FFC Projects

Overview

Applications

Business Guide

Projects

Awards

Dallas/Fort Worth International Airport
Perimeter Taxiway
Demonstration

Karen Buondonno
Kimberlea Price

June 2003

DOT/FAA/CT-TN03/19

Document is available to the public
through the National Technical Information
Service, Springfield, Virginia 22161

U. S. Department of Transportation
Federal Aviation Administration

William J. Hughes Technical Center
Atlantic City International Airport, N. J. 084

NOTICE

This document is disseminated under the sponsorship of the U.S. Department of Transportation in the interest of information exchange. The United States Government assumes no liability for the contents or use thereof. The United States Government does not endorse products or manufacturers. Trade or manufacturers’ names appear herein solely because they are considered essential to the objective of this report. This document does not constitute FAA certification policy.

ACKNOWLEDGEMENTS

The authors wish to acknowledge several people who contributed their expert talent and many long hours to this endeavor. Without their hard work and outstanding support, this demonstration would not have been a success.

We would also like to thank the following organizations for their support and participation:

DFW ATCT controllers, FAA Flight Standards Service (AFS), FAA Office of Runway Safety (ARI), the Airline Pilots Association (ALPA), the Allied Pilots Association (APA), American Airlines, American Eagle Airlines, Atlantic Southeast Airlines, Comair, Delta Airlines, and United Parcel Service Airlines.

TABLE OF CONTENTS

(Skip tp List of Figures)

(Skip to Content)

ACKNOWLEDGMENTS

EXECUTIVE SUMMARY

1. INTRODUCTION

1.1. BACKGROUND

1.2. OBJECTIVES

2. METHOD

2.1. LIMITATIONS AND CONSTRAINTS

3. RESULTS

3.1. SUBJECTIVE DATA

3.1.1. ATC Results

3.1.1.1. ATC Post-Run Questionnaires

3.1.1.2. ATC Post Demonstration Questionnaires

3.1.2. Pilot Results

3.1.2.1. Pilot Debrief Comments

3.1.2.2. Pilot End of Day Questionnaire Ratings

3.2. SUBJECTIVE RESULTS SUMMARY

3.3. OBJECTIVE DATA

3.3.1. Arrival and Departure Data

3.3.2. Communications Data

3.3.3. Communications Summary

3.4. Objective Results Summary

4. CONCLUSION

5. EXPERIMENT WORK GROUP OBSERVATIONS

ACRONYMS

LIST OF FIGURES

(Skip to List of Tables)

(Skip to Content)

Figure 1. DFW current configuration

Figure 2. DFW with proposed PTs

Figure 3. PT Arrivals Standard Taxi Routes

Figure 4. PT Departures Standard Taxi Route

Figure 5. Q1- Rate your ability to move aircraft “to and from the runways” during this run.

Figure 6. Q2- Rate your overall level of situation awareness during this run.

Figure 7. Q3- Rate your situation awareness for current aircraft locations during this run.

Figure 8. Q4- Rate your situation awareness for projected aircraft locations during this run.

Figure 9. Q5- How much coordination was required with other controllers during this run?

Figure 10. Q6- Rate the difficulty of this run.

Figure 11. Q7- What was the level of traffic complexity under your control during this run?

Figure 12. Q8- How would you rate the overall level of efficiency of this operation?

Figure 13. Q9- Rate the performance of the pseudo-pilots in terms of their responding to your control instructions, providing readbacks, etc.

Figure 14. Q1-. What effect, if any, did the new PTs have on the amount of frequency communications?

Figure 15. Q2- Did your communication strategies change when you were able to utilize the PTs?

Figure 16. Q3- What effect, if any, did the PTs have on your control strategies?

Figure 17. Q4- Based upon your experience in the demonstration, do you feel that adding the PTs improves operations at DFW?

Figure 18. Q5- Rate the realism of the overall demonstration experience compared to actual ATC operations

Figure 19. Q6- Rate the realism of the simulated hardware compared to actual equipment.

Figure 20. Q7- Rate the realism of the simulated software compared to actual functionality.

Figure 21. Q8- Rate the realism of the simulated traffic runs compared to actual NAS traffic.

Figure 22. Q9- Rate the realism of the simulated airport compared to the actual airport.

Figure 23. Q1- Based on your experience in the demonstration, do you feel that adding the PTs improves operations at DFW?

Figure 24. Q2- Rate the realism of the overall demonstration experience compared to actual operations.

Figure 25. Q3- Rate the realism of the simulated hardware compared to actual equipment.

Figure 26. Q4- Rate the realism of the simulated software compared to actual functionality.

Figure 27. Q5- Rate the realism of the simulated traffic runs compared to actual NAS traffic.

Figure 28. Q6- Rate the realism of the simulated airport compared to the actual airport.

Figure 29. Overall Departure Rates

Figure 30. 13L Departure Rates

Figure 31. 17R Departure Rates

Figure 32. LE1 Frequency Transmissions Per Hour

Figure 33. LE1 Frequency Time Spent Talking

Figure 34. LE1 Frequency Length of Transmissions

Figure 35. LE1 Frequency Time Between Transmission Starts

Figure 36. LE1 Frequency Number of Words per Hour

Figure 37. LE1 Frequency Number of Words per Transmission

Figure 38. LE1 Frequency Speed of Speech

LIST OF TABLES

(Skip to Content)

Table 1. Summary of the Demonstration Design

Table 2. Summary of Runs

Table 3. ATC Post Run Questionnaire Summary

Table 4. ATC Post Demonstration Questionnaire Summary

Table 5. Pilot End of Day Questionnaire Summary

Table 6. Objective Data

Table 7. Arrivals and Departures that Crossed to/from the West-side.

