Vous êtes sur la page 1sur 4

2013 Third International Conference on Intelligent System Design and Engineering Applications

Rules-based Study of Conflicts Detection and Resolution in Ramps

Gao Wei, Zhang Jia, Wang Taobo


College of Air Traffic Management, Civil Aviation University of China, Tianjin, 300300, China
jzhang_yjs10@cauc.edu.cn

Abstract—The fast growth of flights has made the gate establish a module structure of system, and complete the
resources more scare, thus making taxiing in ramps more realization and effect validation of the model.
complicated, even leading serious flights delays sometimes. In
order to reduce the conflicts in apron area, the optimization of II. ANALYSIS OF TAXIING IN RAMPS
ground operations becomes necessary. Based on the detailed
analysis of the operation rules, an useful knowledge base (KB)
Network in ramps is a path set of aircrafts taxiing. In this
is summarized and the KB knowledge base (KB) model is
proposed. Then taking the XiaMen airlines aprons and the
paper, we define an aircraft is a particle-centered protection
actual flights schedules at XiaMen GaoQi International G = ( N , E ) is physical ground network and
zone, as well as
Airport as an simulation example and realizing the KB model N = {ni i = 1,", k}
on a Java-Mysql platform, the comparison of simulation is a node set of G. N consists of three
results and the actual operations data indicates that the C = {ci i = 1,", l}
modeling method for the ground operations is feasible and subsets, or apron starting point , general
efficient. So the simulation results can provide the immediate A = {ai i = 1,", t} B = {bi i = 1,", g}
operation method and help relieve the workload of apron node ,gate .
controllers. Meanwhile, it can be a relevance to other forecasts E = {ei = (n p , nq ) i = 1," m; n p , nq ∈ N } is ei , or a
of ground operation or gates assignment.
directed line segment from n p to nq . Apart from creating
Keywords-Gate Assignment; Conflicts detection and resolute; topological relations between nodes and line segments, we
Ground operation; Knowledge Base should store relationship between the adjacent point,
including angle and point attribute, into database, to
I. INTRODUCTION accurately describe the state of aircraft.
Due to civil aviation in China keeps fast growing, in An arrival flight will taxi from the main taxiway to the
order to meet the increasing requirements for gates, apron area, then enter the scheduled gate normally or hold
nowadays, domestic airports always ease the pressure by at the end of main taxiway when conflicts occur. Similarly,
expanding the aprons capacity and assigning gates more a departure flight will push and taxi to main taxiway, or will
reasonably. However, it often makes taxiing in apron very hold in the gate under conflicts.
complex and leads some ground delays, even follows by On the summary of all analysis above, conflicts in
large scale flights delays. Thus, we should pay more apron area are classified as follows.
attention to taxiing and taxiing conflicts in apron area. To (1) Conflicts between adjacent pushes.
achieve conflicts detection and resolution, many researchers (2) Conflicts between push of departure flight and
have attempted various methods[1-2]. For example, Hu taxiing of arrival flight.
Minghua studies the optimization of steady flights (3) Conflicts between taxiing, including butt and cross.
schedule[3]. But this model lacks adaptability of dynamic In this paper, we resolve conflicts by reasonable
operation. Stephen Atkins builds a model on dynamic holdings. An arrival flight may hold in apron exit or apron
schedules [4]. But it does not explore taxiing in apron. taxiway, and a departure flight will hold after its push. The
Additionally, most models are put forward, by establishing a typical state of flights will be illuminated later on.
simple dot-line structure of ground network, based on Graph According to feedback of real operation, with similar
Theory and Mixed Integer Programming. Excluding taxiing aircrafts, we set the priority of state, or entering gate>push-
rules in real airport operations, so results on the basis of back>arrival>ready-to-taxiing>ready-to-push. Actually the
network is unreliable. conflict detection and resolution is related to pushing time,
Knowledge base is a feature set of fact, rule and concept, heading, gate, etc. Here we will analyze these factors and
or an organization storing and managing knowledge by establish rules base to detect and resolve conflicts.
describing variable methods[5]. Its knowledge differs from
data or programs, and it is operation rules of the intelligent III. DESIGN OF KNOWLEDGE BASE
system and practical experience of experts. We first analyze
Knowledge Base includes Fact Base and Rule Base.
typical types of taxiing and taxiing rules in apron, and build
Normally polyadic predicates are presentation for the base.
data base of ground network. Next we describe rules
To meet practical needs, the fact base will consist of
reasoning process by Knowledge base models. Then we
basic flight information, such as flight number (FLT No.),

