Abstract

Highly accurate positioning systems and new broadcasting technology have enabled air traffic management concepts where the responsibility for aircraft separation resides on pilots rather than on air traffic controllers. The Formal Methods Group at the National Institute of Aerospace and NASA Langley Research Center has proposed and formally verified an algorithm, called KB3D, for distributed three dimensional conflict resolution. KB3D computes resolution maneuvers where only one component of the velocity vector, i.e., ground speed, vertical speed, or heading, is modified. Although these maneuvers are simple to implement by a pilot, they are not necessarily optimal from a geometrical point of view. In general, optimal resolutions require the combination of all the components of the velocity vector. In this paper, we propose a two dimensional version of KB3D, which we call KB2D, that computes resolution maneuvers that are optimal with respect to ground speed and heading changes. The algorithm has been mechanically verified in the Prototype Verification System (PVS). The verification relies on algebraic proof techniques for the manipulation of the geometrical concepts relevant to the algorithm as well as standard deductive techniques available in PVS. Document type: Part of book or chapter of book

Abstract

International audience; The resolution of Air Traffic Control (ATC) conflicts is a constrained optimization problem : the goal is to propose, for a certain number, n, of aircraft, which might be in conflict in a near future, trajectories that satisfy the separation constraints between aircraft, and minimizes the delays due to the conflict's resolution. The type of conflict resolution trajectories we use allows to split the problem in two steps: first we choose, and freeze, what we call a configuration of the problem, i.e. for each aircraft, the direction in which the aircraft is diverted, and for each pair of aircraft, which of the two aircraft passes first at the crossing point of the two aircraft trajectories. We can then compute the optimal trajectories corresponding to this configuration, by solving a simple linear optimization problem. Thus we can use an Genetic Algorithm, along with a linear optimization algorithm, such as the simplex algorithm: the elements of the population, on which the GA operates, code configurations of the problem, and are evaluated using a linear optimization program.The advantage of this approach is that we get, as well as the fitness of an element of the population, the local optima corresponding to the configuration coded by this element. The GA actually searches for the global optimum among these local optima. We applied this methods to conflicts in which up to 6 aircraft are involved, and obtained really promising results. Document type: Part of book or chapter of book

Abstract

This study compared situation awareness across three flight deck decision aiding modes. Pilots resolved air traffic conflicts using a click and drag software tool. In the automated aiding condition, pilots executed all resolutions generated by the automation. In the interactive condition, automation suggested a maneuver, but pilots had the choice of accepting or modifying the provided resolution. In the manual condition pilots generated resolutions independently. A technique that combines both Situation Global Assessment Technique and Situation Present Awareness Method was used to assess situation awareness. Results showed that situation awareness was better in the Manual and Interactive conditions when compared to the Automated condition. The finding suggests that pilots are able to maintain greater situation awareness when they are actively engaged in the conflict resolution process. Document type: Part of book or chapter of book

Abstract

The conflict resolution task performed by air-traffic controllers appears a suitable task for automation using CBR. This is because human competence seems to involve recognising situations and reusing solutions. In this paper we present our experiences in developing a CBR system to support this conflict resolution task. We discuss the problems of case representation: the macro problem of what should constitute a case and the micro problem of how to characterise a case. We evaluate some alternative case representations and identify a representation with one aircraft per case that is extendible to describe conflicts with multiple aircraft. Document type: Part of book or chapter of book

