Abstract
Air traffic flow management is supported by a highly distributed work system in which airline dispatchers and Federal Aviation Administration (FAA) traffic managers must coordinate. To support asynchronous coordination between a dispatcher and a traffic manager, the FAA has developed software that allows the flight operators to submit multiple, prioritized alternative flight plans. This set of alternative flight plans, submitted along with a filed route, is referred to as a Trajectory Option Set (TOS). And some airlines have now developed initial versions of software capable of generating and submitting such TOSs. This paper reports on cognitive walkthroughs with 5 dispatchers and 3 traffic managers on 5 scenarios designed to evaluate the operational concept, procedures and supporting FAA and airline software. The findings provide guidance for application of the concept of collaborative constraint propagation to support distributed work, as well as 42 recommendations for enhancing associated procedures and supporting software designs.
Keywords
Overview
Air traffic flow management is supported by a highly distributed work system in which airline dispatchers and Federal Aviation Administration (FAA) traffic managers must coordinate to manage the flow of air traffic subject to aircraft constraints and consideration of the business models of the flight operators. One of the challenges of this system is the need for dispatchers and traffic managers to make routing decisions at different lead times relative to a flight’s departure. Note that the traffic manager resides at an FAA en route Air Route Traffic Control Center (ARTCC), such as the Cleveland ARTCC or Washington ARTCC, or an airport traffic control tower. The dispatcher works at a centralized Flight Operations Center for an airline. (This operational concept and the supporting software can be utilized by business and general aviation operators, but the initial application will focus on airline operations.)
Today, flight operators file a single flight plan for each flight, even when there is a potential need for an FAA traffic manager to reroute that flight while it is still on the ground to avoid a dynamic constraint such as weather or traffic volume and expedite its departure. Furthermore, this filed flight plan is erased from the system and is no longer accessible to that traffic manager once a predeparture reroute has been made.
To support asynchronous coordination between the dispatcher and a traffic manager regarding such predeparture routing decisions and to retain memory of filed flight plans, the FAA has developed software that allows the flight operators to submit multiple, prioritized alternative flight plans. This set of alternative flight plans, submitted along with a filed route, is referred to as a Trajectory Option Set. And some airlines have now developed initial versions of software capable of generating and submitting such TOSs.
In preparation for the realization of this new operational process, a cognitive walkthrough was conducted with dispatchers and traffic managers in order to determine whether any final refinements are required for the supporting software and associated procedures. This paper presents the results of these cognitive walkthroughs and discusses the broader implications for collaborative constraint propagation as a design concept to support distributed work.
Approach
Cognitive walkthroughs (Stanton et al., 2005) were conducted for five scenarios focused on the use of the Predeparture Reroute tool (PDRR) by FAA air traffic managers to make predeparture reroutes for flights. As defined in FAA Air Traffic Publication 7210.7:
“Pre-departure reroute (PDRR) is a capability within TFMS [Traffic Flow Management System] that enables ATC [Air Traffic Control] to quickly amend and execute revised departure clearances to mitigate en route constraints or balance en route traffic flows. This capability is especially beneficial during periods of severe weather when departure routes are rapidly opening and closing.” “The ARTCC TMC [enroute Center Traffic Management Coordinator] uses TFMS route amendment dialog (RAD) to define a set of aircraft-specific reroutes that address a certain traffic flow problem …”
This tool makes it possible for a flight operator to submit not only a filed route but also a TOS for each flight. The TOS is a prioritized list of alternative routes for which a flight has been fueled in case a predeparture reroute by a traffic manager is necessary. PDRR is already in widespread use by FAA traffic managers to reroute traffic, but procedures using TOSs have not yet been used outside of limited tests.
Participants
The participants included five individuals with experience as airline dispatchers and as airline ATC Coordinators (one representative each from American Air Lines, JetBlue, United Air Lines and two from Delta Air Lines). These individuals had an average of 22.5 years of experience as dispatchers (range: 14-40 years) and an average of 14.4 years of experience as ATC Coordinators (range: 6-30 years). Note that the timeframes for these two positions overlap, as ATC Coordinators continue to actively dispatch as part of their jobs.
The participants also included three FAA air traffic controllers. One had 10 years of experience as a controller and 8 years as an air traffic manager, with experience at the Memphis ARTCC, the Air Traffic Control System Command Center (ATCSCC) and two airport ATC towers. The second had 20 years of experience as a controller and 10 years as an air traffic manager, with experience at the Cleveland ARTCC and one airport ATC tower. The third had 13 years of experience as a controller at the Washington ARTCC with experience as a controller in 28 sectors, as well as 2 years of experience as a traffic manager.
