K.
Bousson
Avionics and Control Laboratory,
Department of Aerospace Sciences
University of Beira Interior; 6200 Covilha, Portugal
kbousson@mail.telepac.pt
Automation of air traffic control
provides strategies to ensure flight safety in even dense traffic areas. It
requires decisions involving human lives on one hand and enormous costs on the
other. One of the central concerns of air traffic control is the problem of
collision avoidance between aircraft, also known as aircraft conflict
resolution. A conflict occurs when a relatively small airspace domain is
occupied by at least two aircraft, and is defined as the violation of any one
of a set of separation criteria specified in air traffic operation rules. Conflict
resolution consists in developing procedures specifying flight path alterations
for some of the aircraft involved in the conflict such that each aircraft can
keep a minimum safety distance from any other aircraft.
The complexity of automatic collision avoidance stems from the fact that collisions take place not only in space but also in time. Indeed, each aircraft must be separated by a given distance from other aircraft at any time; aircraft which are about to land must be clustered with the objective to make them follow some predefined flight trajectories; each category of aircraft is subject to some speed limit and flight level due to its structural performances and limits; and particular conditions regarding the flight level, maneuvers and speed alterations may be applied to some aircraft.
The airspace is constrained by the fact that
aircraft are forced to fly along predetermined airway. This is obviously not
optimal since it impedes aircraft to fly directly and freely to the destination
and take advantage for instance of flight minimality or even of favorable
winds. Due to the increasing number of aircraft in airspace, the solution for
safe and optimal airspace use would be the give the capability to aircraft to
evolve not only in but also out of predetermined airways in case of
free-flight. Therefore, most of the current activities in air traffic
automation focus on free flight. In the case of traffic in terminal areas,
although the aircraft may theoretically evolve freely in these areas, they do
not do so since their flights need to be guided so that they converge to the
runway for landing without colliding with each other. Effort has been made in
that respect, such as the Center-TRACON Automation System which serves as a
decision support tool for ground controller in an effort to reduce air traffic
control workload and optimize capacity close to highly congested airports.
Another solution for air traffic control would
be to give the capability to aircraft to avoid colliding with each other by
themselves instead of being dependent on ground controllers. In the sake of
achieving that possibility, airborne traffic
alert and collision avoidance systems (TCAS) have been devised and
implemented in most commercial and military aircraft to provide air traffic
controller independent conflict detection and resolution capabilities. The task
of the TCAS is to monitor air traffic in the neighborhood of the aircraft and
provide the pilot with information about vicinity aircraft with which it may
conflict and advisories on how to resolve these conflicts. Meanwhile, existing
versions of TCAS suffer from severe limitations due to the fact that firstly
they only suggest escape maneuver in the vertical plane for TCAS II (or in the
vertical and horizontal planes for the forthcoming TCAS IV) whilst human air
traffic controllers provide avoidance advisories based on heading, speed and
flight level, and secondly the collision avoidance strategies are related to
local (neighboring) traffic alone without taking into account for instance the
global traffic in the overall terminal area, whereas human air traffic
controllers resort to global and
predictive strategies. Finally TCAS is unable to vectorize the aircraft in a
terminal area towards the runway unlikely to human air traffic controllers.
Most of
the work in air traffic collision avoidance are suboptimal in that the
formulation of the problem is not expressed in term of an optimization problem.
An exception is made for instance to where conflict resolution is stated as a
trajectory optimization problem. However, in their work they assume the nominal
trajectory to be given for each aircraft, and their method is to control each
aircraft to follow its nominal trajectory minimizing deviations such that
conflict may be avoided. Their work is much dedicated to en-route traffic
control where nominal trajectories may be pre-specified, and as such the method
employed is rather strategically. As far as terminal air traffic is concerned,
it would be better to generate tactical optimal guidance trajectories according
to the flow of the traffic since aircraft may fly anywhere in terminal areas. The
present paper aims at proposing optimal guidance tactics for air traffic
automation in terminal areas where the air traffic density is usually high and
air traffic control most critical.
This work
copes with aircraft optimal guidance for automatic collision avoidance in
terminal areas. The collision avoidance problem is expressed as an optimization
problem whose solution vector is composed of individual aircraft headings,
velocities and flight levels as guidance information that corresponding
aircraft should follow to automatically avoid collision and at the same time to
converge to a specified landing procedure start point. Simulation results are
presented and show that the proposed method is capable of ensuring collision
avoidance in an optimal way.