Air Traffic Control

Program Leader: Peter Lindsay

As more vehicles take to the air, air traffic control is a constraining factor on the number of aircraft that can be accommodated, and on the trajectories that they fly. Europe, the USA and Australia are all considering fundamentally new ways of managing air traffic, to improve efficiency without compromising safety. In this program, we applied complex systems science to the problem by developing and using air traffic simulators to study new concepts and tools for air traffic management, and developing new approaches to assurance of system-level properties including safety and efficiency.


Agent-based ATC simulation and visualization

Project Leader: Peter Robinson
Researchers: Peter Lindsay, Colin Ramsay, Dale Clutterbuck, Katherine Duczmal, Martijn Mooij, Robert McLeay, Tim Rudge

The aim of this project was to develop a framework for agent based simulations and visualizations of ATC systems. The logic programming language Qu-Prolog, with built-in support for threads and high-level communication, was used to implement the required ATC agents. The intention was to take a proposal for a free-flight ATC system, model the behaviours of the various players of this system, and run simulations to determine how the system will perform - e.g. to determine how likely near misses are.

In 2004 we produced an initial implementation of a Clevel library for supporting simulations, and two prototype visualizations to aid in testing our agent models. Several simple models for controller behaviour were developed as a proof of concept and to check the adequacy of the C-level support library. We also started work on automatic scenario extraction and generation software. The C-level library is used to keep track of aircraft as time evolves and to respond to requests for information such as where aircraft are, how close a pair of aircraft will get, and time to minimum separation. The library also provides an interface to Qu-Prolog, in which we can write intelligent agents. These agents (such as controller agents) can ask for aircraft information from the library to aid in decision making. Katie Duczmal (a summer RA) has developed a more sophisticated controller agent that models conflict detection, conflict resolution and monitoring. Robert McLeay and Dale Clutterbuck (summer RAs) have been porting Qu-Prolog from Unix to Windows in order to make our modelling techniques more accessible to other researchers. (Included Summer Projects 2004/05)

In 2005 we further extended the Qu-Prolog model of a controller and, based on the addition of floating point arithmetic to Qu-Prolog, simplified the various calculations used by the controller agent. We also added the ability to dump information to a file and to run the same scenario multiple times. In combination this allows us to carry out statistical analyses of the behaviour of the controller agent. The results of this project were presented at the ACCS winter school.

Sensitivity analysis of ATC operator performance models

Project Leader: Peter Lindsay
Researchers: David Abramson, Andrew Neal, Colin Enticott, Junhua Wang, Simon Connelly

The SafeHCI project developed an approach to Human Reliability Assessment based on modelling of the cognitive processes involved in the operator’s task. This project investigated the application of Nimrod middleware to SafeHCI-like models in order to run thousands of different scenarios in a distributed fashion over the Grid, and to perform statistical analysis on the results, to identify emergent effects. The approach was intended to enable us to identify how sensitive the models are to changes in parameters, such as traffic patterns.

In 2004 we used the Nimrod middleware to evaluate thousands of ATC scenarios. Through repeated experiments, it was discovered that the original SafeHCI approach - of comprehensively exploring a model's state space - was prohibitively slow and yielded incomplete results. Performance improved immensely when we converted it to use a Monte Carlo approach instead, run with a large number of iterations. A set of scenarios was developed in order to evaluate the model systematically, to encompass various traffic patterns, workload levels, and geometry of aircraft interactions. Each of these scenarios was evaluated against various models of interaction, representing different possible designs for a given interface. The settings for these models were also altered systematically to test sensitivity of the results against how the tools behaved. The result was a proof of concept that the approach can be used to investigate the impact of Human Computer Interface design changes on operator performance.

  • Abramson, D., Dongarra, J., Meek, E., Roe, P., Shi, Z., Simplified grid computing through spreadsheets and NetSolve", Proceedings of the 7th International Conference on High-Performance Computing and Grid in the Asia-Pacific Region, 2004.
  • Beasley, J., Krishnamoorthy, M., Sharaiha, Y., Abramson, D., Displacement problem and dynamically scheduling aircraft landings", The Journal of the Operational Research Society, Vol. 55, 2004, 54-64.
  • Lewis, A., Abramson, D., Peachey, T., RSCS: A parallel simplex algorithm for the Nimrod/O optimization toolset", Third International Symposium on Parallel and Distributed Computing/Third International Workshop on Algorithms, Models and Tools for Parallel Computing on Heterogeneous Networks (ISPDC/HeteroPar'04), 2004, 71-78.

