R-27ER Experiment & Findings

[Fri13]

4 years now we've been talking about how broken the ER/R guidance is...

 

I don't wonder so much how long people have been talking about "problems", but what really baffles me is that what it is in reality. Like R-27 is the main missile for not just Russia, but for so many other country. And the oddity I just see is that if the R-27 is so terrible as it is in DCS, or it is claimed by some of the members, then why it is in use and bought "today".

 

So what ever it is, at least the other acknowledgement would have helped than "Missiles are Work In Progress" in these years period.

 

Maybe the missile really is so terrible that it is just better to use a cannon anyways and is a reason why countries with R-27 in service does everything in their power to not get to situation where it would be required to be in use....

 

So far the DCS community keeps going with very nice testing and results, problem is just there are those who like results and then those who dislike the results.

[GGTharos]

@karambiatros

 

... what would you like it to be, if a non-maneuvering, non-chaffing target comes with a Pk or ~0.6?

 

It should be higher in this case, but if you consider that maneuver isn't taken into account, nor ECM, what exactly would like the pk to be here? It's about half of the 'doing nothing' shot, which seems ok I suppose ... assuming that you want chaff to have a purpose.

 

And realize that ED is also tuning with/vs AI for SP as well.

 

@Fri13

 

In reality it's 'classified', but there are a couple of numbers available. One example of real combat usage has it operating at a stellar 0.05Pk (that's 5%) - but we try not to talk about that conflict since information is thin and even a bit fishy. As a result the best Pk figures we have are mostly for western missiles and Pks tend to be matched to that for comparison. Also realize that air forces used missiles anyway even when in the beginning they had abysmal 0.1 or less pks. The gun was almost never a better weapon and it has certainly not been used much at all in air to air combat lately.

 

However, who/what your target is really changes the whole equation in-game. MP introduces not just players, but also code issues that are difficult to deal with.

Edited April 6, 2017 by GGTharos

[Ironhand]

Yeah im not saying they dont hit but them hitting is an exception rather then a rule. What was it? Less than 30 percent of the time against a non maneuvaring target thats chaffing? From ironhands tests

I need to clear something up. That 30% figure was for the duration of the track which included the weirdness of the third volley. If you only include the first volley, then the Pk in the presence of chaff is 35KM 0% (max range), 30KM 33%, 25 KM 76.6%, and 20KM 76.6%.

 

Yes, more missiles than that missed but they missed in the no-chaff track, too. So you have to subtract those from the total chaff misses as well. Otherwise you're guaranteed to have skewed results due to non-chaff factors.

[DarkFire]

Exactly correct. The only way to make all of this work correctly is a huge under-taking because it involves, IMHO:

 

1) Adding more logic/algorithms for countermeasure characterization than probability of decoy by CM (modified by look down/up, sun in the way, number of CMs in view) and implementing the notion of centroids generated by these CMs for sensors that have low enough resolution to affected in this way (Every heater up to FPA based heaters, vast majority of radar seekers).

 

This will significantly differentiate flares and chaff at minimum, and add infrastructure to differenciate generations of missiles and how they handle CMs. By adding basic, more realistic ECCMs algorithms as well, you will now have to combine CM + maneuver to evade, but at the same time you'll find that some types of missiles (heaters especially) are much easier to decoy, even in AB. If you consider radar guided stuff, it's more resilient (more available data to deal with decoys) but you get into trouble once you add ECM, assuming that would happen.

 

This would be the dream solution I guess, though things would become even more complicated with the eventual introduction of the F-18 and Typhoon since any modelled ECM environment would then have to take in to consideration the likes of the ADM 141/160, towed decoys and even chaff replacement actives like BriteCloud.

 

My opinion has always been that I don't think our desktop PC rigs have enough horsepower to properly model an entire electronic warfare environment, especially not for theatre-wide combined arms operations.

 

Realistically speaking I think we may be stuck with the current chaff-as-flares situation until the point is reached where EW can be sensibly modelled using limited compute resources.

[Ironhand]

...

 

My opinion has always been that I don't think our desktop PC rigs have enough horsepower to properly model an entire electronic warfare environment, especially not for theatre-wide combined arms operations.

 

Realistically speaking I think we may be stuck with the current chaff-as-flares situation until the point is reached where EW can be sensibly modelled using limited compute resources.

When I read what GG had written my first thought was: that'd be great if this was a missile simulator... :)

[karambiatos]

 

Yes, more missiles than that missed but they missed in the no-chaff track, too. So you have to subtract those from the total chaff misses as well. Otherwise you're guaranteed to have skewed results due to non-chaff factors.

Honestly the non-chaff missies are what are the most interesting to me.

