Design and analysis of aircraft conflict resolution algorithms using positive semidefinite programming

Eric Feron, Department of Aeronautics and Astronautics, MIT

The Federal Aviation Administration is promoting Free Flight, a concept that would transfer some authority from Air Traffic Control to the airlines and other airspace users during airborne operations. Under this concept, the current description of the airspace as a network of routes whose nodes are navigation beacons would be replaced by the unstructured three-dimensional space with the Global Positioning System as its principal navigation system. In this talk, we investigate methods to assess the impact of unstructured aircraft routing on the requirement to provide satisfactory separation assurance between aircraft. First, we present an aircraft conflict resolution algorithm based on real-time convex optimization. This algorithm allows us to "bypass" the inherent complexity of the problem. Second, we concentrate on the more important issue of "proving" the stability of the resulting closed-loop conflict avoidance control system. When conflicts involve many aircraft simultaneously, this task appears to be challenging but remains necessary. Initial results based on convex programming are presented.