Heavy pos- and neg-G from AI wingman, at moderate altitude and airspeeds
-
Fork from the amraam thread… here’s some snips of my long-suffering winman in the TR#15 JSOW mission – Sparky32.
Here he pulled positive-10+ G immediately after negative-10+ G, dodging some SAMs. Good thing he’s a robot… Based on positioning and altitude, I suspect he may have been dodging AAA at that time also.
But regardless, I don’t know if this is considered “normal” flight mechanics for the non-afm flight model used by AI, or if it’s something to investigate. I don’t think the afm model can pull half those G’s, in either direction, at those airspeeds.
Repro: stock KTO, TR#15… tell poor Sparky to attack those columns.
-
-
ACMI => https://file.io/qKBAQGMi9CGJ
(final 90 seconds or so) -
@Mav-jp can you check?
-
I think there was a discussion internally if the Tacview G calculation matches the one in BMS…
Don’t have the answer…
-
If that helps a bit, highest measure on both.
-
@airtex2019 try to set some low level flight for AI and give them a support…:D they turn and burn like no other Like acceleration from 250 to 450 in five seconds straight up to the sky…and im talking about A10s
-
@Seifer said in Heavy pos- and neg-G from AI wingman, at moderate altitude and airspeeds:
@Mav-jp can you check?
Yea I will
But tacview G is not reliable
-
OK
https://mega.nz/file/jk9VnY4Y#JSCwj5CTIfXy_XWQ5mF8ZBRPqR9y-1mfVpgODXl2ASI
Please find attached two files
-
is an ACMI
-
is a log file with direct monoprinting of the G force in game.
The time stamps are similar in ACMI and XLog
AS you can see most of the time, both G’s are matching…but …
10:13:50
ACMI SHOWS 6.8 G , Xlog shows 3.3G peak10:13:51
ACMi shows -1.9G , XLog shows Gmin +1.710:13:54
ACMI SHOWS 6.6 G , Xlog shows 2.2G peakNowhere in the xlog you will find negative G’s or > 5.0G G’s…
This is the demonstration that ACMI G’s data is absolutly NOT reliable.
this calculation is guestimated at best -
-
@Mav-jp awesome, thanks sir. I suspected it may be off a little, didn’t expect it to be quite so dramatically different!
-
@Mav-jp But what about their ability to accelerate at seemingly impossible speeds (even from a nasty high AOA?)
-
Why don’t these bogeys G-LOC in accordance with the BMS blackout model?
-
@Stevie did you miss Mav-JP’s analysis?
Nowhere in the xlog you will find negative G’s or > 5.0G G’s…
the key takeaway is some metrics in TacView, like G and lateral-acceleration etc, are computed – derived from changes in position and orientation
maybe an artifact of low sample rates? (I didn’t hear anyone say, but I’d certainly believe if BMS might record AI units at a lower sample rate than humans, to massively reduce acmi file size)
-
@Stevie said in Heavy pos- and neg-G from AI wingman, at moderate altitude and airspeeds:
Why don’t these bogeys G-LOC in accordance with the BMS blackout model?
They do !!
How many times I will have to repeat
Tacview Gs are NOT I repeat NOT real in game !!!
-
@airtex2019 said in Heavy pos- and neg-G from AI wingman, at moderate altitude and airspeeds:
@Stevie did you miss Mav-JP’s analysis?
Nowhere in the xlog you will find negative G’s or > 5.0G G’s…
the key takeaway is some metrics in TacView, like G and lateral-acceleration etc, are computed – derived from changes in position and orientation
maybe an artifact of low sample rates? (I didn’t hear anyone say, but I’d certainly believe if BMS might record AI units at a lower sample rate than humans, to massively reduce acmi file size)
I think there is also possible a tacview wrong computation from the Euler angles accelerations .
Also of course the sampling can have very bad effects on accelerations computation indeed
Ideally we could export Gs in the acmi
-
@WPNS24 said in Heavy pos- and neg-G from AI wingman, at moderate altitude and airspeeds:
@Mav-jp But what about their ability to accelerate at seemingly impossible speeds (even from a nasty high AOA?)
AI are flying the exact same FM and run same code routines as players (OFM)
Can you provide a reliable repro case please ?
-
@Mav-jp - yeah, I saw that. but I also thought I saw some “from game” examples as well?
If we should throw the TACView data out (which doesn’t bother me one way or another), then what are we actually talking about?