978-0-7695-4923-1/12 $26.00 © 2012 IEEE 1238


DOI 10.1109/ISDEA.2012.292
the departure or arrival state, the scheduled gate, etc. X1 and aircraft X2 emerges, F8 will give the corresponding
Furthermore, to be approach to the real situation, we put action modes to aircrafts.
simple reasoning into the base design, for instance, whether F9: State(Z1,Z2). Z1 and Z2 represent states of two
two gates are adjacent. Thus we declare the following aircrafts at the same moment.
composite fields. On the behalf of on-scene commanders, the main
AParam= AParam(X,Y,Z,A,B,C,T,L,P) problems are listed.
Predicate AParam represents featured information of 1) Whether two departure aircrafts will develop a push
every flight, which consists of nine parameters. In real conflict.
operations, some parameters will need no assignment. Tab. 2) When one of two departure aircrafts is pushing or
1 shows definitions of parameters in AParam. has completed pushing, whether the other can push or not.
3) When one of two departure aircrafts has started to
taxi, whether the other can push or not.
Tab1. Definition of AParam formula
4) One of two departure aircrafts has started to taxi,
Parameter Style Implication
X string Flight number (FLT No.). and the other is pushing.
Y char Identification (A/D). 5) One arrival aircraft has entered the apron, and the
Z enum 0 The flight is ready to push in its other departure aircraft does not push.
gate. 6) One arrival aircraft has entered the apron, and the
1 The flight is pushing. other departure aircraft is pushing or has completed the
2 The flight has completed pushing process.
and is ready to taxi.
3 The flight is taxiing. 7) One arrival aircraft is ready to enter the apron, and
4 The flight is holding in apron the other departure aircraft has started taxiing.
taxiway. 8) Two arrival aircrafts enter the ramp from different
5 The flight is entering the gate. apron spots.
A Node A set consists of position and name of Next we describe the eight processes respectively.
nodes of ground network.
B Gate A set consists of position and name of R1: State(0, 0) ∧ SpotIsSame( X 1, X 2) ∧ IsNear ( X 1,
nodes of gate network. X 2) ∧ DistToSpot ( X 1, C1) ∧ DistToSpot ( X 2, C 2)
C Spot A set consists of position and name of ∧Time( X 1, X 2, A1, A2) → IsConflict ( X 1, X 2) →
nodes of apron exits network. ( Maneuver ( X 1, X 2, 0,1) ∨ Maneuver ( X 1, X 2,1, 0))
T int[ ] Departure and arrival time of flight. We will compare pushing time and distance of the two
L int Minimum separation between aircrafts departure aircrafts, and judge whether conflicts will occur.
determined by its types. The strategy to resolve conflicts is to direct that one aircraft
P Node[ ] Taxiing instruction.
holds and the other starts to push. The specific calculation
of conflicts is conducted as the equations below.
Node is an object-oriented type, which is inherited by NN
Gate and Spot. P is an array of Node. toutX 1 + Δt pb + 1 2 = toutX 2 + Δt pb (1)
Vt
Then we use the predicates above to define fact rules
as follows. N1 N 2
F1: Maneuver(X1,X2,z1,z2). At moment T, aircraft X1 toutX 1 − toutX 2 − <ε (2)
maneuvers in z1, and aircraft X2 maneuvers in z2. In the Vt
system, conflict resolutions of aircraft mainly are toutX 1 is the push moment. Δt pb is the required time of
represented by F1.
F2: DistToGate(X,B). At moment T, it represents push. N1 N 2 is the distance of two aircraft nodes. Vt is the
distance from aircraft X to gate B. taxiing speed in ramps. ε is the minimum separation
F3: DistToSpot(X,C). At moment T, it represents requirement. We use equation (1) to calculate the ultimate
distance from aircraft X to apron exit C. conflict and its result is used as the judgment of equation (2).
F4: IsNear(X1,X2). When gate X1 and gate X2 is Similarly, we can describe other calculation mechanism.
adjacent, F4 is true, conversely it is false. R 2 : ( State(0,1) ∨ State(0, 2)) ∧ SpotIsSame( X 1, X 2)
F5: GateDist (X1,X2). It represents distance from the ∧ IsNear ( X 1, X 2) ∧ DistToSpot ( X 1, C1) ∧ DistToSpot
gate occupied by aircraft X1 to the gate occupied by aircraft ( X 2, C 2) ∧ Time( X 1, X 2, A1, A2) → IsConflict ( X 1, X 2)
X2. → ( Maneuver ( X 1, X 2, 0, 2) ∨ Maneuver ( X 1, X 2, 0, 2)
F6: Time(X1,X2,A1,A2). It represents occupied time The difference between R2 and R1 is that the priority
aircraft X1 in Node A1 plus occupied time aircraft X2 in of R2 is higher than R1 and in R2 some aircrafts are pushing.
Node A2. R3 : State(0, 3) ∧ SpotIsSame( X 1, X 2) ∧ GateDist
F7: SpotIsSame(X1,X2). When aircraft X1 and aircraft ( X 1, X 2) ∧ DistToSpot ( X 1, C1) ∧ DistToSpot ( X 2
X2 come in and go out from the same node, F7 is true, , C 2) ∧ Time ( X 1, X 2, A1, A2) → IsConflict ( X 1, X 2)
conversely it is false. To be universal, in the design, we use → Maneuver ( X 1, X 2, 0,3)
fact to judge the coherence of pushing headings, which are R3 represents one aircraft is taxiing in apron taxiway
restricted by runway, terrain, etc. and the other aircraft is ready to push. By the position of
F8: IsConflict(X1,X2). When conflict between aircraft exit and gate, and time, we judge whether the conflict of

