The New Model for Air Traffic Control

Discussion of Aspects of The New Model

Contrails in Stormy Sky

We are pretty nearly up against the stops on capacity in one or two areas of air traffic control and there is urgent need to adopt more modern displays, install data processing systems and system automation, to provide the expansion which increased controller work-load cannot indefinitely sustain.

Air Vice Marshal J B Russell, Controller of National Air Traffic Services, May 1967

Today in many parts of Europe, air traffic control and infrastructure are the limiting factors. Hubs and slot are among the most valuable resources. I assume that everybody agrees that the present capacity of ATC will be reached in 10 years or less.

Capt Werner Knorr, Lufthansa V-P Operations, Flight International, 28 April 1999

I say, get people out of the process to the extent possible. Having worked airplanes - and lots of them - in my 30 plus years as a controller, I have an appreciation of the fallibility of human controllers. I know that I certainly made my share of bad calls, poor decisions and lousy communications - and I was the WORLD’S BEST CONTROLLER. Also, although I loved the job, it is not really something I believe humans do very well. I’ve seen deals caused by bad moods, distractions and arrogance. Please don’t write those human traits into your programs!

A Senior United States FAA Controller, 15 June 2005

A World-Wide Inter-Operable ATM Network

The vision for the New Model Concept is a world-wide inter operable network of computers continuously updating the complete solution to the air traffic situation. This network will be overseen by controllers who will use their skills and experience to guide the network towards better strategic solutions. Real-time data concerning flights will be available to various connected users such as airport and airline operators. An aircraft taking off will be quickly coordinated through the network to the destination airport and its expected landing time will be available within a few seconds. Aircraft will be guided according to their needs and capabilities and while most instructions will be communicated electronically the air ground voice channel will be maintained.

Affordable Air Traffic Control

The evidence for the low cost of the New Model Concept is the Single Sector Demonstrator itself. It has no deficiencies or limitations that could not be rectified with just a few man-years of effort. It has already been tested in a limited multi-sector configuration and every component of the concept has already been coded and tested in some form in an existing operational environment. A network of just ten desktop computers would be capable of controlling the London TMA in a simulated environment. A network of fewer than 1000 computers could meet the basic needs of area control in the 300 or so FIRs world-wide. After costing in all other systems, such as radio and support information displays, main centres, even in busy European areas, should cost no more than €50 million. Smaller centres, with less redundancy, would be ideal as disaster recovery and business continuity solutions.

The Future Shape of Air Traffic Management by Ashley Smout CEO of Airways New Zealand

A Simple and Reliable Human Computer Interface

The New Model Concept breaks new ground by suggesting that the HCI can be made much more reliable by having the controller acknowledge unique entries rather than choose items from a menu or enter alphanumeric data on a keyboard. Because each RT advisory is unique and because the default action is to acknowledge without change the computer can assume the input before it is actually made. There is also much less chance of a controller incorrectly reading a single item than there is of a mistake being made while using menus or a keyboard for data entry. We should be approaching a time in any case when safety critical instructions containing cleared levels are not passed by radio alone.

Fuel Saving Through Timely RT Instructions

One feature of the New Model Concept that is unusual is that, because the focus is not on problem solving, the system is able to manage all aspects of the RT Advisory List preparation and presentation. Instructions for further climb or initial descent from the cruise are effectively bookmarked and presented to the controller in a timely manner. It can be expected that many trajectories will be improved and that fuel savings will result from this approach. Also, because the New Model always calculates these advisories, even a controller who is operating manually can still refer to these instructions from time to time. The ability with a New Model system to predict and manage the use of all resources, particularly those at airports, means that aircraft could carry less contingency fuel and more payload. The New Model Concept also recognises the possibility that under maximum capacity conditions the RT Advisory Instructions might even be spread out temporally so that the controller is never faced with too many instructions in too short a time.

Continuous descent approaches (CDAs) are also recognised as offering significant fuel savings and have been trialled on several occasions. The New Model Concept can compute multiple trajectories to any arbitrary degree of complexity without loosing the picture. Freed also from the need to create routes via fixed reporting points or through fixed standing agreement levels The New Model is the perfect technology for accessing these fuel and time-saving benefits.

