self-separation from the air and ground...

2nd USA/EUROPE AIR TRAFFIC MANAGEMENT R&D SEMINAR Orlando,1st - 4th December 1998

Self-Separation from the Air andGround Perspective

Margaret-Anne Mackintosh, Melisa Dunbar, Sandra Lozito,Patricia Cashion, Alison McGann, Victoria Dulchinos

NASA Ames Research CenterMail Stop 262-4, Moffett Field CA 94035-1000

[email protected]

Rob Ruigrok, Jacco Hoekstra, Ronald Van GentNational Aerospace Laboratory, NLR

Anthony Fokkerweg 2, 1059CM [email protected]

AbstractNASA Ames and NLR have both conducted

research on free flight and aircraft self-separation.This paper describes the research of both NASAAmes and NLR as it pertains to the task of self-separation. Air-Ground integration issues arepresented from the NASA Ames study, and resultsfrom an NLR human-in-the-loop study examiningthe flight deck perspective of self-separation areprovided. The variables that were studied withinthese two investigations were traffic density,convergence angles, maneuvering automation, andnonnominal cases. Data representing conflictdetection times, crew maneuvering procedures,separation losses, and eye gaze patterns arediscussed. Future research areas are provided.

IntroductionThe concept of “free flight” is defined by the

RTCA (1995) as:A safe and efficient flight operating

capability under instrument flight rules (IFR) inwhich the operators have the freedom to select theirpath and speed in real time. Air traffic restrictionsare only imposed to ensure separation, to precludeexceeding airport capacity, to prevent unauthorizedflight through special use airspace, and to ensuresafety of flight. Restrictions are limited in extent andduration to correct the identified problem. Any

activity which removes restrictions represents amove towards free flight.

The goal of the free flight concept is toprovide more flexibility for operators by reducingconstraints within the airspace system. One meansof reducing constraints is to allow for opportunity foraircraft self-separation in certain environments (e.g.,enroute airspace). New technologies are required toprovide for opportunity for self-separation. Theseinclude enhancements to communication,navigation, and surveillance (CNS), as well asimproved conflict probe devices.

The self-separation of aircraft also requiresthe existence of new airspace zones. Similar to theconcept of aircraft zones represented in the TrafficAlert and Collision Avoidance System (TCAS)logic, a system designed for collision avoidancecurrently available on many aircraft, these additionalzones define regions around aircraft that serve as abuffer for collision protection. There are two zonesdiscussed in the free flight concept implementation:the protected zone and alert zone (RTCA, 1995).The protected zone is a representation of the currentoperational separation standards that exist in thedomestic enroute airspace. The protected zone,therefore, is expected to remain free of other aircraft.

The alert zone is a more unique conceptualspace associated with free flight. It is larger than theprotected zone, as it is intended to permit a previewof potential traffic situations, and to allow for worst-case human and system responses (RTCA, 1995).

Self-Separation from Air, Ground and Airline PerspectiveAir Traffic Management R&D Seminar, Orlando FL, 1998

2

The definition of this zone will be influenced byaircraft equipage and performance characteristics aswell as reflect human performance activitiesassociated with aircraft self-separation.

Pilots, controllers, and dispatchers will beimpacted by the technological and proceduralchanges that will accompany the transition towardsfree flight and the increased opportunities foraircraft self-separation. There has been research inthe area of human factors issues as they may relate tothe free flight self-separation concept. In theinvestigation of traffic density and free flight,previous work indicates that density may have aneffect on controller performance (Hilburn, Jorna,Byrne, & Parasuraman, 1997), while having little orno effect on pilot performance (Cashion,Mackintosh, McGann, & Lozito, 1997; Gent,Hoekstra, & Ruigrok, 1998). In addition, researchexploring convergence angles in traffic conflictsindicates longer conflict detection times associatedwith larger angles (Remington, Johnston, Ruthruff,Romera, & Gold, 1998). The impact of both trafficdensity and convergence angle need to be examinedmore fully.

These research findings led to an air-groundintegration simulation at NASA Ames ResearchCenter investigating human performance parameters.This study included both controllers and commericalpilots as participants in an investigation of aircraftself-separation.

NLR also performed a free flight experimentin 1997 to investigate the human factors principles ofAirborne Separation Assurance. This researchconsisted of three substudies: Concept Developmentstudy, Safety Analysis, and a Human-in-the-loopsimulation experiment using NLR’s Research FlightSimulator. The implementation defined by theConcept Development study and the Safety Analysiswas tested in the Human-in-the-loop simulation.

This paper will first describe the NASA Amesresearch on air-ground integration issues followed bythe NLR Human-in-the-loop evaluation. Theunderlying assumptions for both studies included anassumption of Automatic Dependent Surveillance-Broadcast (ADS-B) technology, airborne alertinglogic, and CDTI available to all participants. Also,aircraft self-separation was implemented in anenroute airspace environment. However, there weresome important differences in the implementations

of self-separation and the methodological approachesbetween these two investigations. In general, theNLR study provided more automation technologiesto aid in the task of self-separation than the NASAstudy. For example, resolution advisories, a verticaltraffic display, and an automated means of enactingthe maneuvers were all represented in the NLRresearch. Finally, the NASA study includedcontrollers as participants; these controllers retainedthe ultimate separation responsibility during thescenarios.

The variables of interest for the NASA studywere traffic density and convergence angles forconflicting traffic. The variables for the NLRresearch were traffic density, levels of automationfor maneuvering, and operational conditions(nominal v. nonnominal).

NASA Air/Ground Experiment

METHODFor a more detailed description of the study

methodology please see Cashion et al., 1997.

ParticipantsFlight crews. Participants were ten flight

crews, consisting of both captains and first officersfrom a major US airline. Each member flew theBoeing 747-400 simulator in his or her normal crewposition. All crewmembers were either current onthe B747-400, or retired for not more than sixmonths. Flight crew participants had a mean totalflight time of 18,400 hours and a mean total flighttime on the B747-400 of 1,820 hours.

Controllers. Ten full performance levelcontrollers from the Denver Air Radar TrafficControl Center (ZDIA) participated in this study. Allcontrollers were current on the sector under study.Controller participants had means of 13.3 yearsexperience as controllers and 5.8 years of experienceas full performance level controllers at ZDIA.

