 |
Development and Initial Field Evaluation of FLIGHT DECK Procedures for Flying CTAS Descent Clearances
Everett Palmer, NASA Ames Research Center
Tsuyoshi Goka, Sterling Software
Patricia Cashion, Michael Feary and Nancy Smith, San Jose State UniversityFoundation
Holly Graham, University of California, Davis
ABSTRACT
The Center TRACON Automation System (CTAS) is a new computer-based system thatwill assist air traffic controllers in the management of arrival traffic. TheDescent Advisor (DA), a major component of CTAS, uses an algorithm to predictflight trajectories and arrival times based on an aircraft's cruise airspeed,current air traffic, current atmospheric conditions, type-specific aircraftperformance data and airline preferences. Controllers can use this predictedflight profile to provide a "ground-based FMS" descent clearance that willallow an aircraft to fly an efficient descent while maintaining adequateaircraft separation. New flight deck descent procedures were developed to allowcommercial aircraft to comply with these CTAS descent clearances, inpreparation for a field evaluation of the CTAS Descent Advisor conducted inSeptember 1994. During this evaluation, CTAS descent clearances were issued to97 commercial flights arriving at Denver Stapleton International Airport. Datacollected to evaluate the flight deck descent procedure included questionnaireresponses obtained from participating pilots, and observations recorded in thecockpit during CTAS descents. This paper describes our work on procedure andclearance phraseology development, results obtained during the September FieldEvaluation, and some of the ongoing preparations for eventual CTASdeployment.
BACKGROUND
The CTAS Descent Advisor
The Center-TRACON Automation System (CTAS) is a new, computer-based supportsystem that is being developed by NASA and the FAA to improve the efficiency ofdescents and to increase the rate at which aircraft land at an airport(Erzberger, 1994). One component of CTAS, the Descent Advisor (DA), providesair traffic controllers with precise descent profile and Estimated Time ofArrival (ETA) predictions similar to those provided to pilots by FlightManagement Systems (Williams & Green, 1991).
The Descent Advisor, unlike conventional ATC systems, uses atmosphericmodeling, current air traffic conditions, aircraft performance models, andindividual airline preferences to calculate flight trajectories and predictarrival times for incoming aircraft (Williams & Green, 1991). It can alsosuggest different trajectories for a specific, controller-selected aircraftthat facilitate the sequencing of this aircraft with other arrival traffic andoptimize its fuel efficiency during descent. The controller chooses an arrivaltime for the aircraft and the Descent Advisor calculates a descent trajectoryand descent speed that will place it at an inbound metering fix at thecontroller-specified time. This suggested trajectory is presented to thecontroller in the form of an advisory which, at the controller's discretion, isused as the basis for clearances to the aircraft. This enhances thecontroller's ability to manage air traffic, allowing it to arrive safelyseparated and efficiently sequenced.
The September Field Evaluation
Purpose The Descent Advisor was evaluated in September 1994 at Denver'sStapleton International Airport. This initial field evaluation was conductedto evaluate the performance of the Descent Advisor, to begin development ofprocedures for use in compliance with DA clearances, and to determine themagnitude of errors associated with DA usage. Factors that might reduce theeffectiveness of the DA are atmospheric prediction errors, aircraft performancemodeling errors, radar tracking imprecision, and variations in pilot techniqueand clearance interpretation.
NASA Project Teams Three NASA teams worked together during this fieldevaluation to determine the impact of these factors on DA trajectory andarrival time estimates. The main team of researchers from Ames Research Centerwas concerned with evaluating the DA system performance at the Denver Air RouteTraffic Control Center (ARTCC) and with developing controller procedures forusing the Descent Advisor (please see Sanford, Harwood, & Lee, in thisproceedings). A second team from Langley Research Center was involved inobtaining accurate recordings of CTAS trajectories as flown by NASA's Boeing737 aircraft. Our group, comprised of researchers from the Aviation HumanFactors Branch at NASA Ames, was enlisted to develop and evaluate the DA flightdeck procedures and clearance phraseology that would allow commercial aircraftto fly the DA profiles.
The goals for our project were twofold. While our immediate goal was to supportthe September Field evaluation, our second, longer term goal was to design andevaluate flight deck procedures, clearances and preparation material for futureCTAS deployment. This paper describes our work on procedure and clearancephraseology development for the September Field Evaluation, and the results andobservations obtained during the evaluation. How these results have influencedour continuing work on flight deck procedures, clearance phraseology, and pilottraining material is also discussed.
