Let's talk about IFF

[GGTharos]

That is quite pessimistic :) Would you care to elaborate what you think is particularly complex to implement?

 

Sure.

 

1. Unknowns are prosecuted like Hostiles but with weapon status per ROE. If weapon hold, then AI maneuvers to engage target until contact is identified hostile (then open fire) or friendly (ignore it). This is how Command Modern Air/Naval Operations does it.

 

Inadequate. Maneuvers how exactly?

 

4c. Observation of hostile act against friendlies.

 

Not implemented, not easy to implement at this time.

 

4d. Observation of take-off from hostile base.

 

As above, may as well cheat by using the coalition here.

 

4e. Non Cooperative Target Recognition

 

Not available by any means on some aircraft, and not implemented otherwise. Here may as well cheat - but then again, NCTR can be as ambiguous as IFF, for the same reasons.

 

4f. Any combination of heading, altitude, speed (bigger, lower, equal), zone (inside, outside) or time, as defined as condition in the ROE tab in ME.

 

Capability does not exist, and is also inadequate by itself.

 

4g. Identification by external party to which a radio-/data-link is established (which itself must declare the contact hostile by 4a.-4g.).

 

May as well use the coalition here then.

 

 

One of the main problems is that there is absolutely no notion of collected, remembered knowledge. There is no notion of target tracks, information attached to them or a mechanism for forgetting them. As a result, there is also no AI management or overall intelligence to manage and interpret this information.

AI has no tactical knowledge, plans or capabilities, so it's ability to perform a successful VID is rather limited.

 

You can have all this, but isn't by any means trivial.

[MBot]

1. Inadequate. Maneuvers how exactly?

 

As if the AI would maneuver to shoot someone down, but without shooting. Approach the target and get into firing envelope. In my opinion completely adequate for 95% of the situations. Only thing that might be missing would be some spotter-shooter tactics, but its not as if the AI currently does any group tactics either. What essential maneuver would you miss?

 

4c. Not implemented, not easy to implement at this time.

 

Shooting is a logged event. Has in sight is script function. Correlate shooting, has shooter in sight, has target in sight. I could script it in LUA right now.

 

4d. As above, may as well cheat by using the coalition here.

 

Take-off is a logged event. Has in sight is a script function. Base coalition is global variable. Correlate the 3. Could be scripted in LUA right now.

 

4e. Not available by any means on some aircraft, and not implemented otherwise. Here may as well cheat - but then again, NCTR can be as ambiguous as IFF, for the same reasons.

 

Then either simulate NCTR for AI equally as for the player (which would be the right thing to do) or scratch it altogether.

 

4f. Capability does not exist, and is also inadequate by itself.

 

Everything of this can be easily scripted in LUA right now. A combination of the above can be used to define save passage corridors, save time slots, engaging just westbound Unknowns, engaging just Mach 3+ Unknowns, engage just 20'000m+ Unknowns etc. What else could you possibly want?

 

 

4g. May as well use the coalition here then.

 

No, because some will have to do the initial declaration. F-14 CAP over the Norwegian Sea, E-2C detects bogey approaching CVBG. Could be Soviet Bear-F or Norwegian P-3C. CAP heads out to find out. Once CAP IDs bogey, information cascades through AWACS to everyone on the same frequency.

 

One of the main problems is that there is absolutely no notion of collected, remembered knowledge. There is no notion of target tracks, information attached to them or a mechanism for forgetting them.

 

Nothing particularly difficult to do. I have written LUA scripts with making target tracks, collecting information, forgetting them after an idle time etc.

 

As a result, there is also no AI management or overall intelligence to manage and interpret this information.

AI has no tactical knowledge, plans or capabilities, so it's ability to perform a successful VID is rather limited.

 

How is that any different from what we have right now? AI currently has no tactical knowledge or plans. It just stumbles around, engages and fires on targets based on some criteria defined in the ME. What would be added is that AI would would still stumble around and engage targets based on some criteria defined in the ME, except that it would open fire only once the criteria above have been satisfied, should the ROE be set up to mandate this. And more importantly, air defenses would also be affected equally!

