New AI task search and engage within distance of itself

[twistking]

A very small item, that would help a lot with mission creation:

A "search and engage" task for CAP (but maybe also other tasking types) that is valid for a moving zone of definable size on the the group leader (alternatively for multiple zones of definable size with one for each group member). In practice this would mean that CAP would only search and engage targets near to the CAP flights position.

A similar behaviour could be archieved by adding a new ROE option ("defend aggresively") where the group is only "weapons free" for targets within a certain range. The result would be the same, but i think it's more elegant to have this as a continuous Task type.

In practice you would often want to combine this new task with f.e. a "search and engage in zone", but you could make the zone much smaller, since the flight would still be somewhat aggresive while cruising to it's CAP area.

There is the obvious problem that a CAP zone moving with the group would mean that the CAP flight will never disengage unless the bandit group is fast enough to build distance quickly. This is not terrible, but ideally there was another parameter to control disengage criteria: This would need some creative thinking: Maybe the engagement radius for "cold" bandits could be set manually. If set very small, the CAP flight would be quicker in disengaging from bandits turning cold and not pursue them endlessly as would be otherwise. Again: The goal would be to have the AI in a defensive, but still somewhat aggresive stance (without going full berserk, as they would do with ROE "open fire").

Thanks.

[Exorcet]

This is basically a wish for smarter self defense? If so I agree, but it shouldn't be a CAP task. It needs to be universal.

I think the idea could be simplified by just making it an input, ie "self defense distance = x nautical miles" and then add the type checkboxs "aircraft, helicopter, SAM, tank, ship, etc". Put whatever x you want, and the AI engages anything hostile from the checkboxes within that distance.

To stop the AI from chasing aircraft forever there could be a "maximum distance from waypoint" setting. Similar to self defense distance, but this would determine when the AI stops chasing.

Alternatively, we could fix the problem that Search Then Engage currently has, which I brought up a couple of years ago:

This would allow more control and give us exactly the option you're asking for, a way to set self defense for aircraft. We'd also need ALL planes to be able to access Search Then Engage:

 

[twistking]

2 hours ago, Exorcet said:

This is basically a wish for smarter self defense? If so I agree, but it shouldn't be a CAP task. It needs to be universal.

I think the idea could be simplified by just making it an input, ie "self defense distance = x nautical miles" and then add the type checkboxs "aircraft, helicopter, SAM, tank, ship, etc". Put whatever x you want, and the AI engages anything hostile from the checkboxes within that distance.

To stop the AI from chasing aircraft forever there could be a "maximum distance from waypoint" setting. Similar to self defense distance, but this would determine when the AI stops chasing.

Alternatively, we could fix the problem that Search Then Engage currently has, which I brought up a couple of years ago: [...]

Yes, i guess you could call it smarter self defense; If you put higher values for engage distance, it would function more as a type of CAP though, that's why my suggestion called for a CAP task, but of course it would functionally equally well as a new ROE option. I think your suggestion of an modified search than engage task (that lets you define which part of the flightpath counts for the task) would solve the same general problem... only coming at it from a slightly different direction...

I don't think that "maximum distance from waypoint" would be a good trigger to stop pursuit though. First of all it would need to be "max distance from flightpath", but that might break if flight is not actually travelling along its path, because of other scripts or task (switch waypoint, racetrack orbit, falling back towards path after any location based push task). I think it should be something relative to leader's current position. I think a smaller distance setting for cold bandits with some seconds of timeout would solve that actually... don't you think?

[Exorcet]

On 5/31/2023 at 11:19 AM, twistking said:

I don't think that "maximum distance from waypoint" would be a good trigger to stop pursuit though. First of all it would need to be "max distance from flightpath", but that might break if flight is not actually travelling along its path, because of other scripts or task (switch waypoint, racetrack orbit, falling back towards path after any location based push task). I think it should be something relative to leader's current position. I think a smaller distance setting for cold bandits with some seconds of timeout would solve that actually... don't you think?

The idea is, aircraft generally patrol a specific location, so limiting their hostile behavior to a certain region should make sense most of the time. Distance from enemies also needs to be taken into account, but if that's the only consideration then fighters may end up in endless afterburner chases and run out of fuel. To really, really solve the problem the AI probably needs a smart disengage script. For example a flight of 2 chasing a bandit will disengage one at a time, so that someone is always looking at the bandit, and they will slow down preemptively to create space, rather than keeping full burner all the way to the point where they do decide to turn around. I was just trying to provide a simple method, but there are multiple options here.

[chromium]

Il 31/5/2023 at 13:34, twistking ha scritto:

A "search and engage" task for CAP (but maybe also other tasking types) that is valid for a moving zone of definable size on the the group leader (alternatively for multiple zones of definable size with one for each group member). In practice this would mean that CAP would only search and engage targets near to the CAP flights position

Ok, but I get that DCA behaviour by removing the "CAP" task (which overrides anything) and using the search and engage and then inserting the range constraints. Check attached image & miz file example. You will clearly see that the F-14s won't engage the Mig-21 till it's closer than 30 nm to them. By moving the initial Mig-21 position or adding awacs or else you can also ensure that it's not due to the "discover" distance of the F-14, but due to the range constraints.

