Concept of Future Air Traffic Management and its Real-Time Decision Support utilizing Concept Learning

Abstract

Finding possible solutions for costly aircraft delays and congestion problem is the most important theme for the future air traffic control(ATC). Therefore, intelligent air traffic management (ATM) systems are highly required, which can manage air traffic flows and flight schedules strategically in real time fashion. For this objective, in this paper, an automated decision support system for the efficient ATM in one enroute sector or terminal area (TMA) is designed. This system uses concept learning scheme, a kind of machine learning techniques. The system has capabilities to find a suboptimal solution without interrupting real-time operations, in order to deal with various emergencies and to discover new better scheduling heuristics. Simulation studies show that the proposed scheduling architecture works rather efficiently than the current ATC procedure based on simple heuristic rules such as first-in first-out (FIFO) rule. An intelligent decision support system for ATM in a global airspace consisting of many airports and airroutes is also suggested toward future simulation studies.