FLIGHT DECK PROCEDURE DEVELOPMENT: PROJECT OVERVIEW
Descent Trajectory Requirements
The trajectory provided by the Descent Advisor can be defined by a specifictop-of-descent location along the flight path, a descent speed, and an altitudeand speed constraint at the bottom-of-descent. The DA arrival time predictions,used by controllers to sequence arrival traffic, depend on an aircraft's actualdescent path and speed closely matching this predicted trajectory.
Project Components
Our primary goal was to ensure that flight crews used a descent procedure thatresulted in the desired descent profile with minimal impact on the flightcrews. We identified three tasks we needed to accomplish in order to meet thisgoal: 1) develop a minimal-training, "CTAS/DA Descent Procedure" that wouldproduce a descent trajectory that conforms well with the DA prediction, 2)develop clear and unambiguous phraseology for the DA descent clearances thatcommunicate the needed trajectory information, and 3) compile a briefingpackage to prepare flight crews for the CTAS/DA Descent Procedure andclearances. These tasks needed to be accomplished under a serious timeconstraint, with only two months available to develop the procedures,phraseology and briefing materials, and to obtain the needed approvals from theFAA and the participating airline.
Controller, Pilot and Airline Concerns
The introduction of the Descent Advisor into the air traffic control systemaffects both the controller's and the flight crew's descent related tasks. Acontroller may clear an aircraft for a DA descent as much as 30 or 40 milesbefore the top-of-descent, instead of issuing the descent clearance when thecontroller is ready for the aircraft to descend. Controllers were concernedthat once a descent clearance had been issued, aircraft might initiate theirdescent sooner than expected. Flight crews and airlines, on the other hand,worried that the aircraft's descent path would become more constrained,reducing the flight crew's ability to fly an efficient descent. Since aprimary objective in developing the CTAS/DA Descent Procedure was to gainairline, flight crew and controller acceptance for this new air traffic controltool, we tried to identify and address the concerns of all of the affectedgroups. Representatives from the participating airline, the Denver ARTCC, theDA design engineers, and airline test pilots who flew simulated DA descentsprovided essential information about users' concerns.
Different Aircraft Types
A complicating factor for development was the planned participation ofdifferent fleets of aircraft. The main difference between these fleets was thepresence or absence of Flight Management System (FMS) equipment, which resultsin fundamentally different methods of flying and monitoring the lateral andvertical paths. The FMS-equipped aircraft were Boeing 737-300/500 and Boeing757. The non-FMS aircraft were Boeing 737-200 and Boeing 727. One of our goalsfor the procedure was that it conform to pilot and airline preferred techniquesfor each aircraft type.
FLIGHT DECK PROCEDURES FOR CTAS DESCENTS
Non-FMS-equipped Aircraft Procedures
CTAS descent procedures were developed for the Boeing 727 for an earlier,simulator-based evaluation of the Descent Advisor (Williams & Green, 1991).These procedures provided a starting point for our Non-FMS proceduredevelopment. Minor modifications to these procedures were incorporated, thentested and approved quickly, using the Boeing 727 simulator at United AirlinesDenver Training Center.
FMS-equipped Aircraft Procedures
The FMS procedures took longer to design and were iteratively modified andtested at Ames during August, 1994. Eight commercial airline pilots wererecruited to fly simulated CTAS/DA descents in a Boeing 747-400 simulator, andto critique the procedure and briefing material as it was developed. Thissimulation testing helped to refine the CTAS/DA Descent Procedure and the fieldevaluation briefing material, and identified an important constraint for theclearance phraseology (discussed below).
One of the challenges in developing a descent procedure for FMS aircraftinvolved selection of a point to initiate the descent. The Descent Advisor'sdescent path calculation provides controllers with an explicit top-of-descentpoint. The on-board FMS, however, does not have a mechanism for incorporatingan assigned top-of-descent location into its flight plan calculations; insteadit determines its own top-of-descent based on programmed wind data, aircraftweight, and constraints on the descent path, including descent speed andbottom-of-descent altitude and speed restrictions. Assigning FMS aircraft aspecific top-of-descent point could either rule out use of the FMS to initiatedescent, or require negotiation with air traffic control about thetop-of-descent location.
Since the algorithms used by the FMS and the DA for determining the flight pathwere similar, the two calculated top-of-descent locations should roughlycoincide. The decision was made to have flight crews program the FMS with thecontroller issued bottom-of-descent crossing restriction and descent speed, andto use the FMS calculated top-of-descent point. We would then determine duringthe field evaluation whether knowledge of the DA top-of-descent alone providescontrollers with adequate flight path information.
