Introduction

This paper is has been prepared to provide information on the development of Collaborative Decision-Making (CDM) in the ATM environment. It follows on from a presentation given at the IFATCA EUR Regional Meeting in Brussels in October 2001.

1.1 What is Collaborative Decision Making (CDM)?

CDM is the making of operational and often commercial decisions by the various agents or ‘actors’ within a system such as the Air Transport Industry, based (at least in part) on information which is actively shared with other agents or actors within the same system. The process is collaborative or cybernetic in nature in that the outcomes of the decisions made are fed back into the information pool to maintain its currency and effectiveness for all users.

An alternative description of CDM comes from Eurocontrol and reads as follows:

“Both the collective requirements of all airspace users and the individual aircraft operator’s preferences will be taken into account in determining solutions to events. The open systems environment and better information management will allow a permanent dialogue between the various parties (ATM, Aircraft Operators’ Operations Centres, Pilots and Airport Operations) before departure, and as the flight progresses through the ATM system. This exchange of information will enable the various organisations to continuously update each other on relevant events in real-time and provide the basis for more efficient decision making. Aircraft operators will have up-to-date and accurate information on which to base decisions about their flights, and will be able to apply factors which are not known to ATM, such as fleet management priorities, fuel consumption figures and other aircraft operating parameters, when determining solutions.”

1.2 R & D History

CDM has been recognised in specific R&D programs conducted by the FAA, Eurocontrol and other agencies since the mid nineties. Less formal developments in the area date back considerably further.

CDM now figures prominently in many medium to long term ATM Strategic Plans including those of the FAA, Eurocontrol, Australia and Boeing.

To date Benefits from CDM initiatives have been largely limited to the Air Traffic Flow Management (ATFM) application. However, in the USA at least CDM is being developed in a wider context. RTCA special Committee 169 has developed a set of CDM scenarios and defined the resulting requirements for information exchanges between Air Traffic Management (ATM) and Airline Operations Control (AOC).

The Eurocontrol Experimental Centre has identified 22 potential application for CDM. In a similar manner, the FAA CDM program, formed in 1995 has fielded a number of prototype software applications designed to implement CDM principles.

There is a common acceptance that CDM does not have a single goal but is rather a philosophy of business that offers benefits to all stakeholders in the air transport environment.

Discussion

2.1 A Layered View of Planning Systems

If one considers three of the principle ‘actors’ in the Air Transport Industry, Aircraft Operators, Airport Operators and Air Traffic Service providers, and looks at the planning processes involved in each of these endeavours, a degree of similarity is immediately evident. Figure 1 gives a simple illustration of this principle.

Figure 1

It is logical therefore that shared and cooperatively maintained information relevant to these processes will provide benefits to all concerned. From this it is a small step to see how other ‘actors’ such as aircraft manufacturers will also have an interest in a CDM environment. The strong involvement of Boeing in ATM strategic planning activities is evidence of this.

2.2 Application Areas and Potential Benefits

The potential CDM applications listed here have been identified from initiatives with wide stakeholder representation but led primarily by ATS provider organisations. Applications are listed by primary interest according to the major stakeholder categories shown in Figure 1. Most of these applications have both strategic and tactical implications and offer benefits to more than one of the stakeholder groups.

2.2.1 ATS Providers

2.2.1.1 Flexible Use Airspace

Considered a relatively low-level CDM application and described by the various agencies in terms of specific and local operating environments. In general, a shared understanding of the airspace requirements and individual priorities of the various airspace users will allow optimum flexibility of access to airspace.

2.2.1.2 Facility Planning

Primarily a strategic CDM application which would allow ATS providers to overcome the inertia of long system development and staffing lead times by being involved as far as practicable with the strategic planning processes of the various airspace users. This application also has relevance to most if not all of the stakeholders in the air transport industry.

2.2.1.3 Capacity Planning and management

This application is perhaps the most immediate and active CDM development area. Some of the applications under consideration are:

• Collaborative flow and capacity management
• Traffic planning models
• Coordination between airport slots and ATFM
• Optimisation of Ground Delay Programs (GDPs)
• Re-routing
• User defined priority

2.2.1.4 En Route and Terminal Air Traffic Control

This application includes some of the longer term and controversial strategies as well as some areas where ‘quick wins’ are possible within current systems:

• Optimisation of Arrivals
• In flight traffic management
• Integrated arrivals and departures management
• Autonomous separation in free flight airspace

2.2.2 Airline Operations (AOC)

2.2.2.1 Schedule Management

Developing and maintaining a viable schedule is a major key to efficient airline operations. Making better use of up to date and accurate information from other stakeholders will assist in the following applications:

• Slot acquisition
• Slot Shifting
• Slot substitution on cancellation
• Disruption management and recovery

As with other CDM applications the benefits gained will justify the costs of contributing to the maintenance of the information shared.