1239
two aircrafts will emerge. If it is true, the aircraft in apron
will stop and the taxiing aircraft keeps taxiing.
R 4 : State (1, 3) ∧ SpotIsSame ( X 1, X 2) ∧ GateDist
( X 1, X 2) ∧ DistToSpot ( X 1, C1) ∧ DistToSpot
( X 2, C 2) ∧ Time ( X 1, X 2, A1, A 2) → IsConflict
( X 1, X 2) → Maneuver ( X 1, X 2, 2, 4)
R4 represents the pushing aircraft will keep pushing
when it meets a taxiing aircraft, correspondingly, the latter
one stops.
R 5 : State (0, 3) ∧ SpotIsSame ( X 1, X 2) ∧
DistToSpot ( X 1, C 1) ∧ DistTo Gate( X 2, B 2)
∧ Time ( X 1, X 2, A1, A 2) →
IsConflict ( X 1, X 2) → Maneuver ( X 1, X 2, 0, 5)
R5 represents departure aircrafts will clear to push till
arrival aircrafts have entered gates.
R 6 : State(1,3) ∧ SpotIsSame( X 1, X 2) ∧ DistToSpot
( X 1, C1) ∧ DistToGate( X 2, G 2) ∧ Time( X 1, X 2, A1,
A2) → IsConflict ( X 1, X 2) → Maneuver ( X 1, X 2, 2, 4)
R6 represents the arrival aircraft must stop taxiing
when there is an aircraft being pushing.
R 7 : State (3, 3) ∧ SpotIsSame( X 1, X 2) ∧ DistToSpot
( X 1, C1) ∧ DistToGate( X 2, B 2) ∧ Time( X 1, X 2, A1, A
2) → IsConflict ( X 1, X 2) → Maneuver ( X 1, X 2, 3, 4)
R7 represents arrival aircrafts will enter aprons when 
departure aircrafts taxi out. Figure 1. Flow chart of the system operation
R 8 : State (3, 3) ∧ SpotIsSame ( X 1, X 2) ∧ DistTo
Gate( X 1䯸 B1) ∧ DistTo Gate( X 2䯸 B 2) ∧ Time In this paper, we set an example, Xiamen International
( X 1, X 2, A1, A 2) → IsConflict ( X 1, X 2) → ( Man Airport, to simulate the system feasibility. The northeast part
-euver ( X 1, X 2, 5, 4) ∨ Maneuver ( X 1, X 2, 4, 5)) of the airport is exclusive aprons of Xiamen Airlines, to
R8 represents aircrafts that arrive from different apron simplify the simulation, so we study Gate 61-69 relying on
spots, and we direct the one stops and the other enters its bridges. According to the airport details, we know Gate 62-
gate. 69 can dock B757 and B737, while Gate 61 can only park
B737. Fig.3 is a dot-line structure that abstracts the apron
IV. REALIZATION OF KNOWLEDGE BASE SYSTEM sketch. There are A5 and A6, two exits, represented by A
Here we use Java+Mysql development environment to and B.
complete the system realization. The Knowledge Base
system consists of three components.
1) Fact Base. We build a flight information database
according to operation rules, and build a flight performance
database containing speed (minimum, average, maximum),
geometrical shape(length, wheelbase, wingspan), etc.
2) Knowledge Base. We create several pieces of
JavaBean code to match various rules, analyze data in Fact
Base and then generate resolution strategy.
3) Interactive Interface. It includes interface for flight
feature information input and interface for conflict resolution
results output. The flow chart of the whole process is shown
in the figure 1. 
Figure 2. Dot-line structure of aprons

