Jump to content

AIM-120 still can not chase simple Split S manuever.


opps

Recommended Posts

  • ED Team
18 minutes ago, KlarSnow said:

Keep reading

there is a lot more pages to that

 

Page 81, that's his conclusion on what a 10Khz PRF jittered radar could see at 9000 feet just by randomly picking PRF's, not even optimizing it, just randomizing it.

 

And Page 88 shows that by optimizing it at 9000 feet above the ground his theoretical 10Khz MPRF radar can detect targets out to 23 nautical miles in clutter.

And this is according to Nighthawk and others the bad way of doing this that reduces your detection range. PRF switching as they said should be even better...

 

To be honest this isn't even the best paper about this, it just was one of the first to pop up, and it is fairly easy to understand.

 

I have already watched it all thru. Nothing sufficient new for me. What clutter we talking about?? Sidelobe? Mainlobe harmonics(which indeed may be shifted with PRF chnging)?

Or we talking about mainlobe central line clutter? - (I talking exactly about this thing)

Page 61 directly tells about situation we have see in track from first post.

  • Like 3
Link to comment
Share on other sites

14 minutes ago, Маэстро said:

 

I saw this earlier. it assumes unambiguous range.

 

And why would it assume Unambiguous range if it couldn't get unambiguous range using MPRF... This is the baseline assumption everyone in this conversation besides yourself is starting from.

That matrix on the side looks an awful lot like the MPRF range vs doppler matrix in the video that Krippz posted.

I wonder if those are linked.

3 minutes ago, Маэстро said:

I have already watched it all thru. Nothing sufficient new for me. What clutter we talking about?? Sidelobe? Mainlobe harmonics(which indeed may be shifted with PRF chnging)?

Or we talking about mainlobe central line clutter? - (I talking exactly about this thing)

Page 61 directly tells about situation we have see in track from first post.

Did you read to the end where he runs the simulation with the varying PRF's and determines that he can detect targets out to 23NM in lookdown? He definitively shows that using all of these values and yes he calculates the ground clutter using the same method you did. And yet doing all of that he reaches the conclusion that his theoretical radar with optimized PRF jitter can detect targets in that situation out to 23 NM... 

I guess I'll add that all of these are assuming a good level of sidelobe cancelling, this is an important part of MPRF, and helps a lot with the clutter problems. 

I'll ask you this, you keep saying what clutter, the papers and all of these theories and that patent are telling you they dont have problems with any of it by using these methods. So what clutter are you talking about that is magically still there after applying these methods? The mainlobe? The sidelobes? They are running these calculations and doing the PRF Jitter/Switching regardless and it resolves the range ambiguity regardless of clutter. 


Edited by KlarSnow
  • Like 17
  • Thanks 2
Link to comment
Share on other sites

Just to add a bit of literature here...

Method and system for tracking targets in a pulse doppler radar system
A medium to low PRF pulse doppler radar wherein the effect of the clutter notch filter is eliminated at times when the doppler frequency of the target approximates the doppler frequency of the clutter to continue tracking the target provided that the doppler frequency of the target and clutter are at different ranges.

https://patents.google.com/patent/US4450446

Given that this is a westinghouse patent from 1981 id say it very probably applies to say the APG66 or 68 radars and the techniques likely apply to any later radars. It describes both "track through notch" prediction techniques as well as eliminating clutter notches at high alt as Klar mentioned.

 

  • Like 9

New hotness: I7 9700k 4.8ghz, 32gb ddr4, 2080ti, :joystick: TM Warthog. TrackIR, HP Reverb (formermly CV1)

Old-N-busted: i7 4720HQ ~3.5GHZ, +32GB DDR3 + Nvidia GTX980m (4GB VRAM) :joystick: TM Warthog. TrackIR, Rift CV1 (yes really).

