#202 velocities/uBody-fps, vBody-fps and wBody-fps has a different meaning between yasim and jsbsim

[Anonymous]

Originally created by: jean.pellotier@wanadoo.fr

just by standing with the brakes on the runway, if there is some wind, the properties velocities/uBody-fps, velocities/vBody-fps velocities/wBody-fps is relative to the ground for jsbsim aircraft, but indicate the speed in the air mass for the yasim planes, adding the wind.

this is transmited by mp in motionInfo.linearVel, and then is used in case we need to predict the AI Aircraft position, in AIModel/AIMultiplay.cxx.

I'm trying to use this prediction system to get rid of the jitter, and compensate for the lag for a better formation flight, but the result with yasim aircraft is bad because the speed used for the prediction is not relative to the ground.

jano

[Anonymous]

Originally posted by: bre... (code.google.com)@gmail.com

(No comment was entered for this change.)

Labels: FDM YASIM
Status: Accepted

[Anonymous]

Originally posted by: zakalawe@mac.com

By 'the jitter', do you mean bug 120? If so, it would be much, much better to fix that bug properly, instead of trying to hide the effects using a prediction system. Of course, this bug should also be fixed, but jitter is a real bug, I think - not something to be covered up with filtering.

[Anonymous]

Originally posted by: jean.pellotier@wanadoo.fr

by jitter i mean the mp planes doing the "yoyo", accelerating or decelerating depending on the udp paquets travel time, the thing that make impossible mp refueling if you are on different servers for exemple, this doesn't seem to be related to bug 120 to me.

jano

[Anonymous]

Originally posted by: zakalawe@mac.com

Okay, thanks for clearing that up. next question - assuming this is fixable, is changing it going to break anything in existing aircraft? I can't imagine many instruments or auto-pilots look at these numbers, and of course any core/general code that looks at them, would have to deal with JSBSim too.

[Anonymous]

Originally posted by: zakalawe@mac.com

According to jentron, the Yasim definition is probably the correct one - this may be an actual JSBSim bug, with the air-mass not being considered when on the ground.

[Anonymous]

Originally posted by: jean.pellotier@wanadoo.fr

I didn't found what speed uBody, vBody and wBody should be, a grep in the data show it is used for spinning the wheels (as ground speed, in this case jsbsim is correct), and sometimes I saw TAS (in this case yasim is ok).
If used in motionInfo.linearVel, the jsbsim definition is ok, as it provide a speed independant of the wind, providing the mp plane velocity in ground frame reference.

jano

[Anonymous]

Originally posted by: anders.g... (code.google.com)@gmail.com

IMHO the jsbsim interpretation is the correct one. It is the velocity of the aircraft in the local inertia frame expressed in the body coordinate system.
JSBSim also has properties for u,v,w-aero-fps etc which would be the components of the true airspeed expressed in the body frame.

[Anonymous]

Originally posted by: jean.pellotier@wanadoo.fr

Hi, I did a little tweak to my yasim, to have same speed as jsbsim, see attached patch.
but I don't know if i break something for now :)

jano

[Anonymous]

Originally posted by: tom... (code.google.com)@gmail.com

I have searched for some papers (three randomly chosen papers are listed below) dealing with aerodynamics and all of them state something like the following:

"u,v,w are the components of translational velocity of the cg relative to the earth, along the body axes" [3]

Identical behaviour of JSBsim and YASim would be great to allow for a common calculation of the flight path vector in the HUD.

[1] http://faculty.nps.edu/oayakime/ADSC/Slegers%20-%20Use%20of%20Dynamic%20Incidence%20Angle%20for%20Glide%20Slope%20Control%20of%20Autonomous%20Parafoils.pdf
[2] http://www.dtic.mil/dtic/tr/fulltext/u2/a408904.pdf
[3] http://naca.central.cranfield.ac.uk/reports/arc/rm/3669.pdf

[Anonymous]

Originally posted by: jean.pellotier@wanadoo.fr

back on topic after some time ...

here's a patch that introduce velocities/uBody_wind-fps, vBody_wind-fps, and wBody_wind-fps.
it make jsbsim and yasim reports the correct values for both ?Body-fps and ?Body_wind-fps.

this way depending the goal, the correct prop can be use in the Aircrafts (there are only a little involved), and in the instruments or the HUD.

i didn't have a look in larcim, or in the external fdm, as they are concerned too.

should i make a git branch with this patch? is it a way to address this bug?

jano

[Anonymous]

Originally posted by: jean.pellotier@wanadoo.fr

here are 2 merge request making u,v and wbody-fps ecef velocities in body axis, for yasim and jsbsim
included are the needed changes in the data.
for ada and uiuc, the value used should still be as before, as i didn't explore their internals :)
we will have a motionInfo.linearVel not wind related, and that can now be used in the hla code.

https://gitorious.org/fg/flightgear/merge_requests/47
https://gitorious.org/fg/fgdata/merge_requests/227

[Anonymous]

Originally posted by: kanto... (code.google.com)@gmail.com

Applying these patches on my branch, the results look ok to me.

[Anonymous]

Originally posted by: jean.pellotier@wanadoo.fr

this was commited:

[recfdf354f1607ec45d7ad2291fd550b063edf129] and
[ra6030bc9ce713b620647054d9cc9599521ad7e2d] in flightgear
and
[ref1ec369dbd422e17efb3959821069867f58dba4] for the gsdi necessary modification

[r34749396367cadf6ca6dffd7495a48152ddbc962] for the fgdata part

almost 3 years for a bug correction, :D

Status: Fixed