Edited March 9, 2016 by MBot

[GGTharos]

How is that any different from what we have right now? AI currently has no tactical knowledge or plans. It just stumbles around, engages and fires on targets based on some criteria defined in the ME. What would be added is that AI would would still stumble around and engage targets based on some criteria defined in the ME, except that it would open fire only once the criteria above have been satisfied, should the ROE be set up to mandate this. And more importantly, air defenses would also be affected equally!

 

If you're satisfied with the AI just stumbling around, ok.

Edited March 9, 2016 by GGTharos

[MBot]

If you're satisfied with the AI just stumbling around, ok.

 

Oh come on, what has this to do with IFF? Making AI do smart things is a completely separate topic (which I would fully support and would have a lot to add). The point in this topic is that current AI would do equal smart things with or without targets categorized as Unknown.

[GGTharos]

This just turns them into fish-in-a-barrel IMHO. Anyway, it's neat, in a step-at-time way.

[Hadwell]

Proper IFF assumes you can use multiple methods of interrogating unknown contacts, like ground radar + ship radar + awacs + a few fighters using data link, over a period of time, so you can see what the unknown contact is doing and where he's coming from, to confirm a plane is hostile.

 

and regardless if there's a couple squads out there that run their own private servers (thats still a minority) and a bunch of people who only play single player (i have no idea if this is a minority or not), a very large chunk of people play in a situation where they can only rely on their one single plane to id targets with no information coming from anything but their eyes or built in radar...

[Eddie]

Procedure is what you make of it. IMHO you're not going to see better procedures or communication on public servers, you'll only see it from dedicated squadrons that will do this regardless.

 

While true, those of us wishing to make use of realistic procedures find it very difficult to do so due to some of the limitations in DCS, this being one of them. Yes you can role play it, but as MBot has already mentioned, role play for the sake of it isn't very appealing when you know it really makes no difference to the outcome.

 

Not to mention the training burden caused by having to spend hours correcting bad habits caused by the current status quo.

 

Although I will say I do believe that if the sim did more to encourage and reward the use of realistic procedure (and the documentation and training taught it) I think even the more casual public server crowd would begin to embrace it, at least the basics. But hey, perhaps I'm too hopeful (influenced by my sim history).

 

ALook at birds: EVERYONE turns them off. Why? Same reason you don't turn failures on: You really don't need that intruding on your entertainment time, at least for the vast majority people.

 

I can't speak for everyone of course, only those myself and those I have spoken to directly, but the reason we don't use the "birds" is that the implementation of the effect of a bird strike is unrealistic (catastrophic engine failure every time) as with many of the system failures. Anyway, that's another discussion.

[Eddie]

and regardless if there's a couple squads out there that run their own private servers (thats still a minority) and a bunch of people who only play single player (i have no idea if this is a minority or not),

 

I'm afraid your view is somewhat clouded by your own experience here. Offline players make up by far the vast majority of all DCS players, and I'd say there are more people in organised groups of various types than there are those like you. It's just that these forums have an over representation from the casual public server players which gives the impression it's a larger portion of the community than it is.

[GGTharos]

While true, those of us wishing to make use of realistic procedures find it very difficult to do so due to some of the limitations in DCS, this being one of them. Yes you can role play it, but as MBot has already mentioned, role play for the sake of it isn't very appealing when you know it really makes no difference to the outcome.

 

Not to mention the training burden caused by having to spend hours correcting bad habits caused by the current status quo.

 

I can see this for BFM and perhaps other operational things, but not for an ID matrix. An ID matrix would be something new to learn; there's no bad habit there IMHO. It's just SA -> no SA.

 

Although I will say I do believe that if the sim did more to encourage and reward the use of realistic procedure (and the documentation and training taught it) I think even the more casual public server crowd would begin to embrace it, at least the basics. But hey, perhaps I'm too hopeful (influenced by my sim history).

 

Some will, some won't care. There is a small number of people that care about it. You might think that IFF is easy, but half the time someone new to the game joins, the first thing they do is shoot down a friendly. Sure, it becomes easy later ... just like a startup procedure. :)

 

I can't speak for everyone of course, only those myself and those I have spoken to directly, but the reason we don't use the "birds" is that the implementation of the effect of a bird strike is unrealistic (catastrophic engine failure every time) as with many of the system failures. Anyway, that's another discussion.

