Conflict Probe Benefits to Controllers and Users

Transcription

1 MP96W MITRE PRODUCT Conflict Probe Benefits to Controllers and Users Indications from Field Evaluations August 1996 Daniel J. Brudnicki Alvin L. McFarland Susan M. Schultheis 1996 The MITRE Corporation MITRE Center for Advanced Aviation System Development McLean, Virginia

2 MP 96W MITRE PRODUCT Conflict Probe Benefits to Controllers and Users Indications from Field Evaluations August/1996 Daniel J. Brudnicki Alvin L. McFarland Susan M. Schultheis Sponsor: Federal Aviation Administration Contract No.: DTFA01-93-C Dept. No.: F042 Project No.: 02963A02AB 1996 The MITRE Corporation MITRE Center for Advanced Aviation System Development McLean, Virginia Approved for public release; distribution unlimited.

3 1996 The MITRE Corporation ALL RIGHTS RESERVED This work is the copyright work of The MITRE Corporation. All rights are reserved. The information was produced for the U.S. Government under Contract Number DTFA C-00001, and is subject to Federal Acquisition Regulation Clause , Rights in Data General, Alt. III and Alt. IV.

4 Abstract A prototype version of an automated, flight plan-based conflict probe has been built and installed at Indianapolis Center. This capability is called User Request Evaluation Tool (URET). It was derived from laboratory versions of Automated En Route Air Traffic Control (AERA). URET receives track, flight plans, and flight plan amendments from the Host computer in real time and maintains current flight plan-based trajectories for all aircraft in the center. It predicts aircraft into the future along their flight plans for up to 20 minutes, and alerts the controller to any conflicts found between two or more aircraft, or between an aircraft and a special use airspace. URET also permits a controller to enter a proposed new clearance into the system, and to check it for conflicts prior to issuing it to the pilot. Controllers at Indianapolis Center have evaluated URET on the control room floor during live air traffic control operations. These evaluations began in February, 1996 and continue to present. URET has been favorably received by the controllers, who described various ways in which it could help them carry out their air traffic control duties more safely and more efficiently. The controllers suggested that use of conflict probe could also enable them to provide concrete benefits to users, and they mentioned specific situations that could be impacted. This paper describes the URET prototype and the field evaluations conducted. It summarizes the feedback from the controllers and indicates the types of benefits URET provides to controllers. It also provides quantitative indications and anecdotal examples of the two major types of benefits to users. iii

5 Conflict Probe Benefits to Controllers and Users: Indications from Field Evaluations Daniel J. Brudnicki, Lead Staff Alvin L. McFarland, Principal Staff Susan M. Schultheis, Project Team Manager Center For Advanced Aviation System Development The MITRE Corporation, McLean, Virginia Dan Brudnicki is responsible for assessing the performance of the core URET algorithms. He joined MITRE in 1990, and has evaluated ATC trajectory modeling and conflict prediction algorithms. Previously, he was employed by CAE-Link Flight Simulation Corporation. He has an M.S. in Computer Science. Al McFarland is responsible for test and evaluation of URET. He joined MITRE in 1973, and has contributed to a variety of ATC projects. Previously, he was a pilot with the USAF. He has a PhD in Aeronautical Engineering. Susan Schultheis is responsible for developing the computer human interfaces of URET and leads the field evaluation at Indianapolis Center. She joined MITRE in 1987 and has worked in the area of computer human interfaces, procedures, and training of AERA and URET. Previously, she worked for E- Systems. She has an M.A. degree in Psychology. Introduction User demand for the services provided by the National Airspace System (NAS) continues to increase. Safely satisfying this demand has resulted in controller reliance on Instrument Flight Rules (IFR) Preferred Routes, separation rules, altitude restrictions, flow instructions, etc. Such structure and rigidity translates into increased costs to all airspace users. In response, Federal Aviation Administration (FAA) has been conducting research and development on a number of Air Traffic Management (ATM) decision support capabilities for implementation over the next 10 years. There is also work in progress to define incremental stages in the evolution of these system capabilities. Initial capabilities include automated conflict probe, planning aids, and a supporting controller interface. To reduce the risk of implementing an initial conflict probe, FAA initiated a field evaluation of a prototype version. In January 1995, FAA tasked MITRE s Center for Advanced Aviation System Development (CAASD) to develop a prototype conflict probe, install it at Indianapolis Center so that it could be operated from selected sectors on the control room floor, and conduct an evaluation of it using certified controllers in live operation. This prototype was named User Request Evaluation Tool (URET). URET was developed by adapting an Automated En Route Air Traffic Control (AERA) prototype that had evolved through many years of laboratory evaluations with field controllers. Live evaluations began in February Based on the favorable results of these evaluations, FAA is planning the 1