DesignEach flight crew and controller group

participated in a series of eight experimentalscenarios, which were varied by traffic density andconflict angle. There were two levels of traffic


3

density: low (7 to 8 aircraft) and high (15 to 16aircraft) in 120 nm laterally and ± 4000 feetvertically. Traffic density levels were equivalent onboth the controller’s and flight crew’s trafficdisplays. Within each traffic density, participantswere exposed to four scenario types. In threescenarios, lateral conflicts were created by varyingthe intercept angle between flight crew’s aircraft(ownship) and an intruder aircraft (acute, right, andobtuse angles). In the fourth scenario, which will notbe specifically addressed in this paper, an aircraftpassed close to the ownship, but did not violateminimum separation standards.

Airborne Alerting Logic. This study includeda prototypic, airborne alerting logic designed to aidin airborne self-separation (see Yang & Kuchar,1997 for a complete description of the alertinglogic). This alerting logic overlaid the simulator’sTCAS logic with the goal of creating a seamlessrelationship between the airborne alerting logic andTCAS. Currently, the TCAS display depictssurrounding traffic up to 40 nm from the ownship onthe navigation display. In contrast, the alerting logicin this study extended traffic depiction out to 120 nmin front of and to each side of the ownship and 30nm behind the ownship. This surveillance capabilitywas derived from expected ranges for ADS-B.

The airborne alerting logic provided twoadditional alerting zones beyond that of TCAS. Thefirst level of alert, or “alert zone transgression”(AZT), was triggered for the flight crews when thealerting logic predicted a pending violation of theprotected zones of the aircraft at a higher probabilitylevel (Yang & Kuchar, 1997). Operationally, AZTwas the point at which controller intervention maybe required (RTCA, 1995).

There was no ground-based alerting logicrepresented in this study. Controllers were providedwith minimal conflict alerting information derivedfrom the airborne alerting logic. If no evasivemaneuvers were taken after AZT, the AuthorityTransition point was reached. The AuthorityTransition point represented an increased threat levelbeyond AZT and was visible only on the controller’sdisplay. At this point, the controller could takewhatever action he or she thought was necessary tomaintain aircraft separation, including cancellingfree flight on one or both conflicting aircraft.

Flight Crew Displays and Tools. Traffic wasrepresented on the flight deck navigation display bythe symbol “V” with the apex indicating the aircraftheading. Altitude (relative to ownship or absolutealtitude) was displayed next to each traffic symbol.All traffic was initialized as non-threat aircraft.Figure 1 depicts a low density scenario with allaircraft in a non-threat status.

Figure 1. Flight crew’s traffic display depictingnon-threat aircraftt in a low density scenario

When the probability of a violation ofprotected zone increased as determined by theairborne logic, an AZT was indicated to the flightcrew by the following display changes: 1) A blueline extending from both the ownship and theintruder aircraft symbols. At the end of each linewas a blue circle that represented the currentseparation standard of 5 nm in diameter. Anyoverlap of the circles indicated impending loss oflateral operational separation when aircraft are at thesame altitude; 2) An aural warning "Alert" soundedtwice; 3) The word "ALERT" appeared in blue onthe lower right hand corner of the display, along withthe intruder’s call sign and the time to closest pointof approach. The time to closest point of approachwas the time remaining before aircraft wereprojected to pass in closest proximity to each otheron current flight paths. All display featuresassociated with the aircraft involved in an AZT(aircraft symbol, altitude readout, and call signs) aswell as the display changes related to an AZTappeared in blue to help identify which aircraft werepredicted to conflict. Figure 2 illustrates the displaychanges associated with an AZT. As crews solved aconflict, the alert level degraded from an AZT to a


4

non-threat status as the threat probability wasreduced.

Figure 2. Flight crew’s traffic display depicting anAlert Zone Transgression in a high density scenario

Flight crews also could select certain displayfeatures designed to aid them in self-separation. Asmall box mounted above the Mode Control Panelwas used to manipulate selectable display features.Crews could reduce clutter by toggling a button tode-select the traffic call signs. Another selectablefeature was the temporal predictor. The predictorprovided crews with an estimation, based on currentaircraft state information, of where other aircraftwould be relative to the ownship up to ten minutesinto the future. The selection knob for the temporalpredictors allowed crews continuous control of thepredictor length from zero to ten minutes at onesecond intervals. With the predictors, crews coulddetermine which aircraft might create a potentialconflict prior to an alert level indication. Whenpredictors were selected, they were displayed for allaircraft. Selected predictor time was displayed at thelower right hand corner of the navigation display.The predictors functioned as a display tool only;their use did not incorporate the alerting logic as aconflict detection tool. Finally, crew members couldde-clutter the navigation display by changing thehorizontal map range. Ranges available were thesame as those available on the navigation display ofthe B747-400 (10, 20, 40, 80, 160, 320, and 640 nm).

Controller Displays and Tools. The displaythe controllers used during this study had similarfeatures to those available on their current radardisplay (see Figure 3). Some of these tools arevector lines, five-nautical mile rings around the

aircraft (J rings), and graphical route displays. Theprimary sector of concern for this study was SectorNine in Denver Center (ZDIA), which consists ofoverflight aircraft that are transitioning through thefacility and aircraft that are arrivals into DenverInternational Airport.

Figure 3. Controller’s traffic display with vectorlines selected at 2 min and a minimum separationring for DAL152 (high density scenario)

In addition to the display features that thecontrollers used for traffic monitoring and control,there was also a unique display feature added tofacilitate the controller intervention procedure. Thiswas represented by the flashing text Alert Level 4,along with the aircraft call sign. (Due to the natureof the scenarios, the conflict always involved theownship aircraft.) This alert was depicted in theupper right hand corner of the traffic display. Thefunction of this textual icon was to provide anindication to the controller that the AuthorityTransition point had been reached. The controllerwas instructed that if this visual alert was displayed,he/she should query the flight crews about theirintentions and cancel free flight if necessary.

Procedure/TaskEach crew flew a total of eight different

scenarios, which ranged from 15 to 20 min induration. Crews flew a low and high density versionof four scenario types. The conflicts in the high andlow density versions of each scenario type wereidentical, with the only differences in scenariosbeing that the aircraft had different call signs anddifferent levels of non-conflicting traffic wererepresented. The intruders in all of the scenariosintersected the ownship’s flight path laterally. For


5

all conflicts the ownship and intruder were initialized12 min from the closest point of approach if nomaneuvers were taken.

