Steven M. Green

En Route Descent Advisor Concept for Arrival Metering

The En-route Descent Advisor (EDA) is a set of decision support tool (DST) capabilities for managing complex en route traffic subject to metering restrictions. The goal is to enable controller procedures to evolve from today’s emphasis on sector management towards procedures more oriented towards trajectory management. EDA will help controllers transition traffic from a “Free Flight” (minimally restricted) en route environment into an efficiently organized arrival flow into terminal airspace. EDA assists controllers with high-density arrival metering by providing fuelefficient metering-conformance advisories that are integrated with conflict detection and resolution (CD&R) capabilities. Results of engineering analyses indicate that EDA advisories, based on accurate trajectory-prediction techniques, have the potential to reduce the rate of conflict-probe false alarms and missed alerts by 20% and improve the efficiency of transition airspace operations resulting in an annual nation-wide benefit of

Descent Advisor Preliminary Field Test

291 million.

Abstraction Techniques for Capturing and Comparing Trajectory Predictor Capabilities and Requirements

A field test of the Descent Advisor (DA) automation tool was conducted at the Denver Air Route Traffic Control Center in September 1994. DA is being developed to assist Center controllers in the efficient management and control of arrival traffic. DA generates advisories, based on trajectory predictions, to achieve accurate meter-fix arrival times in a fuel efficient manner while assisting the controller with the prediction and resolution of potential conflicts. The test objectives were: (1) to evaluate the accuracy of DA trajectory predictions for conventional and flight-management system equipped jet transports, (2) to identify significant sources of trajectory prediction error, and (3) to investigate procedural and training issues (both air and ground) associated with DA operations. Various commercial aircraft (97 flights total) and a Boeing 737-100 research aircraft participated in the test. Preliminary results from the primary test set of 24 commercial flights indicate a mean DA arrival time prediction error of 2.4 seconds late with a standard deviation of 13.1 seconds. This paper describes the field test and presents preliminary results for the commercial flights.

TIME-BASED CONFLICT RESOLUTION ALGORITHM AND APPLICATION TO DESCENT CONFLICTS

ion Techniques for Capturing and Comparing Trajectory Predictor Capabilities and Requirements Robert A. Vivona L-3 Communications, Billerica, MA 01821 Karen T. Cate and Steven M. Green NASA Ames Research Center, Moffett Field, CA 94035 Recent research has increased focus on the conceptual design, development and use of airand ground-based aircraft trajectory prediction capabilities to support advanced Air Traffic Management concepts. In both the United States and Europe, the sharing of fourdimensional trajectory information between many automation systems will be necessary for successful operations. Understanding the functional and performance differences between disparate trajectory predictors is critical for enabling this system interoperability. Documented capabilities for four existing trajectory predictors were compared to identify commonalities and differences. For effective comparison, it was first necessary to abstract the prediction capabilities of each trajectory predictor. Three abstraction techniques were developed. The first separated the description of modeled aircraft behavior from the associated mathematical models used to integrate the predicted trajectory. The second defined a conceptual boundary between the trajectory predictor and its client application. The third eliminated the use of domain specific terminology. The abstraction techniques proved not only beneficial for comparing trajectory prediction capabilities, but also for defining trade-offs between the compatibility and accuracy of disparate TPs to achieve system interoperability.

Comparison of Aircraft Trajectory Predictor Capabilities and Impacts on Automation Interoperability

A time-based conflict resolution algorithm was developed to resolve predicted conflicts by time shifting (delaying or advancing) one of the flights prior to conflict. This resolution approach was applied to conflicts in the complex transition phase (from Center to Terminal airspace) involving flights that are also impacted by descent and time restrictions. In order to accurately account for the complex descent dynamics, a high fidelity trajectory model was used to analyze descent trajectories and the geometry of conflicts that occur in the descent phase of flight. The development of an accurate and efficient conflict resolution solution needed an accurate and simple analytical approximation of the descent trajectory. Therefore, linear and quadratic approximations of the trajectory model were assessed for their accuracy and ease of use in the conflict resolution algorithm. The quadratic approximation resulting in an average error of 0.05 nautical miles was more accurate than the linear approximation with an average error of 0.8 nautical miles. However, the linear approximation resulted in a simpler, and therefore more computationally efficient, closed form solution for the minimum time shift required for conflict resolution. It was shown that the time shift computed using the closed-form based on the linear approximation was extremely close to the more accurate time shift computed numerically based on the quadratic approximation. The difference was within 1 second (corresponding to about 0.1 nautical miles) for a relative course angle range between 30 and 150 degrees and under different descent speed profiles. The deviation increased marginally as the conflict occurred closer to the bottom of descent but the increase due to the conflict altitude was negligible. The conflict resolution algorithm and its analysis were applied to conflicts with a number of simplifying assumptions. Namely, the resolution is achieved by maintaining conservatively the horizontal separation requirement between an intruder aircraft with constant altitude and constant ground speed and a descending maneuver

Conflict-free planning for en route spacing - A concept for integrating conflict probe and miles-in-trail

The U.S. NextGen air transportation system is expected to dramatically increase the amount of aircraft trajectory data shared between automation systems operating within the National Airspace System. To illustrate the strengths, weaknesses, and issues of sharing different types of trajectory data, lessons learned from previous work in comparing the capabilities of disparate trajectory predictors (TPs) was leveraged. An abstracted model of top-level internal TP processing, developed to compare the trajectory prediction capabilities of five state-of-the-art TPs, was presented. Detailed results comparing TP constraints, behavior models and math models were presented to illustrate how different TP modeling decisions create significant differences between TP capabilities. The abstracted top-level internal TP processing model was then used to discuss the potential effectiveness of two existing data sharing approaches, intent and 4D trajectory data sharing, and to propose a new approach, behavior model data sharing. Finally, an arrival aircraft conflict scenario used the TP comparison results to illustrate detailed examples of issues that can arise when sharing intent or 4D trajectory data between airborne FMS and ground-based separation assurance automation. The paper concludes that none of the three trajectory data sharing approaches is expected to be the best at achieving NAS automation interoperability under all possible situations and provides motivation for performing future, formal experiments to study the applicability of different data sharing approaches.