Table 8. BL Arrival/Departure Data in 10-minute Increments

Table 9. PT Arrival/Departure Data in 10-minute Increments

Table 10. Description of Inbound Taxi Duration

Table 11. Inbound Taxi Duration

Table 12. Description of Outbound Taxi Duration

Table 13. Outbound Taxi Duration and Departure Runway Occupancy Data

Table 14. Aircraft Stop Rates and Duration

Table 15. Baseline Runway Crossing Data

Table 16. Positions and Frequencies

Table 17. Summary of Communication Results (Controllers and Pilots Combined)

Table 18. Summary of Communication Results for Controllers (only)

Table 19. Summary of Communication Results for Pilots (only)

EXECUTIVE SUMMARY

Currently, the Dallas/Fort Worth International Airport (DFW) typically experiences about 1,700 runway crossings per day which contribute to arrival and departure delays and the potential for runway incursions. In an effort to enhance DFW operations, a perimeter taxiway (PT) concept was proposed which would include new PTs on the East and West sides of the airport. Many fast-time simulations and paper studies have been conducted that support the cost benefit, efficiency, and safety aspects of the proposed airport improvements. However, prior to the Dallas/Fort Worth International Airport Perimeter Taxiway (DAPT) Demonstration, the improvements had not been observed or assessed in an operational setting using high fidelity simulation with human operators. Therefore, a partnership effort involving DFW, the Federal Aviation Administration (FAA), and the National Aeronautics and Space Administration (NASA) was formed to conduct a real-time human-in-the-loop simulation that demonstrated the effect of adding new PTs to DFW. The DAPT Demonstration was conducted in February 2003, at the NASA Ames Research Center (ARC) in Moffett Field, California. The FAA William J. Hughes Technical Center acted as Principal Investigator and provided support for the research team.

The primary objective of this endeavor was to provide the airlines, air traffic controllers, pilots, and their associated unions (i.e., the National Air Traffic Controllers Association, Airline Pilots Associations, and Allied Pilots Association) the opportunity to observe and participate in a demonstration of the proposed airport improvements at high fidelity levels with the goal of gaining their acceptance of PTs. In particular, there were four "views" of special interest for the demonstration 1) the controller view, 2) the pilot-on-taxi view, 3) the pilot-on-arrival view, and 4) the pilot-on-departure view. The secondary objective was to collect and analyze operational data for the purpose of deriving descriptive statistics for runway crossings, taxi times, and pilot and controller transmissions.

NASA ARC's FutureFlight Central (FFC) Facility and Crew Vehicle Systems Research Facility (CVSRF) were used to simulate DFW tower operations and flight deck operations respectively, at high fidelity levels. FFC and CVSRF were integrated and ran simultaneously for all runs. There were four days of demonstrations (including training). East-side, South flow, day time traffic operations at DFW were simulated. Traffic scenarios were created using DFW operations data modified as needed to create future demand levels and the desired traffic mix.

Five Certified Professional Controllers from DFW staffed the FFC simulator. Two taxiway configurations were simulated during 13 runs. The Baseline (BL) condition represented current DFW configuration while the PT condition included the proposed new PTs, the extension of Runways 17C, and a new high speed exit on 17C (exiting to the East).

One staff pilot and seven representatives from the airlines flew the Boeing 747-400 flight simulator. The participating pilots engaged in at least one of the four days of the demonstration. Pilots were encouraged to experience all “views” defined in the objective of the demonstration, in addition to certain predefined typical views.

Controller and pilot subjective ratings, objective data captured from the simulators, and communications data were obtained throughout the demonstration. The objective data captured included taxi time durations, various arrival and departure data, runway occupancy times, inbound and outbound taxi statistics, runway crossing data, and pilot and controller communications data.

In general, the subjective and objective data demonstrated that the PTs would improve operations at DFW if implemented. The results revealed many interesting distinctions between the BL and PT conditions. However, because this was a demonstration, it is imperative to recognize that all results should be used and interpreted with caution.

All controller and pilot participants agreed the demonstration was a good representation of operations at DFW and the proposed new taxiways, and all perceived a marked improvement from BL to PT conditions. The participating controllers believed that the implementation of PTs in the demonstration enabled an overall more efficient operation. They felt the PTs provided for a smoother flow of traffic, afforded better ability to move aircraft to and from the runways, improved situation awareness, and decreased workload demands. Pilot participants thought the PTs improved efficiency and increased safety by reducing the potential for runway incursions. They also speculated that PTs would improve airline performance rates and reduce both pilot and controller workload due to less frequency congestion and a reduction in hold-short instructions.