For the three lateral conflict scenarios theangle of intercept between the ownship and intruderaircraft was manipulated. In the acute angle scenariotype, the intruder intersected the ownship’s flightpath at an angle of approximately 22˚. In the rightangle scenario type, the intruder crossed theownship’s flight path at an angle of approximately90˚. In the obtuse angle scenario type, the intruderintersected the ownships path at an angle ofapproximately 165˚. In all three of these scenariotypes, the ownship had the maneuveringresponsibility (i.e., the intruder was on the right).

In order to increase the difficulty/workload forthe participants, an aircraft blocking the mostcommon avoidance maneuver was included in eachscenario (Cashion et al., 1997). Thus for each of thethree conflict angle scenarios (acute, right andobtuse), a blocker aircraft flew a course parallel tothe ownship approximately 10-12 nm off the rightside of the ownship.

Communications/Negotiations. Twoconfederate pilots represented the intruder andblocker aircraft for communication and wereinstructed to respond to calls from the flight crewand controller but not to initiate calls. Theconfederate pilots were instructed to maneuver ifrequested when the ownship had the right-of-way.When the intruder had the right-of-way, theconfederate crew would maneuver if requested onlyafter the second contact from the ownship. Inaddition, background communication was generatedbetween the two confederate pilots and between oneconfederate pilot and the controller. There were noair-to-air negotiations available on backgroundcommunications. The background communicationwas equal for both traffic density conditions, aboutone call per minute.

Results and DiscussionThe purpose of this study was to examine

flight crew and controller human performance issuesin a self-separation environment. Results indicate anumber of interesting findings regarding how self-separation may impact safety, performance,communication, and workload.

SafetyOne measure of safety collected was the

maintenance of adequate separation between aircraft.Four of the 80 runs resulted in a loss of verticalseparation. Twice, the lost separation occurredbecause crews who had climbed to avoid the conflictdescended back to their original altitude too early.The remaining two losses occurred when crews whohad climbed to avoid the conflict did not reach the2,000 ft vertical separation before incurring the 5 nmlateral separation zone. While a loss of adequateseparation occurred in 5% of the simulation runs,these results should be interpreted with caution giventhe prototypic nature of the alerting logic.

Additionally, data were collected on thenumber of times the Authority Transition point wasreached and the number of times free flight wascancelled. The Authority Transition alert wasreached six times during the 80 runs. Based on thealert and the controller’s assessment of the situation,free flight was canceled in two of these six runs. Inaddition to the cancellations initiated by theAuthority Transition alert, two controllers cancelledfree flight prior to the Authority Transition point,because they were not comfortable that the crew’smaneuver was sufficient to resolve the conflict.Finally, one flight crew requested that the controllercancel free flight because they believed that theintruder aircraft should be required to maneuver. Inthe post-experiment questionnaires, controllers wereasked if they had not been required to wait for theAuthority Transition alert whether they would havecancelled free flight at any time during the scenarios.Interestingly, six of the controllers said that theywould have cancelled free flight during the highdensity scenarios and one stated he/she would havecancelled free flight during the low densityscenarios. This may indicate that someenvironmental factors may affect separationdistances desired by the controllers and these maynot be represented in an airborne alerting logic.These discrepancies in the desired separation minimamay lead to differing expectations for the transfer ofcontrol between air and ground.

PerformanceFlight Crew. While previous research did not

indicate performance differences between high andlow traffic density conflict conditions (Lozito,


6

McGann, Mackintosh, & Cashion, 1997), this studyfound several. Additionally, the results indicateperformance differences related to the angle of theconflict. First, flight crews took significantly longerto detect conflicts in the high density conditions ascompared to low density conditions, F(1,7)=6.25,p<.05 (mean = 48.3 s and SD = 53.4 s for highdensity scenarios, and mean = 20.5 s and SD = 41.0 sin low density scenarios). It should be notedhowever, that crews detected the conflict prior to thealert 95% of the time. Second, for the obtuse angleconflict, crews took significantly longer to initiate amaneuver in high density traffic, and they were morelikely to maneuver after the alert, t(7)=2.87, p<.05(mean = 170 s and SD = 40 s for high density obtusescenarios, and mean = 125 s and SD = 24 s for lowdensity obtuse scenarios). Finally, while no densityeffect was found relative to maneuvering time, crewswere significantly more likely to maneuver using oneparameter in the obtuse angle conflict compared tousing multiple maneuvers to resolve the acute andright angle conflicts [χ2(2)=7.05, p=.03]. In theobtuse angle scenarios crews used 16 one parameterand one two parameter maneuvers. While crewsmade 11 one parameter and 8 two parametermaneuvers in the acute angle scenarios, and 12 oneparameter and 8 two parameter maneuvers in theright angle scenarios.

In order to increase the complexity of theconflicts, an aircraft that blocked the most commonavoidance maneuver was included in each scenario.The presence of this blocker aircraft may explain thedensity-related performance differences. Flightcrews attended to the blocker either to determine itsstatus as an intruder aircraft, or to assess possiblemaneuvers for resolution of the conflict situation.With the blocker aircraft diffusing the crews’attention, as well as impeding the most commonescape maneuver, it appears that the blocker mayhave added the desired complexity. While the flightcrews had the same number of opportunities tomaneuver and resolve the conflict in the low andhigh density conditions, the added complexity of theblocker aircraft may have produced potentialworkload differences between the two conditions.

Another possible explanation for theseperformance differences may be in the dataassociated with the use of map range. Replicatingfindings from the previous simulation (Lozito et al.,

1997), results showed that pilots spent more timeviewing a smaller map range (80 nm) in high densityconditions compared to the low density conditions.Conversely, more time was spent at the larger 160nm range in the low density conditions than in thehigh density conditions. Thus, flight crews appearedto be using the range selection on the navigationdisplay as a filter for the density of traffic depictedon the display. The reduced 80 nm range selectionmay help manage the clutter on the screen andprovide a more detailed view of aircraft in closeproximity. However, this smaller range reduces theopportunity to see traffic as early as possible becauseit no longer includes the 120 nm ADS-B range oftraffic. Hence, the smaller map range may lead tolater detection and less time to resolve the conflict.This may be of particular concern in some of theconvergence angles in which there is less timebetween surveillance range and the closest point ofapproach. For example, the intruder aircraft in theobtuse angle did not appear on the ownship displayuntil 4:19 minutes into the scenario. If the flightcrew had reduced their navigation display to 80 nm,the time of appearance for the intruder aircraft wouldhave been even later. This results in later detectionof the potential conflict and in turn, less time toresolve the conflict. Accordingly, this observationmay explain the findings that, in the obtuse anglescenarios, crews tended to maneuver after the alertand to use only a single parameter to maneuver clearof the conflict.

