Jump to content

Recommended Posts

Posted

So i've been messing around with the radar and have noticed that the track memory never reacquires a target.  Essentially what i'm noticing is the target goes into the notch for the briefest of moments then comes out.  Making it so that he's detected in the next sweep yet it creates a new contact instead of realizing that it's the same contact.  Now it makes sense if the contact accelerates rapidly and can move quite far from where he was last seen and it takes multiple seeps to reacquire (more than 2 or 3) that it won't reacquire.  But in the example i'll provide below were talking about a target that hasn't even moved more than 1km from where he was last seen and is heading in a similar direction.  I just don't see why it shouldn't be able to say hey this target is close to where this contact was last seen, heading in a similar direction, and is within the predicted displacement range.  Is this an actual limitation can the computer just not recorelate tracks to contacts just seems unlikely to me.  Also something else that is concerning is that the moment a bandit hits the blind speed you will lose track and not reacquire, even if you are only in the blind speed for fractions of a second.  

 

 

Example track:

TWS_BUGS.trk

-Target loss at ~7:46:29 in game time target entered notch some time in between the last position update 4 seconds before and the above time.  Target is required at in second sweep at ~07:46:37.

Posted

As far as I am aware, this is a realistic limitation of the AWG-9 and its WCS. You need to remember that the AWG-9 is 60/70s technology and there’s only so much it can do. It certainly capable of correlating consecutive pings as the same target for a track, but once it’s lost track of something it doesn’t have the computation power to ask itself if that is a new track or the same one.

Heavy Fighter Elitist
AIM-120 Best Missiletm
AWG-9 Gaslighter
Diagnosed with terminal Skill Issue

Posted

But shouldn't it be able to merge two track if they are close enough?

Modules: FC3, Mirage 2000C, Harrier AV-8B NA, F-5, AJS-37 Viggen, F-14B, F-14A, Combined Arms, F/A-18C, F-16C, MiG-19P, F-86, MiG-15, FW-190A, Spitfire Mk IX, UH-1 Huey, Su-25, P-51PD, Caucasus map, Nevada map, Persian Gulf map, Marianas map, Syria Map, Super Carrier, Sinai map, Mosquito, P-51, AH-64 Apache, F4U Corsair, WWII Assets Pack

Posted
2 hours ago, captain_dalan said:

But shouldn't it be able to merge two track if they are close enough?

There’s a function on the CAP category that allows you to merge previous tracks with new ones. Not implemented yet.

  • Like 1

BreaKKer

CAG and Commanding Officer of:

Carrier Air Wing Five //  VF-154 Black Knights

 

Posted
7 hours ago, BreaKKer said:

There’s a function on the CAP category that allows you to merge previous tracks with new ones. Not implemented yet.

Ah, good to know! That explains so many things

Modules: FC3, Mirage 2000C, Harrier AV-8B NA, F-5, AJS-37 Viggen, F-14B, F-14A, Combined Arms, F/A-18C, F-16C, MiG-19P, F-86, MiG-15, FW-190A, Spitfire Mk IX, UH-1 Huey, Su-25, P-51PD, Caucasus map, Nevada map, Persian Gulf map, Marianas map, Syria Map, Super Carrier, Sinai map, Mosquito, P-51, AH-64 Apache, F4U Corsair, WWII Assets Pack

Posted
On 4/17/2021 at 1:24 PM, BreaKKer said:

There’s a function on the CAP category that allows you to merge previous tracks with new ones. Not implemented yet.

This is interesting would be nice to have this added.

Posted
On 4/17/2021 at 4:36 PM, Prez said:

As far as I am aware, this is a realistic limitation of the AWG-9 and its WCS. You need to remember that the AWG-9 is 60/70s technology and there’s only so much it can do. It certainly capable of correlating consecutive pings as the same target for a track, but once it’s lost track of something it doesn’t have the computation power to ask itself if that is a new track or the same one.

Heatblur has said that it should miss 5 consecutive sweeps on target before going into trackhold/broken track. There's some basic correllation it would still be able to do based on expected doppler and such in the first brief moments of contact loss.

Posted (edited)

So TWS takes 2 seconds? 2x5=10 seconds? I don't think I've ever seen a lost track get correlated after the second pass (I guess that's 4 seconds). 

Edited by WelshZeCorgi
Posted (edited)
18 hours ago, WelshZeCorgi said:

So TWS takes 2 seconds? 2x5=10 seconds? I don't think I've ever seen a lost track get correlated after the second pass (I guess that's 4 seconds). 

 

Off the top of my head, I believe a trackfile is lost after no return for 8 seconds.

