Airplane Defense Logic

[DWReese]

Apache85,

I know that your are correct about the tactic of the Vietnam era of turning toward the SAM, but could you at least revisit the concept for our game?

Turning toward the SAM is great, but continuing to turn in with each successive missile attack puts the plane in greater danger. I'm literally watching a plane increase its odds of destruction with each successive turn.

Furthermore, when one SAM is launched in the general direction of a bunch of planes, then all of the planes begin turning toward the SAM launcher, likely because they don't know which plane has been targeted. The end result, not one, but all keep getting closer to the SAM.

I know that you are busy, but when things calm down, a "turn and burn" strategy might work just as well, if not better, in some situations.

Thanks in advance.

[Rory Noonan]

ORIGINAL: apache85
If someone can provide a good IFTT summary that covers bases more effectively than the current AI implementation I will be extremely excited to put it forward for implementation.

To clarify this a little, IFTT stands for 'If this, then that'. The kind of answer that i'm asking for is along the lines of:

If X:
....Do A
If Y:
....Do B
If Z:
....If W is not a factor:
.........Do C
....Otherwise:
.........Do E

I'm not trying to be obtuse or obstructionist; but this is a major change to AI behaviour and without some solid direction on what people would like to see changed it's difficult to consider the implications in broader gameplay, let alone come up with a solution that adequately covers the scope of Command (1945-2025+--with constant requests to include WW2!) without the AI for missile evasion rapidly growing to unmanageable size and complexity.

[DWReese]

Apache85,

I do understand the programming concept, and I agree changing things like this can be daunting. I have no idea what could/should be done.

The current model works well, but this one minor issue is troubling. For example, I had three planes that had released their weapons and were all headed back to their base when they began to be fired upon by some SAMs. The planes were about 8 miles away from the SAMs when the missiles were fired. Despite the fact that the missiles were fired in the general direction of the planes, they all went into their "turn toward the SAM protocol." By the end, the surviving planes were now within a mile of the SAM launcher.

This "problem" isn't the most serious in the world, but at some point, when things calm down, it would be nice if your team could brainstorm this topic a little bit and apply a tweak to it, if possible.

Thanks again, and keep up the good work.

[Eggstor]

This piece from Carlo Kopp originally printed in Australian Aviation in 1987 suggests that putting the missile on the aircraft's beam (3 o'clock or 9 o'clock) is an essential start, along with popping chaff/flares and turning on DECM to try to break lock early (something that, if memory serves, is not part of the CMO anti-SAM/anti-AAM model outside of the terminal dice roll).

What happens once visual contact with the missile is made is dependent on what the missile's guidance and range are (or at least the pilot's best guess based on his briefing), and when visual contact is made. How much of that post-visual contact theory is applicable to CMO depends on to what extent CMO models the limitations of the tracking platform's (both the site radar and the missile in the case of a SARH missile) ability to correct for a maneuvering target, and the subsequent recognition by the pilot that the missile has lost lock. An example of that tracking limitation from the AAM realm is the early Sidewinder's inability to keep a lock against highly-maneuverable targets.

A "proper" implementation will involve a lot of reprogramming (and a few cases of opened worms), but an interim general "fix" would seem to be have the aircraft turn (more or less) perpendicular to the launcher in the general direction of where it needs to go and dive.

[Dimitris]

ORIGINAL: Eggstor

This piece from Carlo Kopp originally printed in Australian Aviation in 1987 suggests that putting the missile on the aircraft's beam (3 o'clock or 9 o'clock) is an essential start, along with popping chaff/flares and turning on DECM to try to break lock early (something that, if memory serves, is not part of the CMO anti-SAM/anti-AAM model outside of the terminal dice roll).

What happens once visual contact with the missile is made is dependent on what the missile's guidance and range are (or at least the pilot's best guess based on his briefing), and when visual contact is made. How much of that post-visual contact theory is applicable to CMO depends on to what extent CMO models the limitations of the tracking platform's (both the site radar and the missile in the case of a SARH missile) ability to correct for a maneuvering target, and the subsequent recognition by the pilot that the missile has lost lock. An example of that tracking limitation from the AAM realm is the early Sidewinder's inability to keep a lock against highly-maneuverable targets.