Controllers. Wyndemere (1996) suggests thatconflict geometry may impact the complexity of theconflict. Specifically, conflicts with a smallconvergence angle between the ownship and intruderaircraft are the most complex conflicts to manage.Furthermore, 90˚ intercept conflicts were found to bethe easiest to affect, with the complexity increasingagain as the convergence angle increases to a head-on conflict. Analyses of the controller data revealeda significant interaction between traffic density andconflict angle, F(2,12)=3.92, p<.05 (see Figure 4).Controller participants took significantly longer todetect the obtuse angle conflict in the high densitycondition as compared to the low density condition.Additionally, the controllers detected the acute angleconflict significantly more quickly than either theright or obtuse angle conflict in the high densitycondition.


7

These results correspond to the findings ofRemington et al. (1998). They report that wider angleconflicts are associated with longer detection times.Further, this previous research also provides onepossible explanation for the density differences incontroller conflict detection. Remington et al. (1998)suggest that the number of intervening targetsbetween the two conflicting aircraft may mask thesalience of the conflict. Wyndemere (1996) usedcontroller ratings to determine levels of complexityfor conflict angles and concluded that shallowerangles have the greatest complexity because theyresult in a longer period of potential conflict andrequire action to be taken sooner. However, giventhat larger conflict angles have faster closure rates andappear to be more difficult for controllers to detect ina high density scenario, perhaps a more systematicstudy of angle complexity is required.

Figure 4. Controller detection time for density andconflict angle

It is interesting to note that while the acuteangle conflict was easiest for controllers to detect, itwas not as easy for the flight crews to resolve. Intheir post-experiment questionnaires, the majority ofcontrollers stated that they felt comfortable with alateral separation requirement of 7 nm. However,resolutions for the acute angle conflict were almostalways under 7 nm in lateral separation. Althoughthis separation distance did not violate the required5nm separation standard, it was clear that this lateralseparation was uncomfortable for a number of the

controllers as evidenced by the questionnaire dataand the two free flight cancellations prior to theAuthority Transition point in the acute anglescenarios.

CommunicationAir-to-air communication. Similar to the

findings from the previous study (Cashion et al.,1997) most crews contacted the intruder aircraft inthe three planned conflicts. Specifically, theownship contacted the intruder aircraft in 81% of theruns. Furthermore, the communication occurredprior to any airborne alert in 38 of 46 of runs inwhich the intruder was contacted. Previous research(Lozito et al., 1997), which used a multi-stagealerting logic, found that crews who contacted theintruder did so before the first alert in only two of the37 runs. It was concluded that the alerting logic,and/or display feature changes associated with it,might signal the beginning of the self-separationprocedures for the flight crew. However, based onthe single alerting system used in this study, the startof separation procedures appears to be related moreclosely to the conflict timing provided by thedifferent convergence angles represented. Thesealerting logic differences may need to be examinedfurther. While no density differences forcommunication were found, crews did contact theblocker aircraft 68% of the time. This findingimplies a tendency for the ownship to contact aircraftthat may be either in close proximity to the ownshipor aircraft limiting escape maneuvers. Theseindications of a high potential of inter-crewcommunication in the self-separation environmentand the associated impact on frequency congestionwill need to be addressed.

Air-to-ground communication. Flight crewscontacted the controller at some point in 60% of theruns across the three conflict scenarios. Thesecommunications were usually to inform thecontroller of the maneuver they had already taken toresolve the conflict. Interestingly, in the post-experiment questionnaire, all ten controllerparticipants stated that they would want to beinformed of all maneuvers the flight crew wasmaking, and would want to know prior to the crewinitiating the maneuver. One controller commented,“A constantly changing picture that I can onlyanalyze by my scan (as opposed to an aural cue) is


8

too difficult for extended time periods.” Anotherstated, “I may possess information that they do not,which might determine a better course of action.”Again, these data may present frequency congestionconcerns if the voice channel is used in a self-separation environment.

WorkloadIn the self-report questionnaires, given after

each block of high and low density scenarios, bothcrews and controllers indicated an increasedworkload in the high density conditions over lowdensity conditions. Counter to the findings of Gentet al. (1998) flight crews stated that conflicts weremore difficult to detect, that they felt more timepressure, and that they had an increased workload inthe high traffic density conditions compared to thelow density conditions (t(19)=2.65, p<.05;t(19)=3.33, p<.01; t(19)=2.85, p<.01, respectively).Overall, controller participants indicated that thehigh density scenarios were only of moderatedifficulty. However, supporting the results of earlierstudies (Hilburn, 1997; Remington et al., 1998),controllers did give a significantly higher rating totraffic complexity, t(9)=6.0, p<.001; subjectiveworkload, t(9)=6.09, p<.001; and task difficulty,t(9)=4.0, p<.01, in the high density versus lowdensity traffic conditions.

General ConclusionsThis study was an early attempt at an

integrated examination of flight crew and controllerhuman performance issues in a self-separationenvironment. While this simulation resulted in anumber of interesting findings, there remain severalhuman factors concerns that were not addressed.First, the traffic conditions represented in this studydid not include several elements that could addsubstantial complexity to the scenarios. Theseconditions include weather and winds, special useairspace, mixed equipage, and abnormal situationssuch as aircraft or passenger problems. Severalpilots commented that too much time was spentmonitoring the navigation display for traffic conflictsand that they would not be able to be as vigilantunder abnormal conditions. Second, all aircraft otherthan the ownship were confederates of the study.Consequently, issues related to air carrier differencesand the process of negotiations between carriers

could not be addressed. Third, controllers in thestudy performed a monitoring role and were limitedas to the guidance and instruction that they couldgive crews. It was apparent that this role wasdifficult and a number of controllers noted how itimpaired their scan or picture and increased theirworkload. Finally, flight crews and controllers haveaccess to different sets of information. For example,in one scenario, the blocker aircraft was on arrivalinto Denver International Airport. For the controller,this information was available on the traffic display;however, for the flight crew to gain this sameinformation, the crew was required to contact eitherthe blocker aircraft or the controller. In thediscussion following the simulation, severalcontrollers commented that, in this situation, theywould have started the blocker aircraft down earlyfor its descent, allowing the ownship to make only aminor maneuver off course to resolve the conflict.In this scenario, when the flight crew maneuveredfor the intruder, 38% of crews made a lateralmaneuver and 63% made an altitude change. Thesedifferences in pilot and controller conflictresolutions, given the particular informationprovided to each, need to be systematicallyexamined. Similarly, other questions to be addressedinclude procedural issues for moving betweenconstrained and unconstrained flight, and finalresponsibility for separation.