Conceptual framework for FFATC

Project Leader: Peter Lindsay
Researchers: Binu John, Ariel Liebman, Martijn Mooij,Colin Ramsay, Morgan Smith

The aim of this project was to enable different free flight operational concepts to be modelled and explored. It aimed to develop a hierarchical framework, to capture the complex, dynamic and interdependent nature of control in the National Airspace System (NAS). A suitable framework needs to encompass a wide variety of agents, from pilots, controllers and flow managers through to airports and airline operations centres. We adapted existing conceptual frameworks such as the Boeing 2020 ATM Concept, and extended them for Australian operations.

In 2004, a baseline conceptual framework for the modelling and evaluation of operational concepts was developed. This framework is based on the Boeing 2020 ATM Concept and captures the agent environment interaction at a number of levels of aggregation, using a hierarchical control structure. The Agent-based ATC simulation and visualization project is currently working on implementing this framework.

In 2005 the project underwent conceptual re-alignment and development. In the first instance, the project is now aligned with the ASTRA (Australian Strategic Air Traffic Management Group) 2003 strategic plan. This involves the integration into the free-flight air traffic control concept, of several intermediate steps that go some way towards distributed control. These intermediate operation concepts are user-preferred routes and user-preferred trajectories. The former concept allows airlines/aircraft to choose the 3-D path taken by the aircraft while the latter also allows the choice of the times at which the aircraft arrive at any point along that path. Each of these could have a potentially major impact on emergent system properties. The most significant of those is collision risk, but others include operator workload and delays. In order to support the analysis of the various free-flight air traffic control concepts prototype simulations tools have been developed (see project: ‘Development of a robust ATC simulation codebase’). A preliminary concept has been developed for the comparison of the performance of the centralised air traffic control operational concept to the free-flight air traffic control one. Air traffic control operator agents have been implemented as well as simple pilot agents obeying limited controller instructions. Both the operator agents and the pilot agents are capable of following rudimentary conflict detection and resolution strategies. (Included two 2005/06 summer projects)

Air Traffic Control workload

Project Leader: Andrew Neal
Researchers: Scott Bolland, Gerard Champion, Graham Halford, Mike Humphreys, Ariel Liebman, Peter Lindsay, Shayne Loft, Martijn Mooij, Penelope Sanderson

The aim of this project was to develop a computational model that can measure the flow of traffic through an air sector, and predict the level of workload that an air traffic controller will experience, as well as the overall risk of breakdowns in separation between aircraft. The purpose was to develop a tool that can be used for risk analysis and scenario planning. This was a multidisciplinary project, integrating recent models of human memory and reasoning, with formal methods for the analysis of human-computer systems. The project was funded jointly by an ARC Linkage grant and Airservices Australia and administered through UQ's Key Centre for Human Factors and Applied Cognitive Psychology.

In 2004, we undertook intensive air traffic control training to familiarise all members of the project team with the domain, carried out a systematic review of the technical literature to identify the factors previously found to predict workload, and carried out interviews with controllers, using the critical decision method. We also commenced an analysis of air traffic incidents in preparation for the development of the simulation tool. The scope of the study has been expanded due to the increased commitment from our industry partner.

In 2005, the project focused on three key areas: experiments which provided air traffic control operator behaviour information; the development of an operator model; and the development of software tools for the simulation of aircraft trajectories, emulation of air traffic control operator behaviour, and the computation of workload metrics. The experiments performed cover both ‘part-task’ studies, which focused on specific components of the air traffic control operator's problem solving activities, such as conflict detection conflict resolution, and high fidelity simulations where aircraft trajectories and controller actions were recorded using realistic training simulations. The information gathered during the experimental studies was used, and continues to be used, to develop and calibrate the operator model. Elements of the operator model have been developed and implemented. This includes conflict detection under aircraft uncertainty conditions, and conflict resolution through flight level changes. The tools developed include a prototype for the replay of recorded aircraft trajectory data and the calculation of some workload measures. This prototype was demonstrated to Airservices Australia personnel involved in the project. Furthermore, this tool served as the basis for the development of the ACCS air traffic control simulation toolkit (ATC-ST). The tool was then redeveloped to support general simulation using flight-plans and a complete workload calculation module added.

