In summary, it is expected that AGENT operational frame produces and supports:
- Highly automated ATM environment with minimal human intervention. The project goes a step ahead in the ATM automation by introducing the centred multi-agent system approach against the agent-centred approach, i.e. shifting the decision making process (negotiation & resolution) from the ATC system to the ecosystemâ€™s agents. The new approach is based on the agentsâ€™ capabilities and constraints, but also delegation of the roles, business models and interaction protocols.
- Friendly human interaction within highly automated system. As a safety-critical system, the ATM system will probably never neglect the human factor performance. Thus, ecosystem strives to facilitation of both the airborne and ground-based human interactions throughout development of the novel operational procedures. Based on complexity of the traffic scenarios the procedures will express the facilitation level.
- State space search for an optimal compromised solution. There will be two comparable approaches in the AGENT state space search strategy. Both will comprise ETM and ERM, but complement each other in EDM for the memberships identification. This two-conceptual analysis will further bring a high benefit to the ATM community. Furthermore, ETM- and ERM-based tools with their metrics will fully deploy the reachability tree of feasible solutions and, with Pareto-optimality criterion, generate the best possible set of resolutions. KPIs will indicate validity of the results obtained.
The AGENT impact will be kept viable in the following aspects:
- Demonstration and quantification of the potential for higher levels of automation as well as provision of the benefits in safety, capacity and efficiency of ATM operations. Open Demonstrator will indicate a rate of change/transition time between the safety layers based on a relevant scenario complexity, airspace system requirements and input data from the RBT dynamics. From the generated ETM and ERM inside the ecosystem volume, it will be possible to quantify benefits and costs (cost-benefit analysis) of all resolutions. Qualitative and quantitative assessment will provide the indicators on operational safety, capacity and efficiency. It will facilitate and support better decision making.
- Expression of the potential for improved cost-effectiveness and contributions to decision support techniques for safety-critical situations. To evaluate an effectiveness of decision making tools the following KPIs will be used:
1 – Number of the conflict-free resolutions generated for each ecosystem member;
2 – Speed of decision made (either to keep the original intent or to change it);
3 – Number of structured conflicts per member;
4 – Total number of structured conflicts per ecosystem;
5 – Number of identified induced conflicts per ecosystem;
6 – Percentage of resolved conflicts per ecosystem member;
7 – Total percentage of resolved conflicts per ecosystem;
8 – Percentage of avoided conflicts per ecosystem member;
9 – Total percentage of avoided conflicts per ecosystem;
10 – Number of negotiations per ecosystem member;
11 – Average negotiation time per ecosystem member;
12 – Total negotiation time per ecosystem;
13 – Decision regret factor â€“ relative measure/ratio between the performance KPI of selected manoeuvre and the most optimal trajectory at the current state, per member (alternatively, it could be defined as relative measure/ratio between the most optimal KPIs at the beginning and the end of one negotiation interval, per member).
An extent, to which the expected impact of results will be achieved, is foreseen two-fold:
- Cooperation between the consortium members: The quality of cooperation between SME and academic partners in the scope of developed AGENT tools will certainly produce an impact on the overall project scope, especially if considering different orientations of business and scientific communities. The promotional activities dedicated to the commercial use of Open Demonstrator will be highly desired. For this reason, regular meetings will be regularly organized and detailed reports will be written in order to ensure continuous improvements of project results.
- Collaborative networking: the successful implementation of the tools developed will be shared in a form of the collaborative networking between the AGENT consortium and other organizations (SESAR JU, EUROCONTROL, HALA! Research Network or newly established networks starting from 2017). To promote a unique simulation platform, the consortium will organize dissemination events and participation in workshops and conferences. The real efforts will be placed in engagement a scientific community for the scenarios testing and validation. The community will be provided with a full access to Open Demonstrator.