 

What I've always found strange is how easily trackfiles are lost against non maneuvering/ non-notching aircraft. Seems like a trackfile will be lost momentarily only to have the target reappear right on top of the lost track instantaneously. I mostly play online so a lot of the issue could stem from the ever present desync problems in large dynamic servers. I know its old tech and its most likely "just how it was" but man its annoying trying to actually pull off a six shooter. Doesn't help either that the omniscient AI scatters (even though they'd have zero RWR indication of a missile launch) the second you Fox 3 in TWS making you lose all of your trackfiles.

Edited by Hawkeye91
Posted
1 hour ago, Hawkeye91 said:

What I've always found strange is how easily trackfiles are lost against non maneuvering/ non-notching aircraft. Seems like a trackfile will be lost momentarily only to have the target reappear right on top of the lost track instantaneously.

Yes this is most annoying "feature" of HBs F-14 for me right now, other bugs/wip things Im OK to live with for now, but I highly doubt NAVY would approve AWG-9/AIM-54 system for fleet defense if that system behaved like this. Or if they did, just wow, Im kind of glad they never had to really use it.  

Posted (edited)
19 hours ago, WelshZeCorgi said:

So TWS takes 2 seconds? 2x5=10 seconds? I don't think I've ever seen a lost track get correlated after the second pass (I guess that's 4 seconds). 

 

https://forums.eagle.ru/topic/256009-does-the-aim-54-need-a-constant-tws-track/?do=findComment&comment=4514507

 

I misremembered, @Naquaiistated it marks it as lost after 3 frames and dropped after 7. What I find more interesting is the following: "During that time it can still be correlated if it gets new radar returns fitting the parameters."

 

I don't think I've EVER seen it recorrelate a track after marking it lost. Would be interesting to hear clarification on what behaviour we should see exactly from the RIO seat. I'm curious how tight those parameters 🤔 The average latency of participants on our server tends to be sub 30 ms.

 

Depending on what constitutes a "frame", if it really drops it after a single missed return I'd be inclined to say that is a bug?

 

Edited by Noctrach
  • Like 1
Posted (edited)
4 minutes ago, Noctrach said:

https://forums.eagle.ru/topic/256009-does-the-aim-54-need-a-constant-tws-track/?do=findComment&comment=4514507

 

I misremembered, @Naquaiistated it marks it as lost after 3 frames and dropped after 7. What I find more interesting is the following: "During that time it can still be correlated if it gets new radar returns fitting the parameters."

I don't think I've EVER seen it recorrelate a track after marking it lost. Would be interesting to hear clarification on what behaviour we should see exactly from the RIO seat. I'm curious how tight those parameters 🤔

 

Depending on what constitutes a "frame", if it really drops it after a single missed return I'd be inclined to say that is a bug?

 

 

I bet is more likely that we don't see the trackfile lost until it's already been awhile since the radar picked it up so the trackfile lost indication on the TID would show after 8 or so seconds of no return as opposed to the first time the radar didn't pick it up on a sweep.

Edited by Hawkeye91
Posted
1 minute ago, Hawkeye91 said:

I bet is more likely that we don't see the trackfile lost until it's already been awhile since the radar picked it up so the trackfile lost indication on the TID would show after 8 or so seconds of no return as opposed to the first time the radar didn't pick it up on a sweep.

 

What I'm curious about is whether he meant those 3 frames allow recorrelation, or all the way until frame 7.

If it's marked lost at frame 3, but can recorrelate by frame 7 then I've never seen this happen. (Never seen a broken track "unbreak") It will just paste a new, identical track on top of the one that's lost most of the time.

 

Besides, we can correllate whether sweeps are missed on the DDD. Usually you can tell when you're losing a track because the DDD sweep misses a return, which pretty much always means the track is lost shortly after.

  • Thanks 2
Posted
1 hour ago, Noctrach said:

What I'm curious about is whether he meant those 3 frames allow recorrelation, or all the way until frame 7.

If it's marked lost at frame 3, but can recorrelate by frame 7 then I've never seen this happen. (Never seen a broken track "unbreak") It will just paste a new, identical track on top of the one that's lost most of the time.

 

Besides, we can correllate whether sweeps are missed on the DDD. Usually you can tell when you're losing a track because the DDD sweep misses a return, which pretty much always means the track is lost shortly after.

I wonder too how much things changed in the real aircraft with software updates. Now I know computers back when these planes first came out where verrry primitive, but is it possible they could have updated and changed track correlation logic throughout the life of the aircraft that could better reestablish trackfiles?

Posted
1 hour ago, Noctrach said:

What I'm curious about is whether he meant those 3 frames allow recorrelation, or all the way until frame 7.

If it's marked lost at frame 3, but can recorrelate by frame 7 then I've never seen this happen. (Never seen a broken track "unbreak") It will just paste a new, identical track on top of the one that's lost most of the time.

 

Besides, we can correllate whether sweeps are missed on the DDD. Usually you can tell when you're losing a track because the DDD sweep misses a return, which pretty much always means the track is lost shortly after.