A "proper" implementation will involve a lot of reprogramming (and a few cases of opened worms), but an interim general "fix" would seem to be have the aircraft turn (more or less) perpendicular to the launcher in the general direction of where it needs to go and dive.

Actually, beaming is indeed taken into account, and in two slightly different varieties.

If the missile is guided by a normal pulse radar (or other form of guidance), the pilot tries to beam the missile itself. You can see the effect of this on the weapon engagement calculations, where the impact angle plays an important role in determining the actual "evasion agility": The hardest shot for the missile is a beam-on impact; dead-ahead and rear-aspect intercepts are significantly easier.

If the missile is guided by doppler radar, the pilot tries to beam the guiding sensor instead (either the onboard seeker if the missile is ARH, or the illuminator if SARH), hoping to disappear in the doppler notch. This may or may not work depending on a number of factors.

[DWReese]

By watching the impressive evasive calculations, I am amazed at how well it actually works. You've done a great job with it. Without the evasive tactics, the kill percentage is almost double in most situations.

I do believe that it would be nice, however, if the plane didn't constantly keep getting "closer" to the SAM. At some point, the reduction that the plane gets for its evasive actions is offset by the increased percentage gained as the distance is reduced.

Perhaps some kind of code could be implemented that either assesses the initial threat percentage (the pilot may, or may not really know this), or if the distance is greater than "X percentage", then the plane foregoes the plan of evading and elects to dive and run at full speed. I'm sure that there is absolutely no perfect way to handle this so it will be difficult to do.

You are busy, and this is a low priority issue, so I don't expect anyone to drop everything to look into it. But, if you could make note of it, perhaps you can look into it sometime in the future and that would be great.

Thanks, and keep up the great job.

[Eggstor]

Thanks for the peek inside the engine. Now what we need is a scenario where the behavior in the original post actually happens. I honestly don't recall seeing that particular behavior.

Question - Outside disappearing in the doppler notch, is it possible to break lock after weapon launch and before the weapon engagement calculation?

[serjames]

? it literally happens EACH and every time. Are we playing the same game ? :=)

[Dimitris]

ORIGINAL: Eggstor
Question - Outside disappearing in the doppler notch, is it possible to break lock after weapon launch and before the weapon engagement calculation?

Yes. There are a number of ways:

* If the weapon is remotely guided (incl. SARH/TVM), kill the guiding unit [:)]
* SARH & TVM guidance, as well as TARH seekers, can be disrupted by noise jamming (thought with HOJ this can be a gamble). Noise jamming can also work against the mid-course guidance provider in a datalink+TARH scenario (e.g. you blind the other guy enough that he guides his AMRAAMs to the wrong/obsolete activation point)
* Break line of sight
* Get out of the guiding unit's illumination cone, if SARH. The typical example is shooting a missile at the other guy to force him to beam and turn away. (This can also work fairly well against AMRAAM shooters IF nobody else is painting you. If you can force his radar away from you long enough, and meanwhile change your flight vector sufficiently, his AMRAAMs will fly to your old position estimate)
* [REDACTED - PRO FEATURE ONLY]

[Dimitris]

ORIGINAL: serjames
? it literally happens EACH and every time. Are we playing the same game ? :=)

It does happen a lot, but I suspect this is the natural and logical outcome of the tactical decisions of the sim agents, including fundamental self-preservation choices, rather than a "stupid" decision. At least in air combat.

Let us consider a simple example. A MiG-29 engages an F-16 (w/ AIM-9s only) with an AA-10, and the Falcon has already began turning to beam:



At that point, as the Falcon driver, you have two basic choices:

1) Continue a sharp turn to the right and try to place the missile on your 6 and outrun the shot. Iffy proposition, since you can never be quite certain as to what exactly is coming your way and whether you can outrun it. If you don't, it's coming right up your tailpipe (astern impact --> minimal agility evasion) so there is a very good chance you're going to die. Even if you survive, you have surrendered the initiative to your opponent by allowing him to take another BVR shot at you.