In 2006, a large amount of useful ATC operations room data was collected and analysed for 18 sectors covering much of Australia's busiest upper airspace traffic areas. This data included controller subjective workload ratings, flight plans and flight track data. The data was used to develop and evaluate an improved regression model which correlates the traffic patterns with workload ratings and was then incorporated into a prototype workload prediction tool. Agent-based models were developed to emulate the key aspects of the controller's behaviour, including accepting aircraft into the sector and issuing flight level clearances for climbing and descent, followed by hand-off to the next sector. Finally, a problem-solving algorithm has been developed and coded into intelligent agents in order to simulate conflict detection and resolution by the controller. Additionally, a review of air traffic control workload studies was completed and accepted for publication in Human Factors, one of the most prestigious journals in the field.

In 2007, we completed the collection of data from the operational environment, with a final study being run in Melbourne Centre. Analyses of earlier studies run in Brisbane Centre were completed, and progress was made towards handover of the final workload modelling tool for the industry partner. In the first half of 2008, we will complete the analyses of the data from the Melbourne study and finish debugging of the workload modelling tool.

  • Boag, C., Neal, A., Loft, S., Halford, G., 'An analysis of relational complexity in air traffic control conflict detection task', Ergonomics, Vol. 49, No. 14, 2006, 1508-1526.
  • Bolland, S., Fothergill, S., Humphreys, M., Neal, A., 'Modelling the human air traffic controller, Part II: Emulating controller intervention', Proceedings of the 14th International Symposium on Aviation Psychology, 2007.
  • Loft, S., Sanderson, P., Neal, A., Mooij, M., 'Modelling and predicting mental workload in en route air traffic control: Critical review and broader implications', Human Factors, 2007.
  • Loft, S., Bolland, S., Humphreys, M., 'Modelling the human air traffic controller. Expert-Trainee differences in conflict detection.', Proceedings of the 14th International Symposium on Aviation Psychology, 2007.
  • Loft, S., Neal, A., Humphreys, M., 'The development of a general associative learning account of skill acquisition in a conflict detection task', Journal of Experimental Psychology: Human Perception and Performance
  • Neal, A., Mooij, M., Bolland, S., Xiao, T., Lindsay, P., Boag, C., 'Using multi-level analysis to model the sources of variability in workload within and between sectors', Proceedings of the 14th International Symposium on Aviation Psychology2007.
  • Neal, A., Sanderson, P., Lindsay, P., Loft, S., Mooij, M., Boland, S., Fothergill, S., Liebman, A., ‘ATC workload modelling project: Year 1 progress report’, Key Centre for Human Factors Technical Report, Prepared for Airservices Australia, June 2005.
  • Sanderson, P., Mooij, M., Neal, A., 'Investigating sources of mental workload using a high-fidelity ATC simulator', Proceedings of the 14th International Symposium on Aviation Psychology, 2007.

Safety assessment of ATC human-computer interaction

Project Leaders: Andrew Neal, Jacqueline Wicks
Researchers: Rachel Chitoni, Simon Connelly, Jingru Dai, Peter Lindsay, Colin Ramsay, Junhua Wang

This project aimed to develop a new approach to human reliability assessment and evaluation of human-computer interaction design options by application to air traffic control. The approach was based on modelling the activities (cognitive processes and interactions) involved in en-route control as stochastic processes. The effect of a proposed design intervention could then be investigated by hypothesising its effect on individual activities and conducting simulations to gauge performance over a range of scenarios. The project was a close collaboration between computer scientists from The University of Queensland's School of Information Technology and Electrical Engineering and psychologists from the Key Centre for Human Factors and Applied Cognitive Psychology. We conducted experiments in which human subjects make judgements about, and attempt to manage, air traffic control scenarios presented to them on a computer display. These experiments were conducted with two related goals in mind: the first was to gain a greater understanding of the decision-making processes of air traffic control operators; and the second was to model their behaviour more realistically in the simulations. The work is also a collaboration with Peter Kwantes from Defence Research and Development, Canada.

