Decision Support Research – Specific Decisions We Could Support

• VFR Flight into IMC

• Runway Winds

• Storm Avoidance

Research Topics in Support of These Decisions

• Trending

• Alerting

• Customizing the display based on phase of flight

• Using the computer to put together complex elements into custom views

How the Research Topics Relate to the Supported Decisions

I think this is the list of items we are testing whether should be included in the "minimum service requirements" 

• Trending

  • IMC Decision: Indicate deteriorating/improving conditions via icons on visible part of the map
    • How represent the trend?
    • Trend based on a composite variable (flight category), on on individual components?
    • If trending info is provided on map, details should be provided when pilot touches the station
  • IMC Decision: Indicate deteriorating/improving conditions at important stations not currently in view on the map or airport screens
    
    • Does this fall under 'alerting'? I.e. are we alerting when we post a notice or otherwise indicate off-screen activity that the pilot should be aware of?
  • IMC Decision: Indicate deteriorating/improving conditions in the text reports
    • Could implement this by showing components of report that are deteriorating/improving
    • Should we move to decoded reports?
    • Could add icon representation to places where text reports are shown, and use similar semantics for map trending symbols
  • IMC Decision: Spatial trending, such as motion of storms
  • IMC Decision: radar loop from 2hrs in the past to now. Gives a sense of convective trend and movement.
  • Things we should show trending for:
    • Ceiling/Vis (IMC)
    • Storm avoidance
    • icing
    • turb
    • adverse winds – en route and takeoff and landing
  • For presentation, show them something we could do? A picture will help, but we don't reveal the picture until after facilitating a more theoretical discussion.
    • Map with trending symbols and some indication of trending on the map inset stations
    • Airport screen with a meteorogram-style plot. Ceiling trending down, vis trending up, etc.
  • Could we create a grid of trending using multiple timesteps of C+V where the grid represents diffs/trend btwn timesteps.
  • Maybe at airports you have an up and down arrow for c and for v. Then you go to the airport view and you see more variables.
  • Simple representation for trending where space is limited
  • There are more-complicated indicators of trending. Not sure what these are. 
  • Could pilot 'experts' provide some information about this?
    • What are key items to indicate trending on for 'cruise' phase?
    • What about takeoff/landing?
  • Trending in two senses: 
    • what will conditions be in future? 
    • how reliable is the latest reading?

• Alerting

  Note: We are in the process of defining what alerting is. Thesholding plays a part. Also showing things that are currently off-screen

  • Pilots set thresholds for various hazards and are notified by the app if conditions on the route of travel pass these thresholds (requires a new service that does the heavy lifting for this sort of query)
  • How notify the user?
    • Issue alerts and allow user to configure how they want them to manifest
    • Build in flexible alerting
    • Another angle is to make the Evals about whether a pilot was notified via an alert was beneficial, not how they were notified (popup, color change, text size change, beep, etc)
    • Another angle is diving into the thresholding – which ones can we figure out scientifically, and which ones do we need to leave configurable
    • Schedule a session with Bruce Carmichael
  • Alert en route, or only at airports?
  • How not to repeat alerts, over-alert, or interfere with use of the app? 
  • Pilots seemed alert averse during user interviews. How to make alerting in the MMet app different from whatever has caused this aversion for the pilots?
  • Three ways we could look at alerting:
    • How to depict alerts
    • Thresholding
    • Whether having an alert is valuable – beneficial in that it increases safety
  • Another way to do alerts is to have a list of alerts of time to next encounter with various hazards
  • If we are tracking trends on something, we will most likely be alerting on them.

• Customizing the display based on phase of flight

  Min svc rec question here is "should we recommend that views be customized for each phase of flight" and make some recommendations on what should be different between views.
  
  • IMC Decision: In 'cruise' mode the app could enable a mode or screen that monitors weather on the entire route.
  • IMC Decision: In 'cruise' mode, storm view with forecast of ownShip location along with storm position forecasts.
  • Winds Decision: In 'takeoff/landing' mode, the airport screens could include special views that support this phase of flight
    • Airport diagram with winds plotted.
    • Any other examples?
  • "Wizard" view with progress arrows to show where you are in the process.
  • EXP ADDS: Would there be an argument for having  a loop? 
    • Problem is we are assuming the pilots wants all of this information. And force them to sit there and look at it all.
    • What about having the user pick which products are included in the loop, then start the loop off? Rather than assuming. 
    • Weather scan at each phase of flight
  • It would be VERY helpful to see others' research on workflows in different phases of flight. Bruce? Tenny?

• Using the computer to put together complex elements into custom views

  RETITLE this "Custom Weather Analysis"
  
  • Time to next encounter with each of the listed hazards
  • Winds Decision: winds plotted on map of runways
  • Winds Decision: simple list of headwind and crosswind on each of the runways
  • Winds Decision: resolve the list (above) into parallel and orthogonal vectors relative to each runway (get an example of this from Aaron)
  • Winds Decision: make all of these views runway-centric
  • Winds Decision: If we can't get the 1-minute obs, we can gin something up for the Eval and identify this as a technology gap
  • IMC Decision: ownShip location 'forecast' and plotted on forecast weather
  • IMC Decision: look at wx impacts on entire route of flight (either via grid or point data resolved onto flight route in 4D)
  • IMC Decision: build a service for the mobile app that allows the app to ask "What impacts am I going to encounter on this route of flight during this time?"
  • Winds Decision: if we had access to 1-minute observations at airports, we could provide a service that gave wind averages/variability over the last hour
  • IMC Decision: optimal path through hazards – in case pilot gets too close to something that should have been avoided (requires cost function)