CORDIS - EU research results
CORDIS

Evolutionary ATM. A modelling framework to assess the impact of ATM evolutions

Periodic Reporting for period 2 - EVOAtm (Evolutionary ATM. A modelling framework to assess the impact of ATM evolutions)

Reporting period: 2019-01-01 to 2019-12-31

WHAT IS THE PROBLEM/ISSUE BEING ADDRESSED?
The project proposes a fast-time simulation approach of ATM solutions and their optimization, integrating the human behaviour of controllers, to perform the following analysis:
- Change Impact Assessment on performance
- Support to Design
- Support to Decision-Making

THE THREE SCIENTIFIC PILLARS OF THE MODELLING FRAMEWORK ARE:
- Agent-based modelling
- Evolutionary computing
- Statistical analysis
The agent-based modelling and simulation enables a deep understanding of the emergent (i.e. unpredictable) aspects of the change solution in terms of any performance variation with respect to the baseline (reference scenario).
The Evolutionary-based Optimization enables the automated identification of the possible best combination of parametric characteristics of the change in order to achieve the required performance. This approach allows a concurrent and intelligent execution of several what-if analyses.
Statistical analysis enables the understanding of the most relevant parameters of the change affecting the performance.

WHAT ARE THE OVERALL OBJECTIVES?
The project main goal is to develop the proper methodological approach combining agent-based and evolutionary-computing techniques and to demonstrate the effectiveness of the developed modelling framework by means of a prototype. It uses scenarios whose outcomes in performance are known in order to demonstrate the effectiveness of the approach itself:
- Reference Scenario: Direct Routing according to SESAR1 Solution 32
- Changed scenario: Free Routing according to SESAR1 Solution 33

WHY IS IT IMPORTANT FOR SOCIETY?
EvoATM has a relevant twofold impact.
In ATM industry, the proposed approach would allow:
1) to anticipate the performance assessment at the earliest phases of the solution design lifecycle, reducing the costs of reworks after the real-time simulations;
2) to support the Design
In E-OCVM process, the proposed approach can represent a prescribed methodology to be used for V1 maturity gate and to support the trade-off among different initial new solution concepts. The adoption of the same methodology would allow a sound comparison of the prediction of change impact across all the relevant performance dimensions.
The main results achieved in the reported period are the followings:
ACHIEVED RESULTS
EvoATM prototype has been successfully validated against Italian real traffic data related to both DCT and FRA. It has predicted the performance variation when moving to Free Routing from Direct Routing, in accordance with those recorded in the referring years. Simulation results have been considered plausible by Controllers. The accordance between simulated results and “the real life” allows to state that:
1) EvoATM has demonstrated the capabilities of the proposed approach. By using a known change history, it has shown how the actual performance (including human related performance) could be predicted, since the beginning of the change solution lifecycle, by using the proposed modelling and simulation approach. Hence EvoATM has proposed a methodological framework highly effective in change impact and support to design and decision making requiring less effort of an entire life-cycle to achieve more or less the same results
2) EvoATM has proposed a formalism to facilitate the dialogue between the ATM experts and the modellers and to easily model the human behaviour, that can be re-used
3) EvoATM has proposed a software architecture assuring a modularity and re-usability assuring the saving of costs in model development

In detail in the following some key demonstration cases and results shown a the Project Final Workshop (7-8- November in Naples) are reported:
Demonstration Case: Human performance assessment of the simulation of the High Complexity and Very High Complexity DRA scenario
Analysis: The achieved results clearly report the greater workload of controllers for the VHC scenario with respect to the HC scenario in the simulated environments. The order of magnitude of the metric assessment has been evaluated as likely and plausible by both ENAIRE experts and ENAV experts.

Demonstration Case: Change impact assessment of the FRA solution (with respect to DRA) in a full-day traffic within the Italian airspace. Both the High Complexity scenario and the Very High Complexity scenario are considered for the simulation. The planned traffic is the real full-day planned FRA traffic on 03/07/2019. The reference scenario is the real full-day planned DRA traffic on 07/07/2016
Analysis: The achieved results by means of EvoATM demonstration campaign confirm the trends that were assessed by ENAV in their own evaluation of free routing solution in Italy, as they stated during EvoATM workshop
Demonstration Case: To Analyse the impact of seniority attributes of controllers
Analysis: The results show that the seniority attribute of controllers has a significant impact on their workload . This especially applies for the VHC scenario

Demonstration Case: Tuning of the configuration of elementary sectors in the collapsed sectors to optimize controller workload and minimize the standard deviation of workload measurements
Analysis: The obtained solution has suggested to aggregate specific sectors of the ACCs according to their peak of traffic, producing nearly uniform workload distributions for controllers and with a limited total workload

For dissemination a final workshop on 7th and 8th of November has been organized, where the Advisory Board and relevant stakeholders, such as Industry, ANSPs, and Universities met to discuss the topic Socio Complex Systems Modelling and the Project results. The involvement of speakers from Marine and Transport sectors have enlarged the workshop scope beyond the ATM, investigating the opportunities to exploit the project results also in other sectors.
In addition two events for project dissemination have been organized in order to check the acceptability of the achieved results:
-The presentation of results to ENAIRE controllers
-The presentation of results to ENAV controllers, fast time simulation team, and head of research
Most of the commercial tools provide detailed models of airports and airspace for fast-time gate-to-gate simulation; very few use multi-agent architectures for different actors of the scene, e.g. for airport controllers. Currently No tool seems to embed the human behaviour modelling.
In the recent years some attempts are presented within SESAR Program. Many exploratory research projects (e.g. ELSA, ACCESS, TREE, CASSIOPEIA, MAREA, SPAD, EMERGIA) have adopted the agent-based modelling paradigm to address a wide number of different ATM problems. In these projects, the agent-based simulation is used (usually in combination with other techniques) either as an analysis tool to understand emergent behaviours of the ATM system or to study resilience and disturbance propagation, or even as a tool to determine airport slot auctioning and allocation. No project defines a methodological approach for modelling and simulation to assess the change impact.
In Human Modelling EvoATM has proposed a unique modelling framework of human merging different theories and enabling an easy specification also of the interactions of humans with the technical elements.
EvoATM adopts Evolutionary computing for support to design, no proposal has tried to consider ABM simulation-based optimization for the purpose.
Comparison of Occupancy between simulated and real traffic data in DCT scenario
Change Impact Assessment Workflow
Support to Design Workslow
Project Objectives
Comparison of Timeliness between simulated and real traffic data in DCT scenario
Assessment of Human Workload in terms Executive Controllers communications
Comparison between simulated and real flown trajectory
Formalism Overview