In conclusion, this research has providedinsight into some important air-ground integrationissues. Flight crew performance differencesbetween high and low traffic density conditions werefound in this study that were not realized in theprevious studies (Lozito et al., 1997). The additionof the blocker aircraft to obstruct the most commonmaneuver and add complexity to the scenarios maybe responsible for these density differences.Additionally, several conflict angle differences wereuncovered. The acute angle conflicts were easier forcontrollers to detect, but were not as easy for thecrews to resolve. Furthermore, although the obtuseangle conflicts were detected later by controllers,crews were able to adequately self-separate. Theseangle differences have implications for theoperational setting and should be investigatedfurther. Finally, both air-air and air-groundcommunication issues were revealed. The highfrequency of air-to-air communications and the


9

consequent impact on congestion remains a problemto be addressed. Regarding air-groundcommunications, timing and content of informationrelayed between the flight crew and controllers maybe of particular concern. In sum, these results havehelped define and describe some of the proceduralconcerns for flight crews and air traffic controllers ina self-separation environment.

NLR Human-in-the-Loop Study

IntroductionThe National Aerospace Laboratory (NLR)

conducted a Human-in-the-loop simulationexperiment in 1997 to determine the human factorsissues of operating an aircraft in a future free flightenvironment with Airborne Separation Assurance.

Before the Human-in-the-loop experimentcould be executed on NLR’s Research FlightSimulator (RFS), two separate studies were carriedout. The Conceptual Design study was performed todetermine a feasible free flight concept. Althoughthe RTCA Task Force 3 document (RTCA, 1995)gives a definition of free flight, this definition is notsufficiently detailed for research purposes. The goalof this first study was to determine a feasible freeflight concept and develop it to a level of detail thatcould be implemented in the simulationenvironment. The main result from this study, inwhich several concepts were examined, was thechoice and implementation of the Modified VoltagePotential theory (Hoekstra, Gent, and Ruigrok,1997).

The second study done before the Human-in-the-loop study was a Safety Analysis of the freeflight concept developed in the Conceptual Designstudy. The Safety Analysis showed that thedeveloped free flight concept was at least as safe asthe present day Air Traffic Management (ATM)environment. For more details on the SafetyAnalysis and the safety analysis tool TOPAZ(Traffic Organization and Perturbation AnalyZer),which was used in this study, see Daams, Bakker,and Blom, 1998.

The remainder of this paper will describe theHuman-in-the-loop simulation experiment conductedat NLR and discuss both subjective and objectiveresults obtained from this experiment.

METHODS

ParticipantsFlight crews. Eight flight crews from major

European airlines participated in the NLR Human-in-the-loop Free Flight with Airborne SeparationAssurance experiments. The crew members wereassigned a position for the experiment based on theircurrent rating, position, experience and preference.All crew members were current airline pilots.

DesignExperimental set-up. In the Human-in-the-

loop experiment the traffic density, the level ofautomation (resolution activation) and nominal/non-nominal conditions were varied as the independentvariables.

The traffic densities used in the experimentwere one, two and three times the current meanWestern-European traffic density. More specifically,the densities given in the number of aircraft in a10000 square kilometer area, above FL190 were:

Single density: ~10 aircraftDouble density: ~20 aircraftTriple density: ~30 aircraft

This corresponded to ~10, ~20 and ~30 aircraftrespectively on the lateral navigation display when a120 nm lateral range and +/- 8000 ft vertical rangewas selected.

Three levels of automation were used forresolution activation via the autopilot:1. Manual, in which case the crew had to enter

mode control panel entries themselves.2. Execute combined, in which case the crew

could auto enter both the horizontal andvertical resolution simultaneously by pressinga button on the mode control panel.

3. Execute separate, in which case the crew couldselect to auto enter the horizontal, the verticalor both resolutions simultaneously, by pressingone or two buttons on the mode control panel.

Combinations of traffic densities and levels ofautomation were all tested in nominal and non-nominal conditions. The non-nominal conditionsconsisted of other aircraft failures, own aircraft


10

failures and increased delay times in conflictdetection and resolution.

All aircraft in the scenario were assumed to beequipped with ADS-B equipment and AirborneSeparation Assurance System (ASAS) equipmentconsisting of conflict detection and resolutionalgorithms, Cockpit Displays of Traffic Information(CDTI) and ASAS alerting. Air traffic wasmonitored by Air Traffic Arbitration (replacing ATCon the ground) to assure co-operative behavior. Allaircraft were flying direct routes from origin todestination and only upper airspace was considered.

The experiment lasted two days per crew, inwhich the first half day was used for training. Thefollowing three half days were used for theexperimental runs. All aircrews in the experimentflew 18 experimental runs of 20 min duration in thesimulated Brussels-West sector.

Concept. The concept used in the Human-in-the-loop experiment was the Modified VoltagePotential. Out of several concepts implemented, themodified voltage potential concept was chosen basedon initial route, time and fuel efficiency calculationsand the characteristics of the Modified VoltagePotential (Hoekstra, Gent, & Ruigrok, 1998).

A major benefit of the Modified VoltagePotential is its fail safety. If two aircraft are inconflict with each other, both aircraft calculateresolutions for the conflict as if the other aircraft isnot maneuvering. However, the concept assumesthat both of the aircraft do maneuver, so fail safety isintroduced and the conflict is solved in a co-operative and economic way.

In all conflict situations, several options wereavailable for the aircrew to resolve the conflict.Within the concept, two separate resolutions werepossible which both resolved the conflict on theirown. A horizontal resolution (heading and speedchange) and a vertical resolution (vertical speed andaltitude) was possible. The crew could choose whichresolution fit best to the conflict geometry andcurrent aircraft state. Selecting both horizontal andvertical resolution added another fail safe element tothe concept.

The Modified Voltage Potential theory isbased on algorithms presented by MassachusettsInstitute of Technology Lincoln Laboratory (Eby,1994). The theory is shown in Figure 5.