2) Continue beaming the missile (or in this case the MiG's radar as it's a doppler set) and hope that you'll either get lost in the notch or, failing that, the missile will come at you at a high deflection angle and thus have a harder time hitting you. Unavoidably this will mean reversing your turn and gradually shifting your heading counter-clockwise (watch how the missile is coming to you), ie. _towards_ the MiG that is shooting at you. This is the default behavior in Command.

This grants you two significant pluses:
a) As described, better chance of evading the shot.
b) If you do survive, you are now much closer to your assailant and have sharply reduced (or even entirely eliminated) his standoff advantage.

While typing this, I just realized that a middle-of-road compromise might be workable: Keep the above-described behavior while evading shots from aircraft (OR if the evading aircraft is SEAD-tasked, in which case it does want to get closer to the SAM), but opt for the first approach if evading SAMs (in which case we asume that the aircraft does prefer to stay away from them rather than duel them).

Would this work?

[LetMePickThat]

Maybe the type of evasive could be determined by the range to the launcher (I'm thinking about SAMs here, not aircrafts). If the SAM bty launching the missile have been detected by the evading plane (either via RWR, direct observation, sensor or even remote sensing in the case of planes equipped with the needed datalink), then the plane could make a guess about the likely range of the missile involved and choose the best course of action (flee or beam) accordingly.

In the case the launcher is unknown, then the evasive maneuver could be determined by the distance between the missile and the plane. Maybe the range value at which the change of maneuver is made could be user adjustable, as part of the doctrine.

[ChickenSim]

Summarizing my thoughts on discord, this probably has the potential to open up a number of cans of worms.

The two largest cans in my opinion are those of pilot skill and intent to go hot/cold.

In terms of pilot skill, this definitely brings with it more than just an in-game OODA delay. In CMO/CMANO terms, the higher-skilled pilot brings with them implicitly higher quality intelligence, a better plan, a better brief, and better decision-making and teamwork, with less uncertainty in action. IFTT decision-making trees would naturally need to incorporate the chance that the wrong decision, or none at all (paralysis) is arrived at.

Intent to go hot/cold is similar to the BVR tactics panel, where you can make a choice on the spectrum of cranking and dragging versus following the missile straight into the merge. In determining what an aircraft/pilot decides to do in the event of a SAM launch, there is a similar spectrum based on the range/time-space-distance calculations for deciding whether to just change course slightly to stay out of the WEZ, hit the deck and run, weave away from the threat 135° until you're out of range (attempting to notch with each roll reversal), beam the threat and wait until the missile is defeated to roll back in or get out of the WEZ, weave toward the threat ~45° to attempt to engage at closer range or get past the threat at the real target, (if the TTI of the SAM is greater than the TTI of the pilot's own ARM) to fire back, or have the package roll in hot and hope at least one pilot survives the ingress to take out the threat. Being able to pick these options in Doctrine might be the easiest solution?

And if there's aggressive intent, one forum post probably isn't enough to adequately describe every combination of a section of aircraft hitting the deck to break LOS and executing off-axis pop attacks with mutual support to try and put the SAM into a no-win situation, only able to shoot at one target or the other.

[Eggstor]

ChickenSim put to words my thoughts better than I could have.

[DWReese]

First, let me say thank you for being willing to discuss this like you are doing.

I do like your idea. I think that you are on the right track.

To me, its seems that you could use some sort of math equation (based on probability) to determine the best course of action. For example, if a SAM has an 80 percent kill rate, and the distance is 10 miles, and you can eliminate 15 percent by taking evasive action by turning back toward the SAM, or reduce the kill percentage by 5 percent by attempting to run, then the evasive action would win by a score of 65 to 75. But, eventually, you will get to a point where one will out weigh the other. Also, you should probably include the thinking that IF you survive the FIRST SAM shot, any follow up shot will be from a further distance, thereby increasing the pilot's chance of survival. By constantly moving AWAY from the SAM, your chance of survival goes up. Each time you move closer to the SAM, it goes down.