Abstract

This report presents design principles and algorithms for building a fast-time scheduler of arrival aircraft based on a FCFS (first-come-first-served) scheduling protocol and non FCFS. The algorithms provide the conceptual and computational foundation for the Traffic Management Advisor (TMA) of the Centre/terminal radar approach control facilities (TRACON) automation system, that's it an arrival manager (AMAN concept). The primary objective of the scheduler is to assign arrival aircraft to a favourable landing runway and schedule then to land at times that minimize delays. In order to solve the problem a fast time simulator platform was selected by Eurocontrol (the European Organisation for the Safety of Air Navigation) named AirTOp. Many problems in that module, as for example the lack of information about the code, the fact that for changing one small parameter, the process involved takes more than 2 or 4 months or the fact that EUROCONTROL is paying a special license only to use this specific module has brought to think in a better solution. This solutions entails the creation of a separate module from zero, a plugin, due to AirTOp is providing and API (Application Programming Interface) where the user can create, retrieve or modify the data during the simulations. The creation of this module was made in Java due to AirTOp is programmed in Java and allows easily the creation of a plugin. A plugin is an external module that can use some information from the main software, and allows the user to model and create some new information, such as a new AMAN. In the creation of the AMAN plugin two solution has been tested, in order of complexity to continue programming, an FCFS time schedule algorithm (First Come First Served) and a non-FCFS time schedule algorithm. By the end of this project a full AMAN was created in Java with Eclipse, helping the Air Traffic Controllers to manage the situation with better solutions, it can reduce the average delays up to 20% (depending of the case) and it improves the time constraint feasibility in a 17%. A full deployment and distribution of this plugin has been made for the ATM/RDS/ATS unit in Eurocontrol, unit in charge of the research and development for the Air Traffic Management and Air Traffic Services improvements and innovations. In addition a user guide for the creation of a plugin has been created, explaining all required steps to develop a generic plugin.

Abstract

The increasing in traffic demand has strained air traffic control system and controllers which lead to the need of novel and efficient conflict detection and resolution advisory. In the scope of this thesis, we concentrate on studying challenges in conflict detection and resolution by using machine learning approaches. We have attempted to learn and predict controller behaviors from data using Random Forest. We also propose a novel approach for probabilistic conflict detection by using Heteroscedastic Gaussian Process as predictive models and Bayesian Optimization for probabilistic conflict detection algorithm. Finally, we propose an artificial intelligent agent that is capable of resolving conflicts, in the presence of traffic and uncertainty. The conflict resolution task is formulated as a decision-making problem in large and complex action space, which is applicable for employing reinforcement learning algorithm. Our work includes the development of a learning environment, scenario state representation, reward function, and learning algorithm. Machine learning methods have showed their advantages and potential in conflict detection and resolution related challenges. However, more studies would be conducted to improve their performances such as airspace network representation, multi-agent reinforcement learning or controller's strategy reconstruction from data.; L'augmentation de la demande de trafic a mis à rude épreuve le système de contrôle de la circulation aérienne et les contrôleurs, d'où la nécessité d'un système novateur et efficace de détection et de résolution des conflits. Dans le cadre de cette thèse, nous nous concentrons sur l'étude des défis de la détection et de la résolution des conflits en utilisant des approches d'apprentissage machine. Nous avons essayé d'apprendre et de prédire les comportements des contrôleurs à partir de données utilisant Random Forest. Nous proposons également une nouvelle approche pour la détection probabiliste des conflits en utilisant le Processus Gaussien Hétéroscédastique comme modèles prédictifs et l'optimisation bayésienne pour l'algorithme de détection probabiliste des conflits. Enfin, nous proposons un agent intelligent artificiel capable de résoudre les conflits, en présence de trafic et d'incertitude. La tâche de résolution de conflit est formulée comme un problème de prise de décision dans un espace d'action vaste et complexe, qui est applicable pour l'utilisation d'un algorithme d'apprentissage par renforcement. Notre travail comprend le développement d'un environnement d'apprentissage, la représentation des états des scénarios, la fonction de récompense et l'algorithme d'apprentissage. Les méthodes d'apprentissage machine ont montré leurs avantages et leur potentiel dans la détection et la résolution des conflits. Cependant, d'autres études seraient menées pour améliorer leurs performances, telles que la représentation du réseau de l'espace aérien, l'apprentissage du renforcement multi-agent ou la reconstruction de la stratégie du contrôleur à partir de données.