During 2004 collaborators from the Key Centre for Human Factors and Applied Cognitive Psychology conducted a large number of ATC experiments. These experiments explored the effects of aircraft geometry (speed, angle, order), time pressure, and work-load in order to gain a greater understanding of how humans respond when faced with different ATC scenarios. Important factors have been identified and analysis of this large data resource is continuing into 2005.

The majority of the psychology experiments for this project were completed during 2005. The first working version of the Operator Choice Model to simulate air traffic controller behaviour was successfully developed. It employs data from the psychological experiments that were designed for this purpose. The results were presented at the International Congress on Modelling and Simulation. This novel approach to modelling an agent, where the agent is required to address a variety of challenging problems within a complex system, was well received. A theoretical paper describing the use of the Operator Choice Model for analysing human error in interactive systems was presented to the International Conference on Software Engineering and Formal Methods.

In 2006, Operator Choice Models (OCM) were developed for conflict detection and resolution tasks using experimental data collected from simulator trials with student subjects. By measuring conflict recognition in different settings, the experiments and models showed that operator performance depends critically on the nature of the overall task they are undertaking. Te formal models were further used to explore the effect on performance of different computer-based tools. A parameter sweep was conducted, using the Nimrod tool, to compare predicted operator performance for four different design options on a range of different traffic patterns. Staff from the ACCS's Monash node helped develop means for visualising and interpreting the results. Further to this, a paper was prepared on the use of model checking to assist in the formal categorisation and analysis of patterns of behaviour that lead to task failure.

  • Cerone, A., Connelly, S., Lindsay, P., ‘Formal analysis of human-computer interaction using model-checking’, 3rd IEEE International Conference on Software Engineering and Formal Methods (SEFM 05), September 2005.
  • Wicks, J., Connelly, S., Lindsay, P., Neal, A., Wang, J., Chitoni, R., ‘Simulation of air traffic controllers' behaviour using the operator choice model’, International Congress on Modelling and Simulation, December 2005, 3023–3029.

Operator performance modelling in Microsaint

Project Leader: Peter Lindsay
Researcher: Sean Ness

This summer student project explored the use of the Microsaint simulation tool for modelling operator behaviours. In particular, it developed guidelines for translating from SafeHCI-like models into Microsaint. (Summer Project 2003/04)

  • Ness, S., Feasibility study on implementing the SafeHCI model in Microsaint", ACCS Technical Report, 2004.

Development of robust ATC simulation code base

Project Leader: Ariel Liebman
Researchers: Bangjun Chen, Peter Lindsay, Colin Ramsay

This project aimed to produce a robust codebase for a toolkit which can be used to simulate the motion of aircraft in a 3-D airspace. The toolkit consisted of a core 3-D + Time simulator and interface modules which enable the use of standard flight plan data and airspace specifications. The airspace specification used standard air traffic management approach (e.g. nautical miles (NM) for distance). Aircraft motion was represented to the user graphically with position and altitude being displayed in NM and 100's of feet. The toolkit also contained support for Agents (air traffic control controllers, pilots and so on).

This project commenced in July 2005. The ATC Simulation Toolkit (ATC-ST) grew out of aircraft trajectory replay and simulation software written by Scott Bolland of the Key Centre for Human Factors for the ATC Workload project. The original codebase was completely re-written to make it useable as a simulation library with a well defined API. This toolkit now supports simulation of aircraft trajectories in both replay mode and from initial flight-plans. The latter is of key importance for supporting the free-flight air traffic control research. The trajectory simulation component also supports advanced aircraft characteristics simulation, and the impacts that wind velocity has on aircraft trajectories. Additional functionality includes support for intelligent agents and inter-agent communication. (Included 2005/06 summer project)

In 2006, the main activities have been the development of the toolkit to support active agents representing, for example, pilots and controllers, and the creation of the code base needed for supporting agent interventions to resolve conflicts and to manipulate aircraft timings.