 

@karambiatros

 

... what would you like it to be, if a non-maneuvering, non-chaffing target comes with a Pk or ~0.6?

 

It should be higher in this case, but if you consider that maneuver isn't taken into account, nor ECM, what exactly would like the pk to be here? It's about half of the 'doing nothing' shot, which seems ok I suppose ... assuming that you want chaff to have a purpose.

 

And realize that ED is also tuning with/vs AI for SP as well.

10-20% more likely to hit a non-maneuvering chaffing target from any aspect but near notch, if you want a number.

 

I still fondly remember when i played FC2 where scary, something close to that would be very nice, though perhaps in FC2 chaff was far less useful, perhaps it shouldn't be as effective as it is currently in DCS.

I don't know. It's just that in this point in time it seems to me like it functions exactly like a flare for radar missiles.

 

In my opinion currently in DCS it's, like we have a 6 sided dice when it comes to missiles being spoofed by countermeasures, perhaps we should have 12 or 20 sided dice.

 

Terrible analogy perhaps.

[GGTharos]

I am amused by the dice reference. :)

 

I wouldn't make comparisons between chaff and flares. The reason it appears to function in a similar manner is that they are using the same statistical function.

IIRC there may be some parameters that can be tweaked around the notch, but they don't seem to do what I want them to do.

 

Regarding flares, the missiles are actually too resistant, in the not-so-humble-opinion of certain studies, at close ranges and in some cases at longer ranges. For the types of heaters we have simulated in that game, a certain number of flares = decoyed missile.

 

Although amusingly enough, while the seeker is decoyed the missile might still hit depending on trajectory and a bunch of other factors. So missile decoy probability may be nearly 100%, but hit probability is not 0. :)

 

But anyway, that's an aside. The point here is that an aircraft 1km away spewing flares, AB off when you launch the missile should decoy the missile. It doesn't happen in-game.

[The Black Swan]

I need to clear something up. That 30% figure was for the duration of the track which included the weirdness of the third volley. If you only include the first volley, then the Pk in the presence of chaff is 35KM 0% (max range), 30KM 33%, 25 KM 76.6%, and 20KM 76.6%.

 

Yes, more missiles than that missed but they missed in the no-chaff track, too. So you have to subtract those from the total chaff misses as well. Otherwise you're guaranteed to have skewed results due to non-chaff factors.

 

If those are the figures then i don't see too big of a problem.

 

 

 

I still fondly remember when i played FC2 where scary, something close to that would be very nice, though perhaps in FC2 chaff was far less useful, perhaps it shouldn't be as effective as it is currently in DCS.

I don't know. It's just that in this point in time it seems to me like it functions exactly like a flare for radar missiles.

 

In my opinion currently in DCS it's, like we have a 6 sided dice when it comes to missiles being spoofed by countermeasures, perhaps we should have 12 or 20 sided dice.

 

Terrible analogy perhaps.

 

If Ironhand's figures are even close to accurate then i would guess that the low Pk of missiles in multiplayer has to do with the combination of chaff WITH maneuvering to drag down the missiles performance. Which leads me to the conclusion that more accurate drag simulation for the missiles would alleviate the problem.

 

Missiles which keep near their top speed longer will give less time to react, less time to get to/spend in the notch, and a larger area in which to worry about CM.

 

I guess what I'm saying is, based on Ironhand's tests and my own experience, I'm beginning to think the best way to make missiles "scary" as you put it, is to fix/improve the range and speed problem.

 

A little tweak to the CM might help with the ability to spam chaff to win, but overall...

to achieve the most realism at the smallest performance cost, i think helping the drag issue is the way to go.

[GGTharos]

This would be the dream solution I guess, though things would become even more complicated with the eventual introduction of the F-18 and Typhoon since any modelled ECM environment would then have to take in to consideration the likes of the ADM 141/160, towed decoys and even chaff replacement actives like BriteCloud.

 

This is a non-issue ...

 

My opinion has always been that I don't think our desktop PC rigs have enough horsepower to properly model an entire electronic warfare environment, especially not for theatre-wide combined arms operations.

 

Yes they do. We're not trying to simulate EM here ... there are some basic behaviors that can be simulated with the use of simple-ish properties and functions. It's not a ridiculous resource hog on the PC. It's more like it's complex to design and correctly apply and maintain this thing.

[DarkFire]

This is a non-issue ...

 

Yes they do. We're not trying to simulate EM here ... there are some basic behaviors that can be simulated with the use of simple-ish properties and functions. It's not a ridiculous resource hog on the PC. It's more like it's complex to design and correctly apply and maintain this thing.

