DEVELOPING A TRAINING PROGRAM FOR THE TRAFFIC ALERT AND COLLISION AVOIDANCE SYSTEM IN CONTEXT Elizabeth “Scottie-Beth” Fleming Committee Members: Dr. Amy Pritchett, Chair Dr. Karen Feigh Dr. Ute Fischer Sponsored by the FAA, Tom McCloy as Technical Monitor Overview Develop a training program intended to improve pilots’ understanding of TCAS use for collision avoidance in a range of traffic situations + + + + + Introduction to TCAS and Background Information Design of Training Program Evaluation of the Training Program Results Conclusions 2 Traffic alert and Collision Avoidance System (TCAS) + Pilot always shown the Traffic Situation Display (TSD) + TCAS delivers a two stage advisory and vertical avoidance maneuver Traffic Advisory (TA) - ‘Traffic Traffic’ Resolution Advisory (RA) - ‘Climb Climb’ + Pilot is to follow an RA, even if it conflicts with ATC instructions, unless the pilot believes that safe flight would be jeopardized Federal Aviation Administration (2000). Introduction to TCAS II Version 7. Washington, D.C. 3 Collision Avoidance in a Broader Context Pilots don’t only interact with TCAS, they also receive information from ATC and the environment Notification and Awareness TCAS Traffic Advisory ATC Traffic Callout Visual acquisition of other aircraft Awareness of other aircraft via TSD Advised Maneuvers for Collision Avoidance Pilot’s Response Compliance to TCAS TCAS Resolution Advisory ATC Maneuver Compliance to ATC Personal Assessment and Maneuver Decision Party-line Information 4 Current TCAS Training Requirements: Ground Training + Classroom or computer based + >60 training requirements + Measures performance through quizzes and/or activities To comply to the RA, you should Pull the stick back Push the stick forward Federal Aviation Administration (2001). Advisory Circular No. 120-55B: Air Carrier Operational Approval and Use of TCAS II. Washington, D.C. 5 Current TCAS Training Requirements: Flight Training 6 Observed TCAS Use and Operation o The TCAS Operational Performance Assessment (TOPA) monitored the terminal area of 8 major airports and examined pilot compliance to climb and descend advisories TOPA observed compliance rates of Non-Compliance 41% to Climb RA’s Misunderstanding/ Confusion 59% to Descend RA’s Aggressiveness Olson, W. and J. Olszta (2010). TCAS Operational Performance Assessment in the U.S. National Airspace. IEEE/AIAA Digital Avionics Systems Conference. Olszta, J., & Olson, W. (2011). Characterization and Analysis of Traffic Alert and Collision Avoidance Resolution Advisories Resulting for 500' and 1,000' Vertical Separation. Paper presented at the Ninth USA/Europe Air Traffic Management Research and Development Seminar (ATM 2011), Berlin, Germany. 7 Observed TCAS Use and Operation Pilot reports flying into the red region on the VSI “[My FO and I] incorrectly interpreted the red 'above' target on the VSI and responded improperly. We further reviewed the procedures, agreeing that a person should fly 'away' from the red VSI indication, if instructed via RA.” ACN:785761, 2008 Non-Compliance Misunderstanding/ Confusion Pilot disagrees with advised descend RA “Descending into an airplane that is clearly descending? TCAS software clearly did not give appropriate guidance, nor did it self-correct when the initial guidance was so clearly wrong” ACN: 854982, 2009 Aggressiveness NASA. (2009). Aviation Safety Reporting System. Retrieved August 1, 2010: http://asrs.arc.nasa.gov/ 8 Observed TCAS Use and Operation Far Eastern Air B757 Response to Descend RA (TCAS advised a descent rate of 1500 FPM) Non-Compliance Misunderstanding/ Confusion Aggressiveness Image copied from Lacagnina (2008). Easy Does It. Aero Safety World : http://flightsafety.org/asw/oct08/asw_oct08_p44-47.pdf?dl=1 9 Research Statement and Objectives Develop a training program intended to improve pilots’ understanding of TCAS use for collision avoidance in a range of traffic situations (1) Train pilots to understand TCAS use for collision avoidance in the actual traffic and operational traffic environment (2) Provide pilots with a well-rounded knowledge of different traffic situations that may result in TCAS advisories 10 Approach to Training Design Complement of Two Methods Demonstration Based Training Event Based Training 11 Demonstration Based Training (DBT) + Computer-Based Training ~25 minutes Provides conceptual understanding of TCAS Outlines rules for compliance + 6 segments Introduction to TCAS Traffic Situation Display TCAS Advising Logic Traffic Advisories Resolution Advisories Example Timeline of RA Evolution + Mid-Training Quizzes 12 DBT: Demonstrations of RA’s 13 Event Based Training (EBT) + Presents traffic events that create the requirement to act + Builds context and complexity into each scenario as the flight progresses + Uses a more realistic training environment Fowlkes, J., Dwyer, D., Oser, R., & Salas, E. (1998). Event-Based Approach to Training (EBAT). The International Jounal of Aviation Psychology, 8(3), 209-221. 