Shown in Figure 5 are the ownship aircraft andan intruder aircraft. Each aircraft is protected by aprotected zone of 5 nautical mile radius and a heightof 2000 feet (+1000 ft, - 1000 ft). The predictedprotected zone of the intruder aircraft is shown inFigure 5 at the time both aircraft have approachedeach other at minimum distance. Every predictedintrusion of a protected zone within five minutes isregarded as a conflict. The conflict detectionalgorithms are based on current aircraft states and donot use additional intent information. The resolutionfor the conflict is based on the geometry of theconflict as shown in Figure 5. The vector from theownship’s position at minimum distance to the edgeof the predicted protected zone of the intruderaircraft is the avoidance vector to resolve theconflict. This avoidance vector can be divided in aheading change combined with a speed change asshown. This describes the horizontal resolution.Using the three-dimensional vector, a verticalresolution can be obtained from the conflictgeometry in a similar way.

Figure 5. Geometry of modified voltage potential

Flight Crew Display and Tools. The Human-Machine Interface to support this free flight conceptconsists of a modified navigation display withadditional control panel, additional aural alerts, andmodified autopilot modes. The modifications to thenavigation display are shown in Figure 6.

As can be seen in Figure 6, the conflict andresolution geometry is presented to the aircrewsimilar to the definition of the modified voltagepotential. This gives the aircrew an intuitive anddeterministic picture of why and how to resolve aconflict.

Enhanced TCAS like symbology is used toshow other traffic. The conflicting intruder aircraft is


11

shown in amber (5 to 3 minutes away from intrusionof protected zone) or red (3 to 0 minutes fromintrusion). The predicted protected zone of theintruder aircraft is shown together with the magentaavoidance vector to resolve the conflict. The dottedmagenta lines represent the division of the avoidancevector to heading, speed and altitude/vertical speedchanges. These changes, or resolutions, are presentedon the Primary Flight Display as well.

Figure 6.: Navigation display with conflictsymbology

Airborne Alerting Logic. Besides the alertingon the navigation display, the detection of a conflictis announced to the crew aurally, with a dedicatedblue light in the glareshield and resolution advisoryindications on the Primary Flight Display. Thedisplays used both color coding and aural alerts todepict threat level. The aural alerts used varieddepending on the alerting level of the conflict(amber/red conflict). Additional information on thenavigation display is the distinction of aircraftclosing in or moving away from the ownship,indicated by blue and white aircraft respectively.TCAS warnings were suppressed for experimentalreasons, although TCAS is supposed to be present inthis airborne separation assurance concept as a safetynet.

In total five alert zones can be identified in thisconcept. Aircraft within the ADS-B range of 200nm, a 5 minutes alert zone (amber), a 3 minutes alertzone (red), the protected zone of the aircraft (5nmradius, 2000 ft height) and finally the TCAS timebased Resolution Advisory zone.

Communications/NegotiationsNo communication was required with ground

based stations, although air-ground communicationwith Air Traffic Arbitration (ATA) was available.An inter-air frequency was available for air-to-aircommunication with aircrew of surrounding trafficand conflicting traffic. Simulated communicationbetween ATA and non-nominal behaving aircraft inthe scenario was provided.

Experiment Control. The experiment iscontrolled using the Traffic and ExperimentManager (TEM). The TEM is a traffic generator foraircraft around the subject aircraft, contains allASAS functions for all aircraft in the scenario andcan be controlled using a dedicated graphical userinterface. The TEM is capable of simulating up to400 aircraft simultaneously in real-time.

Results and DiscussionThe purpose of the Human-in-the-loop study

was to identify human factors issues concerned withan Airborne Separation Assurance concept. For thisreason an extreme version of airborne separationassurance was chosen: no Air Traffic Control (ATC)and full responsibility for traffic separation with theaircrew on board of the aircraft.

Subjective and objective measurements,gathered during the experiment, are presented.

Subjective dataThe subjective data were collected with

questionnaires before, during, and after theexperiment. The results are presented related totraffic density. Four main questions were asked tothe subjects concerning acceptability, safety,workload, and operations.

The results from the acceptability question aresummarized in Figure 7. As can be derived from thisfigure, acceptability was around 80% and decreasedslightly with traffic density.


12

0

10

20

30

40

50

60

70

80

90

100

com

plet

ely

unac

cept

able

Und

esira

ble

Acc

epta

ble

Fav

orab

le

Per

fect

inev

ery

way

sub

ject

ive

rati

ng

acc

epta

bilt

y in

per

cen

tag

es

single

double

triple

Figure 7. Acceptability results relative to trafficdensity

The safety results are shown in Figure 8. Thedata show that in around 75% of the runs free flightwas indicated to be as safe or safer than Air TrafficControl. A slight decrease with increasing trafficdensity is noted here as well.

0

10

20

30

40

50

60

70

80

90

100

ATCmuchsafer

ATCsafer

Same FFsafer

FFmuchsafer

Subj

ectiv

e ra

tings

in p

erce

ntag

es

single

double

triple

Figure 8. Safety results relative to traffic density

All participants were asked to enter a mark ona Rating Scale of Mental Effort (RSME) after eachrun. This scale ranged from 0 (“costing no effort”)to 150 (“costing lots and lots of effort”). The overallresult of the subjective workload is shown in Figure9 for nominal and non-nominal conditions, relativeto traffic density. The overall average ratings are ator below the 40 mark, indicating that pilots rated thisconcept as “costing some effort”. Experience from

other experiments (Gent, 1995) demonstrates that intoday’s ATC environment in cruise conditions,average ratings are found to be around 30.

Subject workload

0

20

40

60

80

100

120

140

single double triple

density

nominal

non-nominal

Figure 9. Subjective workload rating (RSME)

The last subjective data concern safety,operations and training. Pilots were given questionson which they could say “True” or “False”. Thesequestions were:1. I think I could safely guarantee the airborne

separation with the set-up just flown.2. I maneuvered more than normally.3. I exceeded passenger comfort levels.4. I need more explicit rules of the road to

guarantee the safety.5. I need more explicit on board procedures to

guarantee the safety.6. I need more training to guarantee safety.

As can be seen from Figure 10, pilotsindicated that especially in triple density,maneuvering is more than in the current ATCsituation. Also 30% of the participants consideredtraining to be an issue.

Objective DataThe objective data collected consist of

duration of conflicts, eye-point-of-gaze data, themaneuvers chosen to solve conflicts, and unintendedintrusions of protected zones.


13

True/False questions related to traffic density

0

10

20

30

40

50

60

70

80

90

100

Can guaranteesafety

Manoeuvredmore

Exceededcomfort

Need morerules

Need moreprocedures

Need moretraining

Per

cen

tag

e an

swer

ed w

ith

tru

e

Single

double

triple

Figure 10. Percentage of true/false questionsanswered with “true”

Figure 11 shows the mean duration of aconflict or conflict time defined as the time fromconflict detection until clear of conflict, in nominaland non-nominal conditions. As can be seen, conflicttimes increase significantly in non-nominalconditions.

