An implementation of a reactive distributed air conflict resolution system

The fast growth of air traffic that began during the 80's raises some questions about what future air traffic control (ATC) systems need to be. A major change will probably be required in Europe within a 30 year time frame. Fortunately, recent developments in air navigation systems provide new opportunities for efficient autonomous navigation. Several studies have been conducted in order to develop automation in ATC, based either on centralized or distributed systems. An important part of the ATC task is to avoid separation losses (which are called "conflicts") between aircraft. The article presents an implementation of a reactive distributed conflict resolution method on a traffic simulator providing a realistic traffic sample. The method is described, and some results regarding its efficiency and its degradation with an increase in traffic are given. In particular, a critical traffic level appears, above which efficiency drops dramatically. This level depends on the method used; the airspace itself is not saturated.