Community Research and Development Information Service - CORDIS

H2020

COPTRA Report Summary

Project ID: 699274
Funded under: H2020-EU.3.4.7.1

Periodic Reporting for period 1 - COPTRA (COmbining Probable TRAjectories)

Reporting period: 2016-03-01 to 2016-08-31

Summary of the context and overall objectives of the project

The prediction of future air traffic situations is central to ATC planning. Its uses range from Air Traffic Controller workload management to ensure capacity, to helping the setup flow management measures in case the demand cannot be accommodated.

Different methods are used to generate forecasts for different time horizons to support relevant decisions; the basic question always being ‘how can the demand be met?’ Long-term answers might be ‘build a new control centre’, ‘train some new controllers’ and so on. The medium-term question might be answered by managing the controller leave roster and planning the ‘sector opening scheme’. The time-frame of main concern of this work is the ATC planning horizon (about 90 mins), with a mix of activated airborne flights (RBT available) and flights still in planning (SBT available). It is however intended that the scheme developed in this work will be generally applicable to all time-frames.

Trajectory Based Operations brings together many different improvements that allow the uncertainty of trajectory prediction to be better managed and reduced. These include downlinking of the Extended Projected Profile from the flight to enrich ground trajectory predictions, sharing of detailed information on the ground through SWIM (IOP), the submission of more detailed flight plan information, the use of 4D contracts during the flight, increasing adoption of Airport CDM, and so on.

For a given location, a prediction of the traffic that will cross it at some moment in the future is a mix of trajectories of flights which are airborne, flights that have been filed but not yet taken off and flight plans that have not yet been filed. Each type of flight has a different level of uncertainty or level of inaccuracy with which the prediction can be made. Current systems and operational processes mostly rely on human judgement and experience to deal with this mix of uncertainty. The result is often that a very imprecise balance is struck between demand and capacity leading to capacity going unused or significant last minute adjustments needing to be made.

In this context, the main concepts defined, modelled and studied by COPTRA are the notions of probabilistic trajectories and traffic situations. The central idea researched is to develop new methods to build the probabilistic traffic prediction by combining the probabilistic trajectories.

Building on the considerable inter-disciplinary expertise in trajectory prediction, applied mathematics and ATC planning accumulated by the project partners, COPTRA proposes an operational concept where the uncertainty of the predicted trajectories (hopefully reduced in TBO) is made explicit at trajectory prediction level and combined using state of the art applied mathematics methods to build a probabilistic traffic situation (i.e., traffic situation where the uncertainty is identified and specifically accounted for). These probabilistic traffic situations will be used to improve the prediction of occupancy counts used in ATC Planning and convey better information to the human operator.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

1. Building probabilistic trajectories (WP02)
Main activities executed throughout the referred period have basically been related to the identification of Uncertainty Quantification (UQ) techniques and how they can be applied to the aircraft trajectory prediction problem. Specifically, trajectory uncertainty quantification involves providing probabilistic definitions (e.g. mean, variance, standard deviation, etc.) of the variability of the state variables that define a prediction. Throughout the referred period, WP02 has constructed a clear understanding of main sources of uncertainty that impact the trajectory predictions uncertainty and their qualitative influences based on the knowledge of available datasets. An initial assessment on how to characterize such inputs variability based on historical data available through DDR2 (Demand Data Repository) has been performed as well. Details of this work and the contributions of each partner are included in the Part B of the 1st Periodic Technical Report.

2. Combining Probabilistic Trajectories (WP03)
WP03 mainly focuses on combining probabilistic trajectories to identify their impacts on traffic flows. As the main idea behind the WP03 is to improve Demand Capacity Balancing (DCB), the traffic flow information provides DCB with probabilistic airspace entry and occupancy counts and their propagation over time. The initial effort in WP03 throughout the referred period was to provide a clear understanding of the inputs and outputs of the WP. Considering the goals of the WP03, initial assessment has been performed and it is decided that two data-driven network models will be developed throughout this WP:
• The Project has developed a complete framework for developing Operations Research tools for enhancing Air Traffic across the European Sky. This takes the form of a Weighted Graph and a mathematical model.
• Queue Network Model enabling to model and simulate the air traffic network through the stochastic inputs, where reflects probabilistic definitions of trajectory, in other words, probabilistic demands to the air sectors.
The details of the initial contributions of the partners for data-driven techniques to be applied to construct these two models are given in the Part B of 1st Periodic Technical Report.

3. Application of Probabilistic Traffic Prediction to ATC Planning (WP04)
WP04 elaborates a set of benefit mechanisms and associated use cases and scenarios that describe how the approach proposed by COPTRA can be exploited operationally. This WP also deals with the adaptation of one of the existing DCB prototype validation tools (NetPerf or TESEO) to measure the improvements.
The initial work performed in WP04 was focused on the definition of the conceptual interfaces between the models for probabilistic trajectories and traffic situations that are developed in work packages 2 and 3, as well as the definition of the benefit mechanisms that support the application of these models to Dynamic Capacity Balancing (DCB).

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

1. Building probabilistic trajectories
Arbitrary Polynomial Chaos Expansion is used to quantify aircraft trajectory prediction uncertainty. It requires individual datasets describing the inputs variability to build the associated one-dimensional expansions and the multi-dimensional expansions describing the outputs variability. This is a data-driven process that exploits historical recorded data to define the stochastic impact of identified uncertainty sources. Take-off mass uncertainty is modelled through model-driven mass estimation techniques based on mass error reduction for certain phases of the flight.

2. Combining probabilistic trajectories in probabilistic traffic forecasts
Queue Network Model (QNM) enables the identification of stochastic parametrisation of the ATM network as to be able to predict the behaviour of the network under disturbances. The QNM studies the required granularity on network model, the Traffic flow propagation, the Topological sector transition queues and air traffic/sector complexity. A complete framework for developing Operations Research tools for enhancing Air Traffic across the European Sky. This takes the form of a Weighted Graph and a mathematical model.

3. Application of probabilistic traffic prediction to ATC planning
Two benefit mechanisms have been identified. These mechanisms are supported by operational use cases. The graph models and analysis proposed offer a better vision and understanding of uncertainty at the network level. The identification of critical flights will allow for a better choice when rerouting planes.
Record Number: 195257 / Last updated on: 2017-02-23