The objective data resulting from the demonstration supported the participants' verbal comments. Both indicated that the PTs would improve operations at DFW if implemented. Arrival rates for the BL and PT conditions remained consistent (by design) but there was a substantial increase in the departure rate per hour for the PT condition. The average inbound taxi duration increased in the PT condition. However, the average outbound taxi duration and associated runway occupancy times showed improvements with PTs compared to BL runs, as did inbound and outbound stop rates and duration times. Furthermore, by design, PTs completely eliminated runway crossings at DFW in the demonstration.

Controller and pilot communications for the most critical frequency were clearly reduced with the addition of PTs. On the Local East 1 (LE1) frequency, significantly fewer transmissions were made (22% relative reduction) with fewer words spoken (27% relative reduction). This resulted in the controllers and pilots spending less time on frequency (24% relative reduction) when compared to BL runs. Words were also spoken slightly slower on average during PT runs. In addition to being operationally relevant, these results were also statistically significant for the LE1 frequency. Such findings were consistent with controller debrief comments; controllers felt that the volume of communications was significantly reduced, and that they used less verbiage because concerns about crossings and reliance on pilot readbacks were alleviated. Many of the positive data results were also apparent in the findings of the other frequencies, but generally to a lesser degree.

Based on the results of the data collected from the demonstration, it was clear that all objectives of the exercise were met successfully. The controllers and pilots were afforded the opportunity to observe and experience the proposed airport improvements with realism and high fidelity, and a considerable amount of valuable data was available for analysis and presented in this report.

1. INTRODUCTION

This report describes the results of a real-time human-in-the-loop (HITL) simulation that demonstrated the effect of adding new perimeter taxiways to the Dallas/Fort Worth International Airport (DFW). The Dallas/Fort Worth International Airport Perimeter Taxiway (DAPT) Demonstration was a partnership effort involving DFW, the Federal Aviation Administration (FAA), and the National Aeronautics and Space Administration (NASA). The DAPT Demonstration was conducted February 10-13, 2003, at the NASA Ames Research Center (ARC) in Moffett Field, California. The data presented in this report are results from controller and pilot subjective ratings, objective data captured from the simulators, and communications data.

This research endeavor was primarily designed to be a demonstration and was not focused on providing data with high fidelity or statistical rigor (i.e. there is limited power for the use of statistical data analysis). The data provides a snapshot of the impact of the proposed DFW perimeter taxiway (PT) operation with human operators (i.e., controllers and pilots) included. It is acknowledged that the data sample is small, participants were limited, and the runs were of variable length. Due to the variable run lengths, objective data was often converted to hourly rates. Inferential statistics were used as appropriate. For most of the data, however, inspection of descriptive statistics (e.g., frequencies, medians, and means) was used as the primary method for evaluation.

Because this was a demonstration, it is imperative to note that all results presented are to be used and interpreted with great caution. Results should not be generalized or accepted as conclusive.

In addition to the this report, an informational video of the demonstration and proposed airport improvements was developed and will be shared with the National Air Traffic Controllers Association (NATCA), International Civil Aviation Organization (ICAO), National Academy of Sciences, International Council of Airports, Airline Pilots Associations (ALPA), the Allied Pilots Association" (APA), and others. The video can be obtained by contacting DFW (perimetertaxiways@dfwairport.com).

1.1. BACKGROUND

Currently DFW typically experiences approximately 1,700 runway crossings per day. The existing configuration at DFW requires that aircraft arriving on the East-side Runway 17C-35C cross the departure Runway 17R-35L, and aircraft arriving on 17L-35R cross both the arrival Runway 17C-35C and the departure Runway 17R-35L. The aircraft arriving on Runway 31R must also cross both Runways 35C and 35L. Similarly, the aircraft arriving on the West-side Runway 13R must cross both the arrival Runway 18R-36L and the departure Runway 18L-36R, and aircraft arriving on 18R-36L must cross the departure Runway 18L-36R. Figure 1 depicts the DFW runways, terminals, three control towers, and existing taxiways and bridges.

Under current operations, the local controller must conduct all runway crossings before the aircraft can be released to the ground controller. This situation increases the local controller's workload in meeting airport demand mainly due to frequency congestion and challenges the local controller to fully utilize the available runways. During major arrival and/or departure pushes, tradeoffs are sometimes made to safely balance all operations. When the local controller maintains the airport departure demand, runway crossings for arriving aircraft can be delayed due to having to cross the departure runway. Similarly, when arrivals stack up at the various runway-crossing points forcing the local controller to meet this demand, departures are 'gapped' to accommodate these crossings. These situations are most evident during the peak traffic times.

In an effort to reduce arrival and departure delays and the number of active runway crossings (with the added benefit of reducing runway incursion potential), a PT concept is proposed. The concept includes new PTs on the East and West sides of the airport, and two new high speed exits each on 17C and 18R. Figure 2 shows an aerial perspective of the proposed new PT concept.

Many fast-time simulations and paper studies have been conducted over the last seven years that support the cost benefit, efficiency, and safety aspects of the proposed airport improvements. It has also been determined that no waivers will be needed for the new taxiways. However, prior to the DAPT Demonstration, the improvements had not been observed or assessed in an operational setting using high fidelity simulation with human operators. In particular, there were four "views" of special interest for the demonstration; 1) the controller view, 2) the pilot-on-taxi view, 3) the pilot-on-arrival view, and 4) the pilot-on-departure view.

