Last update: July 17, 1998

Overview

• Meeting agenda for Thursday, 7/16/98
• Addressing mode awareness issues for current systems
• Training for current systems
• Findings/lessons learned
• Principles/guidelines
• Recommendations
• Training issues
• Procedure/operations for current systems
• Findings/lessons learned
• Recommendations
• Addressing mode awareness issues for future systems
• Design/aircraft certification for future systems
• Findings/Lessons learned
• Specific criteria for design
• Design principles
• Evaluation for future systems
• Evaluation process
• Training for future systems
• Procedure/operations for future systems
• Integration, or pulling it all together
• General lessons learned
• Recommendations
• Questions

Meeting Agenda for Thursday, 7/16/98

Summarize issues:

• Beth’s material, references, web site
• Categories of mode confusion and specific gotcha’s

Addressing mode awareness issues for current systems

• Unless we identify issues that are AD issues,

• Training for current systems

  • Findings/lessons learned

    If a mental model is not provided/trained, the pilots will develop their own

    "Information about specific aircraft may exist but is not necessarily disseminated to all operators of the aircraft" ATA may be a forum for doing so? To manufacturers? POIs? Bizjets? Vendors? IATA? ICAO?

    Training programs are very diverse at different carriers - assumptions about what gets trained must be made carefully.

    Training data could be extremely useful in understanding problems in mode awareness, etc

    There is FAA guidance for training in using autoland in low-vis, but not for using FMS

    Training is still oriented to passing the checkride.

    Training is usually oriented towards a fully functioning system - MEL factors.

  • Principles/guidelines

    Crews need to know the design philosophy and operating policy. Material on design philosophy must come from the manufacturer/vendor.

    Train a consistent policy for dealing with automation based on the carrier’s policy.

    Training philosophy is still based on non-glass aircraft. We need to change this.

    FAA needs better communication between ops, research, a/c cert.

  • Recommendations

    FOQA data - use for looking at mode issues?

    Improve completeness of info provided to pilots - especially for s/w changes?

    Provide training that the pilots can get on their own.

    Look into using CRM to increase mode awareness.

    Airplane buyers should ask for training support as part of buy: e.g., BCAS example

    Make use of WWW as a forum for getting information about specific questions, airplanes, systems

  • Training Issues

    MFF between aircraft types (e.g., 777 and 747-400),

    90-day landing currency

  
  

• Procedure/operations for current systems

  • Findings/lessons learned

    Major problems are experienced when transitioning from manual to automatic flight

    Manual content? (Ops manual, flight manual, etc)

    Documents should be evaluated for usability.

    Lack of procedures on how to cross check modes

    The feedback loop on SDRs is training. Is another mechanism needed?

    Keeping track of information distributed can be hard. Some of it should end up in the flight and training manuals but often it doesn’t.

  • Recommendations

    Explicit/ formal/ documented procedure for assuring mode awareness & cross check of automation

    Ongoing feedback to pilots about in service experience re gotchas

    Hotline for pilots to ask questions? In addition to the web?

Addressing mode awareness issues for future systems

• Design/aircraft certification for future systems

  • Findings/Lessons learned

    "Mode annunciation alone may not be salient enough to communicate mode information that the crew may need to know. So the consequence of missing the mode change should be considered."

    Must include the design assumptions that were made

    Definition of mode awareness is important (what mode, what behavior, is it correct)

    Lots of useful info in different places - can the crew keep up?

    Design/training footprint may be driven by marketing, then design/build team must meet that promise

    Design originally agreed upon at TC board may get changed drastically - affects training.

    Standardization?

  • Specific criteria for design

    Goal should be to have all autoflight mode info in one place? MD-11?

  • Design principles (could include principles for design process)

    "The FMA should communicate the mode engaged, and the target to which it is going, and the source of the target value."    Source needed to know how to control the target.

    HUDs should include mode annunciation in the same or consistent format to the PFD

    "anticipation" information can be useful (speed ghost bug, green arc, etc)

    Related information to perfrom these tasks should be "close" to each other to support the pilot’s scan

  
  

• Evaluation for future systems

  • Evaluation process (Two different kinds - design and certification evaluation)

    The responsibility should be on the applicants to demonstrate the effectiveness of their design to meet the intended function from a flightcrew perspective; e.g., effectively convey the information, such as mode annunciation. The demonstration could be done in a variety of ways, but should be mutually agreed upon.

    Pairwise combination of modes - look at consequences of mistaking them.

    Case of system annunciating one mode when another is active (example of hdg sel when want to follow track, scaling of map display) eg 007,

    We could recommend eval questions, some eval process suggestions

    Closer tie between design eval and cert eval; some structure provided to the process

    Transition matrix to describe mode transitions for flight test eval.

    Getting evaluation started early - eg low fidelity can be done early

  
  

  • Training for future systems

    Checking evaluates the quality of the training

    How do we evaluate training? What tools/level of fidelity for what training?

    Parallel development for PC-based training

    Development of training for new system(s)

    Stuff from above

    Closely integrate with design

    If a box in added, update FSB report for training

    Develop training objectives up front

    Task analysis might be used to help both design and training

    Evaluation can help training also

    Part 91 training requirements (type rated or not)

    Transitioning from manual to automated flight (especially unexpected system behavior when transitioning between the two)

    Nonnormal operation and equipment failures ...

  
  

  • Procedure/operations for future systems

    Major problems are experienced when transitioning from manual to automatic flight

    Manual content? Inconsistent requirements. Provide greater detail on system operation and function (ATA?) (need to determine what level of detail is required from operational perspective)

    Add mode awareness information in manuals

    Operational input early in the design

  
  

  • Integration, or pulling it all together

    Training philosophy must match design philosophy and operations philosophy vice versa

  
  

  • General lessons learned

    Crews regularly struggle with mode awareness.

    Feedback from in-service experience and other domains is critical to design, cert, training, procedure development. E.g. with formal methods/visualization, may need to know what kind of issues to look for

    If you have an idea what issues to look for, methods exist to support analytical evaluation

    Multidisciplinary approach is necessary; e.g., identify issues and determine whether they exist in specific design

    Note for design section: perceived/conceptual complexity and the virtues of simplicity

    Different definitions of mode exist

    Training and procedure improvements are necessary but not sufficient.

    Improved reliability of systems have lead to pilot performance degradation and loss of SA, and mode awareness (fail to monitor the system closely)

  
  

  • Recommendations

    Create a summary of these existing methods for modeling/visualization to support evaluation - tools for designers, etc. How distribute?

    Create a list of known confusion areas e.g., MD8x ALT CAP

  
  

  • Questions

    How do we design/evaluate/certify/train/operate future systems to address mode awareness?

    What is the tradeoff between current precedent and improvement in the future?

    How do we compensate for known violations of HF principles? E.g., separation of displays and controls

    How do we address current systems?

    -    Training

    -    Best practices

    -    Procedures

    Standardization? How much to encourage? Part 23 vs 25

    How do we define performance?

    If I am an aircraft certification official, what do I do to address mode awareness issues?

  
  
  

  ---

  Return to