6 next steps in what could lead to nationwide implementation of URET as integrated capabilities within the (NAS) baseline. This paper describes the URET prototype installed at Indianapolis Center, the field evaluation in progress, results of the evaluation to date, and the benefits to both controllers and airspace users indicated by these results. Description of the URET Prototype The first operational build of URET is envisaged as a D-Side capability, so the URET prototype was also implemented in this manner. While the R-Controller (Radar Controller) continues to monitor the radar display for tactical problems, the D- Controller (Radar Associate) continuously and automatically is provided accurate alert information for aircraft problems in the sector up to 20 minutes in the future. URET presents textual flight information on an Aircraft List and graphical information on a Graphic Plan Display (GPD). Aircraft are automatically added to a sector s Aircraft List 15 minutes before projected entry into the sector, and they are automatically deleted after leaving the sector. The D-Controller uses the Aircraft List and GPD to analyze displayed alerts. The GPD can be used to see the routes of the involved aircraft, the geometry of the detected conflict, the predicted loss of separation, and the results of selected trial problem resolutions. Using these tools, the D-Controller can make a decision about whether to notify the R-Controller immediately or wait and watch the situation. The D- Controller considers both URET data as well as other information that may not be reflected in the URET data in order to take the appropriate action. Figure 1 illustrates how URET was configured at Indianapolis Center. 2

7 URET receives real-time flight plan and track data from the Host through a one-way interface. It also obtains wind, temperature, and pressure from the National Weather Service every three hours. These data are combined with site adaptation and aircraft performance characteristics (e.g., climb and descent gradients as a function of aircraft type, weight, and engine type) to generate 4-dimensional projections, or Trajectories, for all flight plans received. The trajectories are then used as the basis for information presented on the Aircraft List and GPD. Trajectory modeling uses extensive site adaptation data, such as the names and locations of airways, fixes and navigation stations; as well as procedural altitude and speed restrictions for sector and facility boundary crossings. Each trajectory for an active flight is maintained by continuously monitoring the deviation between its track-reported state and its trajectory-predicted state. When they are out of tolerance, the trajectory is corrected. This correction is appropriately applied to the modeled position, altitude, speed, and vertical transition rate of the aircraft URET automatically predicts aircraft forward along their trajectories and searches for potential conflicts. Potential conflicts are declared if aircraft separation is predicted to be less than the separation standard plus a buffer to allow for trajectory modeling uncertainty. If an alert is declared, but the closest horizontal separation is predicted to be greater than the separation standard, the aircraft s entry on the Aircraft List is converted to a yellow color. If the closest predicted horizontal separation is less than the separation standard, the entry is coded in red. If the aircraft is predicted to pass through a Special Use Airspace, 3

