New directions for Airport Collaborative Decision Making (CDM)

What exactly is CDM?

Collaborative Decision Making (CDM) is not a new concept. It is being practiced to a certain degree both in the US and in Europe, focus being on en-route in the former and airports in the latter. Mature as the concept may be, surprisingly we still see experts who seem to believe that CDM is little more than a few wise men sitting together and deciding things for the benefit of the community… Little wonder that they see a role for CDM that is strictly limited to the strategic planning phases. They seem to hang on to this view even in the face of actual CDM implementations at some airports (e.g. Munich) which are anything but limited to the strategic phase. So, what is CDM?
The concept of CDM is very simple. Decisions on all levels must be made not in isolation but based on a shared, common view of the state of the ATM network with full awareness of the consequences of the decisions on every aspect of the operation. Collaborative in this context does not necessarily imply people sitting together or working together remotely. A single person can also make a collaborative decision if the decision is based on the shared information provided by the partners and if it takes into account the impact of the decision on those partners and the ATM network as a whole.

So, CDM starts with the sharing of information and then the making of decisions based on the shared information. As we will see, this understanding of CDM opens up all kinds of interesting possibilities for the future.

What is the scope of CDM?

The current practical scope of CDM came about mainly as a result of the ATM problems prevalent in the environment in which the concept was first introduced. Getting collaborative decisions on manipulating traffic flows was a natural in the US while trying to sort out airport problems in Europe led to Airport CDM (A-CDM) as the first manifestation on the European continent. Of course this division is not black and white. A-CDM works together with the Central Flow Management Unit (CFMU), improving also the en-route situation.
The overall scope is also influenced by the scope of the shareable data available. In respect of A-CDM, this means activities on the airside of the airport.
Is this scoping correct and are the current limits inviolate? Certainly not as we will see.

Help coming from SWIM

The ATM system being envisaged by both SESAR and NextGen builds heavily on System Wide Information Management or SWIM. SWIM ensures that all information generated and needed in the ATM environment is shared and is accessible by those who need it, when they need it and where they need it. Agreed access rights make sure that the sensitivity of information is properly protected.
SWIM is the enabler of net-centric operations.
One of the most important features of SWIM is that it does not establish a scope limit to the information it is able to handle. If there is a need for information from the landside of the airport or from any other source at all, it can encompass it without upsetting the rest of the system. The result is an information driven system with unparalleled flexibility in growth and variety.
In the future, the scope of CDM will no longer be limited by the availability of information. SWIM can deliver anything the future scope of CDM might require.

The ATM environment of the future

CDM will be called upon to deliver its benefits taking due account of the two main characteristics of the future ATM system, namely net-centricity and trajectory based operations (TBO).
The aim of net-centric operations is to increase the information available to the participating entities and to ensure that data is usable by both anticipated and unanticipated users assuming that they possess the requisite access rights. Collaborative decisions are them made in the context of the interconnected community of systems, applications and people, sharing a common situational awareness of the past, present and future state of the ATM network.
Net-centric operations do entail a commitment from the participating entities to share their information and to make use of the information shared by others. In the CDM context, this commitment represents the key to achieving the necessary common situational awareness, interactions between partners on the basis of this awareness and decision making that considers all aspects and consequences of the decisions, not only locally but in the whole of the net-centric environment.
In trajectory based operations (TBO), the 4D business trajectory submitted to the ATM network by an airspace user represents what the airspace user wishes to perform. The trajectory is meant to result in the best business outcome for the given flight, taking into account the applicable constraints. The trajectories are owned by the airspace users and necessary changes to them are defined in terms of constraints with the airspace user concerned deciding the action with which he will satisfy the constraint. This will be the case in all circumstances except short-term tactical ATC and airport operational interventions. By managing the trajectories with the business outcome being taken into account and giving a closer role to the airspace users in the management of their trajectories, the overall cost effectiveness of the ATM network can be increased with also some capacity gains possible through a reduction of controller task-load. Trajectory-based operations represent a move away from the traditional clearance and airspace based paradigm. It requires that the need for data, services, etc. be expressed, with the trajectory (rather than airspace) being central in the descriptions.
In this approach, the trajectory is seen as a continuous line in space and on the ground that commences at a selected point in time, e.g. the first flight of a given aircraft on a given day, with its commencement point in space or on the ground being defined by the position of the aircraft that will execute that trajectory. With the full trajectory being considered, rather than only a segment like en-route-to-en-route, all the airports of this trajectory, involved as destination or alternate, come into view. They then become information consumers and information suppliers in respect of the given trajectory. The supplied and consumed information is managed in the net-centric environment using SWIM, with the trajectory acting as the focus (the “glue”) of the information and necessary decisions. Since TBO is not an airspace based concept, in this view of the trajectory, the airports (like all other entities that are concerned by that trajectory) become integrated in the complete ATM network as they manage the trajectory. This is true also when the aircraft is stationary on the ground. The trajectory continues to evolve along the time axis while it is stationary in the spatial sense. It shall be noted that even such a “stationary” trajectory is consuming resources and as such, impacts capacity. This idling section of the trajectory contains also the conceptual break point in time where the aircraft allocated to the trajectory becomes the “next flight” it will be called upon to perform.