Scenarios and Procedure
Four of the scenarios focused on dynamic weather constraints, each highlighting some different consideration in the use of a TOS. One focused on the use of TOSs to support coordination during space launches.
Sample Scenario
As an example of scenario contents, Figure 1 shows the advisory distributed by the ATCSCC prior to the timeframe when the airline dispatchers would be submitting the filed routes and TOSs for flights in this sample scenario. Figure 3 indicates the departure fixes in the Washington ARTCC relevant to this scenario. Figure 2 shows a sample TOS for a flight in this scenario. Figures 4 and 5 show the actual weather and two hour forecast for this scenario. Additional displays were provided showing the 1 hour forecast and national weather displays.)
Advisory informing flight operators of potential departure constraints.
Sample TOS option list (DCA-LAX).
Major Washington, DC airports (red) and shared departure fixes (5 letter codes).
Actual weather at 1900Z.
Forecast at 1900Z for weather at 2100Z.
Prior to the walkthrough, the dispatchers were provided with the advisory (Figure 1), the actual weather at 1900Z and the forecast weather at 2000 and 2100Z. In addition to the zoomed in views of the weather shown in Figures 4 and 5, they were provided with corresponding weather displays showing national views. They were asked to produce the TOSs for their airlines for a number of destinations with flight departing from Washington Center airports. These results were included in the walkthroughs.
In addition, prior to the walkthrough a Washington Center traffic manager (not the one participating in the walkthrough) was asked to indicate what he would expect in terms of traffic management (departure fix stops; Miles-inTrail restrictions; ground stops; ground delay programs; airspace flow programs) based on the actual weather at 1900Z and the forecast of the weather at 2000 and 2100Z. This traffic manager indicated that he would expect traffic departing via the following fixes to be stopped for this time period (1900-2100Z): WOOLEY, JERES, BUFFR, MCCRAY, RAMAY, OTTTO, CLTCH, SCRAM, JDUBB AND SCOOB. He also indicated that COLIN would likely be stopped except for some departures to airports internal to Washington Center, as most of the airspace to the west forecast to be open would be needed for arrivals. This information was also included in the walkthrough to generate discussion.
The participants viewed the sequence of slides (advisory, weather, airline TOSs, expected Washington Center restrictions based on the weather) as a group and were asked questions following each slide in the sequence. This same basic sequence was followed for each of the five scenarios.
Results and Discussion
42 recommendations were developed based on the results of the 5 cognitive walkthroughs (Smith et al., 2023). This encompassed recommendations to better define procedures for both traffic managers and dispatchers, to refine the designs for the software (PDRR for traffic managers and TOS generators for dispatchers), and to include certain content in the training of traffic managers and dispatchers.
The traffic managers and dispatchers were unanimously in agreement that this operational concept merits implementation. One traffic manager noted: “We could get more done with these TOSs, benefiting the carriers without TOSs as well.”
A dispatcher noted: “If providing this information could help the traffic managers to find a viable route for my flights and reduce departure delays, that would be tremendous. This use of TOSs seems like a good way to communicate our capabilities and preferences for alternative routes for a given flight. This would make it less likely they would keep that flight on the ground or offer it a route making it necessary to go back to the gate and get more fuel or cancel the flight.”
Useful Communication of Constraints
During the walkthrough described above involving weather in the Washington ARTCC, one important insight was that for some flights, without the information in the TOSs, even experienced traffic managers would not have known from past experience whether different flights could accept the more significant reroutes.
As an example, one dispatcher indicated that, based on calculations by his flight planning system and airline priorities, for a flight from Washington National Airport (DCA) to Los Angeles (LAX), he would file for departure via OTTTO in case it was open, but would include departure via SWANN in his TOS as his least preferred but still desirable alternative to avoid a lengthy departure delay (see Figures 2 and 3). For this same scenario another dispatcher included COLIN and SCOOB as alternative fixes.
The traffic managers all indicated that such information would be very helpful as they tried to decide whether and how to reroute departures. More generally, one dispatcher noted:
“Many times we don’t have the fuel and we have to call and tell them we don’t have the fuel. The TCA page [Tactical Customer Advocate website at ATCSCC] is not a good alternative. It is unbelievably slow and burdensome for reroutes at a time when we’re all in battle mode.”