Swarm intelligence for conflict detection and resolution in free-flight environments

Project Leader: Hussein Abbass
Researchers: Sameer Alam, Michael Barlow, Peter Lindsay, Minh Ha Nguywn

Free flight is a revolutionary concept that will enable greater traffic volume and operational flexibility by distributing some of the functionality, including conflict detection and resolution, to airborne systems and pilots. Swarm intelligence is the study of computations in social insects (such as ants, termites and some types of bees and wasps). It is a new branch of distributed artificial intelligence, where computations are carried out by a group of agents working cooperatively to achieve a task. Ant colony optimisation (ACO) is an optimisation technique inspired by the behaviour of real ants. In this project, we developed safe conflict detection and resolution (CDR) algorithms for free flight inspired by natural computations and navigations in colonies of ants. The advantages of such algorithms include being adaptive in a dynamic environment, and being fully distributed.

In 2005, we first investigated a neuro-controller for CDR, which raised a number of safety challenges (Alam et al 2005a).  We then designed an efficient and safe ant-colony based algorithm for CDR that is able to adapt to changes in the surrounding environment. Preliminary experiments of this algorithm can be found in Alam et al 2005b. The current version of the method meets safety standards and constraints for a safe and reliable conflict detection and resolution algorithm. Outcomes will also include recommendations of safety protocols for ‘manoeuvre choices’ in conflict situations.

In 2006, we continued our investigation into developing a methodology for safe manoeuvre. The investigation focused on generating safe trajectories for flights to avoid weather hazards. We developed a method that generates multiple trajectories for the pilot order according to weights that may vary from one situation to another (for example, if the pilot wishes to reach a destination quickly or is happy to have a delay to achieve maximum comfort for passengers). The method guarantees to provide pilots with solutions that are safe according to aviation standards.

In 2007, work proceeded on a fully distributed version of the Air Traffic Operations and Management Simulator (ATOMS) toolset. An evolutionary-based scenario assessment approach was developed, where advanced ATM concepts can be tested and validated. The ATOMS tool was also adapted for use in a study for Airservices Australia to investigate possible benefits of different ATM strategies (including free-flight) on greenhouse gas emissions and noise.

  • Alam, S., Abbass, H., Barlow, M., Lindsay, P., ‘Mapping lessons from ants to free flight: An ant-based weather avoidance algorithm in free flight airspace’,  The International Society for Optical Engineering (SPIE), Microelectronics, MEMS, and Nanotechnology Symposium, 2005, 9.
  • Alam, S., Bui, L., Abbass, H., Barlow, M., 'Pareto metaheuristics for generating safe flight trajectories under weather hazards', 6th International Conference on Simulated Evolution and Learning, 2006; LNCS, Vol. 4247, 829-836.
  • Alam, S., McPartland, M., Barlow, M., Lindsay, P., Abbass, H., ‘Neural evolution for collision detection and resolution in a 2D free flight environment’, Recent Advances in Artificial Life, vol. 3, World Scientific Publishers, 2005, 13–28.
  • Alam, S., Abbass, H., Barlow, M., “ATOMS: Air traffi c operations and management simulator” , IEEE Transactions on Intelligent Transportation, 2008.
  • Alam, S., Nguyen, M., Abbass, H., Barlow, M., “Ants guide future pilots”, Progress in Artificial Life, December 2007; Lecture Notes in Computer Science, Vol. 4828.
  • Alam, S., Nguyen, M., Abbass, H., Barlow, M., “The architecture design and validation of the air raffi c operations and management simulator (ATOMS)”, Proceedings of the 2007 SimTecT, June 2007.
  • Alam, S., Shafi , K., Abbass, H., Barlow, M., “Evolving air traffic scenarios for the evaluation of conflict detection models”, Proceedings of the 2007 EUROCONTROL Innovative Research Workshop, December 2007.
  • Chen, K., Dam, H., Lindsay, P., Abbass, H., “Biasing XCS with domain knowledge for planning flight trajectories in a moving sector free-flight environment”, Proceedings of the 1st IEEE Symposium on Artificial Life, 2007.
  • Greenwood, G., Abbass, H., “A new local search algorithm for continuous spaces based on army ant swarm raids”, Proceedings of the 2007 IEEE Congress on Evolutionary Computation, September 2007; 1097-1102.
  • Nguyen, M., Alam, S., Abbass, H., “Dynamic weather avoidance in a traffi c constrained airspace”, Proceedings of the 2007 EUROCONTROL Innovative Research Workshop, December 2007.
  •  Ziauddin, U., Sarker, R., Abbass, H., “Improving the performance of genetic algorithm in capacitated vehicle routing problem using self imposed constraints”, Proceedings of the 2007 IEEE Symposium on Computational Intelligence in Scheduling, April 2007.