An Experiment Working Group (EWG) was formed to plan, conduct, and analyze the DAPT Demonstration to examine the proposed new PTs. Organizations represented on the EWG were DFW, DFW Tower/TRACON, FAA Southwest Region Charter Program Office (ASW-1C1), FAA Office of System Capacity (ASC-100), NASA ARC, the FAA William J. Hughes Technical Center Simulation and Analysis Group (ACB-330), and NATCA. Other organizations involved in the effort included FAA Flight Standards Service (AFS), FAA Office of Runway Safety (ARI), ALPA, APA, and several airlines including American Airlines, American Eagle Airlines, Atlantic Southeast Airlines, Comair, Delta Airlines, and United Parcel Service Airlines. DFW and the FAA sponsored the study.

1.2. OBJECTIVES

The primary objective of this real-time HITL demonstration was to provide the airlines, air traffic controllers, pilots, and their associated unions (i.e., NATCA, ALPA, APA the opportunity to observe and participate in a demonstration of the proposed airport improvements at high fidelity levels with the goal of gaining their acceptance of PTs.

The secondary objective was to collect and analyze operational data for the purpose of deriving descriptive statistics for runway crossings, taxi times, and pilot and controller transmissions.

2. METHOD

The demonstration was conducted at NASA ARC in Moffett Field, California. FAA ACB-330 acted as Principal Investigator and provided supportfor the research team. NASA ARC's FutureFlight Central (FFC) facility and Crew Vehicle Systems Research Facility (CVSRF) were used to simulate DFW tower operations and flight deck operations respectively, at high fidelity levels. FFC and CVSRF were integrated and ran simultaneously for all runs. Table 1 highlights key aspects of the demonstration design.

For further details and information about the demonstration including methods used, experimental design, laboratory platforms, participants, scenarios, procedures, schedules, etc., please see the DAPT Demonstration Experiment Plan Version 8 (dated 9/6/2002). The document can be obtained by contacting the FAA (karen.buondonno@nasa.gov). The following paragraphs describe the only notable deviations from the experiment plan.

Originally, the demonstration intended to complete a total of 12 data collection runs during which pilots of the B744 flight simulator would fully interact with controllers in the tower simulator. The B744 simulator was to be fully linked to the FFC tower and simulated flights were to be incorporated to interact with the tower for nine of the runs. Each day, pilots were intended to fly the B744 simulator in two data collection runs for a total of six pre-defined flight segments. During each flight segment, the flight crew was to experience one of the following desired “views”: an arriving flight passing over taxiing traffic on the Northeast perimeter; a departing aircraft passing over taxiing traffic on the Southeast perimeter; an aircraft taxiing on the Northeast perimeter with arrivals passing over it; or, an aircraft taxiing on the Southeast perimeter with departing traffic passing over it. Flight segments were intended to last approximately 5-15 minutes per run. The third and final run of each day for the pilots was to be an unstructured “Free Form” run that lasted for 45 minutes. During the Free Form run, the B744 flight simulator was not to be visible to the ATC side of the operation. The flight crew was to be given a menu of options from which they selected to experience a variety of additional conditions of interest. Menu items were to include such options as an arriving flight passing over taxiing traffic on the perimeters, a departing aircraft passing over taxiing traffic on the perimeters, an engine-out departure, IFR or VFR conditions, day or night environments, and eye point adjustments to simulate different aircraft types.

Due to technical difficulties, there were several changes. The original plan called for two of the 12 planned runs to be simulated as nighttime runs in FFC. The EWG decided to eliminate nighttime runs. In the end, there were 13 data collections runs instead of 12, and the runs were of variable length. As planned, the B744 simulator pilots participated in the demonstration at least one of the four days of the pilot demonstration. However, the original two-way link designed for the pilots to fully interact with the tower was degraded, and the link was adjusted to transmit data one-way. Therefore, pilots received information from the tower but the B744 was not visible or audible to the controllers. The experiment design was adjusted to have the pilots run “Free Form” (as discussed above) throughout the entire demonstration. They were encouraged to experience all “views” to be demonstrated from the original scenarios in addition to the “menu items”. Pilots rotated throughout positions during and after each run. Preliminary procedures for PT operations were developed for use in the demonstration and presented in the experiment plan. Prior to the demonstration more detailed operational procedures for standard taxi routes were developed and briefed to the controllers. Therefore, the procedures outlined below serve to replace those found in the DAPT Demonstration Experiment Plan Version 8.

Figure 3 and the following describe the new standard taxi routes for arrivals used by ATC during PT runs.

• ARRIVALS to 17
  • Arrivals from 17L joined the Southeast Perimeter Taxiway from Taxiway P and turned North on Taxiway JS

• ARRIVALS to 17C
  • Non-heavy aircraft joined the Southeast Perimeter Taxiway from Taxiway M and turned North on Taxiway JS
  • Heavy aircraft joined the PT from Taxiway P (heavy aircraft were required to exit the runway to the East due to tail height) and turned North on Taxiway JS

• After joining Taxiway JS aircraft were segregated based on their destination terminal
  • Aircraft parking at Terminals A&C - these aircraft transitioned from Taxiway JS to Taxiway L at Taxiway ER and held short of Taxiway EL
  • Aircraft parking at Terminals E & West side - these aircraft transitioned from Taxiway JS to Taxiway K and held short of Taxiway A

• All arrival aircraft on the Southeast Perimeter Taxiway changed frequencies to monitor GE2 turning North on Taxiway JS

Figure 4 and the following describe the new standard taxi route for departures used by ATC during PT runs.