Thanks again for at least discussing this.

[Rebel Yell]

ORIGINAL: Dimitris



While typing this, I just realized that a middle-of-road compromise might be workable: Keep the above-described behavior while evading shots from aircraft (OR if the evading aircraft is SEAD-tasked, in which case it does want to get closer to the SAM), but opt for the first approach if evading SAMs (in which case we asume that the aircraft does prefer to stay away from them rather than duel them).

Would this work?

I think that would be infinitely preferable.

[LargeDiameterBomb]

I have noticed this "problem" as well with AI behavior. That said I am not presuming to speak for DWReese but his posts have only mentioned SAMs and as he has described the issue in this thread I broadly agree with him.

Dimitri, your suggestion would be a good thing for me if the devs don't want too spend much time on this (Which I wholeheartedly understand as the amount of work in the last months must have been grueling).


To answer Apache85 otherwise a slightly more optimized solution would be

IF missile is an AAM THEN perform evasion maneuver
IF missile is a SAM AND of unknown type AND within 5 nm of a/c at time of detection THEN perform evasive maneuver,
IF aircraft is not destroyed THEN go to X, IF aircraft is destroyed END
X
IF no illuminating radar detected THEN dive and run on approx same course as missile had
when detected until no more missiles pose a threat, continue run for 5 sec after last
missile posed a threat
IF an illuminating radar is detected THEN dive and run from detected illuminating FCR until
no more missiles pose a threat, continue run for 5 sec after last missile posed a threat
IF missile is a SAM AND of unknown type AND more than 5 nm from aircraft at time of detection THEN dive and run on course that allows observation of missile and take evasive maneuver when missile reaches 2,5 nm from aircraft, REPEAT until no more missiles from same launcher is a threat, continue run for 5 sec after last missile posed a threat
IF missile is a SAM AND a SAM launcher has previously been identified (Not detected, only identified. Sic!) in a 60 degree cone extending backwards from the detected missile in a thought-up axis extending from aircraft to detected missile THEN perform rough calculation (With some chance of error, perhaps amount of error depending on pilot skill) assuming identified SAM launcher as threat to decide if chance of outrunning (presumed) missile (type) exists, IF TRUE go to Y, IF FALSE dive and run
Y
IF endgame chance of hit roughly anticipated to more than 5 % dive and run anyway
IF endgame chance of hit roughly anticipated to less than or equal to 5 % perform evasive
maneuver
IF unknown illuminating fire control radar detected THEN aicraft(s) being illuminated dives and runs from illuminating FCR on optimal course
IF illuminting FCR detected and identified AND illuminating radar associated with specific missile type/types THEN use actual missile range to perform rough calculation (With some chance of error) to decide if chance of outrunning presumed missile type exists, IF TRUE then dive and run, IF FALSE go to X

(EDIT: I apologize for not being able to structure the IF-THENs as they should be laid out but the forum software doesn't seem to recognize blank spaces at the start of a new row).

All numeric values are there as suggestions much more than definitive proposals. I only thought this up as I wrote it so I might have missed something and I am in no way sure that my solution is optimal or how easy it would be to implement, but
1) It would allow closing in on old, outdated SAM Systems with a almost zero chance of a hit with a modern aircraft with a (Compared to the SAM system) very modern self protection EW system (DECM in CMO parlance).
2) while always running away if it seems like a possibility to reach outside the missile's envelope in time (Which gives a zero probability of being hit) and
3) at the same time always running away from advanced SAMs
4) Also makes aircraft run away from suddenly appearing illuminating fire control radars mounted on prevously undiscovered SAM systems even when no missile is detected yet which might be a good thing if one is flying above the height of cloud cover and thus cannot visually acquire a missile while
5) Maintaining realism at a reasonable level.