Abstract

_ Best of Session (UTM-6: sUAS Traffic Management and Deconfliction) Award, 38 DASC 38th IEEE/AIAA Digital Avionics Systems Conference This paper studies strategic conflict resolution for air traffic based on sum coloring. We consider two application scenarios: manned and unmanned air traffic, with similar targets: to improve efficiency of operations and to reduce the costs. For the Unmanned Air Vehicles Traffic Management (UTM) we consider also a payment mechanism which incentivizes the operators to share information necessary to find a socially optimal solution. We quantify the potential savings via a series of experiments, showing that our methods drastically outperform the widely used First-Come-First-Serve (FCFS) strategy. Peer Reviewed Award-winning

Abstract

International audience; Air transportation traffic is progressively increasing over the years and dealing with it is an essential task to guarantee fluid flights in the future. Several works already indexed multiple aspects of aviation, among them, the E-MAN system. It introduced the sequencing of arriving traffic, starting from early stages of the En-route phase. This change facilitated the work for the approach controllers but also increased the workload of the En-route controllers. To handle that workload, controllers are now assisted by tools that consider the new constraints introduced by the arrival management system and propose advisories. From that same perspective, our project focuses on an algorithm for a helper tool that will combine both aspects of traffic sequencing in the En-route phase and conflict resolution. With this novel approach, we automatically generate near-to-optimal flight decisions, given that we can modify the speed and the flight level to respect the sequencing constraints and cut down potential conflicts. We categorize the problem as a mathematical optimization case. Thus, we describe a detailed mathematical model which covers all the aspects of the problem. This model gives a basis for the implementation of the flight optimizer. Later, we propose a solution based on a sliding window simulated annealing algorithm which reduces the complexity and takes into account uncertainties. Finally, we successfully test an implementation of the solution with real-life traffic data. It corresponds to flights within France going towards Paris CDG airport over a period of 24 hours. The results demonstrate a total conflicts resolution with satisfying compliance with sequencing constraints.

Abstract

International audience; Air transportation traffic is progressively increasing over the years and dealing with it is an essential task to guarantee fluid flights in the future. Several works already indexed multiple aspects of aviation, among them, the E-MAN system. It introduced the sequencing of arriving traffic, starting from early stages of the En-route phase. This change facilitated the work for the approach controllers but also increased the workload of the En-route controllers. To handle that workload, controllers are now assisted by tools that consider the new constraints introduced by the arrival management system and propose advisories. From that same perspective, our project focuses on an algorithm for a helper tool that will combine both aspects of traffic sequencing in the En-route phase and conflict resolution. With this novel approach, we automatically generate near-to-optimal flight decisions, given that we can modify the speed and the flight level to respect the sequencing constraints and cut down potential conflicts. We categorize the problem as a mathematical optimization case. Thus, we describe a detailed mathematical model which covers all the aspects of the problem. This model gives a basis for the implementation of the flight optimizer. Later, we propose a solution based on a sliding window simulated annealing algorithm which reduces the complexity and takes into account uncertainties. Finally, we successfully test an implementation of the solution with real-life traffic data. It corresponds to flights within France going towards Paris CDG airport over a period of 24 hours. The results demonstrate a total conflicts resolution with satisfying compliance with sequencing constraints.

Abstract

Copyright © 2002 IFAC.The increase in air traffic, forces the current centralized air traffic control systems to be replaced with more decentralized control systems. The idea of free flight is developed with this consideration. In free flight aircrafts communicate and cooperate with each other to solve the conflicts. Each aircraft has its own separate optimization criteria, and aircrafts have to modify their paths considering these criteria. There should be a compromise in the amount of concession that aircrafts make in order to solve the conflict. In this research negotiation is introduced to achieve an agreement on the deviations from the optimal paths. Aircrafts negotiate on the amount of deviations they will perform to modify their paths. A frame for negotiation is developed, negotiation principles are determined, and simulations are performed.