AI AOA significantly more than player AOA given same flight conditions

[Frederf]

The following is a table of AOA readings for a 4-ship human-led flight with excellent skill AI. Lead (human) is config full fuel, ammo, no external stores. #2 is same with #3 having a centerline 490L tank and #4 having 4 missiles, ASO-2, and a 490L tank. Level autopilot was held at 5000m on a DCS standard day. AI AOA values changed (reduced) over time due to their fuel burn but comparisons at various speeds were done at the same time. Values are from F2 info bar for IAS and AOA (°).

 

[TABLE=head]Aircraft|IAS 500 km/h|IAS 600 km/h|IAS 700 km/h|IAS 800 km/h|IAS 900 km/h

1|3.0-3.1|1.7-1.8|0.8-0.9|0.3|-0.1-0.2

2|6.2-6.4|4.0-4.4|2.6-3.0|1.6-2.0|1.0-1.1 @ 880 (100% thrust)

3|6.7|4.2-4.7|2.7-3.2|1.8-2.1|1.6 @ 837 km/h (100% thrust)

4|6.7-7.4|4.5-5.0|3.1-3.4|2.0-2.2| 1.8 @ 837 km/h (100% thrust)[/TABLE]

 

As it stands AI are very down on performance relative to a human pilot of the same config making meaningful AI-containing flights impractical. I can't find a solid explanation of why the AI values are what they are. I thought maybe the two FMs might be calibrated to different scales like:

 

IAS v TAS (at what alt?)

indicated AOA true AOA

fuel weight as kg v L

 

None of those theories has panned out.

[Rex854Warrior]

The following is a table of AOA readings for a 4-ship human-led flight with excellent skill AI. Lead (human) is config full fuel, ammo, no external stores. #2 is same with #3 having a centerline 490L tank and #4 having 4 missiles, ASO-2, and a 490L tank. Level autopilot was held at 5000m on a DCS standard day. AI AOA values changed (reduced) over time due to their fuel burn but comparisons at various speeds were done at the same time. Values are from F2 info bar for IAS and AOA (°).

 

[TABLE=head]Aircraft|IAS 500 km/h|IAS 600 km/h|IAS 700 km/h|IAS 800 km/h|IAS 900 km/h

1|3.0-3.1|1.7-1.8|0.8-0.9|0.3|-0.1-0.2

2|6.2-6.4|4.0-4.4|2.6-3.0|1.6-2.0|1.0-1.1 @ 880 (100% thrust)

3|6.7|4.2-4.7|2.7-3.2|1.8-2.1|1.6 @ 837 km/h (100% thrust)

4|6.7-7.4|4.5-5.0|3.1-3.4|2.0-2.2| 1.8 @ 837 km/h (100% thrust)[/TABLE]

 

As it stands AI are very down on performance relative to a human pilot of the same config making meaningful AI-containing flights impractical. I can't find a solid explanation of why the AI values are what they are. I thought maybe the two FMs might be calibrated to different scales like:

 

IAS v TAS (at what alt?)

indicated AOA true AOA

fuel weight as kg v L

 

None of those theories has panned out.

 

DCS AI uses an SFM and not a EFM or PFM like players do, the AI Mig doesn't have the same flight model as a player MiG, and that is for every aircraft. That's probably not going to change anytime soon.

[Frederf]

I know they use different FMs. They just have to have similar performance which is perfectly possible to code. All the other aircraft have matching player and AI FMs.

[asla36]

Talking about AI/player differences, y u AI can use chaff/flares separately on the MiG-21bis!?!?!?

 

However both these are up to ED, if I haven't missed anything of course.

[Dolphin887]

I posted short replay in other topic, but here it is copied

 

Few quick answers:

 

- UUA shows local angle of attack (note local)

- F2 view in human module will show you relative pitch angle although it says AoA

- local angle of attack is greater than relative pitch angle about 2 times

- exact local AoA calculation is implemented in the code

 

The only correct information about local AoA is what UUA shows.

 

 

Along with that, due to various SFM limitations (such is inabillity to move center of pressure along the medium wing cord and other things), AI aircraft will hardly fly on negative AoA. Particularly precise setup of SFM for AI is not so important as long as SFM closely meets aircraft performances.