We set an example of two aircrafts pushes to introduce


deployment tools. The figure 3 shows the interface for flight
feature information input. Due to dynamics of moments of
aircrafts entering aprons, when inputting information, we
design an acquiring time button. The figure 4 and 5 show
the interface for conflict resolution results export.

1240
how to build a solving knowledge base to resolve physical
conflicts between two aircrafts, and then generate a rule
scheme. Finally we simulate the system in Java+Mysql
development environment. Thus it can be seen whether the
results of deployment is reasonable or not.The system uses
interfaces to make the apron information intuitive and make
air traffic controllers’ directions more convenient.
Meanwhile, it uses rules base to simplify the modeling.
Furthermore, we will add appropriate gate optimization
strategy to the system, so as to improve the ground direction
efficiency.
REFERENCES
[1] Sang Hyun, Kim Eric Feron, John-Paul Clarke. Assigning Gates by
Resolving Physical Conflicts, AIAA 5648,2009
[2] Yu Cheng, Solving Push-Out Conflicts in Apron Taxiway of Airports
by a Network-Based Simulation, Computers ind .Engng Vol
34,No.2,pp:351-369,1998
[3] Hu Minhuang, Yu Zhi, Xiao Guorong. Deploying Model of City
Guide Signs Based on the Intersection Functions. Journal of Traffic
and Transportation Engineering,2006,6(4):96-100. (in Chinese)
 [4] Stephen Atkins, Christopher Briton. Implication of Variability in
Figure 3. Interface for flight feature information input Airport Surface Operations on 4-D Trajectory Planning, AIAA8960,
2008.
[5] Goodchild M F. Geographic Information Systems and Disaggregate
Transportation Modeling. Geographical Systems,1998

Figure 4. Conflict-free interface

Figure 5. Conflict interface

V. CONCLUSIONS
This paper, through comprehensive analysis of factors
that have influence on flight operations in ramps, studied on

1241

Vous aimerez peut-être aussi