NASA Self-Separation from the Air and Ground Perspective Margaret-Anne Mackintosh, Melisa Dunbar, Sandra Lozito, Patricia Cashion, Alison McGann, Victoria Dulchinos NASA Ames Research Center Rob Ruigrok, Jacco Hoekstra, Ronald Van Gent National Aerospace Laboratory, NLR

NASA Introduction NLR: Free Flight with Airborne Separation Assurance –Air perspective NASA Ames: Air-Ground Integration Study –Air and Ground perspective

NASA NLR Human-In-The-Loop Study Introduction NLR: Free Flight with Airborne Separation Assurance – Free Flight Concept Development: Traffic & Experiment Manager off-line simulations Find a suitable base-line concept – Free Flight Safety Analysis: Traffic Organization and Perturbation AnalyZer (TOPAZ) Predict critical non-nominal situations – Free Flight Human-in-the-Loop Simulation Experiment NLR’s Research Flight Simulator Human Factors Issues Validation of concept with Human-in-the-Loop

NASA NLR Human-In-The-Loop Study Methods Probe the limits –No Air Traffic Control –Air crew responsible for traffic separation All aircraft in scenario fully equipped –Automatic Dependent Surveillance - Broadcast (ADS-B) –Conflict Detection –Conflict Resolution –Cockpit Display of Traffic Information (CDTI) Cruise flight only –Direct routing –Optimal cruise altitude

NASA NLR Human-In-The-Loop Study Scenarios 8 crews, 18 runs per crew, 20 minutes per run current airline pilots 2 days including half a day of training Traffic Densities: Single, Double, Triple Level of Automation: Manual, Execute Combined, Execute Separate Non-Nominal: Other aircraft failures/events, Own aircraft failures/events, Delay time increased Traffic Densities: Single, Double, Triple Level of Automation: Manual, Execute Combined, Execute Separate Non-Nominal: Other aircraft failures/events, Own aircraft failures/events, Delay time increased

NASA NLR Human-In-The-Loop Study Concept Modified Voltage Potential Characteristics: –Fail safe –Co-operative –More options –Clear to pilot –Communication not required Similar in vertical plane

NASA NLR Human-In-The-Loop Study Flight Crew Interface Navigation Display –Traffic Symbology –Conflict Detection –Resolution Advisories –Vertical Navigation Display –Extra EFIS Control Panel functionality Modifications to Autopilot –Execute Combined –Execute Separate Aural alerts

NASA NLR Human-In-The-Loop Study Subjective Results: Acceptability Distribution of responses as a function of the three densities, across all sessions, across all subject pilots Acceptability: 91.5% (single), 83.0% (double), 78.7% (triple) Distribution of responses as a function of the three densities, across all sessions, across all subject pilots Acceptability: 91.5% (single), 83.0% (double), 78.7% (triple)

NASA NLR Human-In-The-Loop Study Subjective Results: Safety Distribution of responses as a function of the three densities, across all sessions, across all subject pilots Safety: 88.3% (single), 75.5% (double), 71.3% (triple) Distribution of responses as a function of the three densities, across all sessions, across all subject pilots Safety: 88.3% (single), 75.5% (double), 71.3% (triple)

NASA NLR Human-In-The-Loop Study Subjective Results: NLR Human-In-The-Loop Study Subjective Results: Workload Rating Scale of Mental Effort (RSME) Rating less than 40 (“costing some effort”) over all densities Results similar to cruise phase results in current ATC scenarios Rating Scale of Mental Effort (RSME) Rating less than 40 (“costing some effort”) over all densities Results similar to cruise phase results in current ATC scenarios

NASA NLR Human-In-The-Loop Study Objective Results: EPOG –Primary Flight Display: 8.1 % –Lateral Navigation Display: 48.9 % –Vertical Navigation Display: 7.6 % Eye-Point-Of-Gaze measurements Pilot Flying and Pilot-Not-Flying Percentages of the total fixation duration, averaged over the Pilot Flying and Pilot-Non-Flying, across all sessions: Eye-Point-Of-Gaze measurements Pilot Flying and Pilot-Not-Flying Percentages of the total fixation duration, averaged over the Pilot Flying and Pilot-Non-Flying, across all sessions:

NASA NLR Human-In-The-Loop Study Objective Results: Maneuvers Distribution of maneuvers as a function of the three different modes, across all sessions, across all subject pilots Maneuvers: Heading: 71.0 % Speed: 40.3 % Altitude: 48.7 % Distribution of maneuvers as a function of the three different modes, across all sessions, across all subject pilots Maneuvers: Heading: 71.0 % Speed: 40.3 % Altitude: 48.7 %

NASA NASA Air-Ground Integration Study Methods Boeing simulator and Airspace Operations Lab Flight deck and controller perspectives 8 DIA enroute scenarios (20 minutes in duration) 10 flight crews/10 controllers New display features on flight deck Airborne alert logic (no ground conflict probe) Controller tools similar to those at DIA Controller “monitoring” more than “controlling” Run in March/April 1998

NASA Background/Research Goal Background –RTCA Free Flight document recommends aircraft self- separation in particular situations (e.g., enroute environment) –Requires new conceptual airspace that includes human performance parameters –Aircraft self-separation will require a shift in roles and responsibilities between the users on the ground and in the air Research Goal –To conduct early simulations examining flight deck human performance parameters

NASA NASA Air-Ground Integration Study Scenarios Traffic on flight deck (ADS-B range 120 nms) Traffic on controller’s radar display (DIA Sector 9) Representation of high v. low density/clutter –High = aircraft, low = 6-8 aircraft “Blocker” aircraft preventing most common resolution Conflict event types: high and low density –Obtuse angle –Acute angle –Right angle –Almost intruder

NASA NASA Air-Ground Integration Study Displays Flight deck display –No early alert indication (prior to alert zone transgression) –Alert zone transgression display features –Temporal predictors and call signs selectable Controller Display –Similar features as those currently in DIA (e.g., vector lines, J rings) –Some features from CTAS, but no enhanced functions

NASA NASA Air-Ground Integration Study Flight Crew Results Density and detection time –Flight crews took longer to detect conflicts in high density compared to low density scenarios Conflict Angles and detection time –No differences in detection times between the conflict angles Ratings of conflict detection and time pressure –Significant increase in reported workload and time pressure as a function of traffic density No differences for almost intruder for detection times

NASA NASA Air-Ground Integration Study Pilot Detection Times

NASA NASA Air-Ground Integration Study Controller Results Effects of traffic density and conflict angle on detection times –Interaction between density and angle Longer detection time in obtuse angle high density v. obtuse angle low density Shorter detection time in acute angle high density v. right angle and obtuse angle high density Ratings of workload and task complexity –Significant increase in ratings of workload and complexity as a function of density –No differences for almost intruder detection times

NASA NASA Air-Ground Integration Study Controller Detection Times

NASA General Summary Consistent Findings across Studies –Impact for increasing density density may be exacerbated by other factors existence of abnormal situations (e.g. weather) may limit self- separation –Losses of minimum separation flight crews try to minimize separation between aircraft while maintaining legal separation controllers wanted larger separation than the flight crews maintained (NASA study)

NASA General Summary Unique Findings –Pilots fixate on CDTI 60% of the time and PFD 10% of the time (NLR study) Pilots reported spending too much time on the CDTI (NASA study) –Performance parameter usage Heading was most common parameter used (NLR study) –similar to previous NASA studies Altitude was most common parameter used (NASA study) –inclusion of the “blocker” aircraft in most common lateral escape path

NASA General Summary Unique Findings (NASA) –Conflict angles affect controllers and flight crews controller conflict detect times flight crew timing and type of maneuver –Density and conflict angle may interact –Frequent air-to-air communication

NASA Future Research Issues Addition of abnormal situations for workload realism (e.g., weather, winds, SUA, passenger problems) Assessment of data link for communications to help frequency congestion Simulation including representation of additional carriers and dispatch Information requirements assessment for shared situation awareness