MAGE: Teaching AI to Manage Airspace Complexity, Part 2

This is Year 2 of developing a tool for managing future airspace complexity. The premise is that at each time step, an algorithm decides an aircraft’s next move toward destination. The decision is a solution of a computational problem that considers properties of the environment and aircraft in the vicinity, under uncertainty. In sparse airspace, the solution is fast and optimal. With more neighbors, computing a path takes longer and solutions worsen, all the way to potentially reaching Brownian motion. MAGE (Monitor, Anticipate, Guide, Evolve) uses Machine Learning (ML) models to monitor solutions, predict deterioration, and guide dynamic airspace reconfiguration to maintain efficient operations. In Year 1, we showed that ML can predict approaches to transition from fast and optimal solutions to a loss of tractability and solution quality on a time budget, on simplified problems. In Year 2, the method is scaled up and demonstrated on a more realistic problem. Although the development is grounded in airspace management, the underlying approach applies to all domains where computational decision-making takes place.

In the future airspace operations, it is critical to detect approaches to transitions from efficient to inefficient or even unsafe state, so that complexity of a portion of airspace can be reduced temporarily to restore safe and efficient operations. A tool that enables such analysis and prediction is not only essential for future dense and diverse operations but can be useful in more immediate future of denser operations, still under human control. The methodology that underlies the proposed tool is applicable to any domain where mutually interacting participants make decisions. As the environment becomes more complex, decision-making becomes harder, slows down, and eventually comes to a halt, unless mitigated. A tool that detects an approach to such states before they happen supports any multi-agent decision-making environment, whether in airspace or space exploration applications. Thus, all NASA Mission Directorates that will rely on autonomous multi-agent operations (ARMD, STMD, SMD), as well as commercial and OGA multi-agent system operators will benefit from MAGE, if the approach is successful.

This is early/mid-stage (TRL 4) — the most realistic path in is NASA SBIR/STTR, which funds small businesses and research institutions to develop technology aligned with NASA's needs (equity-free, phased funding). Check whether a current SBIR/STTR solicitation topic overlaps with this project's technology area, or contact the project directly (above) to ask.

None of these are guaranteed paths for this specific project — TechPort itself doesn't have an "apply" button. Reaching out to the contact(s) above with a specific question is usually the fastest way to find out what's actually open.