On Sunday 16 March 2008 03:51, Syd&Sandy wrote:
> Hi all ,
>       I've been trying to change the xmlautopilot to use <prop> and
> <value> for the u_min and u_max properties , and currently have
> quite a mess on my hands right now :) The idea is to have a min
> and max property to control bank-limit / pitch with a panel knob
> ... setting the u_min and u_max from a property seems to be
> working , but I get some strange things happening . The pi-simple
> controller isnt clamped anymore (so i removed the clamp check
> )... and the output goes immediately to the u_min
> value...although u_min and u_max are checked every update... Has
> anyone else attempted this , with good results ? Or , hopefully ,
> already implemented this ? Anyway , I'll keep plugging away at
> it, the answer is probably staring me in the face and I can't see
> it. Is this something that should be implemented anyway ?
>  Cheers

Hi Syd,

one way you could do this with the current autopilot controllers is 
to feed the output from your controller through a gain filter to 
get the range you want.

For example, if you've set u_min/u_max to +/- 40 in your controller 
but want to reduce it to +/- 20, you'd set the gain value on the 
gain filter to 0.5.

If you don't mind re-tuning your controller, it would probably make 
more sense to set u_min & u_max to +/- 1.0, then the gain factor 
would be the required bank or pitch angle limit i.e. for +/- 30 
limits you'd use a gain of 30.

Changing the output clamps does change the overall behaviour of the 
controller, however.  I found that I got more desirable behaviour 
from a pitch controller (output is a hstab deflection) when I set 
the u_min & u_max limits to +/- 0.25 and then passed it through a 
gain filter with a factor of 4 to restore the required +/- 1.0 
range, as opposed to setting the clamps directly to +/- 1.0.

Heh - I'm still not entirely sure why this is, actually having 
fiddled with the code myself, but it came about through an 
experiment where I was trying to increase the effective bandwidth 
through parallelism.  I started off with four identical controllers 
running in parallel, the outputs of which were summed but then I 
realised that I could get the same effect with a single controller 
using the gain filter technique.

LeeE

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Flightgear-devel mailing list
Flightgear-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/flightgear-devel

Reply via email to