Trajectory Prediction Accuracy and Error Sources for Regional Jet Descents

This paper describes a concept for a tool to help air traffic controllers conform to dynamic miles-in-trail (MIT) spacing restrictions. The tool will enable controllers to plan efficient spacing strategies resulting in reduced workload and fuel consumption. Integration of this tool with conflict probe will reduce the probe’s false-alarm and missed-alert rates due to better knowledge of the controller’s intended actions for spacing conformance. Integration will further reduce workload and fuel consumption by reducing the number of corrective clearances needed to achieve flow-rate conformance while avoiding conflicts. The purpose of this paper is to present a near-term concept for applying conflict-probe technology to enable conflict-free planning with efficient MIT-spacing conformance. The paper describes the challenge associated with conflict detection and resolution under flow-rate restrictions, proposes a conflict-probe based concept for the MITspacing problem, and describes a prototype tool and its use through an example traffic scenario.

Characterization Method for Determination of Trajectory Prediction Requirements

The Efficient Descent Advisor (EDA) controller automation tool generates trajectorybased speed, path, and altitude-profile advisories to facilitate efficient, continuous descents into congested terminal airspace. While prior field trials have assessed the trajectoryprediction accuracy for large jet (i.e., Boeing and Airbus) types, smaller (i.e., regional and business) jet types present unique challenges involving different descent procedures and Flight Management System (FMS) capabilities. A small-jet field trial was conducted at Denver in the fall of 2010 with the objective of measuring trajectory prediction accuracy and quantifying the primary sources of error. This paper uses data collected onboard a Bombardier Global 5000 test aircraft to quantify the size and sources of prediction error. Results for en-route descents, from prior to top of descent to the meter fix 60-120 nmi downstream, indicate that the aircraft arrived an average 15 seconds earlier than predicted, with a standard deviation of 10 seconds. Target Mach and CAS deceleration were found to be the two largest error sources. If CAS deceleration error was reduced using a typical, more predictable level flight deceleration then the arrival time prediction error in 2010 would be on par with a 2009 flight trial of Airbus and Boeing revenue flights. Four of the error sources, tracker jumps, CAS deceleration, target Mach, and path distance, lend themselves to significant improvement with modest to no changes to ATC automation and/or procedures. Wind error and its impact on arrival time error was significantly reduced in 2010 compared to a 1994 flight test using NASA’s Boeing 737 test aircraft.

En Route Descent Advisor Multi-Sector Planning Using Active and Provisional Controller Plans

*† ‡ Concepts for trajectory-based operations depend heavily on the performance of the underlying trajectory prediction capability. To address the system-level question “how good of a prediction is good enough?” a fast-ti me simulation method is presented. Modeling and simulation lends itself to capturing the se nsitivity of a concept’s critical performance criteria, to the performance of its supporting trajectory-prediction capability. Given the significant initial cost to develop and va lidate an appropriate simulation tool a characterization method is proposed to provide quicker, yet less precise results in the interim. The analysis characterizes the traj ectory-prediction errors associated with key modeling options for a specific concept. Concept developers can then identify the relative sizes of trajectory prediction errors associated with key modeling options, and qualitatively determine which options would lead to the failure of their concept. The characterization method is demonstrated for a case study involving future airport surface traffic management automation. Of the four sources of error considered in this study, the average variation from the baseline trajectory associat ed with acceleration segments is 10%; the average variation associated with turn modeling is 4%; and the average variation associated with taxi-speed estimation is 29%. These results and the judgment of the concept developer indicate that potential error associated with accelerations segments is unacceptable, the potential error associated with turn modeling is acceptable, and the potential error associated with taxi-speed estimation is of concern and needs a higher fidelity concept simulation to obtain a more precise result. These results point to some specific surfaceautomation trajectory-prediction requirements that can be implement ed right away while modeling and simulation tools are developed to determine other requirements in greater detail.

Field evaluation of Descent Advisor trajectory prediction accuracy

As decision support tools are developed to support controllers in complex air traffic control environments, new approaches to maintaining situation awareness and managing traffic planning must be developed to handle the ever-increasing amounts of alerting and advisory data. Within high-density metering and other environments where flight path changes are the rule, not the exception, and where interactions between these changes are required, current trial planning approaches are limited by potential increases in workload. The Enroute Descent Advisor (EDA) is a set of decision support tool capabilities for managing high-density en route traffic subject to metering restrictions. The EDA system s novel approach builds aircraft plans from combinations of user intent data and builds controller plans from combinations of aircraft plans to effectively maintain situation awareness during traffic planning. By maintaining both active (current) and provisional (proposed) controller plans, EDA supports controllers in coordinated traffic planning both within and between sectors. Ultimately, EDA s multi-sector planning approach will facilitate a transition from current sector-oriented operations to a new trajectory-oriented paradigm, enabling new levels of efficiency and collaboration in air traffic control.