Link to comment
Share on other sites

  • ED Team
17 minutes ago, BlackPixxel said:

I don't think anyone said that HPRF/MPRF ambiguity cannot be solved for most conditions. For a target only the range bins that correspond to its doppler frequency need to be checked, and there will only be a few peaks. So solving the ambiguity will be simple.

But when the target is in the notch, the range bins that corrensponds to its doppler will also be full of ground return. So you cannot just search for single peaks and compare their positions over multiple PRFs unless the target really stands out due to favourable geometry of the setup.

Exactly!

11 minutes ago, KlarSnow said:

And why would it assume Unambiguous range if it couldn't get unambiguous range using MPRF... This is the baseline assumption everyone in this conversation besides yourself is starting from.

That matrix on the side looks an awful lot like the MPRF range vs doppler matrix in the video that Krippz posted.

I wonder if those are linked.

 

Без имени-2.jpg

 

 

PS. Let's continue tomorrow I need to go now.

  • Like 1
  • Thanks 3
Link to comment
Share on other sites

No they wont because they are range bins... why would they have any of the main lobe ground return in them. IF they are sorted by range and in the situation presented at the start of this thread. There is no Main Lobe or side lobe clutter in the 2.5 Nautical mile range bin. So if you sort it by range and receive the unambiguous range benefits of MPRF with PRF Jitter, how is the Main lobe or side lobe doppler that is 20 miles away getting into the 2.5 nautical mile range bin. And if it is, how is it of such magnitude that it is overwhelming the target return.

None of this is happening on a single pulse.... This is all integrated over several pulses on several PRF's

MPRF Radars sort by Range bins and track by Range bins first, then they see if the doppler return is in there. Thats what the range vs doppler matrix is. If you have them all laid out again, why if there is no ground return coincident with target range, would any of the ground clutter be in the same range bin as the target.

The side of this as well is that by sorting the entire return into range bins the total clutter level in each bin is lowered. Think instead of the massive MLC spike you get in HPRF, you instead have as many little MLC spikes as you have range and velocity bins. Makes it much much easier to see the target which will have all of its return in one bin, vs the entire spread out clutter spectrum.

If you are doing a single PRF MPRF radar as is apparently being implemented, then yeah, none of this works because you cant solve any of the ambiguity problems and there will totally be doppler in every single range bin no matter what you do.


Edited by KlarSnow
  • Like 20
Link to comment
Share on other sites

19 minutes ago, KlarSnow said:

No they wont because they are range bins... why would they have any of the main lobe ground return in them. IF they are sorted by range and in the situation presented at the start of this thread. There is no Main Lobe or side lobe clutter in the 2.5 Nautical mile range bin. So if you sort it by range and receive the unambiguous range benefits of MPRF with PRF Jitter, how is the Main lobe or side lobe doppler that is 20 miles away getting into the 2.5 nautical mile range bin. And if it is, how is it of such magnitude that it is overwhelming the target return.

None of this is happening on a single pulse.... This is all integrated over several pulses on several PRF's

 

Yeah i'm still a little fuzzy on this but don't rage bins come before the Doppler filters?

Link to comment
Share on other sites

2 hours ago, Маэстро said:

Ok, then please, tell how you will resolve ambuguties and seprate target from clutter signal if all range bins filled with clutter? 

A long time ago

According to razbam's galinette DCS was only returning a constant value for the RCS even when side on.  I linked him to your post and I think he said he was going to talk to you.  Are you sure this is implemented?

  • Like 1
Link to comment
Share on other sites

19 hours ago, GGTharos said:

The manuals say that supporting the 120 beyond pitbull has no effect on Pk.  Of course, that's probably not entirely true.

That is not correct. The manuals all state the the data link improves Pk in situations where the missile loses track. Reference the HAF-34 for the F-16 for example. 

  • Like 1
Link to comment
Share on other sites

13 minutes ago, KenobiOrder said:

That is not correct. The manuals all state the the data link improves Pk in situations where the missile loses track. Reference the HAF-34 for the F-16 for example. 

