Constrained Project Scheduling Problem and integrates aircraft turnaround operations with existing approaches for aircraft, crew and passenger recovery. The model is supposed to provide tactical decision support for airline operations controllers at look-ahead times of more than two hours prior to a scheduled hub bank. System-inherent uncertainties about process deviations and potential future disruptions are incorporated into the optimization via stochastic turnaround process times and the novel concept of stochastic delay cost functions. These functions estimate the costs of delay propagation and derive flight-specific downstream recovery capacities from historical operations data, such that scarce resources at the hub airport can be allocated to the most critical turnarounds.
The model is applied to the case study of a network carrier that aims at minimizing its tactical costs from several disturbance scenarios. The case study analysis reveals that optimal recovery solutions are very sensitive to the type, scope and intensity of a disturbance, such that there is neither a general optimal solution for different types of disturbance nor for disturbances of the same kind. Thus, airlines require a flexible and efficient optimization method, which considers the complex interdependencies among their constrained resources and generates context-specific solutions. To determine the efficiency of such an optimization method, its achieved network resilience should be studied in comparison to current procedures over longer periods of operation.
For the sample of analysed scenarios in this dissertation, it can be concluded that stand reallocation, ramp direct services, quick-turnaround procedures and flight retiming are very efficient recovery options when only a few flights obtain low and medium delays, i.e., 95% of the season. For disturbances which induce high delay into the entire airline network, a full integration of all considered recovery options is required to achieve a substantial reduction of tactical costs. Thereby, especially arrival and departure slot swapping are valuable options for the airline to redistribute its assigned ATFM delays onto those aircraft that have the least critical constraints in their downstream rotations.
The consideration of uncertainties in the downstream airline network reveals that an optimization based on deterministic delay costs may overestimate the tactical costs for the airline. Optimal recovery solutions based on stochastic delay costs differ significantly from the deterministic approach and are observed to result in less passenger rebooking at the hub airport.
Furthermore, the proposed schedule recovery model can define flight priorities and internal slot values for the airline. Results show that the priorities can be communicated confidentially to ATFM by using flight delay margins, while slot values may support future inter-airline slot trading mechanisms.