 

Unfortunately I don't think there's a more graceful engine failure model in-game. Generally the purpose of birds was to encourage lawnmowers to stop mowing the lawn. :)

[Pitot]

I once talked about implementing IFF in DCS and people responded with flame: "But the IFF codes are A SECRET!!!" which is so shallow...

Why wouldn't ED develop their own IFF codes and implement them in the aircrafts, hide it God knows where and how in the game code and voila - it works. Give us transponders for advanced models, and Mission Editor fields to enter frequencies for FC3 aircrafts. No reason not to do it. Surely, it takes time to be implemented, but hey... We have proven to be a resilient bunch when it comes to waiting for ED to do their work and develop new thingies, so we can wait for that too :D No problem :D Just do it :D

[QuiGon]

I once talked about implementing IFF in DCS and people responded with flame: "But the IFF codes are A SECRET!!!" which is so shallow...

Why wouldn't ED develop their own IFF codes and implement them in the aircrafts, hide it God knows where and how in the game code and voila - it works. Give us transponders for advanced models, and Mission Editor fields to enter frequencies for FC3 aircrafts. No reason not to do it. Surely, it takes time to be implemented, but hey... We have proven to be a resilient bunch when it comes to waiting for ED to do their work and develop new thingies, so we can wait for that too :D No problem :D Just do it :D

 

I don't think IFF-Codes are necessary in DCS, because that might make the game a bit too complicated if we would have to deal with that. I would already be happy if in case of IFF-interrogation the game would actually check if the targets IFF-transponder is on (and maybe set to the correct mode). The way it is now, the game just doesn't bother about IFF-transponders at all, which makes it impossible for friendlies to be shown as enemies under any circumstances :(

[Zaz0]

Or (I think) easier, make 3rd neutral coalition playable, so they avoid to create an funtional "IFF" for all aircrafts, and this way contacts that not responding to IFF calls could be bandits or neutrals...

[Dudikoff]

Or (I think) easier, make 3rd neutral coalition playable, so they avoid to create an funtional "IFF" for all aircrafts, and this way contacts that not responding to IFF calls could be bandits or neutrals...

 

This is definitely needed in the game in general, not only to give the simplified IFF modeling some depth.

Edited June 20, 2016 by Dudikoff

[Pitot]

I don't think IFF-Codes are necessary in DCS, because that might make the game a bit too complicated if we would have to deal with that. I would already be happy if in case of IFF-interrogation the game would actually check if the targets IFF-transponder is on (and maybe set to the correct mode). The way it is now, the game just doesn't bother about IFF-transponders at all, which makes it impossible for friendlies to be shown as enemies under any circumstances :(

 

Sure mate, but going trough 600 pages of manual, and knowing well the instruments and procedures, and performance limits of detailed modules in DCS is also complicated. Very. What I am saying is that DCS players are mostly intelligent people who wouldn't find it hard even with codes implemented. But yeah, what you have said and comments between the quoted one and this one - I totally agree with them. Multiple coalitions sound good.

[QuiGon]

Sure mate, but going trough 600 pages of manual, and knowing well the instruments and procedures, and performance limits of detailed modules in DCS is also complicated. Very. What I am saying is that DCS players are mostly intelligent people who wouldn't find it hard even with codes implemented. But yeah, what you have said and comments between the quoted one and this one - I totally agree with them. Multiple coalitions sound good.

 

Yes, I also don't think that the actual usage of codes would be a problem, technically, but I think it would be pretty confusing to handle different codes in MP, especially on public servers. That's what I had in mind when I said that ;)

[Pitot]

I agree that MP servers are, at it's current state, pretty bad environment for such operations.

[PapaSmoke]

I don't know if its just taken as read that people already know this (in which case please tell me to poke it), but IFF is only ever really used as part of a wider decision making process to identify tracks and generate the Recognised Air Picture anyway. By no means all IFF-equipped aircraft are equipped to interrogate IFF, for most it is a passive system that is switched on and 'forgotten' for the sortie, save a Parrot/India check at check-in.

The requirement to simulate the IFF systems seems to be a separate discussion from the need to identify friend from target in a multiplayer mixed-type team vs team furball because of the restrictions arising from LoD. I've never played a big 'air-quake' (sorry if that's not the right term) so I can't input on that, but I can input on some of the reasons that implementing a wholly realistic IFF system is absolutely not small beer for the developers.