It is clearly stated that once the missile is autonomous, what the launching aircraft does has no effect on Pk.   Like I said, it's a -34 and details in the classified -34 supplements or other documents may give a more nuanced picture.

[sIGPIC][/sIGPIC]

Reminder: SAM = Speed Bump :D

I used to play flight sims like you, but then I took a slammer to the knee - Yoda

Link to comment
Share on other sites

12 minutes ago, GGTharos said:

It is clearly stated that once the missile is autonomous, what the launching aircraft does has no effect on Pk.   Like I said, it's a -34 and details in the classified -34 supplements or other documents may give a more nuanced picture.

No it does not. I am looking at it right now in two separate documents. It specifically states that maintaining datalink has no affect on missile Pk after the missile goes Pitbull UNLESS the missile loses track. 


Edited by KenobiOrder
  • Like 1
Link to comment
Share on other sites

2 hours ago, KenobiOrder said:

No it does not. I am looking at it right now in two separate documents. It specifically states that maintaining datalink has no affect on missile Pk after the missile goes Pitbull UNLESS the missile loses track. 

 

89401B73-DF7C-4829-A756-C4EE3E42DCD8.pngA paper that explicitly says the opposite, and has data later on to back up why (seeker accuracy and SnR at close range is better than supporting radars,especially in a degraded environment) and as GGTharos said there are ample manuals that match this paper as well. How old are your documents.

https://trace.tennessee.edu/cgi/viewcontent.cgi?article=6680&context=utk_gradthes

  • Like 5
Link to comment
Share on other sites

On 6/9/2022 at 1:49 AM, Santi871 said:

unknown.png

Can you help me understand why it's getting confused by the terrain returns that are 6 range bins away? (assuming 1km wide range bins)

 

MPRF doesn't have clean range bins like LPRF. Range is ambiguous and is resolved with PRF switching and Chinese remainder theorem. Because range clutter is ambiguous, it gets superimposed over all the range bins.

---------------

11 hours ago, Vatikus said:

Already set at time stamp which talks about how active a2a missiles do it

 

I know this says "MPRF", but the parameters say otherwise... (>.>) 

10 usec PRI (100,000Hz PRF) and 30m range res.

------------------

 

ED is in the right here from what I see.

(Though they need to fix a lot of other radar aspects. For instance)

 

The only way the AIM-120 could re-acquire / track the target through a look-down and doppler notch would be with a LPRF mode precisely for this purpose. Who knows if its actually capable of this.  MPRF would likely succeed in tracking the target if the target went into the notch in a 'look-up' condition.

 

I'm sure target RCS does come into play, along with radar beam width and sidelobes. But I wouldn't trust any 'back of the envelope' math to determine the SNR required. A simulation of this would be necessary IMO. 

 

What about dogfights?

With a Fighter radar in "MPRF", the Main Beam is much tighter, and has reduced sidelobes because it can afford to taper the antenna. This means the SNR would be much better than with the AIM-120's radar. Also, a fighter radar may have the capability to lower the PRF into the "LPRF" regime during a lock. There's room to argue that a fighter radar could have decent performance in tracking through a look-down + notch. At least at close range.

 

 

Remember guys: 

HPRF filters out clutter via doppler.

LPRF filters out clutter via range. 

MPRF cannot filter out range or doppler clutter. But if the target 'pops out' above the clutter in either range or doppler, MPRF will pick it up. Hence its use in 'all aspect' detection and ACM.  


Edited by Beamscanner
  • Like 2
Link to comment
Share on other sites

Then I bring it back to the points raised earlier, there are many many methods for reacquiring and dealing with momentary loss of track that have to do with the autopilot, and as has been discussed in here, he is using a single PRF MPRF which is functionally worthless, you can't sort anything out of the ambiguities with that, so the doppler notch will be massive, the clutter problems will be massive. I'm sorry but saying you are developing a single PRF MPRF and then dismissing all research that shows that multiple PRF MPRF resolves many range and clutter ambiguities, implies a massive lack of understanding of what is going on here.

