Executive Summary

Дата канвертавання24.04.2016
Памер60.99 Kb.

Ground Delay Program Enhancement Team

Orlando, FL

January 13-15, 2009

Executive Summary

The Collaborative Decision Making (CDM) Ground Delay Program Enhancement (GDPE) sub-team conduced a meeting on January 13-15, 2008 at JetBlue University in Orlando, FL. Attendees are listed at the end of this meeting summary.

The key objective of the meeting was to discuss Unified Ground Delay Program (UDP), Integration of Traffic Management Advisor (TMA) with Flight Schedule Monitor (FSM), and Control by Time of Arrival (CbTA).
These meeting notes will be reviewed by the GDPE Lead and posted to the CDM web site: Click Here
The next GDPE meeting will be a telcon held on February 9 at 1:00 PM Eastern. This meeting will be to discuss action items and receive an update from the CbTA sub-team.

Ed Gannon, FAA GDPE Lead, welcomed everyone to the GDPE meeting and reviewed the agenda for the meeting. Attendance was taken and the meeting began.

The TMA workgroup was not able to send a representative to the meeting. It was suggested to change our meeting in March to coincide with TMA in Memphis. Andy Beach from FedEx said he would be able to arrange a room.
Day 1 – January 13, 2009

Unified GDP

Ken Howard gave a presentation about Unified GDPs. The goal of UDP is to provide one set of algorithms for issuing and managing delay programs in a manner that is stable, equitable, efficient, incentivized and adjustable. The primary focus is on how pop-up flights are handled. The UDP concept combines the best ideas of a Delay Assignment Program (DAS) and a General Aviation Airport Program (GAAP) and allows the programs to run in a “mixed” mode.

The next step consists of running controlled experiments using the Jupiter Simulation Environment (JSE). Some of the issues that still need to be addressed are:

  • How exactly should UDP work?

  • What scenarios to run?

  • How should the experiments be run?

  • What are good metrics?

Currently, FSM reserves slots based on historical pop-up estimates. Estimates are created by looking ahead and anticipating how many pop-ups will occur in a given time interval. There are two types of slots: Class A and Class B. There was a concern that CDM participants would get all the reserved slots before general aviation (GA) flights because the GA flight plans are being delayed by the Flight Service Stations (FSS). This was resolved by the two types of slots so one set is for CDM participants and one for non-participants. This issue is expected to be resolved with En Route Automation Modernization (ERAM) release in spring 2011.

FSM allocates known flights to all remaining slots. The GDPE WG needs to decide how they want to treat pop-up flights. Should a few slots remain unassigned for pop-up flights? Should FSM allocate the uncontrollable flights before reserving slots or should we start to distinguish exempt (uncontrollable) flights from priority (distance/tier) flights? It was questioned if this would be a critical issue. After experimenting with different scenarios, these questions will need to be revisited.
The Traffic Flow Management System (TFMS) uses an algorithm to allocate pop-up flights to a slot. The intent is to control pop-up flights that file after a GDP is in place. It was questioned how exempt pop-ups should be treated. Should they be treated differently and get first available slot or just have priority? It was made clear that pop-up flights will never take an airline assigned slot unless the airline releases that slot.
TFMS compresses reserved slots that might go unused. It was questioned if the current Adaptive Compression (AC) logic should be used with UDP. Where should the slots be compressed to? They could go to the end of the program but there might not be a pop-up flight during that time. This still needs to be discussed.
Another issue that still needs to be discussed is if UDP should provide incentive for early notification. Gaming does occur in the current system. Providing an incentive to file flight plans earlier may help. Gaming is beyond the scope of UDP. Incentives should be considered.
The airlines were asked how they want to treat pop-up flights. Research is needed to discover periods of high pop-up demand. Pop-ups should not be given a slot during high demand which will force them to file early or they can take a delay.
Different scenarios were discussed to see if the concept of UDP would improve the situation. EWR, ORD, A05 and A08 are possible places DAS programs can occur with long programs, large delay and multiple revisions. November 7, 2008 was also given as a day that had discussion of putting out a GAAP program. Other ideas included:

  • Regular AFP with high pop-up rate

  • Super Bowl events

  • NASCAR events

  • Golf events

  • GAAP transition into DAS

The experiments can be done with the Jupiter program using real scenarios. It will be able to model different programs, handle departure times, subbing and revisions.

Metrics were discussed on how to collect performance data for a UDP. Since this is a new concept, there are currently no numbers on which to base the metrics. After the experiments are done, the GDPE WG can decide what is acceptable to obtain the overall goal of UDPs.
Planned schedule:

  • 2/9 – Scenarios, metrics, and prototype requirements documents

  • 6/09 – Prototype software ready

  • 6/09 to 9/09 – Run experiments and collect data

  • 9/09 – Run human in the loop (HITL) testing

  • 11/09 – Document results

Developers will continue as scheduled and will bring the results back to the GDPE WG. This is a high profile item for the program office. If there are any other scenarios that should be included in the experiments, please email Ken Howard.