2.2.2.2 Meteorological Information exchange

In addition to weather forecasts there is much to be gained by sharing weather information gathered from aircraft in flight as well as airport surface data. Strategic and tactical route planning is an obvious application for this data.

2.2.2.3 Flight plan and flight progress information distribution

It is axiomatic that aircraft operators have the best information on a flight’s intent while ATC has the best information on the constraints that may apply to it. Using CDM principles to integrate this information promises efficiency gains for both parties.

2.2.3 Airport Operations

Airports are also holders of constraint information. Decisions about the management of scarce ground resources influence and are influenced by events and decisions in other stakeholder domains. Specific applications include:

• Collaborative stand and gate management
• Collaborative departure sequencing
• Ground Service Equipment (GSE) management

2.3 Design and Implementation Issues

Considerable work has been completed on the nature, format and availability of the information that can be (or should be) shared effectively in CDM applications. Of particular interest is a 1998 study led by the Eurocontrol Experimental Centre which was essentially a gap analysis of the information needs of ATC authorities, airlines and airport operators. The results of this study are published in detail as part of the report on the FASTER project jointly funded by Eurocontrol and Aerospatiale. In the same year the Eurocontrol Experimental Centre published a report identifying potential CDM applications. From this work a number of issues are highlighted which will need careful attention if the CDM concept is to be implemented effectively.

2.3.1 Costs and Benefits

In assessing the cost/benefit of a particular collaborative application it is not sufficient to simply specify that the benefits should outweigh the costs for each stakeholder. Using this approach would probably ensure that CDM applications rarely progress beyond the basic R&D stage. There are two problem areas.

2.3.1.1 Performance Metrics

First there is the problem of deciding on an appropriate set of metrics by which the proposed application can be assessed. The complexity of this problem is discussed in detail in a paper presented by the FAA to the 3rd USA/Europe ATM R&D seminar in Naples in June 2000. This paper sets out a number of proposed metrics for assessing Ground Delay Program optimisation initiatives.

2.3.1.2 Return on investment

Taken individually, many CDM initiatives will offer uneven levels of cost and benefit across the various stakeholder groups. Indeed there may well be instances where the cost to an individual stakeholder outweighs the benefits available. It will be necessary to consider groups of CDM initiatives together in order to provide a realistic picture of actual cost/benefit.

2.3.2 Prioritisation and Equity

In general ATC will process aircraft on a first come first served basis within some broader priority structure eg scheduled over non-scheduled services. This regime has stood the test of time as providing both equity and efficient use of airspace and runways. Inevitably different system users will have different internal priorities. For example, it is not unusual for airlines to have particular flights, a delay to which will have far greater cost in terms of overall schedule impact than others. This implies that in order to give the best possible service the ATC system may need to consider a more complex prioritisation scheme. It is not difficult to demonstrate that in systems of this type, efficiency will decrease as levels of priority increase. This area will need careful attention if the CDM philosophy is to be fully applied in the ATFM area.

2.3.3 Security

Quite apart from the obvious issue of security and integrity of information transmission and storage is the issue of commercial confidentiality. How much information can a system user share without compromising its commercial opportunities? Conversely, how little information can a system user share and still remain an effective contributor to the collaborative nature of the process. There will undoubtedly be some intense corporate ‘soul searching’ as these applications evolve.

2.3.4 Safety

The paramount requirement of ATM systems is the safety of operations. Increasing the numbers of ‘actors’ involved in a decision making process will logically increase complexity. A rigorous safety case approach will be required to determine and treat any potential for CDM applications to impinge on the ATM risk profile.

Some CDM applications will undoubtedly suggest the sharing or delegating some safety critical decisions that would normally be the sole preserve of the ATC system (eg. ASAS applications). There is continuing vigorous debate on how such applications might be managed with regard to such aspects as transfer of responsibility and liability.

2.4 Information Management

In many enterprises there is a strong trend towards information based application development as opposed to process based development. This move is evident in the most recent examples of ATM strategic planning. The CDM concept is a prime example of this trend.