It is obvious that if we wish to appreciate the actual impact of the evolution of a trajectory on the ATM network as a whole, we need to look at the trajectory taking all the segments it may contain in a given day (or other appropriate length of time). In this view the only limitation is the length of the known (published) segments but it is not suggested that trajectories should be considered beyond what is operationally sensible.

Is there a 5th dimension of the trajectory?

In the new ATM paradigm, the concept of the four dimensional management of trajectories plays a pivotal, fundamental role. It is this 4D (3 spatial and 1 time) view that brings the necessary improvements in accuracy and hence predictability.
Based on the SESAR concept of “business/mission” trajectory one might consider defining a fifth dimension of the trajectory. This fifth dimension is the economic value of the trajectory to its owner, the aircraft operator. Assuming that a number of conditions are fulfilled as described below, it can be used to collaboratively prioritise trajectories based on their economic value. This value is very important as it can drive several aspects of the operation, like requests for moving up or down in sequences, user established prioritisations, solutions to constraints, etc.

The economic value of the trajectory evolves from a hypothetical figure when it is first published, to one that can be expressed in actual money terms by the time the trajectory becomes the so called reference business trajectory. The overall value accrues from a number of different elements which include the other flights a given aircraft is scheduled to carry out, the number of high-yield passengers on a flight, the number of transfer passengers and the kind of transfers, etc. Clearly, some of these factors become known only fairly late before the flight takes place and hence the precise value of the fifth dimension may not be known until shortly before push-back. The value is obvious for the aircraft operator who is the owner of the given trajectory but there is also an identifiable value for the network as a whole. For instance, major distortions to a high value trajectory are likely to trigger mitigating actions sooner than in the case of a trajectory of lower value.
Collaborative prioritisation based on economic value completes the picture by making visible the commercial dimension of the operation. Important considerations are that this economic dimension must be defined and used in a way that does not penalise any operator or generate a competitive advantage for the owner or others. It must serve solely as an aid for better, more focused management of the trajectories, an additional element in collaborative decision making. Also, the economic value of a trajectory will never override tactical ATC or airport operational decisions even while they must take the fifth dimension into account whenever possible.
The fifth trajectory dimension is not an agreed concept element or something that is essential for the future ATM network. It does however require effective CDM to work well should a decision be made to try it out.

The need for a process based, service oriented view

Current CDM implementations and the concept itself are based on defined processes, A-CDM being anchored in the aircraft turn-round process. Service orientation in air traffic management is however new, as it is indeed new for the SESAR concept itself.

Service orientation is a means of separating the business aspects and the IT aspects of the enterprise, thereby enabling the business aspects to drive IT rather than the other way round (as is often the case in legacy systems).
From the perspective of ATM and airport collaborative decision making, we might say that there are two main processes to be considered. One process is the management of the trajectory the other the management of passenger and baggage flow. From the perspective of an airline or handling company, there are several other important processes (like crew to aircraft, fuel and catering to aircraft, etc.) but from a CDM perspective, these can be ignored. Their effects will be incorporated into the trajectory management process in the form of eventual “distortions” to the trajectory (mainly but not exclusively the time dimension).
In order to be executed, the operational processes require various services. They also fall into several categories, related to one or the other (or several) of the processes mentioned above. Later on in this document I will describe certain novel services that will become possible as a result of the net-centric environment. Obviously, there are certain services that exist already to-day (even if they are not called services in the legacy environment) and even more services will be defined in the future as the full power of the SWIM based net-centric environment takes hold.
The services in turn support end-user applications which are the operational interface to the outside world (for humans) or other systems, as appropriate. Decision making and two-way access to information is realised via the end-user applications enabled by the services.