The last two IF-THENs are kind of optional though (Running away from illuminators) an probably would require an option in doctrine to disable it as it would be easy to make an opponents aircraft run away just by turning on an illuminator, which in the game would be detected by all aircraft in the vicinity of said illuminator and just not the one or two aircraft in realty being painted (Ie being inside the highly directional illuminating radar's beam) as CMO doesn't model the highly directional characteristics of illuminators as regards detection by RWR/ESM equipment.
I am also aware that many SAM systems with the same illuminator use different missiles with different ranges (Eg Buk and Buk-M1) but I think there is a case for using the real value of the missile range since what equipment the enemy has access to would most likely be known in almost all real life cases. It does become a problem with regards to realism if a side in the game uses for instance both Buk and Buk-M1 systems though, but such occasions are fairly uncommon.

Also, running at high altitude if the calculations indicate the aircraft can escape the SAM that way might be an idea to use against missiles with early 2000s or later active seekers since background clutter probably won't give much of a protection for an aircraft being targeted by such a missile (In other words, almost all missiles with the characteristic "Capable versus Sea-skimmer" in-game) anyway.

As a side note, it's somewhat unfortunate that DECM systems does not function in a more realistic way in CMO, ie the same way as OECM systems so the system could break the lock-on from the tracking radar and then the missile would quickly veer of course, which the pilot would notice and which would not lead to any evasive maneuvering. For instance an AN/ALQ-211 self protection ECM system coming up against an non-upgraded SA-2f would probably be able to break lock almost all of the time, except at very short ranges.
It would be very handy if a DECM system would be able to jam a single or perhaps a few emitters, the latter on more modern aircraft or aircraft with two pilots, with much lower jamming power than ECM systems but enough to break lock-on for incoming missiles. That would make this problem much easier to solve.

I of course understand that the developers have to make a lot of compromises to get a functional and smooth running game with today's hardware and an excellent job they have done!

[BrianinMinnie]

"While typing this, I just realized that a middle-of-road compromise might be workable: Keep the above-described behavior while evading shots from aircraft (OR if the evading aircraft is SEAD-tasked, in which case it does want to get closer to the SAM), but opt for the first approach if evading SAMs (in which case we assume that the aircraft does prefer to stay away from them rather than duel them)."

Yes!!!

in the immortal words of Ben Franklin........... "Make it so"

[BrianinMinnie]

Dude!!

I have noticed this "problem" as well with AI behavior. That said I am not presuming to speak for DWReese but his posts have only mentioned SAMs and as he has described the issue in this thread I broadly agree with him.

Dimitri, your suggestion would be a good thing for me if the devs don't want too spend much time on this (Which I wholeheartedly understand as the amount of work in the last months must have been grueling).


To answer Apache85 otherwise a slightly more optimized solution would be

IF missile is an AAM THEN perform evasion maneuver
IF missile is a SAM AND of unknown type AND within 5 nm of a/c at time of detection THEN perform evasive maneuver,
IF aircraft is not destroyed THEN go to X, IF aircraft is destroyed END
X
IF no illuminating radar detected THEN dive and run on approx same course as missile had
when detected until no more missiles pose a threat, continue run for 5 sec after last
missile posed a threat
IF an illuminating radar is detected THEN dive and run from detected illuminating FCR until
no more missiles pose a threat, continue run for 5 sec after last missile posed a threat
IF missile is a SAM AND of unknown type AND more than 5 nm from aircraft at time of detection THEN dive and run on course that allows observation of missile and take evasive maneuver when missile reaches 2,5 nm from aircraft, REPEAT until no more missiles from same launcher is a threat, continue run for 5 sec after last missile posed a threat
IF missile is a SAM AND a SAM launcher has previously been identified (Not detected, only identified. Sic!) in a 60 degree cone extending backwards from the detected missile in a thought-up axis extending from aircraft to detected missile THEN perform rough calculation (With some chance of error, perhaps amount of error depending on pilot skill) assuming identified SAM launcher as threat to decide if chance of outrunning (presumed) missile (type) exists, IF TRUE go to Y, IF FALSE dive and run
Y
IF endgame chance of hit roughly anticipated to more than 5 % dive and run anyway
IF endgame chance of hit roughly anticipated to less than or equal to 5 % perform evasive
maneuver
IF unknown illuminating fire control radar detected THEN aicraft(s) being illuminated dives and runs from illuminating FCR on optimal course
IF illuminting FCR detected and identified AND illuminating radar associated with specific missile type/types THEN use actual missile range to perform rough calculation (With some chance of error) to decide if chance of outrunning presumed missile type exists, IF TRUE then dive and run, IF FALSE go to X