Mike Berners-Lee and Small World Consulting

Consistent and Comprehensive

The New Model Concept is the first approach to be fully consistent in that it proposes the same fundamental mechanisms for supporting the controller in every ATC situation. Although the presentation to the controller may need to be different and the resolution algorithms may have to be adapted the concept can be applied equally well to radar or procedural environments. Thus Area Control, Military Area Control, Terminal Control, Approach Control, and even Tower and Ground Movement Control facilities could all be assisted by inter-operating implementations of the New Model Concept. A radar or ADS surveillance input is not necessary for the New Model to work and in a procedural environment regular position reports combined with dead-reckoning would be used, just as they are today, to predict trajectories. From within an airfield tower even aircraft in the circuit can be handled if their positions are occasionally confirmed by visual contact.

Modeless Operation

A significant advantage of the current operational concept is that it is modeless. The controller operates in an endlessly repeating loop of monitoring, solution generation and instruction application. Only occasionally is it necessary to change the information displayed. Almost no mental energy or time is lost by the controller in changing contexts. This modeless approach is also part of the New Model Concept in that there are no menus of windows to be opened up and examined under different circumstances. Everything necessary is contained in the RT Advisory Form. Having established this principle the New Model Concept does not prohibit the use of other displays for use when time permits. The Separation Monitor is one example of an optional display. Anyone who doubts the fundamental logic should consider that in the London TMA an aircraft requires an instruction on average about once every 3 minutes. For ten aircraft that means that each instruction has to be devised, issued and acknowledged in 18 seconds. There is just no time to examine different windows when the traffic levels are high.

No More Conflict Alerts

The idea of conflict alerts which seem so central to most other controller productivity tools are banished from the New Model Concept. This is because the existence of two predicted trajectories that do not maintain separation must indicate a discrepancy in the trajectory descriptions. One or other (or both) of the trajectories must be adjusted or the air traffic system will have failed in its primary task. And for as long as any such discrepancies are present other controllers and other systems will be drawing false conclusions and producing erroneous output. This effect, when multiplied across all sectors, guarantees that human centric conflict resolution is not the way forward. It should be clear that the only way to resolve this situation is to have the computer calculate suitable conflict-free trajectories. The human controller just cannot think or enter conclusions fast enough. Even imminent conflicts are not indicated. Instead, the New Model Concept calculates immediate advisory manoeuvres for each aircraft and may be unique in proposing that this approach will work just as well for multiple aircraft as it does for two. The New Model Concept is Separation-Oriented rather than Conflict-Oriented.

Self Testing Capability

Another unique feature of the New Model Concept is that an implementation can be tested fully automatically. Another smaller system is needed to model a population of aircraft and simulate the surveillance system input. If suitable means are included to create endless scenarios of traffic then the two systems can be connected together and left to run. If we assume that there is a recording and exception reporting capability then the systems can run completely unattended. In fast time many months of activity could be tested in a few days. For the first time an ATC system could be tested in a manner similar to the testing of an aircraft airframe in a fatigue rig. The algorithm can also be tested against live operational data without supervision as it is easy for the system to detect and report on any situation for which there appears to be no resolution. Few ATC systems in the past have been tested this comprehensively.

Allied Applications

The New Model Concept is ideally suited to support a number of allied applications in a cost effective way.

Phased Implementation

With its ability to operate autonomously the New Model Concept could be introduced into service in a completely evolutionary manner. There is no need to change the current concept of operations overnight and the risks associated with such steps are almost entirely avoided. As provided for above, an early implementation of the New Model would simply shadow the live operation with its performance being assessed by controllers and systems engineers. Where differences between the two environments were noted improvements to the resolution algorithms might be made. Only when the performance of the system was considered to be adequate would trials be conducted on lightly loaded sectors. Following such trials to gain confidence and experience the maximum number of aircraft permitted in a single sector might be cautiously increased. By proceeding in this way the benefits of full deployment should be achieved within the shortest possible timeframe.

The New Model or Alphabet Soup?

Over many years the starting point in the search for a means of automating the Air Traffic Control task has been an analysis of the work done by the controller and a breakdown of the typical flight into phases. Although this approach is fundamentally correct in engineering terms it does depend for its success upon the establishment of a single coherent model of the problem. No such simple model has ever existed for the ATC toolset and a bewildering array of Windows, Concepts, Managers, Tools, Facilities, Alerts, Monitors, Warnings and Advisors has resulted from the work done so far. No human controller can ever be expected to translate the multiplicity of outputs into a few coherent instructions for each aircraft in a safe and timely manner. Also, crucially, clear functional interfaces between the components never seem to be defined.