Evaluation of future Air Traffic Management concepts

Project Leader: Peter Lindsay
Researchers: Ariel Liebman, Martijn Mooij, Greg McDonald, Colin Ramsay,  Peter Robinson, Miguel Vilaplana

This project used modelling and simulation to explore new operational concepts for air traffic management (ATM) using the ATC simulation toolkit developed by the Centre. The 2007 Australian ATM Strategic Plan proposed a notion of User Preferred Trajectories, whereby airspace users have more direct influence over the 4-D trajectory that they fly. This in turn is expected to lead to large efficiency gains. The challenge was to understand how to implement this concept, which will involve a fundamental change of conceptual viewpoint, from distance-based air traffic control to time-based trajectory management.

In 2006, the project developed an agent-based modelling framework for exploring a timing-based approach to air traffic control. The idea was to try to integrate conflict resolution strategies into a 4D trajectory-based approach, rather than treating them as separate activities as is currently the case. The different control intervention options available to an agent are assessed not only on whether they maintain safe separation between aircraft, but also whether the aircraft will meet certain time requirements along their trajectories. Our simulations showed that the approach is generally feasible, provided certain parameters (such as operator workload and traffic complexity) are kept within certain bounds. The simulations also revealed areas where the concept will need refinement. The investigation involved a detailed study of actual controller interventions and development of realistic scenarios on which to evaluate the approach.

In 2007, two major studies were conducted. The first was a study of sector transit times in Australian airspace based on actual trajectories fl own. Three cases were compared: actual transit times using track data; transit times extrapolated from flight plans using our ATC Simulator software; and transit times from flight plans using agents emulating ATC interventions. The study demonstrated that the simulator is broadly accurate, in as much as the general behaviours were broadly the same in all three cases. The main differences seem to be due to lack of wind modelling in the Simulator, but another issue was timing of interventions, where they occurred in real life, since this information was not available to us. The second study focused on feeder sectors and used agent-based models to investigate the likely effects of increased traffic volumes on delays. A nonlinear relationship between traffic density and airspace capacity emerged where delays and other system level metrics increase suddenly when a certain threshold was reached. The study was repeated with the addition of a simple decision support tool enabling the controller agent to optimise for arrival timing accuracy. It was found that this capability significantly reduced delays without imposing penalties on other system properties such as safety.

In 2008 the project’s focus was on analysing and writing up the result of the timing intervention study that was begun in 2007. The study modelled the effect of increased traffic volumes on delays, caused by ATC flow and separation interventions in feeder sectors. A non-linear increase was observed, together with large jumps in system metrics such as workload when certain thresholds were reached. A decision support tool concept was modelled and found to significantly reduce delays, but to have little or no effect on the thresholds observed. Work also began on a study of idle-throttle Continuous Descent Approaches, in collaboration with Boeing RTE in Madrid, using Boeing trajectory simulation tools to study the robustness of such trajectories against late changes to Required Time of Arrival. This fundamental scientific research is expected to inform the redevelopment of airspace design to enable optimally efficient, quieter trajectories as aircraft approach airports.

  • Lindsay, P., 'Evaluation of a simple timing-based intervention heuristic for trajectory-based Air Traffic Management', 5th EUROCONTROL Innovative Research Workshop Proceedings, 2006.
  • Cerone, A., Connelly, S., Lindsay, P., “Formal analysis of human operator behavioural patterns in interactive surveillance systems”, Journal of Software and Systems Modelling, Vol. 7, No. 3, 2008, 273–86.

Propagation of uncertainty in trajectory computations

Project Leaders: Peter Lindsay, Miguel Vilaplana
Researchers: Ariel Liebman, Greg McDonald, Colin Ramsay,

The ability to accurately predict the 4-D trajectory that an aircraft intends to follow is important for a wide range of ATM applications, including conflict detection, traffic planning and air/ground coordination. Boeing Research and Technology Europe (BRTE) commissioned the ACCS to develop a theoretical framework for modelling uncertainties associated with the trajectory prediction process. The work began in 2007 with a preliminary review of the literature.