(EDIT: I apologize for not being able to structure the IF-THENs as they should be laid out but the forum software doesn't seem to recognize blank spaces at the start of a new row).

All numeric values are there as suggestions much more than definitive proposals. I only thought this up as I wrote it so I might have missed something and I am in no way sure that my solution is optimal or how easy it would be to implement, but
1) It would allow closing in on old, outdated SAM Systems with a almost zero chance of a hit with a modern aircraft with a (Compared to the SAM system) very modern self protection EW system (DECM in CMO parlance).
2) while always running away if it seems like a possibility to reach outside the missile's envelope in time (Which gives a zero probability of being hit) and
3) at the same time always running away from advanced SAMs
4) Also makes aircraft run away from suddenly appearing illuminating fire control radars mounted on prevously undiscovered SAM systems even when no missile is detected yet which might be a good thing if one is flying above the height of cloud cover and thus cannot visually acquire a missile while
5) Maintaining realism at a reasonable level.

The last two IF-THENs are kind of optional though (Running away from illuminators) an probably would require an option in doctrine to disable it as it would be easy to make an opponents aircraft run away just by turning on an illuminator, which in the game would be detected by all aircraft in the vicinity of said illuminator and just not the one or two aircraft in realty being painted (Ie being inside the highly directional illuminating radar's beam) as CMO doesn't model the highly directional characteristics of illuminators as regards detection by RWR/ESM equipment.
I am also aware that many SAM systems with the same illuminator use different missiles with different ranges (Eg Buk and Buk-M1) but I think there is a case for using the real value of the missile range since what equipment the enemy has access to would most likely be known in almost all real life cases. It does become a problem with regards to realism if a side in the game uses for instance both Buk and Buk-M1 systems though, but such occasions are fairly uncommon.

Also, running at high altitude if the calculations indicate the aircraft can escape the SAM that way might be an idea to use against missiles with early 2000s or later active seekers since background clutter probably won't give much of a protection for an aircraft being targeted by such a missile (In other words, almost all missiles with the characteristic "Capable versus Sea-skimmer" in-game) anyway.

As a side note, it's somewhat unfortunate that DECM systems does not function in a more realistic way in CMO, ie the same way as OECM systems so the system could break the lock-on from the tracking radar and then the missile would quickly veer of course, which the pilot would notice and which would not lead to any evasive maneuvering. For instance an AN/ALQ-211 self protection ECM system coming up against an non-upgraded SA-2f would probably be able to break lock almost all of the time, except at very short ranges.
It would be very handy if a DECM system would be able to jam a single or perhaps a few emitters, the latter on more modern aircraft or aircraft with two pilots, with much lower jamming power than ECM systems but enough to break lock-on for incoming missiles. That would make this problem much easier to solve."

It's thinking like that, that makes this game what it is and what it can be, you worked that out nicely. I have no idea if it's doable but it sounds great, well done.

[JPFisher55]

i don't know if it is related to this topic, but I have always wanted aircraft to RTB at full afterburner rather than full military speed. By doing so, they would sooner get out of danger. Also, when surrounded by enemy aircraft, I find it useful to disable auto evasion and RTB an aircraft at full afterburner. This gives them a better survival chance than trying to evade all the missiles and giving the enemy aircraft much more chances to down my aircraft

[c3k]

At combat ranges, full afterburner time is measured (in some cases) in just single-digit minutes.

I like the ideas for IF/THEN with various missiles. Getting away from the launcher is THE highest priority...unless you're just about to be hit by a missile.