Machine Intelligence in Air Traffic Management (L'Intelligence Artificielle dans la Gestion du Traffic Aerien).

摘要

Airspace congestion is a reality and, whatever the causes, the resulting problems are known only too well. And yet, demand continues to grow. Forecasting experts agree that traffic densities will double in 10 to 15 year's time. Others envisage, in the slightly longer term, the bringing into service of supersonic aircraft cruising at Mach 3 and even hypersonic speeds. In today's chain of control it is the human link, i.e. the controller, who is sovereign. He is in a sense an artist dealing with situations on a case-by-case basis, drawing on his talent, intuition, experience and judgement. These are all inputs which are difficult for him to get across to us and which are hard for us to put in a form which can be expressed mathematically. The arrival of state-of-the-art technologies has opened up new visions revealing ways of rolling back the present day limitations of the ATC system, namely the restricted runway or en route capacity and the rapid saturation of the human controller's ability to handle increasing situation complexity. This Symposium, namely the 56th Symposium organized by the Guidance and Control Panel of AGARD, intended to place the emphasis on the potential use of Machine Intelligence in the overall control loop covering all sub-loops-management, flow management, control, guidance, conflict alert and collision avoidance-and assessing the benefits which should or might result for the aviation community-designers and users.