And even if the various enabling technologies such as AGDL, ADS, TCAS, MODE-S are set aside then just a few of the ideas are:

Arrival Manager (AMAN)
Departure Manager (DMAN) )
Cockpit Display of Traffic Information (CDTI)
Conflict Resolution Advisor (CORA)
Electronic Flight Data (EFD) or Electronic Paper Strips (EPS)
En-Route Manager (EMAN)
Final Approach Vectoring Tool
Flexible Use of Airspace Concept (FUA)
Free Flight (FF)
Gate-To-Gate (GTG)
Interoperability (IOP)
Medium Term Conflict Detection (MTCD)
Metering & Sequencing
Minimum Sector Altitude Warning (MSA) or Minimum Safe Altitude Warning
Monitoring Aid (MONA) or Flight Path Monitoring (FPM)
Multi-Sector Planning Tool (MSP)
Runway Manager (RM)
Separation Monitoring Facility (SMF)
Short Term Conflict Alert (STCA)
Surface Manager (SMAN)
Traffic Load Monitor (TLM)
Trajectory Prediction Tool (TP)

The New Model aims to provide a single unified conceptual framework to realise the true goals of all of these entities.

Central Flow Management - CFMU

The Eurocontrol Central Flow Management Unit in Brussels (CFMU) has been one of Europe’s success stories since it came into full time operation in 1996. It works by first requiring every flight departing from within Europe to request a take-off slot and, in effect, a route slot by filing a FLIGHTPLAN. The maximum capacities of airports and airspace sectors at any moment are also notified to the CFMU by the regional air traffic authorities. By comparing the demand with the available capacity and then by adjusting the slot times or filed routes the computers calculate an acceptable mapping of one to the other. The output is a slot time and an agreed route for each flight. It is of particular note that this is simply resource-based traffic management and that conflicts are not considered until the tactical phase.

The New Model Concept proposes that this mechanism can now actually be extended right down to the tactical controller level. The only difference in the future may be that, with significantly higher peak capacities enabled by the computer support within each sector, CFMU may ultimately only have to check that runway space is available at the departure and destination airports before issuing a slot time.

Eurocontrol Library Brochures

The Myth of Optimisation

Many system specifications ask that the resulting computer program should optimise some features of the outputs. Rarely, however, are the specific parameters that should be optimised actually identified. For an ATC system it is nonsense to require that trajectories are optimised since optimisation can only take place within a specified domain and with respect to a single parameter. For a projected trajectory the parameter might be time taken or fuel consumed but both of these cannot be optimised together unless a conversion function is supplied. With such a function a resultant quality factor can be computed for different trajectories and the trajectory with the best quality factor will win. Multiple individual flights may then also be optimised if they do not interact. If flights interact then it is also necessary to specify how flights should be grouped together in space and in time. These things can be done but optimisation is not the trivial task that it may at first appear. But, once the difficulties are understood, with suitably specified parameters a New Model system would be perfectly able to assess the efficiency of single trajectories or of entire airspace solutions. No other interactive controller tool could ever perform such computationally intensive optimisation.

Intervene as Little as Possible but as Much as Necessary

It would seem obvious that this maxim should be a part of any air traffic control solution. This maxim would permit free flight where no conflicts occur and a gradual change to more positive control where traffic densities are higher. A New Model System could adapt seamlessly to any changing situation. There should be no need to apply geographic or temporal boundaries to air operations where all aircraft agree to participate. In the future, when all aircraft are equipped with some kind of link to the ground, it is possible to imagine the world’s airspace with no fixed boundaries. Trajectories, even for General Aviation and VFR flights, would be set up on request and blocks of airspace would be reserved as required. Aircraft that consumed fewer resources, perhaps by having a RNP-1 (Required Navigation Performance) capability instead of RNP-10, could receive more favourable treatment. Non-RVSM (Reduced Vertical Separation Minimum) equipped aircraft would still be able to use RVSM airspace at quiet times. The New Model Concept is a flexible concept.