8 the entry is coded in white. If the aircraft has no predicted conflicts, its entry is coded in green. The color coding of the entry will change as the conflict situation changes. For any conflict the controller can display the conflict situation on the GPD and can call up an alert display to obtain more details about the conflict. In addition to this automatic conflict probe performed on the aircraft s current trajectory (called the Current Plan), URET has the capability to allow the controller to evaluate a proposed clearance (called a Trial Plan) to an aircraft. The controller might use this capability to evaluate a pilot request, or to evaluate his proposed resolution of a current conflict. The controller receives the same kind of indication of conflicts from Trial Plans as from Current Plans. The URET Field Evaluation at Indianapolis Center The initial evaluations of URET at Indianapolis Center have been conducted with four evaluation teams of full-performance-level air traffic controllers from two Areas. Each team consists of three controllers. Prior to the start of evaluations, each team received 16 hours of URET training. Evaluation sessions were conducted during one week per month. Each evaluation session consisted of a one hour operation segment and a one hour debrief segment. The operation segment of each session was conducted at the sector position. To date, evaluations have been conducted during live sector operations at the Pocket City, Louisville, Falmouth, and Rebel high altitude sectors. During the operation segment of each evaluation session, the URET prototype workstation was positioned at the sector on a moveable cart. During each evaluation session, one controller (referred to as the URET Operator) from the evaluation team operated the prototype while the other two controllers in the team staffed the positions of R- and D-controllers at the sector. During an evaluation week, the members of the team would rotate through the position of URET Operator so all would have an opportunity to evaluate the prototype. The URET operator was plugged in to the sector radio frequency and could hear communications between the R-Controller and pilots and between the D-Controller and other sector controllers. During the operation segment of the evaluation, the R- and D- controllers performed their typical duties at the sector while the URET Operator shadowed sector activities by evaluating conflict probe results for Current Plans and creating Trial Plans to evaluate pilot requests or clearances issued. During the operation segment of each evaluation session, a MITRE Evaluation Facilitator observed the operation at the sector. The facilitator observed the prototype and sector operation and recorded notes on the operation. The facilitator was also able to monitor audio communication at the sector during the session. While the prototype was being used, the sequence of interaction between the operator and the prototype as well as characteristics of aircraft processed by URET were recorded by data capture software. These data yielded preliminary summaries of the aircraft processed 4

9 and patterns of interaction between the operator and the prototype. These data were correlated with the subjective feedback captured during the debrief segments. Immediately following the operation segment of the evaluation session, the evaluation team participated with the Evaluation Facilitator in a one hour debrief segment. During the debrief segment, the facilitator asked evaluation questions from a prepared questionnaire and recorded the team s responses. The URET Operator was the primary respondent during these sessions and the other team members provided operational context from the sector operation. There were three main categories of issues that were explored in the evaluations. These were the operational accuracy of the trajectory modeling and conflict probe capabilities; the operational utility and suitability of the capabilities and their implementation; and the benefits to the controllers and airspace users that could be achieved if the capabilities were implemented. Detailed questions for each of these categories were discussed with the controllers during each debrief segment and the same questions were repeated in some cases across segments to capture changes in opinions as prototype experience increased. There was a very clear consensus among all controller evaluation participants that the prototype capabilities are sufficiently accurate and suitable to support conflict detection, conflict resolution, and planning at the D position. During the course of the evaluation sessions, controllers were able to verify that situations predicted by the probe would develop as predicted if no action was taken. On more than one occasion, plans were displayed as problem-free when controllers believed otherwise. When controllers monitored these situations, they observed the plan to actually be problem free. They saw no operationally significant modeling inaccuracies across aircraft types or operational situations. There was consensus that the presentation of information (Aircraft List and GPD) is better suited and more accurate than information available from paper flight data. Controllers consistently stated that URET capabilities should be implemented at the sector as a D position tool and should replace the paper strips. The D controller could then use probe results for the purpose of planning sector operations, coordinating with other sectors and centers, and pointing out problems to the R controller. Controllers said this information would be more useful than paper flight strips. At the same time, the requirement for physical manipulation of the paper strips could be eliminated. The D controller could anticipate R controller actions, and use Trial Planning to check them in advance. "Things would be done before the R-side is ever talking to the aircraft". There was consensus that using URET at the sector would strengthen the sector team and provide the D controller with a meaningful and useful capability. Benefits of Conflict Probe to Controllers The initial URET evaluation yielded the consensus among the controller participants that URET capabilities, if implemented at the D position will provide controllers with several operational advantages over today s system: 5