The briefing package for the DA Descent Procedure explained the specificparameters and expectations the descent had to meet. These includedmaintaining cruise speed until reaching the top-of-descent, descent withengines at idle and maintaining the cleared descent speed within +/- 10 knotsuntil necessary to slow down for the crossing restriction.
Descent Procedure Clearance Phraseology
Clearances for the CTAS/DA Descent Procedure had two primary functions: 1) torequest a flight crew's participation in the field evaluation, and 2) toprovide the information needed to specify the descent profile, includingdescent speed, bottom-of-descent crossing restrictions and, for non-FMSaircraft, a top-of-descent location. Clearance phraseology was based on FAAair traffic control procedure document 7110.65, with changes for usabilityprovided by CTAS engineers and a Denver ARTCC controller representative.
Three complicating factors affected clearance development. The first was thefact that aircraft participating in the field evaluation would cross threedifferent "sector" boundaries within the Denver ARTCC air space while receivingCTAS/DA descent information. Each of these sectors is managed by a differentair traffic controller, who cannot issue a clearance that takes effect outsideof his airspace. The second factor was the need to limit the amount ofinformation contained in each clearance. It became clear to us during the747-400 simulation testing that both flight crews and controllers havedifficulty remembering more than 3 or 4 distinct pieces of new information whenit is packaged in an unfamiliar clearance. Morrow and Rodvold (1993) reportmarked increases in pilot readback errors, repeat or clarification requests,and procedural deviations when the number of commands contained within a singleATC clearance are increased. Finally, there was a need to limit the number ofcommunications between the aircraft and the controller, both because itincreases pilot and controller workload and because it prevents the controllerfrom communicating with other aircraft.
The compromise solution for the field evaluation was to break communicationinto three different clearances, each issued from a different air trafficcontrol sector:
1. Initial CTAS Clearance: "NASA 123, expect CTAS Descent Procedure, plan to cross DRAKO at 17,000 feet and 250 knots."
Upon entering Denver ARTCC airspace an aircraft received this "heads up"clearance, issued by the first sector controller approximately 150-250 milesfrom Denver. This clearance alerted pilots that they would be participating inthe CTAS Field evaluation, and provided them an opportunity to decline. Theclearance also included information about their probable bottom-of-descent(DRAKO) crossing restrictions, needed to plan the DA descent profile.
2. First Descent Clearance (non-FMS Aircraft): "NASA 123 descend and maintain FL 240. For CTAS Descent Procedure, begin descent at 71 miles from the Denver VORTAC; descend at 260 knots, if unable, advise."First Descent Clearance (FMS-equipped Aircraft): "NASA 123, descend at pilot'sdiscretion, maintain FL 240; for CTAS Descent Procedure, descend at 260 knots;if unable advise."
This clearance assigned aircraft a descent speed and permitted them to initiatetheir descent. For non-FMS pilots the top-of-descent location was explicitlyassigned, while FMS-equipped aircraft were cleared for a "pilot's discretion"(PD) descent, with the assumption that they would begin descent at the FMScalculated top-of-descent point. This descent clearance extended to the highaltitude sector controller's clearance limit (FL 240).
3. Continuation Clearance: "NASA 123, cross DRAKO at and maintain 17,000 feet at 250 knots, Denver altimeter is 30.05, maintain CTAS Descent Procedure."
This final clearance, issued by the low altitude sector controller, allowed theaircraft to complete the descent. It also made clear that thebottom-of-descent crossing restrictions for the CTAS/DA Descent Procedure(contained in the Initial CTAS Clearance) were now active.
Descent Procedure Briefing Package for the September Field Evaluation
Since time and resource constraints ruled out training via simulator orinstructor briefings, a paper briefing package was chosen as the only effectivetraining vehicle. This briefing package included procedure description pageswith a text supplement, a cover sheet discussing the evaluation's purpose, andCTAS/DA benefits for an efficient descent. These benefits includeduninterrupted descents and fewer speed changes when crossing sector boundaries.The package also contained a letter of endorsement signed by NASA, FAA andairline representatives; and a questionnaire soliciting pilot feedback.
The procedure description page needed to adequately convey information aboutconformance to the DA vertical path: when and where to begin the descent, howto transition from cruise speed to descent speed, the acceptable tolerance onthe descent speed, and information about crossing restrictions at the bottom ofthe descent, including when to begin speed changes while still on the descentpath. This information needed to be presented clearly enough that flightcrews could successfully execute the CTAS descent immediately after reading thebriefing package. In addition, sample clearances were included to prepare thecrews for the type of clearances they would receive.