Airside/landside division, does it make sense?

The airside/landside division at airports is not something that was introduced in reaction to an operational need. It has everything to do with security, a vitally important aim in itself but not something that should necessarily limit the scope of collaborative decision making. The fact that the definition and actual location of the airside/landside boundary changes from place to place, further (and unnecessarily) complicates things.

We will recall that currently A-CDM is limited to airside processes and there is considerable opposition to considering what is happening on the landside. In other words, an artificial, more or less administrative boundary is limiting the full exploitation of CDM’s potential. This will have to change in the future for two reasons: in the net-centric environment there is no shortage of information from anywhere, including the landside; secondly, as will be shown, the whole idea of airside and landside should be made invisible from a CDM perspective.
Of the two main processes of concern to A-CDM, passenger and baggage flow extends on both sides of the airside/landside boundary, irrespective of its actual location at any particular airport. The processes require various services and we concerned with the decisions associated with the delivery of those services. The sequence in which the services are delivered does depend on local circumstances (e.g. the Passenger Screening service may be applied to the passenger flow just after check in or shortly before boarding) but the sequence is not necessarily affected by the location of the airside/landside boundary.
The conditions and restrictions applicable to those delivering services may be different depending on where the service is delivered but this does not affect the service itself.
It is evident that the existence of a designated airside and landside on any airport is of no consequence for the processes CDM has to consider. From the CDM perspective, the airport can be as a single continuum in which the processes are executed through the application of the required services, without the processes and services being divided by an airside/landside boundary.
This “single airport continuum” approach does not preclude the existence of an airside/landside boundary if maintaining or establishing one is found necessary for other purposes but it should never be allowed to be a limiting factor on the effective execution of the various processes.

Mr. Smith proceed to your gate or we will off-load your luggage…

At many airports, the above text has become a mantra repeated endlessly and one can only hope that the Mr. Smiths and others like him still listen to what is being said. Clearly, the problem of passengers getting “lost” between check-in and the gate has become a major issue.
True, you can always off-load the bags and leave the passenger behind but this is not the most customer friendly solution not to mention the fact that even with a strict and timely decision a departure delay is difficult to avoid.
For CDM aficionados however the above announcement indicates a brave new field where CDM, building on the basically unlimited scope of SWIM, can introduce totally new services, some of them completely automated, to get a better handle on the flow of passengers and their luggage.
Part of the problem of straying passengers comes of course from a basic conflict of interest between airports and their airline customers. With more and more of the airport revenue coming from concessions and the supermarket and mall style shopping areas popping up at airports everywhere, it in the airport’s interest to keep passengers in the shopping areas for as long as possible in the hope of enticing them to spend money on everything from expensive sunglasses to even more expensive sandwiches. And they do, with passion and enthusiasm… so much so that a lot of them almost forget the purpose of their coming to the airport in the first place… to catch an aircraft.
Airlines of course would like to see their customers come to the gate in a timely manner, not too early but certainly not too late… things we humans are not that good in without effective guidance.
But if we go beyond the shopping mall and consider how many random events there are that can potentially impact a passenger’s ability to arrive at the airport on time (an accident on the access road, industrial action of the train drivers, traffic jams caused by ice on the road to name just a few), we might say it is a wonder that so few of them arrive actually too late…
However, the fact remains that an airline has precious little knowledge of, or control over, what its passengers are doing until they discover that one or several of them are missing when it is time to board. With home check-in commonplace these days, the time span of the big unknown has only increased.
Freeing CDM from its shackles and limited scope can help here and the result is a major improvement of operational predictability and efficiency.

Embracing new CDM partners

The upwards open scope of SWIM should give us a hint as to how the scope of CDM can also be seen as limited only by the actual need for decision making information.
To properly appreciate the potential of a true SWIM environment, we do need to adjust our thinking. An information driven system needs a major jump in how we see the world.