10 Consistently longer lead time between conflict detection and loss of separation resulting in enhanced safety The performance of conflict detection does not degrade as the amount and complexity of traffic increases Reduced reliance on flight strip data for conflict detection Reduced uncertainty about potential conflict situations due to the accurate performance of automated conflict detection Increased confidence in the long term effects of plan amendments Reduced need for sector-to-sector clearance approval coordination Benefits of Conflict Probe to Users The advantages to controllers translate into advantages for the airspace users as well. The increased conflict detection accuracy will remove some of the uncertainty from the separation process and will result in aircraft being maneuvered less frequently than they are today. As the burden of comparing flight data and performing calculations to detect problems shifts from the controller to automation, it is likely that many of the restrictions in place to assist the controller with separation can be relaxed. The controllers in the evaluation felt that the URET tools would allow them to operate safely without some of the restrictions, and they suggested several specific ones that might be eliminated. The next following paragraphs assess the benefit of eliminating these restrictions. With use of conflict probe, aircraft will more often be able to operate on preferred routings and remain at preferred altitudes. The evaluations suggest that the maneuvers aircraft receive to resolve conflicts will be less disruptive because the problems are detected earlier. When controllers begin to solve problems strategically, there will be more time for negotiation between controllers and pilots to develop clearances that meet the objectives of both. Also, aircraft will be maneuvered less often since controllers will be able to avoid clearing aircraft into problems that they could not normally detect today. The on-going evaluations at Indianapolis Center will focus on validating these potentially beneficial aspects of conflict probe in sector operations. Since URET provides an effective conflict probe capability that enables air traffic controllers to anticipate when and where conflicts will occur, the need to restrict traffic to predictable and potentially undesirable patterns by imposing sector boundary restrictions is minimized. In order to quantify the benefits of removing restrictions, an analysis was conducted using a subset of common altitude restrictions applied to Indianapolis Center arrivals. Flight profiles were constructed for typical routes and aircraft types, once with the appropriate restriction applied and again with the restriction removed. The resultant fuel burns were computed and compared Since altitude restrictions generally require an aircraft to descend at a point earlier than that desired for optimal fuel usage, this comparison represents the potential savings to the user. For example, a flight from Baltimore to Louisville, a distance of 458 nm., must be at or below 28,000 ft. as it crosses 6

11 the Charleston/Falmouth sector boundary in Indianapolis Center, roughly 150 nm. from its destination. A DC9 will burn 6,554 lb. of fuel on this trip. If that restriction is removed, the same flight will burn 6,466 lb., a savings of 88 lb. or 1.3%. The overall results of the analysis are consistent with the example above. A total of 23 cases were examined for a combination of routes (Washington National to Cincinnati, Atlanta to Indianapolis, Baltimore to Louisville, Albuquerque to Cincinnati, and Fort Worth to Cincinnati) and the aircraft types that would typically fly those routes. Fuel savings were shown to fall approximately between 100 and 200 lb., or 0.5 to 1.5%. The average was 130 lb., or 1.1%. Based on inspection, similar benefits could be expected from the removal of additional sector boundary restrictions. These savings are clearly significant, particularly considering the fact that the analysis was limited in scope to internal sector boundary crossing restrictions. An even greater benefit could be accrued by removing restrictions between adjacent centers due to the longer flying times and distances involved. A second type of user benefit is the ability to fly off-airways on user-preferred routes. A specific situation observed in an Indianapolis Center traffic sample from March 29th, 1995 illustrates the benefit. A cluster of eleven aircraft bound for Dallas/Ft. Worth, and all on airway J6 were observed in the vicinity of Charleston, WV. This situation is shown in Figure 2. 7