The questionnaire was a major index of CTAS performance and crew feedback.Additionally, an observer log was developed for NASA cockpit observers, tomaintain observation data integrity, and ensure that vital pieces ofinformation were gathered for later analysis.
THE SEPTEMBER FIELD EVALUATION
Preparation
A list of candidate flights for the September Field Evaluation was compiledbased on planned entry into Denver from the northwest arrival gate (DRAKO) andplanned arrival times during light traffic (no scheduled arrivals withinapproximately 10 minutes of the participating flight). These requirements alsodetermined key departure airports for these flights. Briefing packages weredistributed at these airports with the help of United Airlines personnel, whoalso helped to inform crews of the field evaluation.
Scenario
The following describes a typical scenario for participating flight crews inthe September Field Evaluation. Crews were informed about the field evaluationand given the Descent Procedure Briefing Package by their flight operationsmanager. Upon entering Denver ARTCC airspace, air traffic control confirmedthat the crew was prepared to participate in the DA field evaluation. Next thecrew received the descent clearance. Flight crews on FMS-equipped aircraftentered the crossing restrictions and descent speed into the CDU, and used theFMS to initiate and fly the descent. Crews on non-FMS aircraft used theappropriate navaids to determine when they had reached the clearance-defineddescent point. All crews used thrust and drag corrections when necessary tomaintain the assigned descent speed and path, and continued their descentthrough the feeder gate and on to Denver.
During the first week of the test, NASA observers flew in the cockpit on allparticipating flights. After the first week it was decided that the briefingpackage provided flight crews with sufficient preparation for participation inthe field evaluation, and many of the later flights did not have a NASAobserver on board. Researchers met many of these flights in Denver to debriefthe crews and to collect questionnaires and flight data information.
Results
In order to determine the success of the Descent Procedure and clearancephraseology, the questionnaire and flight observations were analyzed. By theend of the field trial, 97 aircraft had flown the CTAS Descent Procedure.Thirty-nine of these flights were accompanied by NASA cockpit observers.
Sixty-four questionnaires were collected. The questionnaire asked pilots toidentify any problems with either the procedure or the briefing package, tooffer suggestions for improvement, and to describe any adjustments needed tomeet the assigned speed and altitude constraints. The following is a summary ofthe questionnaire results.
Approximately 95% of pilots answered "no" to the question: "Was the clearancedifficult to interpret?" Ninety-three percent stated they did not feel a timepressure in complying with the clearance, and 75% said they could have compliedwith less time. Among suggestions for improving the procedure or briefingmaterial were: "less wording in the clearance", "[the package] needs to be ona [Jeppesen] plate", and "airspeed and altitudes need to be given sooner".Pilots frequently observed that "for FMS aircraft this was a routineprocedure". Overall pilot reactions were very positive.
Less than 10% of pilots said they "needed to make unusual corrections to meetthe constraints at the bottom-of-descent." However, 57% of non-FMS pilots and51% of FMS pilots responded that they needed to make either thrust or dragcorrections to maintain the descent speed and profile. The needed correctionsvaried by aircraft type: of those who stated what kind of correction wasrequired, 11 out of 13 FMS pilots added drag, while 8 out of 10 non-FMS pilotsadded thrust.
Pilots were also asked if they anticipated any problems with routinely flyingthis type of clearance. Some of the comments we received were: "there will bea problem in [thrust] increases due to anti-ice," "[what about] thunderstormdeviations," "the rate of descent was between 3000-4000 fpm, which can beuncomfortable."
Discussion
Pilots identified some of the important issues for CTAS that have yet to befully resolved; for example, pilots questioned the effect of weather conditionson CTAS descent calculations. One discrepancy between the FMS and DAcalculated trajectories can be attributed to the fact that CTAS has access tomore current wind profile data than the FMS. To eliminate this discrepancyduring the next field test, updated wind trend information will be provided tothe aircraft using the Aeronautical Communication Addressing and ReportingSystem (ACARS). The pilot can then enter this "uplinked" information into theFMS, providing the same wind profile data to both the FMS and CTAS.
