On Mon, 8 Jul 2002 16:54:12 -0500 (CDT)
  "Curtis L. Olson" <[EMAIL PROTECTED]> wrote:

>However, we would also need to be able to turn off the auto-failure
>generation module and allow an instructor (or a script) have complete
>control over the failures.  This way an instructor could use the sim
>to train for specific failure scenarios and have complete control over
>when, where, and what happens.

For shuttle training, the instructors have scripts, but 
they still enter the mals by hand. Also, pretty much any 
sim variable can be accessed and changed by hand. The 
functionality to change a set of variables and then have 
them take effect at a particular time is also there. For a 
complex machine like the space shuttle, malfunctions are 
often not the product of simply turning on or off a 
boolean (although many are). Some are so complicated that 
they cannot really be programmed, but can give the correct 
"symptoms" if an instructor has the power to do as he can 
in the shuttle mission simulator and access any variable.

The ability to access any variable (perhaps via the 
property tree) could open up a whole range of surprise 
training scenarios. This was demonstrated to me (and Curt, 
you may have seen something similar in your encounters 
with "real" commercial training sims) on a session where I 
was testing some changes I had made in the shuttle sim 
code at Johnson Space Center. I had made a series of nice 
landings (OK: "survivable" landings) in the shuttle 
simulator early one morning in testing my code when the 
operator thought I was getting a bit cocky. Just as I 
flared during the last landing he gave me a 100 knot 
tailwind. If there would have been a black box, it would 
have gotten from me only a "What the ... !" before I 
pancaked in.

When I worked F-16 FCS modeling we had failures like CADC 
failures (air data computer), stuck switches, etc. Some of 
the failures we might want to eventually model might 
involve simulating sensors - i.e. right now we assume we 
are getting air data, rates, velocities, etc. as-is from 
the FDM. If we get a bad sensor it might read 0.0, or 
stick at a value, or something else depending on the 
associated electronics. I wonder if we might need to model 
a sensor class someday.

Jon

_______________________________________________
Flightgear-devel mailing list
[EMAIL PROTECTED]
http://mail.flightgear.org/mailman/listinfo/flightgear-devel

Reply via email to