• DEPARTURES
  • Aircraft taxiing to Runway 13L for departure taxied North on Taxiway K, transitioned to Taxiway J via Taxiway Y and joined the Northeast Perimeter Taxiway. These aircraft held short of Taxiway N and changed frequencies to contact LE2 after crossing Taxiway M

   

   

2.1. LIMITATIONS AND CONSTRAINTS

Simulation is a powerful tool for analyzing, designing, and operating complex systems. It enables hypotheses testing without having to compromise safety in the real world. It is a cost-efficient method to check your understanding of the surrounding world and can help produce better results faster for a research question. It can also be a very effective way to show how an operation works while stimulating creative thinking about how it can be improved. However, all simulation techniques make assumptions about the environments they are representing. It is very important to understand and realize the impact of such assumptions as they also often include limitations and constraints that must be considered when examining the results and conclusions.

The DAPT Demonstration employed a real-time method of simulation, that is, human participants (i.e., controllers and pilots) interacted with and reacted to the simulated aspects of the operational environment in real-time. Because it was purposely designed to be a demonstration (i.e., less data rigor and limited experimental design) it is particularly important to recognize and consider the implication of its limitations. The following is a list of the limitations and constraints experienced in the DAPT Demonstration (note: all participants were advised of the potential for these irregularities prior to the start of the exercise).

• The Digital Bright Radar Indicator Tower Equipment (D-BRITE) was available to controllers but was not as informational as the field version (e.g., no time share, no groundspeed, no Heavy designator, departures do not tag until 2.5 nm South),

• The Airport Surface Detection Equipment (ASDE) -3 orientation was off by about 90 degrees (North-South orientation),

• The ATC communications system was a “touch screen” emulation of the field system. There was no intercab communications and there were no West-side coordination calls since only the East-side tower operations were simulated,

• There was a slight delay (0.5 second) inherent in the digital communications system,

• Pseudo-pilot software anomalies occasionally caused aircraft to appear to stop or jump while taxiing,

• Pseudo-pilots were responsible for “flying” multiple aircraft in the simulation. Their task load demand caused an increase in controller repetition of clearances and calls, and pilot voices for different aircraft were often the same,

• Visual cues occasionally appeared odd to the controllers. For example, objects appeared slightly farther or closer than normal and controllers occasionally had difficulty discerning aircraft type,

• Technical glitches in the software caused a few aircraft to “wheelbarrow” (i.e., nose down landing) down the runway on arrival, or “spin” on their tail at the ramp. These aircraft were removed from the runs when encountered,

• The aircraft simulator is a high fidelity representation of a B744. Because there are so few Boeing 747 aircraft at DFW, the eye point of the aircraft was lowered to better represent the experience of a McDonnell Douglas 80 (MD-80),

• In the aircraft simulator, the visual software limited the out-the-window view to the 16 closest aircraft occasionally causing surrounding aircraft to mysteriously appear or disappear,

• A Traffic Alert and Collision Avoidance System (TCAS) issue was identified in the simulator cockpit during the demonstration. Because pilots flew “free form” the whole demonstration (e.g., moving about freely, invisible to FFC controllers, hovering, parking on the end of the runway, etc.), unlikely traffic situations were showing up on the display and distracting the pilots. Since it was felt that TCAS was not crucial to the experience of the participating pilots, TCAS was turned off to reduce the distraction, and,

• There were technical issues with the simulation software that caused several runs to be terminated prematurely. Four of the 13 runs in the dataset terminated prior to the approximate 45-minute design time for the exercises. Based on pre-set criteria, two of those runs were too short (i.e., less than 30 minutes) to be included in the data analyses.

Though the list may seem long, in general, these limitations were normal for a demonstration of this complexity. For example, though it may seem to skew the results because there was increased controller repetition of clearances, it happened in both conditions (PT and BL) so the comparison of interest was not significantly affected. It is certainly important to identify such potential sources of bias, but in actuality, those listed above only minimally affected the data and the experience of the participants. When asked, the participating controllers and pilots indicated these limitations and constraints only slightly affected their experience.

3. RESULTS

There were 13 runs in the demonstration that included six Baseline (BL) runs and seven PT runs. Table 2 describes the order, condition, and duration of each run. Though data for the 13 runs have been recorded, retained, and analyzed; two of the runs (Runs 1 and 9) did not meet the pre-set 30-minute minimum run length criterion to be included in the final data results. Shorter runs would not accurately capture the affect of surges, lulls, or build up in the traffic flow. For example, a short run would not experience the typical cumulative build up of delay which could distort measurements such as taxi durations, runway crossings, stop durations, frequency congestion, etc. Also, due to an isolated technical issue, communications data for one PT run (Run 2) was not captured. All other Run 2 data were included in the results.

* Not included in results reported due to run lengths less than 30 minutes.

3.1. SUBJECTIVE DATA

Questionnaires were distributed to participating controllers and pilots to elicit opinions about their demonstration experience. Responses from controller and pilot participants are presented in both descriptive and graphical formats in the following sections. Debrief sessions and comments on questionnaires were summarized and included where appropriate, with particular emphasis on interesting or recurring themes.