IFF (really ought to be called I, because it'll only ever positively ID friends with the serviceable kit and the right codes) is only ever used as part of ID criteria that dovetail into the RoE. This is a complex process and there's a lot riding on it, hence why BVR engagements are rare - taking that shot is a HUGE call. ID criteria tend to be complex and based on the ever-changing Int situation. Programming AI that can flexibly and reliably behave within that (and have the weaknesses of that system exploited against it) would be a huge deal. Furthermore, it will rarely be the dude in the jet that gets to say, "yep, he's hostile - Fox 3". The responsibility for that usually (almost always) sits further up the chain and IFF, NCTR and a lot of other sensors and Int provide clues for the person who holds that risk.

Also, where do you draw the line with realism? Because folks will soon get fed up with spending 40 minutes trying to get the Mode 4 to load.

 

To cut a long story short, my thoughts are these:

Anything approaching accurate IFF implementation would require huge alterations to the AI

IFF is only a small part of the way that bogeys become hostiles

Accurately implemented IFF systems and a "shoot/don't shoot" system for air-quake are separate entities

[QuiGon]

I don't know if its just taken as read that people already know this (in which case please tell me to poke it), but IFF is only ever really used as part of a wider decision making process to identify tracks and generate the Recognised Air Picture anyway. By no means all IFF-equipped aircraft are equipped to interrogate IFF, for most it is a passive system that is switched on and 'forgotten' for the sortie, save a Parrot/India check at check-in.

The requirement to simulate the IFF systems seems to be a separate discussion from the need to identify friend from target in a multiplayer mixed-type team vs team furball because of the restrictions arising from LoD. I've never played a big 'air-quake' (sorry if that's not the right term) so I can't input on that, but I can input on some of the reasons that implementing a wholly realistic IFF system is absolutely not small beer for the developers.

IFF (really ought to be called I, because it'll only ever positively ID friends with the serviceable kit and the right codes) is only ever used as part of ID criteria that dovetail into the RoE. This is a complex process and there's a lot riding on it, hence why BVR engagements are rare - taking that shot is a HUGE call. ID criteria tend to be complex and based on the ever-changing Int situation. Programming AI that can flexibly and reliably behave within that (and have the weaknesses of that system exploited against it) would be a huge deal. Furthermore, it will rarely be the dude in the jet that gets to say, "yep, he's hostile - Fox 3". The responsibility for that usually (almost always) sits further up the chain and IFF, NCTR and a lot of other sensors and Int provide clues for the person who holds that risk.

Also, where do you draw the line with realism? Because folks will soon get fed up with spending 40 minutes trying to get the Mode 4 to load.

 

To cut a long story short, my thoughts are these:

Anything approaching accurate IFF implementation would require huge alterations to the AI

IFF is only a small part of the way that bogeys become hostiles

Accurately implemented IFF systems and a "shoot/don't shoot" system for air-quake are separate entities

 

I'm totally with you. I also don't see a true-to-reality implementation of IFF in DCS feasible in the foreseeable future. That's why I propesed a solution (see post #1 of this thread) of which I think it would be better than what we have now without taking up to much work.

 

About your thoughts:

- "Anything approaching accurate IFF implementation would require huge alterations to the AI"

I would just leave the AI as it is (IFF-transponder always "on")

- "IFF is only a small part of the way that bogeys become hostiles"

That's exactly why I proposed that solution in post #1, because it would make IFF unreliable. The way it is now IFF (meaning radar-IFF) is 100% reliable. You interrogate your target and you will get a correct response if the target is friendly or not. With my proposed implementation of IFF it would be possible, that the target is friendly but doesn't respond to the IFF-call because the pilot has not turned on his IFF-transponder.

[Tarambuko]

It would be great if we can utilize the existing IFF panels in a meaningful way or at least have something simple and dirty until a 'real' solution is implemented:

 

- Include in the mission description a required code, frequency, etc. to which aircraft's transponder shall be tuned.

- Enable aircraft transponder's controls (knobs, buttons, levers, etc.) so that player can 'tune' on the required code/frequency.

- Keep everything in the game 'as it is' in terms of unit identification as friendly or enemy with only one exception (see below).

