Date of final exam: 19/01/2002E-mail: l.pallottino@ing.unipi.it
Tutor: Prof. A. Bicchi, Università di Pisa
___________________________________________________________________________________________________________
___________________________________________________________________________________________________________
AIRCRAFT CONFLICT RESOLUTION IN “FREE FLIGHT” AIR TRAFFIC
MANAGEMENT SYSTEMS: MODELS AND OPTIMAL SOLUTIONSAdvisor:
Prof. A. Bicchi, Università di Pisa
Summary of the thesis
This thesis deals with conflict resolution in Free Flight air traffic management systems. Unfortunately, recent tragedies, due to terroristic attacks in New York and Washington of September 11th 2001, have temporary changed the air traffic situation in the entire world. Actually, air traffic demand is decreasing drastically and air companies have high risks of failure. Hopefully, situation will go back to normal in few years but still many things have to be changed to increase safety of flight.
Current air traffic management system has technological and management limits. Those limits, with the forecasted increasing of air traffic annual demand of about 3-5% in US and 5-6% in Europe, would compromise efficiency and safety of the system in few years. Current control strategies are based on centralized non-automated decisions: control centers manage landing and take-off, maintain safety distances between aircraft and assign flight trajectories. The increase of air traffic will bring to a workload and a big responsibility for controllers and then will bring to an increasing risk of errors, a difficult management of control in case of delays and a decreasing of safety. Furthermore, in current control systems aircraft are allowed to fly only along fixed airways and as a consequence, available airspace is drastically restricted.
Other kinds of control strategies, opposite with respect to the centralized control, are under development. One of this new strategies is Free Flight. In a Free Flight strategy, pilots are free to choose their own trajectories as autonomous agents, based on tasks like minimizing time of flight, minimizing fuel consumption, avoid overloaded airspace or avoid sever weather. Choice of trajectories is subject to a non-conflict constraint when more aircraft fly in the same airspace. In other words aircraft have to maintain a minimum distance between each other in order to avoid conflicts. This is the case of coordination control (that can be collaborative or non-collaborative) between aircraft in the same airspace. From this point of view control of air traffic move from centralized centers to aircraft. Indeed, in this case, aircraft are responsible to solve conflict in their surrounding airspace. Hence, on-ground control centers are just responsible for aircraft during landing and take off.
Decentralized control is the fundamental aspect in Free Flight and seems to be the normal evolution of current air traffic control systems towards Free Flight. On-ground control centers are innovating instruments through a gradual automatization. In following years, pilots will gain freedom in planning trajectories during flight, this will gradually bring to Free Flight. Major task in Free Flight is to assure safety and to give pilots several decision support tools to detect and solve possible conflicts.
Within this thesis, en-route phase of flight in case of centralized and decentralized control is considered. We focus on aircraft conflict resolution problem in case of Free Flight. Collaborative conflict resolution will be considered. Each aircraft has its own task as minimizing time of flight, in case of collaborative environment this task is secondary with respect to the common goal of safety: a minimum distance has to be maintained between aircraft. In a collaborative decentralized framework, each vehicle shares information on position and direction of its flight with the other surrounding aircraft and they collaborate (as elements of the same team) in order to obtain collision free trajectories.
Different kinds of models have been obtained based on different types of problems (centralized or decentralized control) and of maneuvers which are allowed to avoid conflicts. Three different types of conflict avoidance maneuvers has been considered: maneuvers with bounded steering radius, maneuvers with instantaneous change of flight direction and maneuvers with instantaneous speed change. Models obtained for previous cases are nonlinear optimal control models (bounded steering radius case) of mixed integer linear programming models (instantaneous deviation of motion or of linear velocity) and hybrid systems model (decentralized control).
The main results of this thesis can be summarized by:
· Bounded steering radius maneuvers: necessary conditions for optimality using Pontryagin Minimum Principle and optimal control theory have been obtained; both constrained and unconstrained collision-free trajectories have been characterized; a conflict resolution algorithm has been developed based on obtained theoretical results.
· Instantaneous heading angle change maneuvers and linear velocity change maneuvers: Mixed Integer Programming models have been obtained through geometrical construction of conflict avoidance constraints; optimal conflict-free trajectory for both types of maneuvers have been obtained: minimum deviation from original path for heading angle deviation maneuvers and maximum velocity for instantaneous speed change maneuvers; due to the linear formulation of the two problems, solutions may be obtained within c.p.u. seconds with standard optimization software, hence those approaches may be used as part of a real or fast-time simulation.
· The decentralized control strategy: the conflict resolution problem in a decentralized control strategy has been considered and formulated as a hybrid system; safety verification of the model has been studied for the instantaneous heading angle deviation maneuvers; for the case of 3 aircraft, safety of the system has been demonstrated.
_______________________________________