Air combat AI tactics

Matrix Games Forums

Forums Register Login Photo Gallery Member List Search Calendars FAQ

My Profile Inbox Address Book My Subscription My Forums Log Out

Air combat AI tactics

View related threads: (in this forum | in all forums)

Logged in as: Guest

Users viewing this topic: none

Printable Version

All Forums >> [New Releases from Matrix Games] >> Command: Modern Operations series >> The War Room >> Air combat AI tactics

Page: [1]

Message

<< Older Topic Newer Topic >>

Air combat AI tactics - 10/21/2017 4:07:55 AM

wildcolonialboy

Posts: 19
Joined: 10/18/2017
Status: offline

Greetings

I was just messing around with 1-v-1 AAW scenarios, in this case F-22 Raptors (2006 model with AIM-9M and AIM-120C) and a 1980s version Su-27.

I decided to see what would happen if I removed all the radar-guided medium range weapons and left them with only one IR-guided missile. I placed the F-22 at 15,000 feet and the Su-27 at 1,000 feet, within each others AIM-9 / R-73 range. I wanted to see what effect the difference in altitude would have, particularly because on my understanding of E-M theory, the aircraft with the greater altitude would have a significant advantage (the missile from the lower-altitude aircraft would have to expend greater energy to climb, the missile from the higher-altitude aircraft would have an energy advantage).

The F-22 fired first and the Su-27 turned to evade before it got off its own shot. The AIM-9M missed, the final PH was only 17% (adjusted for distance 44%, high deflection impact no effect on agility, final agility modifier -27%). I got the Su-27 to swing around and fire off its R-73, and then something strange happened.The F-22 went into a very steep dive, I think it actually managed to get below the R-73 as the R-73 couldn't track it fast enough and manoeuvre before the F-22 zoomed down past it in the vertical plane.

I was surprised the F-22 did this. Surely this would cause it to lose the energy advantage given it could just switch on the afterburns, zoom upwards and get the R-73 to chase it? In any case, the R-73 also missed with a final PH of 21% (adjusted for distance 89%, high-deflection impact no affect on agility, final agility modifier -38%, Sea-skimmer modifier -30%).

So my questions are, why did the distance modifier favour the R-73 so much given they were at comparable ranges, and if anything the AIM-9M had the E-M advantage? What is the sea-skimmer modifier (is this that at extremely low altitude, the air is denser meaning more drag for a missile + easier for the seeker to be confused by the either RF or IR reflections from the surface of the ocean)? Finally, is the AI's tactic to hit the deck (which makes sense if a missile has been on a lofted trajectory and above it) no matter what the altitude of the missile relative to aircraft altitude realistic and correct in the circumstances, or is that something for a future update?

Post #: 1

RE: Air combat AI tactics - 10/21/2017 3:07:18 PM

Cik

Posts: 671
Joined: 10/5/2016
Status: offline

aircraft dive and beam when fired upon to maximize their own survival chances (being low increases their agility and decreases missile energy due to the higher drag while the missile flies it's intercept)

so while theoretically you could zoom to evade a missile (though i would never do this because the follow-up shot the enemy would easily get would waste you) that will never happen in C:MANO because all shots (unless deliberately forced to not be) are lethal; that is to say that an aircraft will not fire unless he knows that the missile is going to at least get to it's terminal attack. while in sims you (or they) often fire missiles that will not get near the enemy unless they continue right at you, this will not happen in command: if that R-73 comes off the rail you can expect it to "hit" (get to make a PK% roll) so the AI maximizes it's survival chance against that roll instead of trying to outrun it in the kinetic sense.

so IIRC, the R-73 has a greater range(?) and the distance modifier is less. in addition the R-73 likely has a greater PK to start with. the sea-skimmer modifier is a modifier when attacking targets that are very low to the ground, which the missile was not designed specifically for (unlike some ship missile defenses etc.)

the design is correct assuming you always expect missiles to "hit" you, like in command. while going to the deck is not always wise in EVERY situation, it's usually pretty wise in most- i would never climb when under attack by any kind of missile, because generally speaking it would gain range against me and i would lose maneuverability (due to loss in kinetic energy and thinner air weakening my control surfaces) also, if the missile is between me and the ground, it has a much easier time seeing me (because it is looking at me against cold space and not me against 20,000 trees and a bunch of garbage emitting/reflecting heat and radar energy)

the only modification i'd make to the current behavior is making an assessment based on nearby known triple-A whether or not to go to the deck. the real issue with descending to 80 ft AGL is not "this will help me dodge this missile" which is always pretty much true, it's "this will help me dodge this missile and i will be in range of 20,000 ironsighted AAA guns around hanoi" which is often true and what ends up getting you instead of the SA-2 with 1% PK through all of your jammers and chaff and agility.

< Message edited by Cik -- 10/21/2017 3:10:59 PM >

(in reply to wildcolonialboy)

Post #: 2

RE: Air combat AI tactics - 11/19/2017 10:42:37 AM

Luckschaden

Posts: 39
Joined: 11/9/2017
Status: offline

quote:

the design is correct assuming you always expect missiles to "hit" you, like in command. while going to the deck is not always wise in EVERY situation, it's usually pretty wise in most- i would never climb when under attack by any kind of missile, because generally speaking it would gain range against me and i would lose maneuverability (due to loss in kinetic energy and thinner air weakening my control surfaces) also, if the missile is between me and the ground, it has a much easier time seeing me (because it is looking at me against cold space and not me against 20,000 trees and a bunch of garbage emitting/reflecting heat and radar energy)

Interesting. Are these things modelled in CMANO?
I usually try running away to increase the distance modifier, or going to terrain following min alt and over hills/mountains to try and shake it off.

I also try to have the final approach of the missile be perpendicular to my aircraft,but it's a bit hard to time.
If left to themselves, the AI seems to turn too soon, so the missile turns too and approaches them from 6 o clock.
Would manually flying a question mark shape work? I.e. Flying towards the missile, and then flying a curve?

(in reply to Cik)

Post #: 3

RE: Air combat AI tactics - 11/19/2017 2:21:13 PM

Dimitris

Posts: 13282
Joined: 7/31/2005
Status: offline

quote:

ORIGINAL: Luckschaden
Interesting. Are these things modelled in CMANO?

Yes, every factor he mentions.

quote:

I usually try running away to increase the distance modifier, or going to terrain following min alt and over hills/mountains to try and shake it off.

Both valid.

quote:

I also try to have the final approach of the missile be perpendicular to my aircraft,but it's a bit hard to time.
If left to themselves, the AI seems to turn too soon, so the missile turns too and approaches them from 6 o clock.
Would manually flying a question mark shape work? I.e. Flying towards the missile, and then flying a curve?

It might, but it's tricky to get it right when you're not literally "in the cockpit". You also have to consider the vertical element. The AI actually does a better job than you most of the time because it also understands the altitude separation which can be hard to grasp in a top-down map. (A big reason why beaming can still result in a tail-on impact is that the missile usually comes down on the aircraft from higher altitude, thus literally chasing it down).

_____________________________

Facebook
Twitter

(in reply to Luckschaden)

Post #: 4

RE: Air combat AI tactics - 11/19/2017 3:08:01 PM

T Rav

Posts: 387
Joined: 5/29/2004
Status: offline

I continue to be astounded by the detail in CMANO and the level of support by the devs. The only reason why I haven't bought Shifting Sands yet is because I'm only an hour into the second scenario of Chains of War... Thank you and keep doing what it is that you do.

(in reply to Dimitris)

Post #: 5