The missile should not completely lose guidance and go utterly dumb when the target enters the notch, and Signal to Noise ratio still plays a part. Using MPRF spreads the clutter signal through all of the range bins. The notch still exists but again, if you are close enough, it can still see the target because the amplitude is high enough. 

My issue right now is less the range gating vs clutter issue, its more the assumptions that are going into this "advanced missile API" and the follow on effects because they betray an utter misunderstanding of what should be some very basic MPRF concepts. 

My other issue is that the dev working on this is not engaging with any of the research, he is dismissing it and literally cherry picking the setup that is describing the problems in the first part of the paper and then ignoring the solution provided at the end because he has "seen it all before". 

The target diving through the notch should not be an unexpected event for a missile that operates with a pulse doppler radar. It should be an expected event that the missile has countermeasures for, many of them being discussed here that minimize or allow it to reacquire after the notch. The way it is right now the missile is completely defeated by a maneuver that is easily predictable, and quite easily resolved.

We are basically pretending that the missile is developed solely for clear air look up employment, that is the only way these design decisions make any sense.


Edited by KlarSnow
  • Like 14
Link to comment
Share on other sites

Using a single PRF is a huge benefit FYI... it means you get to increase the number of pulses integrated (increasing SNR) AND you get a higher update rate on your track (improving tracking algorithm). (normal MPRF search has to cycle through 7-8 PRFs to resolve all range and doppler regions of interest)

It should comfort you to know that a lot of radars do this while tracking.

 

 

As for everything else, its pure speculation. Sure, we can assume that the missile to programmed to 'coast' the track in memory. But it cant know a split S from a normal 90 degree notch. 

 


Edited by Beamscanner
  • Like 2
Link to comment
Share on other sites

11 minutes ago, Beamscanner said:

Using a single PRF is a huge benefit FYI... it means you get to increase the number of pulses integrated (increasing SNR) AND you get a higher update rate on your track (improving tracking algorithm). 

 

As for everything else, its pure speculation. Sure, we can assume that the missile to programmed to 'coast' the track in memory. But it cant know a split S from a normal 90 degree notch. 

 

But using a single PRF then means you have a massive clutter and ambiguity problem exactly as described, that excellent SnR from a single PRF would be great, for long range look up, not for a missile that is always trying to loft and acquire at relatively short ranges in lookdown, and is designed to engage low altitude targets like cruise missiles. He has developed a system that cannot resolve targets in the typical engagement geometries, and has no clutter reduction capabilities. So of course it is going to have a lot of problems. 

Multiple PRF's also reduces the size of the notch in MPRF, making it much smaller than in HPRF, again this adds to all of the guidance and autopilot stuff, if the notch in MPRF was say only 30 knots wide, instead of 100 knots wide, that massively simplifies the notch reacquisition problem.


Edited by KlarSnow
  • Like 8
Link to comment
Share on other sites

5 minutes ago, KlarSnow said:

But using a single PRF then means you have a massive clutter and ambiguity problem exactly as described, that excellent SnR from a single PRF would be great, for long range look up, not for a missile that is always trying to loft and acquire at relatively short ranges in lookdown, and is designed to engage low altitude targets like cruise missiles. He has developed a system that cannot resolve targets, and has no clutter reduction capabilities. So of course it is going to have a lot of problems. 

Nearly all tracking radars switch to a single PRF during track. The PRF is adaptively switched when needed to prevent eclipsing the target in range or doppler. 

 

Using MPRF search (cycling through 8 PRFs) provides a SNR 1/8 of what you'd get from an adaptive PRF track (single PRF). And it reduces the target update rate. 

 

There's no comparison. Once you've found the target, and you don't care about finding other targets, adaptive track is the way to go. 


