On Friday, June 14, 2002, at 04:19 am, Jim Wilson wrote:
> Armed vs. Active modes: > Active modes are the only ones that affect control surfaces. Armed > modes > select an impending active mode that may or may not get triggered, > depending > on if the condition for that mode is satisfied. Note that if you have > armed > for a higher altitude but have the aircraft set maintain pitch down, > you will > continue to descend and never trigger the armed altitude hold mode. > Note that > often a flight computer can ensure that an armed mode is reachable by > coordinating multiple settings, but the autopilot by itself does not > guarantee this. Okay, this sounds fine. > > c. Heading Select (HDG-SEL) uses ailerons to turn toward heading bug (or > setting from FMC). Switches to active Heading Hold when target reached. > Changing heading bug after Heading Hold is active requires initiating > HDG-SEL > again (won't automatically go there like it does now). This needs to be configurable, most HDG modes I've 'used' on big jets remain live, i.e if you spin the heading bug, they track immediately. It's possible there is a ALT mode out there that does the same thing (though that would be pretty crazy, I admit) > > d. Heading NAV Arm (HDG-NAV1-ARM, HDG-NAV2-ARM) Maintain HDG (Heading > Hold) or > Wing Leveler (WL) until radial is intercepted. Upon interception, > automatically switch to active HDG-NAV1 or HDG-NAV2. Note that on ILS > this > will be synonymous with APR ARM and APR mode which will also arm for GS. The intercept conditions need to be programmable. Also, the mode may not engage at all based on criteria in some systems. Example: the 777's NAV mode will not engage if the CDI is greater than some value (2nm?) *and* the current heading it not an intercept. (at least, according to the simulation I have used) > AXIS 2) Vertical control modes: look fine > > AXIS 3) Yaw damper: I am anxious that this be a totally separate system from the autopilot, it's a very independent system apart from the fact it drives control surfaces and the switch is usually on the MCP :-) If we ever go as far as systems modeling at the block level, that's especially the case (yaw dampers usually have separate gyros, and sometimes separate inputs to the control surface actuator servos) > > Engage: > > Upon engagement the autopilot will begin processing any preselected > modes > (limited to one Arm and one Active mode per axis at a time). If no > mode is > preselected then vertical will default to PITCH and lateral will > default to > HDG-HLD. It may be desirable to have these defaults configurable, > since some > autopilots may default to either ALT HLD for vertical or WL for lateral. Definitely. Also note that we're going to need some ability to reject engagement of modes altogether, so there is really a multi-stage set of the conditions for each mode: - conditions to arm the mode - conditions to activate the mode - conditions to transition to another mode (largely an FMC problem, but it does interact I suspect) > The autopilot will be able to disengage. Certain conditions will > require it. > Suggestions on how to specify conditions would be welcome. Note that > the > current autopilot will reduce pitch to prevent a stall (as "min climb" > is > approached). I believe that the more correct behavior would be to > disengage > once airspeed reaches a certain minimum. Some aircraft have other > automated > stall prevention measures (e.g. automatic nose dropping on aircraft > that are > incapable of stall recovery), but those involve separate systems that go > beyond the scope of typical autopilot functionality. Actually, I believe some auto-pilots will just 'kill you' in this situation :-) So it needs to be configurable for certain. > Notes: > > The current autopilot supports some degree of throttle control...not > sure if > this should be included or if it belongs in a separate class. In any > case > thrust control gets involved with some aircraft. I would prefer a separate class, as you say, this can get quite involved. > > A variety of flight management modes should be configurable with these > basic > autopilot functions. For example if the Boeing "FL CH" mode is > selected on an > FMC then the FMC can command the autopilot for "ALT ARM" and "SPD", and > until > a thrust computer class is developed, set the throttle controls with an > approximation for throttle in climb or descent mode. Yup, all sounds rather plausible., > > This is the second time I've outlined autopilot functions, but what I'm > striving for here is to actually outline how the class methods will be > constructed, and the way that the modes will interact. Now is the time > to > set me straight if something doesn't look right, or looks too > inflexible :-) This looks like 90% of what we need (i.e, I can't see anything missing, but we haven't built it yet :-) H&H James -- What I like about deadlines is the lovely whooshing sound they make as they rush past. -- Douglas Adams _______________________________________________ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