In 2008 the literature on aircraft trajectory-prediction methods was extensively reviewed and the results were synthesised into a theoretical framework that separated out the different types of uncertainty. Important factors included: environmental conditions such as wind, aircraft performance characteristics; operational decisions such as the pilot procedure invoked and the precise time at which it is invoked; the prediction process itself, and the data and algorithms on which it is based. A series of experiments were undertaken using BRTE’s trajectory description language and toolset to demonstrate how different uncertainty factors can be separated out and their effects studied independently and in conjunction.

  • Lindsay, P.A., Liebman, A., Ramsay, C., “Review of previous work on trajectory prediction uncertainty”, Report to Boeing Research & Technology Europe, April 2008.
  • Ramsay, C., “Automated support for study of the effects of uncertainty in trajectory computation”, Report to Boeing Research & Technology Europe, April 2008.
  • Ramsay, C., Lindsay, P.A., Liebman, A., “A theoretical framework for the qualitative and quantitative study of the effects of uncertainty in trajectory computation”, Report to Boeing Research & Technology Europe, April 2008.

Using network simulation and visualisation tools for air traffic management research

Project Leader: Ariel Liebman
Researcher: Daniel Bradley

The aim of this project was to model air traffic management using existing network simulation software developed to model gene regulatory networks. This project aimed to provide a network simulation approach to understanding air traffic management and air traffic congestion; and to increase understanding of appropriate abstraction levels suitable for a unified simulation approach for complex systems.

In 2008 we developed a discrete, spatial, hexagonal network model and simulation system, that utilised an ant-trail inspired network generator to generate possible flight-paths through the network.

Distributed conflict resolution for a free flight environment

Project Leader: Peter Robinson
Researcher: Keith Clark

This project investigated the use of agents to cooperatively resolve conflicts while trying to optimise overall system efficiency. The idea behind this is that plane agents would propose conflict resolution plans together with the cost of carrying out the plans to the planes they are in conflict with. Through negotiation the plane agents would decide on a collection of plans to resolve the conflicts.

In 2008 we developed an agent-based system where the agents cooperate in conflict resolution. The system is currently being evaluated and a paper being written.

Modelling interdependency in airport operations

Project Leader: Hussein Abbass
Researcher: Jiangjun Tang

This aim of this project was to understand interdependency in airport operations. For example, the way in which airport security can affect customs operations. Understanding the information flow is key to measuring the impact of interdependency on air traffic.

In 2008 we were very successful in capturing interdependency in network operations. We captured key types of information flow, and also produced a service oriented diagram.

  • Tang, J., Alam, S., Lokan, C., Abbass, H., “Modelling and evolutionary multi-objective evaluation of interdependencies and work processes in airport operations”, World Summit on Genetic and Evolutionary Computation (GEC), 2009.

General ATC publications

  • Chang, R., Lindsay, P., ‘A simulator for exploring autonomous control of multiple UAVs at non-radar controlled airstrips’, 2005 Intelligent Sensors, Sensor Networks & Information Processing Conference, December 2005, 391–396.
  • Chen, K., Dam, H., Lindsay, P., Abbass, H., 'Biasing XCS with domain knowledge for planning flight trajectories in a moving sector free flight environment', Proceedings of the 1st IEEE Symposium on Artificial Life, 2007.
  • Chen, K.Y., Lindsay, P.A., Robinson, P.J., Abbass, H.A., “A hierarchical conflict resolution method for multi-agent path planning,” Proceedings 2009 IEEE Congress on Evolutionary Computation (CEC 2009), Trondheim, Norway, 2009.
  • Connelly, S., Lindsay, P., Gallagher, M., “An agent based approach to examining shared situation awareness”, 12th International Conference on Engineering of Complex Computer Systems (ICECCS 2007), July 2007, 138-147.
  • Dam, H., Abbass, H., Lokan, C., Yao X., “Negative correlation learning for neural-based learning classifier systems”, IEEE Transactions on Data and Knowledge Engineering, Vol. 20, No. , 2008, 26–39.
  • Rojanavasu, P., Dam, H., Abbass, H., Lokan, C., Pinngern, O., “A self-organized, distributed, and adaptive rule-based induction system”, IEEE Transactions on Neural Networks, 2009.

World-class basic and applied inter-disciplinary research on questions fundamental to understanding, designing and managing complex systems
© 2009 The ARC Centre for Complex Systems, Australia