Yeah I haven't seen one recorellate either in the track this is what happens.

Posted (edited)

I'd be primarily interested in hearing roughly what parameters the AWG-9 would use for correllation.

 

E.g. if it's closure and altitude, then the current logic of having a two-ship resolve into a broken track on top of two new trackfiles at 35-40 miles is suspect imo.

At least one of those two should correlate with the original return even if TWS couldn't make it out to be two aircraft.

 

I get the whole "weak vs fighters" bit but the different statements by HB on the radar's logic/capabilities don't seem to add up to what we see in-game entirely.

As it stands the AWG-9 is "a 60s radar utterly incapable of resolving broken trackfiles", which is fine to me, but not what they seem to be saying.

Edited by Noctrach
  • Like 1
Posted

I there is a general misapprehension of what primitive means in terms of the radar in these aircraft. Track While Scan correlation logic and track memory are not particularly advanced capabilities. They certainly provide the high level of functionality required for a pulse doppler aircraft radar, but everyone seems to forget that these kinds of capabilities are literally 40-60 years old.

 

If the real AWG-9 functioned like the one we have in game, than no one would ever have considered it capable to performing its design mission. It would not even have been suitable for fleet defense agaisnt missiles or bombers. The current in-game radar cant seem to keep a track to save its life even under the simplest of conditions. It cant seem to correlate tracks properly, and this is probably part of why the OP noticed that the radar never seems to require a lost track.

 

Do some reading on how track memory works or how TWS track correlation works. I highly recommend Skolniks radar handbook or Stimsons Introduction to Airborne radar. The aircraft computer basically needs to calculate the statistical distance of received contacts from previously known ones by comparing them based on weighted parameters of said tracks etc. You dont need a mainframe to do this. Consider that while your home computer is orders of magnitude more powerful than the computers in a F-14, it is doing something like this while running the rest of the DCS game and whatever else you have running in the background on your PC.

Posted
On 4/25/2021 at 4:16 AM, KenobiOrder said:

I there is a general misapprehension of what primitive means in terms of the radar in these aircraft. Track While Scan correlation logic and track memory are not particularly advanced capabilities. They certainly provide the high level of functionality required for a pulse doppler aircraft radar, but everyone seems to forget that these kinds of capabilities are literally 40-60 years old.

 

If the real AWG-9 functioned like the one we have in game, than no one would ever have considered it capable to performing its design mission. It would not even have been suitable for fleet defense agaisnt missiles or bombers. The current in-game radar cant seem to keep a track to save its life even under the simplest of conditions. It cant seem to correlate tracks properly, and this is probably part of why the OP noticed that the radar never seems to require a lost track.

 

Do some reading on how track memory works or how TWS track correlation works. I highly recommend Skolniks radar handbook or Stimsons Introduction to Airborne radar. The aircraft computer basically needs to calculate the statistical distance of received contacts from previously known ones by comparing them based on weighted parameters of said tracks etc. You dont need a mainframe to do this. Consider that while your home computer is orders of magnitude more powerful than the computers in a F-14, it is doing something like this while running the rest of the DCS game and whatever else you have running in the background on your PC.

I've recently started bouncing back and forth between multiplayer and single player and the radar is by FAR more reliable in single player. I think multiplayer has a huge sync problem and this screws with the AWG9 simulation. In multiplayer even with an STT lock its basically impossible to hold the lock long enough to get a phoenix to hit from almost any range outside of 20nm even if you fly low enough below the target to have the filter automatically turned off, it just randomly loses the lock and will immediately show up again when Jester goes back to TWS.

  • Like 1
Posted

Its almost like we need a Kalman filter inbetween the netcode and radar simulation (no idea if such a thing has been attempted / implemented already)

 

Using the raw netcode as your radar simulation input all someone has to do is lag for 2 or 3 seconds warping across the screen and your track / STT is lost and wont correlate. Even when they catch up and warp back to where, from their perspective, they were all along. Your trackfile is still hosed.

 

Taking some kind of filter on where the target appears to be for a few seconds while they are lagging and just substituting that raw data with extrapolated data (almost another TWS correlation 🙂 ), possibly with some kind of error correction checking the track into the past - so if eventually they do really break the lock but it occurred during a lag spike you could backdate the TWS being lost and drop the track.

 

None of this would be simple though and edge cases abound.

Posted

I've wondered about this on multiple occasions but so far the only response to this question that we have had from the developers is "Working as intended". 

Specs: Win10, i5-13600KF, 32GB DDR4 RAM 3200XMP, 1 TB M2 NVMe SSD, KFA2 RTX3090, VR G2 Headset, Warthog Throttle+Saitek Pedals+MSFFB2  Joystick. 

  • Recently Browsing   0 members

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