- If the player hasn't configured the transponder (set the knobs and switches in the required positions according to the mission summary) before taking off then he shall be considered an enemy unit by all fractions as soon as he's airborne.

[QuiGon]

It would be great if we can utilize the existing IFF panels in a meaningful way or at least have something simple and dirty until a 'real' solution is implemented:

 

- Include in the mission description a required code, frequency, etc. to which aircraft's transponder shall be tuned.

- Enable aircraft transponder's controls (knobs, buttons, levers, etc.) so that player can 'tune' on the required code/frequency.

- Keep everything in the game 'as it is' in terms of unit identification as friendly or enemy with only one exception (see below).

- If the player hasn't configured the transponder (set the knobs and switches in the required positions according to the mission summary) before taking off then he shall be considered an enemy unit by all fractions as soon as he's airborne.

 

Yeah, that's pretty much what I had in mind. The IFF-controls (the knobs and switches) are already working on most aircraft, meaning they are clickable, but they don't have any effect so far. A simple IFF system as you said (and I proposed in the first post) should be no problem to implement in DCS.

 

I have just one correction on your suggestions. The last suggestion should be:

- If the player hasn't configured the transponder (set the knobs and switches in the required positions according to the mission summary) before taking off then he shall be considered as not friendly (unkown) by all factions as soon as he's airborne.

[Tarambuko]

- If the player hasn't configured the transponder (set the knobs and switches in the required positions according to the mission summary) before taking off then he shall be considered as not friendly (unkown) by all factions as soon as he's airborne.

 

I don't know what is the current AI logic regarding 'unknown' units. Is there a status 'unknown' or the units can be only friendly, enemy, and neutral?

And if there is 'unknown' status defined then will the bots attack unknown units - an unset transponder could make the player 'invisible' for the enemy bots.

[QuiGon]

I don't know what is the current AI logic regarding 'unknown' units. Is there a status 'unknown' or the units can be only friendly, enemy, and neutral?

And if there is 'unknown' status defined then will the bots attack unknown units - an unset transponder could make the player 'invisible' for the enemy bots.

 

You're thinking about AI, alright. I had especially multiplayer in mind :D

 

The AI does indeed only know hostile or friendly. There is no unkown state for AI units, as they always know to what side a unit belongs.

[bn880]

This is a great suggestion IMO, and would avoid the god mode for some fighters. Making it much more interesting and detailed. Also it's interesting to quote AI, as it seems for the most part it can be simulated to act like a human in this regard. (you can add mission options for their IFF knowledge in general, some theaters they'd just know, some they would need to confirm, add appropriate delays, done.)

[GGTharos]

No, he is considered 'friendly' by the friendly faction because they saw where he took off from, they pinged him with a SAM or other available IFF-testing facility to verify that the equipment works, and they would inform the pilot.

 

Likewise, the enemy sees aircraft taking off from a certain airfield, it's hostile.

 

Unless you really, badly want to pretend that we actually have civillian traffic or something - but then, the AI probably doesn't get their codes wrong :)

 

I have just one correction on your suggestions. The last suggestion should be:

- If the player hasn't configured the transponder (set the knobs and switches in the required positions according to the mission summary) before taking off then he shall be considered as not friendly (unkown) by all factions as soon as he's airborne.

[RED]

I think this thread (especially first post by OP) is still very important. I really hope ED doesn't put a FC3 level system into the simulated Hornet. We really miss out on ROE that create interesting fights. ROE are pretty useless if you can ID the contact from 70 miles with your radar. We wouldn't need to use the full potential of the targeting pods - the A-A modes - either. A better IFF-system would also allow older generations like F-5, Mig21 or F-4 to enter WVR fights and have a chance. Thinking about online events like Red Flag it would be really disappointing to not be able to fly with proper ROE because the Hornet can magically IFF every aircraft for 100% sure like the FC3 ones.

Even for Single Player it will be very strange to IFF aircraft that not even have any IFF equipment onboard and would compromise the experience there. With every high fidelity fighter aircraft -M2k, FA-18, F-14, ... - this will get more important.

 

I didn't read the whole thread, so maybe some stuff was already covered. Didn't want to open a new one although the Hornet deserves one. :smilewink: