In The Name of God
A new method for Conflict Detection and Resolution in
Air Traffic Management
By:
Farnaz Derakhshan
Hojjat Emami
University of Tabriz
Faculty of Electrical and Computer Engineering
Semantic Cities @ AAAI 2012
July 23rd, 2012
Contents
Content
1
Introduction
2
Conflict Detection and Resolution Process
3
Graph Coloring Problem (GCP)
4
Imperialist Competitive Algorithm (ICA)
5
Our Proposed Model
6
Test Results & Conclusion
7
References
Semantic Cities @ AAAI 2012
2-28
Introduction (Air Traffic Control)
 Air traffic:
 “Aircraft operating in the air or on an airport surface, exclusive of loading ramps
and parking areas”
(Federal Aviation Regulations and Aeronautical Information Manual, 2010 edition)
 Air Traffic Control:
 A service operated by appropriate authority to promote the safe, orderly, and
expeditious flow of air traffic
(Federal Aviation Regulations and Aeronautical Information Manual, 2010 edition)
Introduction
Semantic Cities @ AAAI 2012
3-28
Introduction (Air Traffic Control)
 Air traffic control is a complicated task, involving multiple and dynamic controls and
have high level of granularity.
 having a reliable, safe and efficient air traffic management is a fundamental and
critical need in aviation industry.
Introduction
Semantic Cities @ AAAI 2012
4-28
Introduction (Free Flight)
 Free flight is a new concept presented potentially to solve problems in the current air
traffic management system.
 Free flight method has many advantages (such as less fuel consumption, minimum
delays, reduction of the workload of the air traffic control centers, high efficiency,
and has distributed nature)
 The most notable problem in free flight method is the occurrence of conflicts
between different aircrafts’ flights.
Introduction
Semantic Cities @ AAAI 2012
5-28
Introduction
 One of the fundamental challenges in the current air traffic management and
especially in the free flight is detection and resolution of conflicts.
 We presented a new model for conflict detection and resolution between aircrafts in
airspace using graph coloring problem.
 We mapped the conflict detection and resolution problem to graph coloring problem.
 To solve graph coloring problem we used imperialist competitive algorithm.
Introduction
Semantic Cities @ AAAI 2012
6-28
Conflict Detection and Resolution Process
 Conflict:
 Conflict is the event in which two or more than two aircrafts experience a loss of
minimum separation from each other
 Conflict Detection:
 The process of deciding when conflict (conflict between aircrafts) will occur
 Conflict Resolution:
 specifying what action and how should be to resolve conflicts
Conflict Detection and Resolution Process
Semantic Cities @ AAAI 2012
7-28
Conflict Detection and Resolution (General view)
Environment (airspace)
Operating area- Traffic
Information – weather
conditions and …
Conflict Detection
Operator
or Agent
No
Conflict
Detected?
Figure.1: Conflict Detection and
Resolution (General view)
Yes
Conflict Resolution
Conflict Detection and Resolution (General view)
Semantic Cities @ AAAI 2012
8-28
Graph Coloring Problem (GCP)
 GCP is an optimization problem which finds an optimal coloring for a given graph G.
(Jensen and Toft 1995)
 GCP is one of the NP-hard problems
 GCP is a practical method of representing many real world problems including:
 Time scheduling
 Frequency assignment
 Register allocation
 Circuit board testing
…
Graph Coloring Problem (GCP)
Semantic Cities @ AAAI 2012
9-28
Graph Coloring Problem (Cont)
 Given an undirected graph G=(V, E) with a set of vertices V and a set of edges E;
 A K-coloring of G includes assigning a color to each vertex of V, such that
neighboring vertices have different colors (labels).
 GCP implemented by using a conflict minimization algorithm
1
2
Chromatic number = 2
1
2
3
4
|K| = 2 (Colors : Red, Green)
3
4
An Undirected Graph before Coloring
|V| = 4
|E| = 4
Colored Graph
Figure.2: an example of coloring a graph
Graph Coloring Problem (Cont)
Semantic Cities @ AAAI 2012
10-28
Search Strategies
Complete
Search Strategies
Heuristic
Deterministic
Point-based
(Non-Deterministic)
Evolutionary Programming
Population-based
Evolutionary Strategy
Genetic Algorithms
Genetic Programming
Evolutionary
Algorithms
Imperialist Competitive Algorithm
Estimation of Distribution Algorithm
Search Strategies
Semantic Cities @ AAAI 2012
11-28
Imperialist Competitive Algorithm (ICA)
 Imperialist Competitive Algorithm (ICA) is Novel Socio-politically Motivated
Optimization Strategy.
 Proposed by Atashpaz-Gargari and Lucas, 2007.
 is inspired by sociopolitical process of imperialism !!
 since in late inception, it has been used in many applications.
 has shown good convergence and global minimum achievement.
 has a lot to do with.
Introduction(ICA)
Semantic Cities @ AAAI 2012
12-28
A Big Picture of ICA
Figure.3: A big picture of ICA (Gargari and Lucas, 2007)
A Big Picture
Semantic Cities @ AAAI 2012
13-28
Flowchart of the ICA
start
Eliminate this empire
Initialize the Empires
Yes
*
No
Is there an empire
With no colonies
move the colonies
toward the imperialist
stop condition
satisfied
Yes
is there
a colony in an empire
which has lower cost
than that of
imperialist
No
pick the weakest colony from the
weakest empire and give it to
the empire that has the most
likelihood to posses it
Done
No
End
Yes
Exchange the positions of that
Imperialist & colony
*
Compute the total cost of all empires
Figure.4: Flowchart of the ICA
Flowchart of the ICA
Semantic Cities @ AAAI 2012
14-28
Our Proposed Model
 In our model, the main strategy is based on: "Prevention is better than cure“
 we mapped the conflict detection and resolution problem to graph coloring problem.
 we map the aircrafts congestion area to a corresponding graph based on aircrafts
condition in airspace.
 imperialist competitive algorithm has shown great performance in both convergence
rate and better global optimal achievement, therefore we used this algorithm to solve
graph coloring problem rather than other evolutionary algorithms.
 In our model, a Global approach is used for resolving the multiple conflicts.
(the entire traffic situation is examined simultaneously, and it has the high capabilities)
Our Proposed Model
Semantic Cities @ AAAI 2012
15-28
Our Proposed Model
start
Monitor the Environment
Map the Congestion Area to a Graph
(Making Adjacency Matrix)
Define Problem Parameters
And other Traffic Parameters
Solving the GCP using ICA
(Conflict Resolution process)
Project current states to future states
by using nominal propagation method
Send new Flight Paths Plan to Existing
Aircrafts in Congestion Area
Detect the Congestion Area
Compute Distance between all Aircrafts
in Congestion Area
End
Figure.5: Block diagram of our proposed model
Our Proposed Model
Semantic Cities @ AAAI 2012
16-28
Creating the Graph
 here, the nodes of graph indicate aircrafts in the congestion area and every edge between two
nodes indicates a probable conflicts in the future time step, if the aircrafts continue their current
flight plan.
Figure.6: Graphical display of an example
conflict detection & resolution scenario
(creating the graph)
Creating the Graph
Semantic Cities @ AAAI 2012
17-28
Conflict Detection
 we use a simple conflict prediction method.
 we used nominal state propagation method for prediction conflicts between aircrafts.
 the current position, heading and speed of existing aircrafts in congestion area used
for mapping the current location to the future states.
 in the future states if distance between two aircraft is less than a predefined reliable
distance threshold we say a conflict is going to occur.
(i.e. when in future states the protected zones of two aircraft is overlapped then
system report a conflict)
 here, we focused only horizontal plan dimension
Conflict Detection
Semantic Cities @ AAAI 2012
18-28
Conflict Resolution
 we map the congestion area to a state space graph.
 we replace solving the conflicts between different aircrafts with coloring the
corresponding graph.
 we used imperialist competitive algorithm to solve graph coloring problem
 a colored plan for graph is a reliable solution for conflicts problem
 in the conflict resolution process, for each aircraft, we allocate a flight path in which
this aircraft will has a reliable distance with other aircrafts and there is no risk of
conflict.
 In our model, we can assume each flight path as five directional options namely:
main line, deviation to right of the main line, deviation to left of the main line, top
of the main line and bottom of the main line.
Conflict Resolution
Semantic Cities @ AAAI 2012
19-28
Test Results
 To evaluate our proposed model, we use random flights model (Archibald et al. 2008)
 here all aircrafts are constrained to fly at the same altitude and at a constant speed.
 Small and instantaneous heading changes for each aircraft are the only maneuvers
of resolving conflicts.
 We used supposed scenarios that these samples contain 2, 3, 4, 5, 6, 8, 12, 16, 20 and
30 aircrafts in congestion area.
 In air traffic control we deal with a multi objective problem.
 In this paper, our goal is providing a safe, reliable and efficient strategy for solving
conflicts between aircrafts.
Test Results
Semantic Cities @ AAAI 2012
20-28
Performance Metrics
 The ideal state in conflict resolution model is that the aircrafts are able to track their
destination without deviation or with minimal deviation from their original path.
 Maneuvers that are used in the conflict resolution methods, causes the aircraft to be
diverted from the ideal and optimal mainstream.
 System efficiency metric:
 measures the degree to which the aircrafts in the system are able to follow direct and linear
flight paths to their destinations.

In this paper, we define a performance criterion for each aircraft same as follows:
Performance Metrics
Semantic Cities @ AAAI 2012
21-28
Performance Metrics (cont)
 here, we define a performance criterion for each aircraft same as follows:
ideal and optimal flight path for an aircraft
the amount of deviation from mainstream of aircrafts
 We try to select routes with lowest cost when we redirect the aircrafts’ main routes
(i.e. the lowest deviation from the main flight path for each aircraft)
Performance Metrics (cont)
Semantic Cities @ AAAI 2012
22-28
Performance Metrics (cont)
 System Efficiency:
Total number of aircrafts
Performance of each aircraft
 how much this criterion is closer to “1” indicates good performance of the system
and how much this criterion is closer to “0” indicates poor performance
 our proposed model for solving GCP for higher dimensions (for a great number of
aircrafts that are in the congestion area) acts as a good way.
Performance Metrics (cont)
Semantic Cities @ AAAI 2012
23-28
Test Results
Table 1: Test results of applying the algorithm onto input graphs (congestion areas) with
specified parameters
Test Results
Semantic Cities @ AAAI 2012
24-28
Conclusion
 one of the fundamental challenges in the current air traffic management and
especially in free flight method is detection and resolution of conflicts.
 in our approach, we mapped conflict detection and resolution problem to graph
coloring problem, then use imperialist competitive algorithm to solve GCP.
 Our proposed model provides an efficient and reliable solution for solving conflicts
in air traffic management
 in conflict resolution process we tried to select routes with lowest cost when the
aircrafts’ redirected from main route, therefore, we have least delay in flights and
the minimum consumption of resources (e.g. in fuel consumption).
 our proposed model use multiple strategies for the resolution of conflicts.
Conclusion
Semantic Cities @ AAAI 2012
25-28
References
1. Federal Aviation Regulations and Aeronautical Information Manual, 2010 edition. 2009 ASA, Inc. Newcastle,
Washington.
2. Rong, J., Geng, Sh., Valasek, J., R. Ioerger, T. 2002. Air traffic conflict negotiation and resolution using an onboard multi
agent system. Digital Avionics systems Conf. Irvine, CA.
3. Kuchar, J., and Yang, L. C. 2000. A Review of CD&R Modeling Methods. IEEE Transactions on Intelligent
Transportation Systems, Vol. 1, No. 4, December.
4. Agogino, A., and Tumer, K. 2008. Adaptive Management of Air Traffic Flow: A Multiagent Coordination Approach.
Proceedings of the Twenty-Third AAAI Conf. on Artificial Intelligence.
5. Agogino, A., and Tumer, K. 2009. Improving air traffic management with a learning multiagent system. IEEE Intell.
Syst., vol. 24, no. 1, pp. 18–21, Jan/Feb.
6. Atashpaz-Gargari, E., Lucas, C. 2007. Imperialist Competitive Algorithm: An algorithm for optimization inspired by
imperialistic competition. IEEE Congress on Evolutionary Computation, 4661–4667.
7. Eby, M. S. and III, W. E. K. 1999. Free flight separation assurance using distributed algorithms. In Proceedings of
aerospace conference. Aspen, Co.
8. Archibald k., Hill c., Jepsen a., Stirling c., and Frost l. 2008. A satisfying approach to aircraft conflict resolution. IEEE
transactions on systems, man, and cybernetics—part c: applications and reviews, vol. 38, no.
References
Semantic Cities @ AAAI 2012
26-28
References
9.
Agogino, A. K., and Tumer, K. 2009. Learning indirect actions in complex domains: Action suggestions for air traffic
control. Advances in Complex Systems, 12, 493–512.
10. Whalen, S. 2002. Graph Coloring with Genetic Algorithms. http://mat.gsia.cmu.edu/COLOR/color.html.
11. Garey, M.R. and Johnson, D.S. 1999. Computers and intractability: a guide to the theory of NP-completeness, W.H.
Freeman and Company, New York.
12. De Werra D. 1990. Heuristics for Graph Coloring. Computing Suppl. 7, pp. 191-208.
13. Goldberg, D. E. 1980. Genetic Algorithms in Search, Optimization and Machine Learning. Addison-Wesley, Reading,
MA.
14. Valkanas, G., Natsiavas,P., Bassiliades,N. 2009. A collision detection and resolution multi agent approach using utility
functions, Fourth Balkan Conf. in Informatics.
15. Chao, W. and Jing, G. 2009. Analysis of Air Traffic Flow Control through Agent-Based Modelling and Simulation.
International Conf. on Computer Modelling and Simulation, IEEE.
16. Jensen T.R., Toft B. 1995. Graph Coloring Problems, Wiley interscience Series in Discrete Mathematics and Optimization.
References
Semantic Cities @ AAAI 2012
27-28
The End & Thanks
T
hanks for your attention
The End & Thanks
!