Edited by Beamscanner
  • Like 4
Link to comment
Share on other sites

42 minutes ago, Beamscanner said:

Nearly all tracking radars switch to a single PRF during track. The PRF is adaptively switched when needed to prevent eclipsing the target in range or doppler. 

 

Using MPRF search (cycling through 8 PRFs) provides a SNR 1/8 of that of using and adaptive PRF track (single PRF). And it reduces the target update rate. 

Ok, so his MPRF should be switching in order to minimize these issues... again you are agreeing that it shouldn't be a static single PRF that never changes. which is what Maestro has claimed is the only solution. This is my issue with his description, he has set a static MPRF PRF, and is not simulating any kind of adaptive adjusting or switching in order to help any of these situations, if he did that, it would resolve many of these issues, Oh no I have A SnR issue or eclipsing issue at this particular PRF, so swap to the next one that doesn't have that, and keep doing that all the way in to the edges of the MLC, which again, will be smaller compared to HPRF because the clutter zones are more spread out due to harmonics.

Again this is all covered in that MPRF PRF paper I posted earlier, it discusses how different PRF's have varying levels of clutter issues and eclipsing, and how swapping between them can allow you to track through one PRF's issues.

He didn't even engage with that part of the paper, just dismissed it out of hand while pointing at the setup to the problem as proving his point.

I also very much do not understand the point of modelling all of this stuff with the radar in utterly minute detail which adds and adds to the development time and opens up so many issues with starting assumptions, how does all of this stuff work with chaff, how does all of this stuff work with jamming. There should be an end goal for the system, that quite honestly can be abstracted, it feels very much like there is no end goal, and we are iterating for years at this point with no real end goal for how the missile should work in sight, and constantly varying levels of missile effectiveness and needlessly obfuscated mechanics.

 

This compounds with every other system and missile that is in some varying state of needing an overhaul. The Russian missiles should have a lot of this complexity and should not be nearly as susceptible to all of this as well, but they are stuck on the old missile API until this one is "done". The Aim-54 should have a lot more complexity to it but my understanding is Heatblur cant do anything more complex than currently because this guidance API isn't "done". 

I get that all of this takes time, but it has been over 3 years since ED has started reworking this stuff and they have been bouncing back and forth between various things and have not clearly explained what exactly the end goal of what we should see looks like.

This combined with what appears to be some very questionable assumptions about how radars in general work (this applies to far more than the amraam) leads to a very general feeling of disconnect between the community and the developers.

This is a game, what is the desired endstate of gameplay that you are working towards, this drives all of your assumptions about how these things you are speculating on work.

What are you trying to make here?

What is the criteria of done?

It is obviously working as intended as he has explained, but is that how its going to sit? is the next update going to break it more, are any of the possible solutions we are talking about or theories going to eventually be implemented?

Basically what this entire development of various features on the AMRAAM feels like is pointless iteration that either has no discernable or in many cases a deleterious effect on the gameplay.

A "Feature" is added, the community points out the problems, an iteration on it occurs, the community points out more issues, nobody knows what the desired endstate is, and this cycle continues with features getting added or removed or tweaked and nobody has a definition of what "realistic" results for this stuff should be. This has extended and extended this process and lead to whiplashes back and forth on the capability of weapons and the effectiveness of various parts of how the gameplay works.

How is this all supposed to work in the end? If we knew that we could appropriately assess bugs vs not implemented yet vs working as intended. Right now there is no clear layout of how all of this works.


Edited by KlarSnow
  • Like 17
  • Thanks 3
Link to comment
Share on other sites

22 minutes ago, Beamscanner said:

As for everything else, its pure speculation. Sure, we can assume that the missile to programmed to 'coast' the track in memory. But it cant know a split S from a normal 90 degree notch.