0

5

10

15

20

25

30

35

nominal nonnominal

conf

lict t

ime

in s

econ

ds

Series1

Figure 11. Mean conflict times in nominal and non-nominal conditions

During the experiments, the pilots were fittedwith Eye-Point-Of-Gaze equipment to determine thelocation of interest. The total fixation duration onthe Lateral Navigation Display of the pilot flying andpilot-not-flying across all sessions averaged around50%. The numbers for the Primary Flight Display

and the Vertical Navigation Display are around 10%each.

From the resolution maneuvers data in manualmode, see Figure 12, it becomes evident that headingchanges to resolve conflicts are favored over speedand altitude changes. This correlates with thefixation times on the Lateral Navigation Display.Aircrews were allowed to overrule the automaticmodes and also were allowed to use the manualmode, although the automatic mode was suggested.Therefore, the numbers in the execute combined andexecute separate mode differ than what wasexpected. As seen in Figure 12, speed was oftenoverruled in the automatic mode.

0

10

20

30

40

50

60

70

80

90

100

heading speed altitude

Per

cent

ages

of u

sed

para

met

er

manual

execute combined

executeseparately

Figure 12. Percentages of use of each parameter toresolve conflicts as a function of the mode ofoperation

Finally, all intrusions of the protected zone ofthe subject aircraft were logged. One of the non-nominal conditions introduced in the scenarios wasan aircraft performing an emergency descent throughthe protected zone of the subject aircraft. Apart fromthese deliberate intrusions, Table 1 shows theintrusions, which were not prescribed by thescenario. Table 1 shows the minimum separationdistance, minimum separation altitude and intrusionduration. As can be seen, the intrusions are mainlygrazes of the protected zone, either vertically orhorizontally.

These grazes occurred in cases of suddenmaneuvering of aircraft already close to the subjectaircraft, either due to reaching top of descent of theother aircraft or lateral maneuvering of the other


14

aircraft due to clear-of-conflict situations. Somegrazes occurred in non-nominal conditions (NN)where own conflict detection and/or resolution werefailed.

Session min. sep.

distance (nm)

min. sep.

altitude (ft.)

intrusion

duration (s)

107 E T N 4.29 815 18

309 M D N 3.42 914 8

701 M S N 4.55 499 114

701 M S N 4.77 119 341

706 M S NN no data no data 4

709 E T N no data no data 12

711 E T NN 4.23 692 38

713 A T NN no data no data 12

715 A D NN no data no data 9

718 A D N no data no data 18

814 M T N 4.77 9.7 76

817 M S NN 4.83 618 20

Table 1: Intrusions of the protected zone.

General Conclusions

The aircrews participating in the experimentwere given the new and extra task of trafficseparation assurance, in a traffic environment up tothree times as dense as today, with new displays anddisplay features, new procedures, new cockpitautomation and without extensive training.Therefore, the hypothesis was that the concept wouldbe rated less than acceptable, less safe than today’sATM environment and that workload would increasea considerable amount.

Subjective data showed an acceptability ratingof 80%. Seventy-five percent of the runs wereperceived by the pilots to be as safe or safer than AirTraffic Control. Workload was not rated higher incruise conditions than in today’s ATM environment.These figures included triple density scenarios andnon-nominal conditions. The true/false questionsindicated that in the majority of the runs, the amountof maneuvering was more than in today’s ATMenvironment, in the opinion of the pilot participants.

The objective data show a significant increaseof conflict times in non-nominal conditions. Thiscan be explained by the fact that during non-nominalconditions other traffic did not always maneuver co-operatively or own aircraft conflict detection and/orresolution was sometimes failed.

The eye-point-of-gaze data show that fixationduration on the Lateral Navigation Display wasaround 50%. This is higher than in today’s cruiseoperation and could therefore be of concern. Incurrent operations, the fixation duration times havebeen measured to be about 20% on the navigationdisplay. However, it is not clear whether these highfixation duration times are required to operate in thisconcept. As the participants were unfamiliar with thedisplays and the task of separation assurance, onecould argue that the novelty and desire to be in theloop resulted in the high fixation duration times.Further investigation will be needed to clarify thisissue. The low fixation duration times on the VerticalNavigation Display are another concern. Thequestion is whether the Vertical Navigation Displayis useful for conflict detection and resolutionpresentation, or that the pilots were not yet trainedenough to optimally use all possibilities.

The resolution maneuver data show a clearpreference for heading changes, although altitudechanges are far more economic and less disrupting tothe intended route than horizontal maneuver (ValentiClari, 1998). The eye-point-of-gaze data confirmthis tendency with the low fixation duration times onthe Vertical Navigation Display. From theseobjective data, it seems that more training is requiredto optimally use all tools on-board.

The intrusions of the protected zone are allgrazes, either due to sudden maneuvering by aircraftclose by the subject aircraft or the fact that theconflict resolution algorithms aim at the edge of thepredicted protected zone of the intruder aircraft,without safety margin.

The overall conclusion of the Human-in-the-loop simulation experiment is that the feasibility ofthe given Airborne Separation Assurance concept fora future free flight environment could not be refuted.Both objective results and the results from theseparate safety analysis (Daams et al., 1998) couldnot reject this conclusion.

Out of the many issues raised during theexperiment, the main issue was that means should beprovided to prevent intrusions of protected zones onshort term due to sudden maneuvers. Future researchshould focus on the required fixation duration on alldisplays to complete the task of self-separation.Passenger comfort in relation to the amount ofmaneuvers will have to be addressed.


15

General Summary

Both NASA and the NLR conducted researchpertaining to aircraft self-separation in the enrouteenvironment, an operational concept within freeflight that is fairly advanced. The underlyingassumptions for both studies included an assumptionof ADS-B technology, airborne alerting logic, andCDTI available to all participants. However, therewere some important differences between theimplementations of self-separation, as well as themethodological approaches, between these twoinvestigations. In general, the NLR study providedmore automation technologies to aid in the task ofself-separation. For example, resolution advisories,a vertical navigation display, and an automatedmeans of enacting the maneuvers were allrepresented in the NLR research. Finally, the NASAstudy included controllers as participants; thesecontrollers retained the ultimate separationresponsibility during the scenarios.