14 Simulator Study in Integrated Flightdeck – ATC Environment Audio Communications (Aviation Intercom) SideStick ATC->TSD VGA Simulation Architecture Touch screen First Officer PFD PartyLine B747-400 Simulator (RFS) ND Charts & Checklists TSD PFD Air Traffic Simulator (TGF) Air Traffic Transcripts Captain Eyetracker ATC ND TCAS Logic Experimenter/Instructor TCAS Alerts Coded Log of Flights Pritchett, A., Fleming, E., Cleveland, W., Zoetrum, J., Popescu, V., & Thakkar, D. (2012). Pilot Interaction with TCAS and Air Traffic Control. Paper presented at the 2nd International Conference on Application and Theory of Automation in Command and Control Systems (ATACCS), London. 15 Structure of Events: Descend RA Example Training Objective Context of Event Instrument Meteorological Conditions (daytime, clouds, no winds) ATC provides no traffic information Accurate interpretation o Conflict caused by IFR of and response traffic enroute (most to TCAS likely on departure) Descend RA. RA maneuvering should not violate ATC o instructions No conflicting ATC or party-line information Performance Measures Feedback Session Pilot disengages autopilot and flight directors Pilot responds to advisory with appropriate vertical If the pilot did speed Pilot ensures vertical speed is not meet any particular not excessive performance Pilot notifies ATC of response measure, to TCAS advisory as the review the maneuver is performed correct Pilot reengages autopilot and response in flight directors regards to that measure Pilot notifies ATC of TCAS advisory and response after clear of conflict Pilot returns to original clearance (if needed) 16 Evaluating the training program’s impact Overview of Study + + + + Baseline Study Comparison to identify the impact of training Training Study 6 scenarios with 2 traffic events per scenario Traffic events defined by RA type ATC information Traffic density 18 Overview of Study Pre-Training Data Collection (50 minutes) Pre-Experiment Questionnaire Pre-Experiment Quiz Introduction to TCAS TCAS Training Program (50 minutes) Demonstration Based Training Short Section Quizzes Event Based Training Evaluating TCAS Training Program (120 minutes) Debrief (15 minutes) Flight Scenarios Post Scenario Questionnaires Post- Experiment Questionnaire 19 Does the training program improve pilot performance in response to TCAS advisories and increase pilot understanding of TCAS? Skill-Based Behavior Rule-Based Behavior Knowledge-Based Behavior 20 Does the training program improve pilot performance in response to TCAS advisories and increase pilot understanding of TCAS? Decrease time to achieve compliance Decrease time to disconnect autopilot Reduce aggressive response features Increase percentage of RA duration in compliance Increase appropriate response post-Clear of Conflict Increase understanding of TCAS Increase trust in TCAS Skill-Based Behavior Rule-Based Behavior KnowledgeBased Behavior 21 Assessing the Pilot’s Response: Skill-Based Behaviors Pilot’s Response Vertical Speed TCAS RA Maneuver 5 seconds RA Climb X Time Pilot TCAS TCAS assumed Autopilot assumed constant vertical Disconnect First Achieves Compliance RA rate Time ¼ g pull-up Skill-Based Rule-Based Behavior Behavior 2 ½ sec TCAS weakens required vertical rate KnowledgeBased Behavior time Clear of Conflict 22 Mean Time to Comply: During Training pMM = Significance for the mixed model ps2 = Significance of the variance pm = Significance of the means Skill-Based Behavior Rule-Based Behavior KnowledgeBased Behavior 23 Autopilot Disconnect: Post Training + During training, no significant differences + Post training Autopilot disconnect time decreased Pilot response was more consistent for one event Skill-Based Behavior Rule-Based Behavior KnowledgeBased Behavior 24 Impact of Training Program Decrease time to achieve compliance Decrease time to disconnect autopilot Reduce aggressive response features Increase percentage of RA duration in compliance Increase appropriate response post-Clear of Conflict Increase understanding of TCAS Increase trust in TCAS Skill-Based Behavior Rule-Based Behavior KnowledgeBased Behavior 25 Assessing the Pilot’s Response: Aggressiveness Pilot’s Response TCAS RA Maneuver Maximum vertical rate Vertical Speed *Altitude Deviation Maximum vertical rate difference 5 seconds RA Climb TCAS TCAS assumed assumed constant vertical ¼ g pull-up RA rate Skill-Based Behavior Vertical rate difference 2 ½ sec TCAS weakens required vertical rate Rule-Based Behavior KnowledgeBased Behavior time Clear of Conflict 26 Assessing the Pilot’s Response: Compliance Pilot’s Response TCAS RA Maneuver Not in Compliance Vertical Speed In Compliance 5 seconds RA Climb TCAS TCAS assumed assumed constant vertical ¼ g pull-up RA rate Skill-Based Behavior 2 ½ sec TCAS weakens required vertical rate Rule-Based Behavior KnowledgeBased Behavior time Clear of Conflict 27 Assessing the Pilot’s Response: Return to Clearance Pilot holds new altitude Contacts ATC and request new clearance or ask for further instructions Altitude Pilot’s Response TCAS RA Maneuver Pilot begins descent back to originally cleared altitude May inform ATC of response to RA Cleared altitude at time of RA time 5 seconds RA Climb TCAS TCAS assumed assumed constant vertical ¼ g pull-up RA rate Skill-Based Behavior TCAS weakens required vertical rate Rule-Based Behavior Clear of Conflict KnowledgeBased Behavior 28 Before Training + 89% of the pilots did not know that an RA should cause less than 500 feet of altitude deviation + 56% of the pilots responded that they would hold current altitude achieved after responding to an RA, as opposed to returning to their clearance + When asked about airline procedures for following TCAS, all 18 pilots noted the need for complying with an RA BUT 28% of the pilots commented compliance wasn’t necessary is there was a TCAS “malfunction” or if the RA would cause an “unsafe situation” “[Pilots] must always comply with a TCAS RA unless [aircraft] performance is hindered (i.e. operating single engine) or [there is] an obvious TCAS malfunction (ie you can see traffic and it is not a threat)” Skill-Based Behavior Rule-Based Behavior KnowledgeBased Behavior 29 Aggressiveness + During training, aggressive response features decreased + Post training, same decreasing trend observed Altitude Deviation Average Vertical Rate Difference Maximum Vertical Rate Difference Maximum Vertical Rate Difference Skill-Based Behavior Rule-Based Behavior All decreased KnowledgeBased Behavior 30 Percentage Compliance + During training, percentage compliance decreased in training event with “Climb RA” (93.1% compared to 99.6%) + Post training No significant differences in means observed But, trained pilots had a more consistent response Skill-Based Behavior Rule-Based Behavior KnowledgeBased Behavior 31 Return to Clearance Pilot returns to original clearance Yes No Skill-Based Behavior Rule-Based Behavior KnowledgeBased Behavior 32 Impact of Training Program Decrease time to achieve compliance Decrease time to disconnect autopilot Reduce aggressive response features Increase percentage of RA duration in compliance Increase appropriate response post-Clear of Conflict Increase understanding of TCAS Increase trust in TCAS Skill-Based Behavior Rule-Based Behavior KnowledgeBased Behavior 33 Before Training + 94% pilots agreed with statement “I understand TCAS maneuvers when they are issued”, but… Only 50% of the pilots correctly identified the assumptions made by TCAS advisory logic When asked to interpret TSD symbols, only 50% of the pilots got all parts of the associated questions correct Skill-Based Behavior Rule-Based Behavior KnowledgeBased Behavior 34 Post Training Understanding of TCAS “My understanding of TCAS has increased” 61% 33% 0 0 6% Strongly Disagreed Disagreed Neutral Agreed Strongly Agreed + 50% pilots reported an increase in understanding TCAS logic + 27% pilots claimed to have learned about different types of TCAS RA’s (notably, “Crossing RA’s”) “Types of RA's were not previously taught. We were taught simply to comply” Skill-Based Behavior Rule-Based Behavior KnowledgeBased Behavior 35 Post Training Trust in TCAS “I am more likely to trust TCAS after completing today’s training” 39% 39% 11% 11% 0 Strongly Disagreed Disagreed Neutral Agreed Strongly Agreed “My trust in TCAS was already at a maximum so I wouldn’t be ‘more’ likely to trust it” Skill-Based Behavior Rule-Based Behavior KnowledgeBased Behavior 36 Post-Training Decrease time to achieve compliance Decrease time to disconnect autopilot Reduce aggressive response features Increase percentage of RA duration in compliance Increase appropriate response post-Clear of Conflict Increase understanding of TCAS Increase trust in TCAS 37 Conclusions and Future Work Conclusions + Pilots may not need more training, but instead need better training Integrating DBT and EBT methods permits ground-based and flight training material to be more cohesive EBT structure allows for the design of purposeful training events + Current FAA mandated TCAS training objectives may not fully reflect all training areas Language needed to address the reduction of excessive responses to RA’s Future Work + What facilities and technologies would be required to implement this type of training program? + What implications arise when considering training design versus system design? Would incorporating human factors considerations in initial design stages decrease the amount of required training? 38 39 Conclusions and Future Work Conclusions + Pilots may not need more training, but instead need better training Integrating DBT and EBT methods permits ground-based and flight training material to be more cohesive EBT structure allows for the design of purposeful training events + Current FAA mandated TCAS training objectives may not fully reflect all training areas Language needed to address the reduction of excessive responses to RA’s Future Work + What facilities and technologies would be required to implement this type of training program? + What implications arise when considering training design versus system design? Would incorporating human factors considerations in initial design stages decrease the amount of required training? 40 Acknowledgements + + + + + + + + + + Work sponsored by the FAA, Tom McCloy as Technical Monitor 34 pilot participants Dr. Amy Pritchett Dr. Karen Feigh Dr. Ute Fischer Dr. Wesley Olson, MIT Lincoln Labs Wayne Gallo, FAA Roger Sultan, FAA Kylie Garey TCAS Team: William Cleveland, Vlad Popescu, Justin Mullins, Anil Bozan, Henry Tran, Jack Ridderhof, Alyssa Whitlock, Colin Ludwig, Dhruv Thaakar, Jonathan Zoetrum, Jelle Wissink + CEC Lab Members 41 42