12 These eleven aircraft took off from eight different airports in the northeast - Boston, Bradley, Laguardia, Newark, Philadelphia, Baltimore/Washington, Washington National, and Dulles. It is surprising to find that these aircraft are essentially sequenced for arrival at DFW over Charleston, which is more than 800 miles from DFW. Eight of these aircraft are evidently participating in a hub bank at DFW, and given this, a clustering of flights is to be expected. The opportunity for user benefit is the fact that all of these aircraft have been forced to fly on J6 even though they took off from eight different airports. We consulted the Digital Aeronautical Chart Supplement to determine if there were any Preferred IFR High Altitude Routes designated for these flights. We found that each of the eight departure airports had a route specified for DFW, the entire route was spelled out, and the route coincided with J6 through Indianapolis Center. The Preferred IFR Routes require aircraft to fly these routes every day, even when the weather conditions are such that the aircraft would consume less fuel and arrive earlier on different, user-preferred routes. URET provides a graphical wind grid display as a tool for the controllers. Because we observed that all of the aircraft in the cluster were flying low ground speeds, we used this tool to see the winds these aircraft were experiencing. It was apparent that there was a strong jetstream wind aligned almost as a direct headwind on J6. The strongest winds in this field occur almost along J6. On J6, winds are 120 to 125 kts. 250 nm. north of Charleston, the strongest winds are less than 50 kts and it appeared that a more efficient route might be found for some of the aircraft. To obtain a quick indication of the potential benefit from flying user-preferred routes in this situation, we created a single alternative flight plan for AAL1375, the aircraft that originated at Boston. We submitted both the Preferred IFR Route and the alternative that we generated with reference to the wind grid to the URET system, and compared the times of arrival. Even though the alternative route had a ground track 29 miles longer, it had a flight time 4 minutes and 33 seconds shorter than the Preferred IFR Route. This represents a 2% savings in flying time. Had we generated a true minimum-wind-miles route, the time savings might have been greater. The Preferred IFR Route and the alternative are shown in Figure 3. 8

13 This situation illustrates an important aspect of what the users mean by free flight. The time savings from the alternative flight plan show why they wish to conduct free flight. Users would like the freedom to file a different flight plan every day to capitalize on the different winds, instead of having to fly the Preferred IFR Route every day. The Preferred IFR Route system was put in place to provide structure and predictability for the ATM system. URET can model aircraft trajectories and reliably detect traffic conflicts without the rigidity and structure of the Preferred IFR Route system. Laboratory evaluations conducted with field controllers indicated that a conflict probe like that in URET can allow controllers to handle scenarios where a substantial number of aircraft are off airways. We plan to assess this capability further in future URET field evaluations. And we plan to use a flight planning system to generate minimum-wind-miles routes for each of the aircraft in this scenario. We will use winds from different days, to get an indication of savings that could be experienced throughout the year. 9

14 Distribution List F010 R. Shaver (AC7) F040 A. Sinha (AC7) S. Koslow (AC7) ATC Data File F041 R. Amato (AC6) J. Freedman (AC4) G. Leone (AC5) P. Ostwald (AC4) F. Petroski (AC6) S. Stalnaker (AC4) F042 J. White (AC4) A. McFarland (AC5) (100) All F042 ATS and Above (40) F042 Library (15) F043 A. Cascone (AC5) J. Celio (AC5) T. Garceau (AC6) J. Lebron (AC4) F. Leiber (AC3) M. Liggan (AC5) E. Newberger (AC5) M. Mock (AC4) M. Walker (AC3) F. Willingham (AC6) B. Wright (AC5) 1

15 F044 R. Swensson (AC6) F045 A. Smith (AC5) K. Thompson (AC4) B. Wasser (AC3) E. Wilhelm (AC5) F066 J103 Records Resources (3) PSI A. Barnes W. Young B. Rushing Approved for Project Distribution: M. White (AC4) W. Kuhn (AC4) F068 C. Jackson (AC4) C. Shiotsuki (AC3) C. Wanke (AC4) E. P. Carrigan Program Manager, F042 DI-2