2.4.1 The ATM Information Pool

The virtual ATM information pool can be said to be made up of a number of distinct subject domains including:

• Weather information
• Air navigation and hazard information
• Time reference information
• Geographical information
• Airspace information
• Air navigation system information
• Traffic flow information
• Mission/payload information
• Flight information
• Airport information
• Aircraft information

2.4.2 Information Attributes

All of the domains will have distinct but overlapping attribute sets that define such things as:

• Volatility
• Service level requirement
• Ownership
• Access
• Safety criticality
• Commercial criticality

These attributes will in turn determine how the information is collected, stored and accessed in the applications in which it is used.

2.4.3 Technology issues

Information technology itself has evolved dramatically in the years since the CDM concept emerged. Hardware and software for the storage, transmission and manipulation of data have developed the extent that they now potentially offer a cost effective basis for the implementation of many CDM applications. Internet and Intranet systems together with web based technologies are being used to host some of the less safety critical applications such as Ground Delay Program management. More critical applications may need to wait until the ATN is more fully implemented.

2.5 A Case Study – The Evolution of CDM in the Australian ATM Context

The following case study is intended as an example of how a CDM application can be implemented in response to a specific need using existing technologies.

2.5.1 Background

The Kingsford Smith Airport in Sydney Australia services approximately 300,000 scheduled aircraft movements annually at rates up to 1000 per day over the 17 hour non curfew period. The airport is limited to no more than 80 scheduled movements per hour for environmental reasons. In fine weather the parallel runway system and ground infrastructure is more than capable of handling this demand with minimum delay provided that the schedule is evenly distributed in any given hour. ILS conditions and or strong crosswinds can reduce capacity to as few as 40 movements per hour. The majority of aircraft using Sydney Airport are on flight legs of less than 3 hours duration.

FLOWMAN

First attempts at a strategic demand management system for Sydney Airport were made in the mid eighties when demand was exceeding capacity during certain hours of the day before the parallel runway was constructed. The system developed locally, relied on published schedules. The schedule data was used to allocate landing ‘slots’ during the hours during which, depending on weather, demand exceeded capacity. Slot information was communicated to operating companies on the evening prior to operation. Capacity prediction was based on an enhanced meteorological forecasting capability derived from a special met unit co-sited in the ATC centre.

The system was hosted on the Airservices corporate intranet and serviced as a commercial application.

The system evolved to serve its purpose but suffered considerably from clustered and inaccurate schedules and communications inflexibility.

2.5.2 Slot Allocation

The opening of a parallel runway at Sydney Airport increased capacity but brought with it an environmental uproar that resulted in, among other thing, legislated demand regulation. This meant that no more than 80 runway movements could be scheduled in any hour and that no more than 8 of those slots could be allocated in any five minute period. The fact that an independent agency was required to be put in place to administer those slots gave rise in 1998 to a true CDM application, that of cooperative scheduling. The process is supported by a set of government regulations and procedures enforceable by fines.

2.5.3 Central Traffic Management System (CTMS)

This new environment paved the way for a much needed improvement in the strategic demand management procedures in place. Quite apart from the fact that the FLOWMAN system was not Y2K compliant, there was strong pressure from system users to reduce the amount of airborne holding into Sydney.

2.5.3.1 Development

The CTMS system was developed using standard commercial hardware and software. As with FLOWMAN, CTMS is hosted on the corporate computing network. This time considerable enhancements were made to the network to accommodate the special service needs of the new system.

The initial CTMS rollout was a single node system servicing Sydney alone and developing a ground delay program for all ports serving Sydney within 3 hours flying time. However, an important part of the design brief was that the system should be progressively upgraded to provide a network demand management function for the major Australian East Coast airports.

The Skyflow application, which forms the core of CTMS, is essentially a database with a variety of user interfaces and reporting mechanisms. Data elements include:

• Updated schedules
• Slot lists
• Flow programme
• Aircraft performance
• Forecast capacity values
• Typical taxy times for arrivals and departures at key airports

The system architecture is such that the various may either use the user interfaces provided or may develop their own for integration with in house systems.

A multi layered fail-soft arrangement includes alternative means of collecting processing and distributing information down to basic telephone and fax transmissions.