Day 2 – January 14, 2009


Rick Oiesen gave a presentation on TMA terminology and GDPs. The freeze horizon is very important in TMA because it is the defining point where an aircraft is controlled or not by TMA.

TMA is still in the early stages of development. The next release will allow more accurate arrival times which will decrease the distance of the freeze horizon. TMA also provides air traffic control specialist with more options versus a GDP. If an aircraft gets stuck outside the freeze horizon waiting for an open slot, the specialist can be proactive and fit him in. This is very similar to an Enroute Spacing Program (ESP) delay but more precise.
There are still a few issues being worked out with TMA. For example if there is a partial slot, TMA will not release an aircraft unless it has a whole slot. A TMC can release an aircraft if they see this happen but may cause some delay. Another issue is the man power required to run TMA. Each aircraft has to be scheduled for a departure time. Some facilities do not have the man power so they can only run it part of the day.
TMA separation matrix for runway threshold was discussed. The airport rate is not being used because they would over deliver the airport if they actually used the rate. Instead filters and buffers are being used to add additional space to each aircraft which is set by each facility as they feel needed.
There is a National TMA workgroup, but they are not working to resolve issues with TMA and GDPs. The GDPE WG needs to get together with the TMA workgroup to discuss this issue.
EWR TMA Briefing

Rick Klarmann gave a presentation on TMA at EWR. The example presented was from December 22, 2008 where EWR had a GDP with first tier exempt. The first tier aircraft receive their delay from TMA whereas further out aircraft will receive delay from the GDP. There was discussion on different issues that occurred during this scenario. TMA is a different system at each facility and independent of each other. The runway is the major constraint when TMA has an adjacent facility feeding to them.

A movie was played on the representation of FSM at EWR on December 25, 2008. It showed the intervals of how the aircraft were landing. With no program in place, the airport was landing under the rate. After a GDP was established, EWR was able to hit there landing rate. There was discussion on whether an airport is being under or over delivered.

TMA Strategic vs Tactical

Rick Oiesen continued his briefing about TMA. GDPs are a strategic decision that looks further in time to decide what needs to be done. GDP handles airport congestion by reducing the number of flights to a level the airport can handle several hours in advance. TMA is a tactical decision that deals with decision making close in time. It handles the flow of traffic delivered by a GDP and merges flights in an efficient way. These two programs are not tightly coupled. Additional discussion is needed to determine how they should work together.

When a GDP is issued, it might be designed to provide:

  • Strategic exact delivery: predicted demand is matched to predicted capacity.

  • Strategic over delivery: This might be designed to keep pressure on the runway.

  • Strategic under delivery: This might be designed to allow for the possibility of a capacity shortfall due to bad weather.

In a GDP, tactical decisions are made where the actual demand must be dealt with. If an aircraft is being controlled by a TRACON, they have to decide whether the flight will hold or continue.


  • Strategic: When computing a typical GDP, aim for some under delivery. This will minimize the possibility of over delivery and can be compensated for at the tactical level.

  • Tactically: Whatever the actual demand and airport acceptance rate, use close-in flights to achieve exact tactical delivery.

Catch phrase: Strategically under deliver while tactically delivering exactly.

The idea is to slightly under deliver airports and use close in flights or pop-up flights to effectively use the runway capacity. If an airport is tactically over delivered there is nothing that can be done to prevent a delay. However, if it is under delivered, flights can take off early.
Both the FAA and NAS users will need to change procedures to make close-in flights much more flexible. Users will need to be ready to release close-in flights early, i.e. before Expect Departure Clearance Times (EDCTs) but after P-time.
Under delivery would allow the system more flexibility. There was concern that keeping pressure on a fix or runway could cause a snowball effect and would result in a more difficult situation manage. It was suggested that all GDPs should be using this concept. It is also site specific because the West coast facilities normally always hit there rates so if they were told to under deliver, they would have a potential of not using full capacity of the runway.
TMA Scheduling

Dwight McConnell briefed his proposal about coupled scheduling with TMA. It is in the early stages and facilities are working out issues. The proposal is automatic scheduling for TMA. The TMA WG is discussing this proposal this week. It was discussed again that the TMA WG and the GDPE WG need to meet to discuss all the issues. Volpe and Metron volunteered to participate in a small meeting with key players. Ed will request a meeting before the March WG meeting

Action Item: Request a meeting with the TMA WG or a member to participate in the next GDPE WG meeting

Assigned to: Ed Gannon, FAA

Due: March 10, 2009
A comment was made about the lack on communication between TMA and the airlines. There is no electronic transfer for the airlines to see what is going on. There is no ability to get times that can be dumped into the Operational Information System (OIS) or other system for the airlines to access.
Control by Time of Arrival (CbTA)

Mike Brennan gave presentation about CbTA. The general concept is to get to the destination runway at your designated time. Flexibility is needed during the en route portion of the flight to meet the arrival time. Not all flights will be involved due to avionics on the aircraft and lack of participation. Most major carriers already have the capabilities.