Additional weather factors include thunderstorms, turbulence and icingconditions. Thunderstorms may require lateral deviations, while turbulence mayrequire vertical deviations; both will affect the descent trajectorycalculation. If ice is encountered, thrust is increased as a consequence ofproviding thermal anti-ice protection. However, DA predictions are based on anidle power descent. Although this thrust increase can be accounted for in theFMS, no method exists for reliably informing the DA when thermal anti-icing isplanned to be used. Thus the problem caused by icing conditions closelyresembles the problem associated with winds: different information is availablefor the DA and FMS trajectory calculations.
The field evaluation shed light on several other issues. One proceduralproblem observed was pilot misunderstanding about when to perform thetransition from cruise to descent airspeed. Although the procedure descriptioninstructed pilots to maintain cruise mach until reaching the top-of-descent, itdid not explain when to transition to the cleared descent airspeed, whichresulted in misunderstandings among some of the flight crews.
As mentioned earlier, the procedure was identified in clearances as the "CTASDescent Procedure". This was partially in order to introduce the aviationcommunity to the system as a whole. This caused an anticipated problem, aconfusion between the acronyms: "CTAS" and "TCAS". While no conceptual orprocedural misunderstandings resulted from this confusion, the term used toidentify the procedure may be changed in future field trials.
FMS Aircraft One important issue for controllers was giving FMS aircraftpilot's discretion (PD) descent clearances. As expected, pilots and theairline were very pleased that the procedure allowed them to use the FMS to flythe descent. Controllers, however, were uncomfortable issuing a PD descentclearance to FMS-equipped aircraft during the field evaluation, and felt thatit would be impossible in moderate to heavy traffic conditions. The reason forthis is that PD clearances permit the aircraft to begin descent anytime afterthe clearance is issued, requiring the controller to protect an extensive blockof airspace.
Non-FMS Aircraft The CTAS clearance referenced the top-of-descent point fornon-FMS aircraft using the Denver VOR ("...begin descent 71 miles from theDenver VORTAC"). This caused several different problems for flight crews due tothe atypical nature of referencing a VOR that was not being used for navigationto determine a top-of-descent point. One example of the resulting confusion wasdemonstrated by a flight whose crew calculated the "along path" distance toDenver (which included a 60 degree heading change), instead of the "straightline" distance. This calculation placed the top-of-descent point roughly 10miles late, causing the aircraft to be high on the descent path.Understandably, in response to the question about anticipated problems flyingthis procedure, one of the pilots replied "give us easily defined descentpoints."
FUTURE DIRECTIONS
Preparations are currently underway to conduct additional field evaluations ofthe Descent Advisor. As a result of the September Field Evaluation, theprocedure, phraseology and briefing package have changed considerably. Inresponse to pilot comments on the excessive length of the briefing package, thedescription pages have decreased from 6 (8.5 x 11) pages to 1 (5.5 by 8.5) pagewith a procedural description on the front and a description of the next fieldevaluation on the back. In addition, future tests will involve three airlinesand a variety of both FMS, non-FMS jet aircraft and turboprop commuteraircraft, so the distribution methods used in the last test are no longerpractical. To solve these problems the procedure description page will bedistributed through Jeppesen or through standard company distribution routeswith the rest of the crews' instrument charts.
The most significant procedural change was the decision to give FMS aircraft acontroller issued top-of-descent point, as a result of controllerdissatisfaction with giving FMS aircraft a pilot's discretion descent duringthe September Field Evaluation. Development of the new pilot procedures andphraseology is ongoing. FMS procedures are being evaluated using simulators atAmes and the United Airlines training facility. Finally, a users' committeecomposed of controller, airline, and pilot representatives has been formed.This committee provides a forum for these groups to express their views andplan for eventual CTAS deployment.
ACKNOWLEDGMENTS
We would like to thank Randy Kelley and John Rowbottom of United Airlines, andJim King, a CTAS consultant at Denver ARTCC, for their generous assistance andsupport during the September Field Evaluation.
REFERENCES
Erzberger, H. (1994). Concerning the Center-TRACON Automation System (CTAS).Presented to the Federal Aviation Administration Research and DevelopmentAdvisory Committe, July 12, 1994, Washington, D.C.
Morrow, D. & Rodvold, M. (1993). The influence of ATC message length andtiming on pilot communication. (NASA Contractor Report #177621).
Sanford, B. D., Harwood, K., & Lee, K. K. (1995). Tailoring advancedtechnologies for air traffic control: The importance of the developmentprocess. The Eighth International Symposium on Aviation Psychology, Columbus,Ohio.
Williams, D. H. & Green, S. M. (1991). Airborne four-dimensional flightmanagement in a time-based air traffic control environment, (NASA TechnicalMemorandum #4249).