The independent variables examined by thetwo studies also were different. The variables ofinterest for the NASA study were traffic density andconvergence angles for conflicting traffic. Thevariables for the NLR research were traffic density,levels of automation for maneuvering, andoperational conditions (nominal v. nonnominal).

In summarizing the research findings from theNASA and the NLR studies, there were some thingsin common. Both studies found some impact ofincreasing amounts of traffic on human performance.In the NASA study, both controllers and flightcrewparticipants generally took longer to detect conflictsin higher densities. In addition, self-report responsespertaining to subjective workload also indicatedhigher ratings for higher traffic density for bothcontrollers and pilots. In the NLR study, theresearchers uncovered some trends related to trafficdensity in their subjective data. Although thechanges were typically not dramatic, there was anincrease in the subjective workload ratings, and therewere decreases in acceptability and safety ratingsassociated with higher traffic densities. The impactof traffic density may also be exacerbated by otherfactors. Abnormal situations that may furtherrestrict self-separation operations, for exampleweather phenomenon and sudden changes in aircraft

maneuvering which may cause short-term conflictsand intrusions, should be explored more thoroughly.

In both the NASA and NLR studies, therewere cases in which separation was lost between twoaircraft (the separation standards were similar tothose proposed for future ATM operations). As astrategy, flight crew participants often attempted tominimize the separation between aircraft while stillmaintaining legal separation. Findings from theNASA study indicate that controllers would feelmore comfortable with a larger separation than theflight crews often obtained. The NLR study did notinclude the use of controllers as participants. Furtherresearch needs to address the separation needs forthe controllers and the pilots, and should identifypotential modifications to the alerting schemes toreflect the operators’ requirements within self-separation.

The NLR study also found interesting resultspertaining to eye gaze fixations and CDTItechnology. Specifically, pilots appeared to fixateupon the CDTI (both the lateral and verticalcomponents of the navigation display) approximately60% of the total eye gaze time. By contrast, the timespent fixating on the primary flight display was onlyabout 10% of their total time. The NASA findingssuggest that their pilot participants felt that they werespending too much time attending to the CDTI,possibly supporting the results from the NLR study.

There were some different findings for crewmaneuvering from the two investigations. Bothstudies revealed that crews often used more than asingle performance parameter (e.g., altitude) toresolve a conflict. In the NLR work, the mostcommon parameter used was heading. This isconsistent with previous studies conducted at NASA(Cashion et al., 1997). However, in the currentNASA study, altitude was the most commonparameter used to solve traffic conflicts. This waslikely due to the introduction of an aircraft blockingthe most common lateral maneuver (the blocker) thatmade the use of the altitude solution more desirable.

Finally, two other potential humanperformance issues were revealed in the NASAstudy. Conflict angles seemed to impact thecontroller conflict detection and the timing and typeof maneuvering used by the flight crews. In somecases, traffic density may interact with the effects ofconvergence angle. Additionally, there were many


16

air-to-air negotiations and communications. Thisamount of communication could be problematic in afree flight environment in which there may besimultaneous negotiations between several crews.System constraints may restrict the availability of theradio channels, blocking it when a controller may berequired to contact and instruct a crew in the eventthat self-separation become hazardous.

In sum, these research projects have helpeddefine and describe some of the procedural concernsfor flight crews and air traffic controllers in a self-separation environment. Future research is neededto further examine the human performancecharacteristics in a broader spectrum of self-separation environments.

References

Cashion, P., Mackintosh, M., McGann, A., &Lozito, S. (1997). A study of commercial flightcrew self-separation. Proceedings of the 16th

AIAA/IEEE Digital Avionics Systems Conference,Vol. 2, (pp. 6.3-19 – 6.3-25). New Jersey: Instituteof Electrical and Electronic Engineers.

Daams, J., Bakker, G.J., & Blom, H.A.P.(1998). Safety evaluation of an initial Free Flightscenario with TOPAZ (Traffic Organization andPerturbation AnalyZer). NLR TR-98098.Amsterdam: National Aerospace Laboratory.

Eby, M.S. (1994). A self-organizationalapproach for resolving air traffic conflicts, TheLincoln Laboratory Journal, 7,(2).

Gent, R.N.W.H. van. (1995). Human factorsissues with airborne data link: Towards increasedcrew acceptance for both en-route and terminal flightoperations. NLR TP 95666. Amsterdam: NationalAerospace Laboratory.

Gent, R. N. H. W., van, Hoekstra, J. M., &Ruigrok, R. C. J. (1998). Free flight with airborneseparation assurance. Proceedings from theInternational Conference on Human-ComputerInteraction in Aeronautics, (pp. 63-69). Montréal,Québec: l’École Polytechique Montréal

Hilburn, B., Jorna, P.G.A.M., Byrne, E. A., &Parasuraman, R. (1997). The effect of adaptive airtraffic control (ATC) decision aiding on controllermental workload. In M. Mouloua, & J. M. Konce(Eds.), Human automation interaction: Research andpractice, (pps. 84-91). Mahwah, NJ: Erlbaum.

Hoekstra, J.M., Gent, R.N.W.H. van, &Ruigrok, R.C.J. (1997). Offline validation of aconceptual design for free flight in cruise. NLRContract Report. Amsterdam: National AerospaceLaboratory..

Hoekstra, J.M., Gent, R.N.W.H. van, &Ruigrok, R.C.J. (1998). Conceptual design of freeflight with airborne separation assurance. AIAA-98-4239.

Lozito, S., McGann, A., Mackintosh, M., &Cashion, P. (1997). Free flight and self-separationfrom the flight deck perspective. The First UnitedStates/European Air Traffic Management Researchand Development Seminar, Sacley, June 17-20,1997.

Remington, R., Johnston, J., Ruthruff, E.,Romera, M., & Gold, M. (in press). The effect offree flight on the detection of traffic conflicts byexperienced controllers. NASA Technical Report.Moffett Field, CA: NASA.

RTCA. (January, 1995). Final report of theRTCA board of directors’ select committee on freeflight. Washington, DC: RTCA, Incorporated.

Valenti Clari, M.S.V. (1998). Cost benefitanalysis of conflict resolution maneouvres in freeflight. Delft University of Technology, GraduationReport.

Wyndemere. (1996). An evaluation of airtraffic control complexity. NASA Contractor ReportNAS2-14284. Moffett Field, CA: NASA AmesResearch Center.

Yang, L., & Kuchar, J. (1997). Prototypeconflict alerting logic for free flight. (AIAA 97-0220). Reston, VA: American Institute ofAeronautics and Astronautics.

self-separation from the air and ground...

Documents