A second example arose in one of the other weather scenarios and in the space launch scenario. In these scenarios deep water routes were viable in terms of the weather, but as one dispatcher noted: “I can have the same type of aircraft flying the same city pair but one is overwater equipped and the other isn’t. The air traffic controller doesn’t know. If a flight can do an overwater route, let me tell them in a TOS.” The traffic managers were similarly positive about the value of such information about aircraft capabilities: “This use of TOSs would save a ton of time. That would be it [for the flights with TOSs], instead of talking with the airline or calling down to the sup for that area.”
Use of a Cognitive Task Analyses to Generate Recommendations
The walkthroughs produced cognitive task analyses for both traffic managers and dispatchers. For the traffic manager, task analyses for two different kinds of scenarios were generated. Below is the analysis for one of these scenarios using a DCA-LAX flight as an example for illustration.
Scenario
An example of a typical task flow is described in the following scenario. The traffic manager looks at each flight as it starts to push back, determining whether a reroute is desirable to significantly reduce its departure delay because leaving the flight to depart via its filed departure fix (such as a flight filed to depart via OTTTO) will result in a significant departure delay. A semantic level cognitive task analysis for this scenario is as follows - assuming an advisory has already been sent out to the flight operators (See Figure 1):
The traffic manager uses existing tools and procedures to assess the situation and determine if a pre-departure reroutes are desirable. The traffic manager determines that certain fixes are restricted or stopped and develops a strategy for initiating reroutes. In the Washington Center scenario, the traffic manager sees that, for all of the departure fixes with the exception of northeast fixes (AGARD, SWANN and PALEO) and WATRS (Atlantic) routes), departures are stopped as a flight pushes back.
The traffic manager determines which flight to evaluate next (such as a flight from DCA-LAX).
The traffic manager determines that a reroute is desirable for that flight because its filed departure fix (OTTTO) is stopped. The displays indicate that it has a TOS, so the RAD for that flight is opened to look at its TOS.
For this flight, the traffic manager skims down the prioritized list of alternative routes in the TOS (see Figure 2) and picks the first alternative route that has an open departure fix and that is judged to be a good reroute (SWANN).
The traffic manager selects this alternative route and makes the route amendment.
Examples of the resultant recommendations were:
Traffic managers sometimes make a reroute but then want to make a second reroute back to the filed route because the weather changes before the flight departs. The filed route should therefore be shown in the TOS.
Airline TOS generators would benefit by making FAA advisories machine readable (see Figure 2), so that this software could ingest FAA forecasts regarding possible fix constraints and recommended routing solutions.
To handle departure from airports like those in Washington Center where there are many departure fixes, the TOS should be allowed to have a significant number of alternative routes. (For the Washington ARTCC that number is around 7, one for the filed route and 6 for the number of different clusters, thus allowing one per cluster.)
Collectively the FAA and airline software should help the traffic manager when judging whether a route in the TOS is viable or requires coordination to determine whether it is viable. At least for the initial implementation of this process, this can be accommodated by restricting routes in the TOS to those in one of several FAA databases that contain routes that have been precoordinated across the relevant enroute ARTCCs based on their traffic flows. The downside of this solution, however, is that an analysis prepared for one of the walkthrough scenarios by one of the airlines of 5 sample city pairs indicates that such precoordinated routes add from 98-3480 lbs. of extra fuel burn relative to corresponding user-preferred trajectories.
More General Conclusions
Classic systems engineering suggests two design strategies to deal with system complexity: (1) Design complex systems into nearly independent subtasks; (2) Incorporate hierarchical control into system design. In addition, these authors have previously recommended that design strategies should allocate subtasks to individuals (and organizations) in such a way as to match the locus of control, authority and responsibility for a subtask with the necessary existing knowledge and skills; availability of necessary processing capacity (human/technological); access to data; goals and priorities; and capabilities to act (Smith & Bass, 2017).
This study illustrates the value of applying collaborative constraint propagation as a design solution to extend beyond classical systems engineering in a system where decision making is asynchronous and where relevant data, knowledge, processing capabilities and expertise are distributed across individuals and organizations that are not co-located. It further illustrates how the encoding of tribal knowledge can help to deal with expertise that is distributed.
Footnotes
Acknowledgements
A great deal of appreciation is due to the Collaborative Decision Making Program Flow Evaluation Team, as well as to the participating dispatchers and traffic managers. The views and opinions expressed in this article are the authors’ own and do not necessarily reflect those of the Federal Aviation Administration, the Department of Transportation, or the United States government.