2.5.3.2 Business Rules

On top of the regulatory framework of the Slot system, it was important to develop a comprehensive set of business rules for the use of CTMS. These were developed after a considerable period of negotiations between stakeholders. Business rules are supported by an ongoing system of compliance reporting to the system steering group.

2.5.3.3 Data Communication

Maximum use has been made of Internet capabilities in providing flexible and robust access to the CTMS system.

The standard CTMS interface is a web-based product and is accessed through a dedicated Internet web server. Some larger users have arranged direct connections with the application server in Canberra.

2.5.3.4 Lessons Learned

The CTMS system is a good example of a collaborative decision making application that has been developed to take advantage of a changed environment. System development has highlighted the need to reinforce the collaborative aspects its operation. Strong and mutually acceptable business rules are essential. It is also essential to provide a mechanism that provides visibility to all stakeholders of compliance with those rules.

The differing requirements that different applications have of common information resources have caused some difficulty. For example, The notification of ‘off slot’ operation required by the legislated slot management scheme is in the order of plus or minus 30 minutes. This is far too course for effective strategic and pre-tactical. flow management. These problems are particularly evident in the case of smaller airlines operating short flights to Sydney.

Integration with adjacent systems requires careful planning. ATC need to manage the interaction between the CTMS system and the MAESTRO tactical flow management system in use. Airline operators have had to integrate their CTMS interface with their existing operations control functions. By and large this integration has been effective and efficient.

Conclusions

CDM applications in the air transport industry are in use now and have probably been used for some considerable time without formal classification. A more scientific approach to the development of the concept has arisen from a realisation among stakeholders that there are significant benefits to be gained. This approach will lead in turn, to the use of CDM as a fundamental principle in the development of the future ATM environment.

3.1 CDM Implications for ATS

CDM offers a philosophy by which ATS can improve its quality of service as long as this can be done without compromising safety. In many instances, particularly those in which ATS are provided by a government or quasi government organisation, it will fall to those providers to instigate CDM applications and manage the necessary data resources. Data management in CDM applications will almost certainly be complicated by the move towards corporatisation and privatisation in the provision of ATS.

3.2 CDM and the Controller

The work of ATCs will undoubtedly be affected by the increasing use of CDM principles in ATM.

3.2.1 Current and Near Term

Most of the current and near term applications of CDM concentrate on flow and capacity management issues. There are already examples where basic separation functions can be complicated by the need of individual aircraft to meet slots in order to maintain a viable operation. This problem will inevitably become more widespread as traffic volumes increase.

3.2.2 Longer Term

Longer term CDM developments will affect the way in which separation itself is provided. The development of Airborne Separation Assurance (ASAS) and Cockpit Display of Traffic Information (CDTI) are clear indicators of this trend. Cooperative separation decisions present some complex problems for ATC in both the technical and legal/regulatory aspects of their function.

Recommendations

That this paper be accepted as information material.

References

Miart Eric (Eurocontrol Brussels) Collaborative Decision Making – Presentation to IFATCA EUR meeting Zagreb Oct 2001.

ATM Strategy for 2000+, EATCHIP Doc – FCO.ET1.STO7.DEL02, Issue 2; Eurocontrol 1998.

Wetherly, J, Wambsganns, M, Kollman, K, Kupper, K – Collaborative Decision Making in Aviation Transportation: Improving Decision Making Through Shared Information – FAA 1997.

Martin, P, Hudgell, A, Vial, S, Bouge, N, Dubois, N, de Jonge, H – Potential Applications of Collaborative Planning and Decision Making – Final Report – Eurocontrol Experimental Centre 1998.

Operational Concepts and Data Elements Required to Improve ATM/AOC Ground/Ground information Exchange to facilitate Collaborative Decision making, RTCA/DO-241; RTCA, 1997.

Martin, P, Hudgell, A, Bouge, N, Vial, S – Improved Information Sharing: A Step towards The Realisation of Collaborative Decision Making – Eurocontrol Experimental Centre 1998.

Martin, P, Hudgell, A, Bouge, N, Vial, S – Future ATFM-AO-Airport Synergies Towards Enhanced Operations (FASTER) – Eurocontrol/Aerospatiale 1998.

Ball, M, Hoffman, R, Knorr, D, Wetherly, J, Wambsganns, M – Assessing the benefits of Collaborative Descion Making in Air Traffic Management – FAA, Metron, University of Maryland June 2000.

EATMS Invariant Process Model, Ref. ECTF/DP/012, Issue 1.0; Eurocontrol 1996.