 

Fair enough. What little I do know about modelling that sort of thing on a PC is 20+ years out of date. I guess there are ways to approximate some EM behaviours, after all the relevant mathematics has been a known thing since at least the 60's. If it can be simulated to an extent that balances computational performance with realism, then hopefully it's something that ED will consider at some point.

[Ironhand]

Honestly the non-chaff missies are what are the most interesting to me...

If you study the videos and, especially, look at the TRK files, I think you'll find that the majority of those misses occurred due to the launching aircraft maneuvering in a way that lost the lock. That's what it seemed to me. Keep in mind that all of those tracks were generated from the same initial track file. So all of the AI start positions were almost identical. The only major change occurred in the placement of the target aircraft in relationship to them. So you are likely to see the same AI deficits throughout.

[SinusoidDelta]

I need to take a close look at the track files still. We have almost no idea how ED guiding or controlling the missile.

 

The game isn't performing a total multiphysics simulation. From where we are we know little to nothing about how ED is simulating the process. What we do seems....basic.

 

I need to watch the tracks in detail again but the first thing that comes to mind is english-bias of lack therof.

[SinusoidDelta]

To clarify regarding English Bias - Maybe someone more knowledgeable than myself can comment if this has any relevance to this discussion. Note the commment about accelerating laterally.

Edited April 7, 2017 by SinusoidDelta

[Ironhand]

To clarify regarding English Bias - Maybe someone more knowledgeable than myself can comment if this has any relevance to this discussion. Note the commment about accelerating laterally.

So that's how it's done. I doubt that calculations for all of that are being made in the sim. Rather a sort of shorthand is being used to mimic the results of those calculations.

[GGTharos]

The sim uses a PN function. There is no additional calculation except for lofting.

 

Also, 'wing deflection' is out-dated...torque is used instead to command g.

[SinusoidDelta]

The sim uses a PN function. There is no additional calculation except for lofting.

 

Also, 'wing deflection' is out-dated...torque is used instead to command g.

 

What are you posting in reference to?

 

And outdated compared to what?

[GGTharos]

In reference to your post, specifics as to what happens in the game.

 

Wing deflection command is outdated IRL since the 50's or 60's. IIRC. Wing torque command is used instead.

Edited April 9, 2017 by GGTharos

[SinusoidDelta]

In reference to your post, specifics as to what happens in the game.

 

Wing deflection command is outdated IRL since the 50's or 60's. IIRC. Wing torque command is used instead.

 

I see, for pitching moments and such :doh:

 

The rest of this post are just my thoughts from observing this thread, not directed at you GG

 

In DCS, immediately after launch, missiles almost always make nonsensical energy depleting maneuvers. The code is is trying to navigate proportionally but the commanded maneuvers result in a negative feedback loop, As an observer its a fairly straight forward problem. Inaccurate state estimation, and inaccurate control command/response.

 

I honestly think many of these issues could be fixed with some fundamental engineering techniques with respect to missile design even if not explicitly known to be used in a particular missile. Most fundamentals for modern control systems have been publicly known since the 1970's, such as Kalman filters ( or 1930's IIRC for the monte-carlo method). Recursive calculations of matrices such as covariance or kalman gain use a negligible amounts of computing power.

 

With the exception of lag/network issues, the absolute state of any target in the ME environment is always known. Sparing the details, the most challenging engineering problems in missile design don't exist here in DCS. I would say it's reasonable to think missiles in such a simulation would outperform consistently compared to real life. Frankly, I'm still baffled DCS missiles perform this badly and have not improved. I'm starting to lose hope that they will.

Edited April 9, 2017 by SinusoidDelta

[GGTharos]

Ok, I guess I lost the plot. :D

[ShuRugal]

Frankly, I'm still baffled DCS missiles perform this badly and have not improved. I'm starting to lose hope that they will.

 

I think this thread is nearly 500 pieces of supporting evidence for the last statement, there.

[Schmidtfire]

Is the R-27ER supposed to give an RWR warning on launch? Or should it give a launch warning when it is "x" distance from the target, almost like when the AiM-120 goes pitbull? I read somewhere that the R-27ER has some type of midcourse link (radio?) with the launch platform.

[Sweep]

HPRF STT + signal change (missile uplink) is probably going to trigger your RWR...

[GGTharos]

Yes, it should. It is not launched without an STT lock on target, so there are no silent launches.

 

Is the R-27ER supposed to give an RWR warning on launch? Or should it give a launch warning when it is "x" distance from the target, almost like when the AiM-120 goes pitbull? I read somewhere that the R-27ER has some type of midcourse link (radio?) with the launch platform.

[Weltensegler]

couple patches ago I did a test with a flanker launching at me and I still got missile warnings when the ER was defeated and past me and the bandit had turned cold :D ...go figure

[red_coreSix]

RWR memory.