All questionnaires, including ATC Post Run, ATC Post Demonstration, and End of Day Pilot Questionnaires were designed using 7-point Likert scales. Therefore, all rankings ranged from 1 to 7; however, the anchors varied according to the accompanying statement or question. In the following sections, anchors are provided both in the graphs and discussion of each specific question.

Data analysis for the questionnaires consisted of deriving descriptive statistics for each individual question. For the purpose of reporting responses, the overall median scores were used to describe the data. The median score is the most appropriate measure of central tendency when using ordinal data or when scores are not normally distributed. The median is the value above or below which one half of the observations fall. When there is an even number of observations, no unique center value exists, so the mean of the two middle observations is taken as the median value. In the charts and tables below, the frequency and median are provided to further describe the distribution and allow for an assessment of the responses.

3.1.1. ATC Results

3.1.1.1. ATC Post Run Questionnaires

Post Run Questionnaires were administered to participating controllers after each run. Overall ratings for the Post Run Questionnaires were positive and, in general, the controllers perceived a marked improvement from PT to BL conditions. Table 3 provides a summary of the questions and results. More detailed results and summaries for individual questions (or groups of questions) follow.

1--n = number of observances ( e.g., controllers who answered, pilots who answered, runs included)

2--Because there are various interpretations of the term"situation awareness," for this demonstration, the participants were instructed that to have good situation awareness was to maintain awareness of the present state of events (at the lower end of the scale) and the be able to predict and anticipate future events (at a higher end of the scale) in the dynamic environment. In other words, a rating of 1 to 3 would indicate more of a reactionary control strategy perhaps due to traffic volume, frequency congestion, etc., whereas a higher rating of 5 to 7 would reflect an approach that was more proactive in nature.

Question 1

Controllers reported that they felt better able to move aircraft “to and from the runways” in the PT condition than during BL runs. The median rating for the PT condition was 7 or “extremely good,” while the BL median score was 5. Controllers generally believed that the elimination of runway crossings better enabled them to smoothly transition aircraft to their respective gates and/or to the runways. This was particularly true when taxiing turboprops to 13L. Controller comments indicated that during PT conditions they felt workload was lighter and aircraft flows were “smooth and steady.”

Questions 2-4

For Question 2 participants reported that their overall level of situation awareness improved as a result of PT implementation. This can be seen by a median response of 7 for PT conditions as compared to the BL median of 6. In their comments they attributed this to the reduced complexity of scanning tasks that required them to ensure runways were clear to cross. With PTs they were able to re-focus their attention to other tasks since there were no runway crossing queues. This was particularly true for the Local Controllers. Responses to Question 3 indicated that situation awareness was also perceived to improve for current aircraft locations under the PT condition (median = 7) as compared to a BL score of 6. Question 4 responses to situation awareness concerning projected aircraft location did not show an improvement or degradation with PTs. Both of these ratings had a median of 6.

Question 5

The amount of controller-to-controller coordination required received a median score of 1 or “very little” for PT runs, and a median score of 1.5 for BL runs. Controllers remarked that due to the nature of the tower control environment, the need for controller-to-controller coordination is normally minimal.

Questions 6-7

Responses to Question 6 show that ATC participants generally perceived the BL runs to be more difficult than PT runs. The median score for BL difficulty was 6, while the median for PT difficulty was 4. Ratings of traffic complexity from Question 7 remained fairly stable for both BL and PT runs (median = 5.5 and median = 5 respectively) indicating that the complexity was perceived as moderate to high for all runs. It is interesting to note that these two questions had responses that ranged from 1 to 7 over the course of the demonstration indicating that different controllers experienced varying levels of difficulty and complexity. Since the runs were all built with the same or similar traffic, this could be due to several things such as differences in roles and responsibilities between the positions, or simply varying opinions on the meaning of “difficult and complex.”

Question 8

Controllers believed that the PT operations were more efficient than the BL condition. PT efficiency was rated as “extremely good”, with a median score of 7. BL runs were perceived as less efficient with a median score of 5, indicating acceptability somewhat above average. PT ratings were consistent with recorded comments that indicated the controllers felt PTs eased operational demands, improved situation awareness by reducing the complexity of scanning activities, provided for a smooth flow of traffic, decreased workload demands, and allowed for more effective strategies to be implemented (such as sequencing departures more efficiently in order to increase departure rates). It is interesting to note the distribution of responses once again. Both sets of responses actually had a wide distribution on the rating scale, but BL ratings were more evenly distributed from 3 to 7, while PT ratings swayed more prominently to the higher rankings.

Question 9

Controllers rated pseudo-pilot performance regarding their response to control instructions during the demonstration. They rated a median score of 7 (“extremely good”) for the PT condition, and a median score of 5 (moderate to high) for the BL condition. The decline in scores from PT to BL could be attributed to the fact that fewer readbacks and controller commands were required in the PT environment. Controllers commented that they felt the pseudo-pilots did a very good job overall.

3.1.1.2. ATC Post Demonstration Questionnaires

Post Demonstration Questionnaires were administered to participating controllers at the conclusions of the demonstration. All of the controllers believed PTs would be advantageous to implement at DFW and that the demonstration provided a good representation of operations. Table 4 provides a summary of the questions and results. More detailed results and summaries for individual questions (or groups of questions) follow.

Question 1

Controllers perceived that PTs reduced the amount of frequency communications in comparison to the BL scenarios. Their median response was 2, indicating a marked improvement. This rating is consistent with verbal feedback provided by the controllers. Along with several comments about reduced frequency communications, one controller felt, “workload and frequency congestion was lower due to reductions in hold-short instructions and readbacks”. Furthermore, controllers reported that PTs eliminated the need for calls to turboprops from the GE1 position.

Question 2

A median response of 6 indicated that controllers felt that communication strategies changed quite a bit when PTs were available for use. However, no feedback was provided to specify how in fact they had changed. Inferences can be made that fewer controller-to-pilot transmissions and less frequency congestion allowed for more efficient communication strategies.

Question 3

Participant responses to whether PTs imposed positive or negative changes in control strategies resulted in a median response of 6, indicating that controllers believed PTs had an overall positive effect. Controller comments revealed that they felt they were able to increase departure rates because the need for ‘gapping’ for runway crossings was eliminated. The controllers reported that without gapping restraints they were able to sequence aircraft more efficiently, resulting in more ‘nose-to-tail’ departures. In addition, the elimination of runway crossings and the resulting ease of taxiing aircraft to their destinations (particularly for turboprops going to 13L) also improved control strategies.

Question 4

Nearly all controllers thought that adding PTs improved operations at DFW “a great deal,” which was a median response of 7. Controllers further felt that PTs reduced frequency communications and that the operation was much smoother and less work intensive. In their opinion, the elimination of aircraft crossings reduced workload demands, decreased scanning complexity, and allowed controllers to sequence departures more efficiently in order to increase departure rates. Common comments were that PT’s offered “greater efficiency,” created a “smooth and steady” environment, and “cut workload in half.”

Questions 5-9

Question 5 realism ratings for the overall demonstration ranged from 4 to 6. The median response from controllers was a 6, on the high end of realistic representation. Questions 6 and 7 addressed the realism of hardware and software components, which received median scores of 5 (moderate to high realism), as did the realism of the simulated airport environment (Question 9). The traffic sample realism ratings addressed in Question 8 were not as favorable; the median response for simulated traffic runs compared to actual NAS traffic was 4. Controller comments indicated that the lower scores were due to some of the following difficulties. Controllers had some difficulty in discriminating the types of the most distant aircraft, largely due to the resolution of the screens. One controller’s opinion was that increased traffic contributed to the problem. (Note: Traffic was intentionally increased by 20 to 30% to emulate future demand levels). Another confounding difficulty that was reported by the controllers was that pilots did not respond to crossing clearances as quickly as they would be able to in actual conditions. They thought that large workload demands on pseudo-pilots (who were “flying” multiple aircraft at one time), unrealistic repetition of controller clearances, and increased calls all contributed to crossing delays. Controllers felt these complications might skew the BL run data, making them less representative of actual operations. In addition, the ASDE produced more clutter than actual operations, making the screen less readable and more confusing to the controllers. Controllers developed a strategy to enlist GE3’s assistance by writing down the call signs for arrivals coming off the PTs for GE2.

3.1.2. Pilot Results

A total of seven pilots participated in the DAPT Demonstration at the CVSRF. All pilots were asked to complete a Biographical Questionnaire to provide researchers with information about their range of skill and other attributes. The results indicated that pilot participants varied widely in terms of demographics, skill levels, and experience. Of the seven participants, five were active Federal Aviation Regulations (FAR) Part 121 pilots. The remaining two inactive pilots held administrative positions and had a vested interest in PT operations. Participant ages ranged from 33 to 56, and all were male. The experience of the part 121 pilots ranged from 0 to 600 hours total hours experience in the past 12 months. Time as commercial and military aircraft pilots ranged from 0 to 30 years. In addition to demographic information, pilots were asked to rate their skill levels, current level of stress, and level of motivation to participate in the study using Likert scales ranging from 1 to 7 (anchors were adjusted as appropriate). Pilots’ self-assessed skill levels ranged from 2 to 7 (1 = Not Skilled, 7 = Extremely Skilled). Their level of stress ranged from 2 to 4 (1 = Not Stressed, 7 = Extremely Stressed) indicating that outside stressors should not have affected the pilots’ ability to effectively participate in the demonstration. All reported they were largely motivated to participate in the study with scores ranging from 4 to 7 (1 = Not Motivated, 7 = Extremely Motivated).

Pilots were encouraged to experience the three “views” outlined in the test plan; specifically, pilot-on-taxi, pilot-on-arrival, and pilot-on-departure. In addition, the pilot community had specific concerns about aircraft landing overhead of taxiing perimeter traffic, and aircraft departing overhead of taxiing perimeter traffic. In order to alleviate these concerns, all participating pilots requested views of a “worst case” scenario for the pilot takeoff view, specifically an engine loss at maximum gross weight takeoff. Participants were reportedly comfortable that traffic cleared PTs by several hundred feet on departure. The pilots also set out to ease concerns regarding the clearance between aircraft landing over the Northeast Perimeter Taxiway and the aircraft taxiing on the PT. To experience this perspective, they "froze" the B744 simulator directly above the northern perimeter on the 17C glideslope during final approach to Runway 17C. Then they switched viewpoints and froze as a taxiing aircraft directly below the approaching aircraft so they could experience overhead crossings. From the perspective of the aircraft taxiing on the PT, participants noted the height of the arriving aircraft above them. They also noted the clearance between arriving aircraft on both 17C and 17R and the PTs. Pilots felt that adequate distance existed between the aircraft taxiing on the PTs and landing traffic. As a whole, all pilot participants were satisfied and comfortable with what they observed. One participant did comment he thought that despite the adequate distance between aircraft, passengers and pilots alike may need to adjust to the new experience of aircraft passing overhead.

3.1.2.1. Pilot Debrief Comments

All pilots reported being satisfied that the goals of the demonstration were met. Two of the seven were disappointed that FFC and CVSRF were not integrated, while the remaining five reported that integration would have deprived them of more beneficial use of their time in the simulator. All pilots believed that the PTs would be an improvement to current operations in terms of efficiency and safety, but were awaiting data analyses results to confirm. Several participants said they felt that even if taxi times were identical between BL and PT conditions, PTs would eliminate risks and decrease controller workload, making a safer and more efficient operation. The general perception was that PTs would save both fuel and time. Consensus was that controller and pilot workload and communications would also benefit through less radio traffic and a reduction in hold-short instructions.

In general, the pilots all held positive and confident opinions about the benefits of adding PTs. Some pilots also gave their opinions on building the PTs. For example, one pilot expressed that he would like PTs sooner than several years from now. Another pilot felt that “the virtual elimination of runway incursions justifies the expense”; whereas another speculated that it would be difficult to justify the expense and complications of building the PTs in today’s environment.

The majority of the participants expressed positive comments, not only about the high fidelity and overall impressions of the demonstration, but also concerning the ramifications of the demonstration. Based on their experience in the demonstration, the pilots believed the PT concept may be of benefit to other facilities as well.

3.1.2.2. Pilot End of Day Questionnaire Ratings

End of Day Questionnaires were administered to participating pilots at the end of each demonstration day (pilots typically participated for one day). In general, pilots believed PTs would be advantageous to implement at DFW and that the demonstration was a good representation of operations. Table 5 provides a summary of the questions and results. More detailed results and summaries for individual questions (or groups of questions) follow.

Question 1

A median score of 7 indicated that pilots felt adding PTs would improve operations at DFW “a great deal”. This is consistent with the positive comments expressed during debrief sessions. Pilots unanimously felt that PTs would not only improve the efficiency of DFW, but would also reduce the potential for runway incursions and enhance safety and airline performance rates.

Question 2-6

Results from Question 2 indicate that pilot participants felt the overall realism of the demonstration experience was moderately to highly realistic (median = 5) in comparison to actual operations. In Question 3, hardware components received high scores for realism (median = 6), while software received moderate to high scores (median = 5) in Question 4. The traffic sample realism ratings addressed in Questions 5 and 6 were favorable. Pilots felt that the traffic runs were extremely realistic (median = 7), and that the simulated airport environment was highly realistic (median = 6).

Question 2-6

Results from Question 2 indicate that pilot participants felt the overall realism of the demonstration experience was moderately to highly realistic (median = 5) in comparison to actual operations. In Question 3, hardware components received high scores for realism (median = 6), while software received moderate to high scores (median = 5) in Question 4. The traffic sample realism ratings addressed in Questions 5 and 6 were favorable. Pilots felt that the traffic runs were extremely realistic (median = 7), and that the simulated airport environment was highly realistic (median = 6).

3.2 Subjective Results Summary

The subjective data collected from participating controllers and pilots indicated that the primary objective of the exercise was met. That is, the participants were afforded the opportunity to observe and experience the proposed airport improvements with high fidelity and realism. The controllers and pilots indicated they felt the overall demonstration realism was good. In particular, they rated the realism level of the hardware, software, traffic, and the airport as moderately high to high.

The participating controllers gave all positive feedback on the proposed new PTs. Based on their experience they unanimously indicated that PTs would improve operations at DFW. They believed that the implementation of PTs in the demonstration enabled a more efficient operation. They felt the PTs provided for a smoother flow of traffic, afforded better overall ability to move aircraft to and from the runways, improved situation awareness, and decreased workload demands. Departure rates increased and aircraft were sequenced more efficiently since the need to create ‘gaps’ for runway crossings was eliminated. Furthermore, the controllers said that the complexity of scanning activities was reduced due to the elimination of runway crossing queues. The result of this was an enhanced awareness of current aircraft locations and the opportunity to re-focus attention to other tasks. They also reported their communications workload was reduced due to less frequency congestion resulting from a reduction in hold-short instructions and pilot readbacks.

Pilot participants thought the PTs improved efficiency and increased safety by reducing the potential for runway incursions. They also speculated that PTs would improve airline performance rates and reduce both pilot and controller workload due to less frequency congestion and a reduction in hold-short instructions.

3.3. Objective Data

Objective data related to arrival and departure information, and voice communications were collected. To allow for exploring the effect of adding PTs to DFW operations, all data and results are presented and compared by condition (BL or PT). Table 6 summarizes the data presented.

3.3.1. Arrival and Departure Data

The following sections present departure and arrival information obtained from the demonstration. The data are presented for each condition (BL