It was questioned if CbTA would be needed with TMA. Once an aircraft reaches the freeze horizon, TMA takes over. The aircraft will have no say on whether they can make their time or not.
There was discussion on whether pilots would participate in the concept. Dispatchers said they would not unless they were instructed by ATC. Controllers would have to know the whole route of flight for the aircraft to make its CbTA. Other constraints during a flight were also a concern, i.e. if a pilot is unable to make their arrival time due to weather, it will discourage pilots. Currently the CbTA changes according to the flight. If a pilot is unable to make their time frequently, the program will lose its credibility.
The main issue is the pilot’s buy in - if pilots do not cooperate then CbTA will not work. The GDPE WG needs to decide if there is enough benefit to continue with or without the pilots agreement.
CbTA needs to be designed to work with TMA. Different scenarios were discussed on how they would work together. EDCT structure would remain in place. Some questions that still need to be answer include:

  • What type of airline involvement would there be?

  • What type of pilot involvement would there be?

  • What benefit is there if we continue?

  • How would we adjust scheduled times?

The next step is to design a test plan for proof of the concept. The objective would be to prove that pilots can meet specific metering objectives en route and at arrival. Airlines need to be involved to see what is feasible for them. Ed will ask Ellen King or Mark Libby about CTA enforcement. Scott Bayless will talk to Air Line Pilots Association (ALPA) representatives to find out if pilots would be willing participate in CbTA.

Action Item: Contact Ellen King or Mark Libby about CTA enforcement.

Assigned to: Ed Gannon, FAA

Due: March 10, 2009
Action Item: Talk to ALPA representatives to find out if pilots would be willing to participate in CbTA.

Assigned to: Scott Bayless, MITRE

Due: March 10, 2009
Next CbTA telcon will be held on February 2 at 1pm Eastern.

Day 3 – January 15, 2009

FAT and Thin ATC Slots

Charlie Mead briefed the group on the concept of FAT and Thin slots. If a flight is constrained by two or more points, it may be mathematically impossible to find a set of times that fit a flight. The FAT Slot concept allows for the flight to make speed adjustments to allow it greater flexibility to meet a slot. It would give the pilot more control over their speed, however, they may not like the idea of slowing down. TMA allows ATC to use speed to met times so pilots can do the same thing to create slots. The concept allows a slot to be more flexible. Slots can also be swapped with other aircraft so the user can met their time. How would these slots be assigned? They would be free for all. This is only a concept for the GDPE WG to use.

Additional Discussion

The GDPE WG has many concepts that can be used. CbTA can be used to the freeze horizon then transferred to TMA. CbTA would help smooth the flow into TMA. Dispatchers think CbTA will work if they can get the pilots to participate. Once a pilot is off the ground, they can adjust the autopilot to meet their time. There is slightly different terminology. Airlines use required time of arrival (RTA).

Omar Baradi made a comment that TMA will be transferring to Sysops Program Office so he can find out what the plan is for TMA and help the groups get together.
Slot assignments need to be more flexible by changing the algorithm with GDPs. The first problem is that long flights get delays and non-exempt flights receive a longer delay. The second issue is if an airline has to delay a flight (beyond the GDP delay) it should have the ability to reduce the delay. Customers can only work with the slots they have. The proposal is to change the algorithm to take exempt and non-exempt flights together and have an equal distribution of the delay within each airline. This would allow airlines to pin down a couple critical flights without moving them and allowing the rest of the flights take the delay. The need to change the algorithm came from the need of other TFM programs changing and allowing flexibility to slots in FSM to work better.

      • February 2: CbTA telcon at 1pm Eastern

      • February 9: GoTo/telcon at 1pm Eastern to discuss action items and CbTA update.

      • March 10-12: GDPE WG meeting in Dallas, TX.

Action Items

Issue Date



Date Due


Ed Gannon

Contact Ellen King or Mark Libby about CTA enforcement.



Ed Gannon

Request a meeting with the TMA WG or a member to participate in the next GDPE WG meeting



Scott Bayless

Talk to ALPA representatives to find out if pilots would be willing to participate in CbTA.







Baradi, Omar




Bayless, Scott




Beach, Andrew




Beatty, Roger



Belayachi, Anwar




Brennan, Michael




Gannon, Ed




Gibertson, Brett



Guensch, Craig




Holmes, John

Air Tran


Howard, Ken




Klarmann, Rick




Koogle, Scott




Korey, Julia



Lehky, Miro




McConnell, Dwight



Mead, Charles




Namendorf, Mike




Oiesen, Rick




Ooten, Ron




Smith, Danielle




GDPE Team Meeting

January 13-15, 2009

База данных защищена авторским правом ©shkola.of.by 2016
звярнуцца да адміністрацыі

    Галоўная старонка