 

TestDCA.miz

Edited June 3, 2023 by chromium

[twistking]

3 hours ago, chromium said:

Ok, but I get that DCA behaviour by removing the "CAP" task (which overrides anything) and using the search and engage and then inserting the range constraints. Check attached image & miz file example. You will clearly see that the F-14s won't engage the Mig-21 till it's closer than 30 nm to them. By moving the initial Mig-21 position or adding awacs or else you can also ensure that it's not due to the "discover" distance of the F-14, but due to the range constraints.

I don't think that is a good enough solution. The search and engage task is valid for the whole flightpath and not only for those legs of the path that are part of the CAP area. If AWACS/EWR is available, CAPs will often leave their station to hunt targets very far away, that just happen to cross the transit flightpath of the CAP. Therefore i think the "search and engage" task is absolutely useless for bigger, somewhat realistic scenarios. For these only "search and engage in area" is viable... but with it, you run into the issue of the flight not being aggressive enough on transit.

[chromium]

8 ore fa, twistking ha scritto:

The search and engage task is valid for the whole flightpath and not only for those legs of the path that are part of the CAP area.

False. It's valid only after you add it in the adv.wpt option, and you can use as stop condition a last waypoint.

About the rest, I don't agree but these are only different point of view, not technical issues.

[twistking]

5 hours ago, chromium said:

False. It's valid only after you add it in the adv.wpt option, and you can use as stop condition a last waypoint.

About the rest, I don't agree but these are only different point of view, not technical issues.

ohh... ok. i somehow have missed that. thanks for the heads-up.

just a quick question: how would i set up a stop condition at a certain waypoint? i'm generally familiar with the ME and tasking/triggers, so if you could only point me in the general direction, it would be helpful already (i'm only aware of the stop condition in the task specific option, but that only allows for mission time, duration, or flag if i remember correctly)...

Edited June 4, 2023 by twistking

[chromium]

Here it is:

PS: There wasn't that option in the past, don't know exactly when was added... probably 2.5.something.

In the attached miz there's a DCA for 30 minutes between wpt 2 & 3 with a 30 nm engagement range (from the airplanes).

 

TestDCA2.miz

[twistking]

2 hours ago, chromium said:

Here it is:

 

PS: There wasn't that option in the past, don't know exactly when was added... probably 2.5.something.

In the attached miz there's a DCA for 30 minutes between wpt 2 & 3 with a 30 nm engagement range (from the airplanes).

Thanks a lot! Yes, they must have added it after i got frustrated with the task and started using "engage in area" exclusively

Still think that there should be smarter ROE options. ROE "only shoot back" (or whatever it is called) is just not good enough and will get AI killed because they react to late...

[Exorcet]

13 hours ago, chromium said:

False. It's valid only after you add it in the adv.wpt option, and you can use as stop condition a last waypoint.

This is incorrect and is the basis of my thread.

 

EDIT -

STE still causes fighters to engage well beyond the stop condition in 2023:

STE_retest.miz

Edited June 4, 2023 by Exorcet

[chromium]

41 minuti fa, Exorcet ha scritto:

This is incorrect and is the basis of my thread.

No wait, what I'm saying (as you also say in your thread) is that the "defence" behaviour can be activated enabling the S&E in a specific waypoint, an then deactivating it in another one o with any desired stop condition (including last wpt). But, once activated and till it's working, it will command the flight to act as you described using the path as offset area. But I consider that more as a feature than a drawback.

Edited June 4, 2023 by chromium

[Exorcet]

3 minutes ago, chromium said:

No wait, what I'm saying (as you also say in your thread) is that the "defence" behaviour can be activated enabling the S&E in a specific waypoint, an then deactivating it in another one o with any desired stop condition. But, once activated, it will work as you described using the path as offset area.

OK I see.

3 minutes ago, chromium said:

 

But I consider that more as a feature than a drawback.

My interpretation was that @twistking found this to be a drawback, as I do, since it can make the AI engage in unexpected ways or at unexpected times. Using Search then Engage in Zone is an alternative, but it's limited to circles, which creates it own problems.

[chromium]

Adesso, Exorcet ha scritto:

OK I see.

My interpretation was that @twistking found this to be a drawback, as I do, since it can make the AI engage in unexpected ways or at unexpected times. Using Search then Engage in Zone is an alternative, but it's limited to circles, which creates it own problems.

But for this level of detail, you might probably want to go to scripting to get the best solution. But that can be a pain when starting ad kind of addictive when using it XD.

[Exorcet]

1 minute ago, chromium said:

But for this level of detail, you might probably want to go to scripting to get the best solution. But that can be a pain when starting ad kind of addictive when using it XD.

Yes, this is why I feel like changing the "stop behavior" or adding a separate limit for the STE boundary by waypoint is preferable.

[twistking]

yes @Exorcet is correct. if the AI will still consider the whole flightpath, this doesn't solve my problem.

right now, i believe it would be best to just add a new defensive ROE option, that would allow the AI to defend itself when threatened - not only when shot at. with this one could set up a CAP zone with "search and engage in area" and would not have to worry about transit...