Catching up with ARTIMATION: Furthering research on transparent AI for enhanced air traffic management
The ARTIMATION(opens in new window) project was launched in January 2021 to provide a transparent and explainable AI model that can ensure safe and dependable decision support in air traffic management. Although the project ended 2 years later, its research team has since continued to analyse the experimental results, making improvements along the way.
Better, clearer, more transparent
ARTIMATION researchers have made several improvements to their AI algorithms that will increase interpretability. For example, an article(opens in new window) published in February 2024 presents enhancements in the interpretability of the Extreme Gradient Boosting, or XGBoost, machine learning model. The team has also made progress with the project’s two use cases: flight take-off time delay prediction (FToTDP) and conflict detection and resolution (CDR). A paper(opens in new window) published on FToTDP in June 2023 explores the use of explainable AI in clarifying flight take-off time delays predicted by machine learning models to air traffic controllers. Another paper(opens in new window) on CDR, published in January 2024, examines decision-making in air traffic control and explores the impact of offering additional explanations alongside a conflict resolution algorithm to improve decision-making. EU support played an important part in boosting ARTIMATION’s research, development and innovation. Within a 2-year period, the project team was able to explore novel AI techniques and foster collaborations with leading institutions within the consortium, boosting the project’s capacity to develop trustworthy AI systems with greater explainability. “This financial backing was crucial in helping ARTIMATION achieve its goals, scale its operations and position itself as a competitive player in the AI research landscape,” comments Mobyen Uddin Ahmed, professor at Sweden’s Mälardalen University that coordinated the project. The work produced during ARTIMATION (TRANSPARENT ARTIFICIAL INTELLIGENCE AND AUTOMATION TO AIR TRAFFIC MANAGEMENT SYSTEMS) has led to many innovative ideas for solving air traffic management problems using AI. This has earned the team further funding from Sweden’s innovation agency Vinnova. It has also led to more EU funding through a project called TRUSTY, which is using trustworthy and transparent AI to increase the efficiency, data capacity and overall durability of remote digital towers.
Keywords
ARTIMATION, air traffic management, control tower, AI, explainable AI, transparent AI