In this expanded scope CDM and SWIM environment, a lot of new, raw data is suddenly available. On the user end, we then have end-user applications that put the data into context, creating information that we can use in any number of ways.
Just imagine that you have the railway company, the taxi companies, the police, the motorway maintenance outfit, the GSM operators, the airport parking garage, etc., etc. all on board as CDM partners, sharing their data. End-user applications of the airport, of airlines or the taxi folks themselves, are constantly shifting through the information and draw the appropriate conclusions… Lots of GSM calls going to a single airline number? Fewer than average taxis heading to the airport at a given moment? Parking garage full? Or just empty?? Trains running late? Queues unduly long at the security check point? Or there is no queue at all when in fact there should be one?
It is possible to deduce a lot of planning information from these seemingly disconnected elements and if they are available, there will be developers who create the applications needed. The nice thing about this is that we are not talking about a one way street here. If a given conclusion is strong enough and there is still time, passengers can be warned via sms and push email to come earlier or do other things to ensure proper flow towards the aircraft. Some airlines already offer an sms warning when home check in opens for their flight. The new CDM scenario will open a much more effective range of interaction with the passengers.
The future system must embrace all CDM partners who have useful information to share, regardless of how close or apparently remote they are from traditional ATM.

Examples of novel services

In the following a number of novel services will be described without any claim of being comprehensive or thoroughly validated. They are given here as examples which may very well be implemented in the not too distant future. The selection focuses on passenger flow. Similar services could be defined for example for the luggage flow also.

Airport Status Report Service

Under trajectory-based operations (TBO) it is no longer sufficient to have an indication of when delays are likely. A delay is the representation of only one aspect of the 4D trajectory (time) and distortions in the other dimensions may be even more important under certain circumstances. The Airport Status Report Service provides a trigger that can be graduated to represent potential trajectory distortions in different dimensions or combinations thereof and as such, is a much finer tool to trigger intervention than simple congestion warnings. The benefits would appear even during the transition period towards full TBO, an important consideration in the quest for early benefits.

This service is coupled with the Airport Operations Plan where available and is also feeding into the Network Operations Plan.
The Airport Status is an index developed on the basis of appropriate algorithms taking present and future circumstances into account, each having a weight attached to it. Each index/airport combination indicates the degree to which the known business trajectories would suffer an impact. The indexes are updated dynamically but with sufficient latency to ensure stability of the system.
The index can serve as the primary trigger for users of the airport to consider eventual collaborative modifications to their trajectories. The modifications are coordinated and agreed through the planner end-user applications. These facilitate real collaborative decisions on the most effective, minimum impact changes.
The agreed changes are published and the Queue Management Service evaluates their impact. The aim is to achieve, if at all possible, a new index for the airport, representing a lower level of overall trajectory distortion.
The Airport Status Report Service would be subject to all the characteristics and limitations that its users agree to, including for instance the maximum acceptable distortion to any one trajectory.

Timely Passenger Delivery Service

Passengers who fail to arrive at the boarding gate on time are a significant source of disruption especially if checked luggage needs to be off-loaded as a result. Even if a given late passenger has no checked luggage and hence the flight can depart without him without delay, the resulting situation of an irate customer having to be re-accommodated is something that is better avoided.

Innovative information management techniques, building on the net-centric environment, can be used to motivate passengers to proceed to the boarding gate in a timely manner. The means available to influence behaviour (both desired and incidental) while the passenger is in transit from his or her off-airport departure location are limited. The means expand however once the passenger is on the airport and in particular when he or she enters the passengers-only area.
The business model of most airports prefers to keep passengers in the concession areas to maximise revenue from the concessions. This preference also helps in managing passenger flow towards the gates, avoiding congestion in the gate areas. The tools used to influence passenger behaviour are rather simple, usually not being more than posting the departure gate numbers a short time before the designated boarding time. Passengers do not receive an active reminder that the gate has been posted. The result is often missing passengers at the gate.
While the interests of the airlines and the airports appear to be at odds in this context, the net-centric environment and its expanded partner scope offers a service based solution to the problem. Reconciling the difference in a sensible manner is important also because time spent in the concession area is an important part of the passenger experience and keeping it pleasant while also ensuring timely departures is in the airline interest too.
The Timely Passenger Delivery Service is a service to be offered by the airport operator to airlines on terms to be agreed locally. It is an example of the practical application of the collaborative decision making principles in the expanded scope of CDM in the future.