The AIM-120 has an INS so by default its going to know the targets location in 3D space.  You could easily look for a split S based off of this information.  We know from documentation that there is a method by which there are adjustments made to compensate for the split-S on the AIM-7.  I highly doubt the amraam would not have something similar if not even better.

Link to comment
Share on other sites

  • ED Team
3 hours ago, KlarSnow said:

How is this all supposed to work in the end? If we knew that we could appropriately assess bugs vs not implemented yet vs working as intended. Right now there is no clear layout of how all of this works.

Hence the reason for the white paper, and for the interaction here. To show the design consideration, what we can and cant do based on info we have, etc. I see a lot of assumptions throughout all this, so I don't know where it will all land or how some of you will accept that some of this stuff is just unknown. I see some who want this spectacular missile so to get kill after kill, and some that want to see it be trashed easy. It will end up at some point in between. 

  • Like 3

64Sig.png
Forum RulesMy YouTube • My Discord - NineLine#0440• **How to Report a Bug**

1146563203_makefg(6).png.82dab0a01be3a361522f3fff75916ba4.png  80141746_makefg(1).png.6fa028f2fe35222644e87c786da1fabb.png  28661714_makefg(2).png.b3816386a8f83b0cceab6cb43ae2477e.png  389390805_makefg(3).png.bca83a238dd2aaf235ea3ce2873b55bc.png  216757889_makefg(4).png.35cb826069cdae5c1a164a94deaff377.png  1359338181_makefg(5).png.e6135dea01fa097e5d841ee5fb3c2dc5.png

Link to comment
Share on other sites

I think the question of what the end-state should be is one that is very apt. I can barely follow the most basic concepts discussed in this thread, and I appreciate the level of knowledge you guys can bring to the table on this matter. But as KlarSnow mentioned, without a clearly defined end-state, it's hard to say if we're moving in a good direction or not. 

 

I'm just an average DCS pilot, and not qualified to say what's right or wrong, but from picking up bits and pieces of information from pilots, I get the impression that reliable missile defence in the real world is GENERALLY achieved through defeating missiles kinematically or through EW, not by performing basic aerobatics or last-ditch maneuvers. I think generally this is the end-state most people expect, based on what we can piece together from real world tactics and training snippets. Regardless of whether that's the case or not in real life, it's certainly not reflected in DCS currently. 

  • Like 12
Link to comment
Share on other sites

16 minutes ago, TordinVarglund said:

I think the question of what the end-state should be is one that is very apt. I can barely follow the most basic concepts discussed in this thread, and I appreciate the level of knowledge you guys can bring to the table on this matter. But as KlarSnow mentioned, without a clearly defined end-state, it's hard to say if we're moving in a good direction or not. 

 

I'm just an average DCS pilot, and not qualified to say what's right or wrong, but from picking up bits and pieces of information from pilots, I get the impression that reliable missile defence in the real world is GENERALLY achieved through defeating missiles kinematically or through EW, not by performing basic aerobatics or last-ditch maneuvers. I think generally this is the end-state most people expect, based on what we can piece together from real world tactics and training snippets. Regardless of whether that's the case or not in real life, it's certainly not reflected in DCS currently. 

Agree.

While we can speculate forever and ever: talking to various SME's, piecing together what we can from tactics, etc. of the era DCS is simulating we know from various sources that the real world method to defeating a modern ARH missile is through kinematics and not these last ditch maneuvers. These last ditch maneuvers should certainly work to some extent, but the end goal should be a place where it is not repeatable or reliable to defeat missiles through these measures, certainly not reliable enough that it becomes option #1.

  • Like 11

 1A100.png?format=1500w  

Virtual CVW-8 - The mission of Virtual Carrier Air Wing EIGHT is to provide its members with an organization committed to presenting an authentic representation of U.S. Navy Carrier Air Wing operations in training and combat environments based on the real world experience of its real fighter pilots, air intercept controllers, airbosses, and many others.

 

Link to comment
Share on other sites

Guest
This topic is now closed to further replies.
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...