Re: [Jsbsim-devel] Position/orientation-hold

[Agostino De M. <ago...@un...>]

Quoting David Culp <dav...@co...>:

> On Saturday 29 December 2007 08:25, Jon S Berndt wrote:
>> This sounds like a good idea. Could be that if the hold-position flag
>> was set, then the final position and/or orientation of the aircraft
>> would be set to its previous value, overriding any newly calculated
>> value. The new "previous value" would then be set (to what it already
>> is), and the position/orientation would hold.
>
> Yes, that should do the trick.  Actually there are three types of position
> freeze that I can think of:
>
> 1)  Lat/Lon Freeze:  the orientation and altitude can change, but  =20
> not lat/lon.
>
> 2)  Position Freeze:  the orientation can change, but not lat/lon/altitude=
.
>
> 3)  Full Freeze:  the orientation and position freeze.
>
> The first one is the kind of freeze I see used in the high-end sims I deal
> with.  This allows one to manually trim up for a given configuration prior=
 to
> being "cut loose".  In this case the airplane feels like it's flying becau=
se
> the altitude is allowed to change.
>
> The second one I've not seen before.  Off-hand I can't think of a benefit.
>
> The third one would give a useful wind tunnel effect for data collection a=
nd
> testing.
>
> Dave


Could we also think of a kind of freezing that allows a 3dof simulation?

This can be done obviously at the model configuration level by blanking
some sections in the aerodynamic model.
But I think that sometimes there's the need to verify the longitudinal
motion aerodynamics of aircraft complete models, without necessarily
editing them.
Having already a full model, the user could set the 3dof flag to true,
enabling only the longitudinal motion: lat, lon, altitude, pitch angle.

This feature could become useful when we test simple models, or
when we want to verify aerodynamic data taken from textbooks,
or when we compare alternative models and we don't want to care about
the effects of coupling between longitudinal and lateral-directional
characteristics.

I guess that this could be done simply by setting the Y-force,
the rolling moments, and the yawing moments, of any type, to zero in the
time stepping cycle.

Agostino