ASTAIR defined, developed and validated a TRL1 concept for automated airport surface movement management, covering engine-off and conventional taxiing at major European airports. Its aim was to assess how AI-based planning, gradual automation and human-machine collaboration could improve predictability, safety, capacity, environmental performance and operator workload.
The project first consolidated operational knowledge through workshops and interviews with stakeholders from Paris-CDG, Frankfurt and Amsterdam Schiphol. This identified current apron and ground movement procedures, constraints, roles, data sources and automation use cases. The result was the ASTAIR Concept Outline, defining “Autonomous Taxi Management” and proposing SESAR Solution 0501. Remaining work toward TRL2 was also identified, especially on safety, economics, integration and further operational validation.
ASTAIR then developed support algorithms based on AEON path-planning methods, extended with airport traffic rules, runway configurations, vehicle constraints, operational conditions, controller preferences and spatiotemporal constraints. These were embedded in a multi-agent planning framework for aircraft, tugs and other ground vehicles, producing adaptive methods for conflict-free routes and speed profiles compatible with human supervision.
Human-automation teaming was addressed through HMI prototypes enabling controllers to supervise, inspect, adjust and override automated plans. The tools supported gradual automation, task sharing, future trajectory inspection and non-nominal event management. Reviews with ATCOs and validation activities confirmed the usefulness of this approach while highlighting further needs on explainability, alert prioritisation, role clarity and integration of future-traffic views into controller workstations.
The demonstrator integrated HMI and algorithmic components into simulation environments for Paris-CDG and Amsterdam Schiphol, using airport maps, stands, traffic rules and automated movement generation. A Wizard-of-Oz mode supported real-time validation where full algorithm integration was limited by computing constraints. The resulting platform enabled representative airport scenarios for human-in-the-loop and fast-time simulations.
Validation included stakeholder workshops, prototype demonstrations and a final campaign combining real-time simulation, fast-time simulation and expert review. Data came from questionnaires, observations, debriefings, logs, historical traffic and expert feedback. Results showed the concept was technically and operationally feasible at TRL1, with effective human-AI collaboration, conflict-free routing, improved predictability and promising efficiency benefits. ATCOs found the tools useful, especially for routine tasks and trajectory inspection, though cognitive load, interface integration and AI rationale clarity need further work.
Key limitations remain: pilot- and driver-in-the-loop validation, full HMI-algorithm integration, degraded automation modes, handover after failures, safety nets, cockpit and vehicle datalink, task allocation, AI transparency, regulation and liability. Overall, ASTAIR delivered a validated TRL1 concept, SESAR solution, routing and tug-management algorithms, HMI prototypes, an integrated validation platform, and evidence on feasibility, benefits and future research priorities.