With this service, the airport undertakes to introduce and maintain effective, agreed measures designed to motivate passengers to move towards the boarding gate on a timely basis, the motivation being unobtrusive but geared to the urgency of the particular situation.
In practical implementation, an airport providing this service would need to make full use of the net-centric environment to gather information from hitherto uninvolved partners like the Passenger Screening services (queue situation), concession operators, etc. The airport would need to ensure their becoming partners in the net-centric environment and also the definition of hitherto unused dynamic information items (e.g. flights closing soon, etc.).
The measures that can be taken as part of the content of this service might include:
• Individual tagging (RFID) of passengers and personalized voice and visual announcements
• Scanning the boarding passes at the entrance of all major concessions (something usually done on exit anyway) and generating an immediate warning to the passenger if flight departure is within a set time period
• Having dedicated displays showing those flights still missing passengers and which will be closing soon
The list of flights to which passengers need to hurry or risk being left at the airport and having their luggage off-loaded can then be used in several innovative ways (some of which were listed above) to trigger the passengers to move. The same information with slightly different time horizon can be used to even catch passengers who have not yet arrived at the airport. A generic version of this type of service is available already where, on request and on passenger cost, a text message is sent on status changes of the flight concerned. The Timely Passenger Delivery Service is more targeted, focusing on ensuring on-time departure rather than just providing raw information.

In support of this service, the concept of Control Over Passenger Behaviour Points could be defined. The parts of the passenger flow process, delimited by these points, are designated according to the level of uncertainty of the passenger position along the process and the level of influence the provider of the service has on the behaviour of the passenger. This latter is dependent also on the means available in the particular circumstances for exerting the required influence. As an example, the period while the passenger is in transit to the airport is considered as being of high uncertainty/low influence/few means. A passenger in the concession area is on the part of the process that is considered as being of medium uncertainty/high influence/multiple means. The points delimiting the sections are identified and agreed by the CDM partners.
The concept of Control Over Passenger Behaviour Points enables the provider of the service to properly estimate the effectiveness and cost of the service required along different parts of the passenger flow process and to generate an offer accordingly. The users of the service can then decide on an agreed, common basis whether the service is cost effective for them and which part(s) of the service they want to sign up for.

Passenger Flow Information Service

The passenger flow process is subject to several possible disturbances at different points of the flow. The effects of a disturbance depend to a large extent on the organisation of the process which determines where the source of the disturbance is located along the process. Tightened security or a scanner failure will impact the process differently if the passenger screening is centralised or boarding gate based.

With the expanded scope of future CDM, influences as diverse as airport access road conditions and train driver industrial action can be included to generate a good picture of the evolution of the passenger flow process.
The time scope is not limited to the present or immediate future. Planned industrial action several weeks or even months in the future are also considered by the service.
In terms of granularity, the service aims primarily to generate state (current and future) information based on the global aspects of the passenger flow process. This means that the basic Passenger Flow Information Service does not differentiate between the effects on the operation of individual partners. Only the (potential) global distortions are signalled. There is no reason however why a partner should not request a more detailed version of the service and why it should not be provided under the right conditions.

The information generated by this service is published into the shared information space to be made use of by different end-user applications. These may be as simple as a warning displayed at a working position or they may take the form of more advanced, intelligent applications that are able to interpret the state information and initiate defined, context sensitive actions. The actions concerned will be mainly those aimed at mitigating the effect of the distortion (e.g. opening additional Passenger Screening lanes, hiring alternative transport, etc.) or, if that is not possible, start action aimed at minimising the effects of trajectory distortion that may potentially be caused by the disturbance. The users of the service may also be systems which support other operational processes.
The Passenger Flow Information Service will require the implementation of facilities that can monitor the passenger flow at different, required points, correlating the movement data and trends to generate the appropriate conclusions.
It is important to remember that passenger tracking may involve issues related to the need for protecting personal data. Such issues will have to be addressed and properly resolved.

In conclusion

It is a fallacy to pretend that the predictability of ATM operations is not affected by events, information and processes that have hitherto been disregarded for various reasons. The current land-side services are among those that are meeting opposition when it comes to defining future CDM developments.
At the same time, the fine tuning of CDM in the traditional sense will soon reach the area of diminishing returns and hence further improvements will not be found in those areas. Expanding the scope is the most efficient manner of finding new improvement potential.
We have seen that SWIM and net-centric operations open up a whole new stage for CDM enhancements. Even the classical airside/landside division may very well disappear from the CDM perspective as it has no usable meaning in an information driven environment.

CDM is a very powerful concept and has no built in limitations. If information is made available from traditional as well as new partners, quality decision making can grow in line with the growth of the partner scope. The net result is much improved overall decision making in ATM itself.
Note: Part of this material originally appeared in the Level 4 CDM (L4CDM) study (2008) produced for EUROCONTROL by BluSky Services, Altran and Slot Consulting Ltd.