Does The Aim-54 need a Constant TWS Track?

[THE KING]

So as of now, if a TWS track is lost after an Aim-54 is fired and before it goes active, the missile will go stupid and continue to fly in a straight line. I remember reading a post addressing this issue a while back but can’t seem to find it. Is this a guidance bug or is this how it is supposed to behave? I would think the missile would guide onto the targets last calculated position but it does not.

[near_blind]

Without doing the long winded explanation again, yes. The AIM-54 needs a constant TWS track until it goes active.

[THE KING]

Here we go.

My interpretation of this is that the missile will guide onto the targets last known location and go active at a certain time. In the sim as of now, I do not observe this behavior. Missile will just keep flying on a straight line. My apologies if I interpreted this incorrectly. 

Edited December 20, 2020 by THE KING

[DarkStar79]

yea and this dosnt seem to work at all atm..

[Naquaii]

It does but due to DCS engine limitations we can't point the missile itself, we can only control the seeker state. What happens is that the code checks whether the intended target is within the seeker area of where the missile would be if going for the held track and if it is it activates the missile.

 

A truer implementation would be to tell the missile exactly where to go and then activate the seeker if within the scan zone there but that is not currently possible.

 

Against a maneuvering target this is unlikely to result in a hit ofc but that is also true of the real functionality.

Edited December 20, 2020 by Naquaii

[DarkStar79]

11 minutes ago, Naquaii said:

It does but due to DCS engine limitations we can't point the missile itself, we can only control the seeker state. What happens is that the code checks whether the intended target is within the seeker area of where the missile would be if going for the held track and if it is it activates the missile.

 

A truer implementation would be to tell the missile exactly where to go and then activate the seeker if within the scan zone there but that is not currently possible.

 

Against a maneuvering target this is unlikely to result in a hit ofc but that is also true of the real functionality.

If the awg 9 looses the track, and then reaquire the track on the same target.  Will it then start to go on that target again?

11 minutes ago, Naquaii said:

 

 

[Naquaii]

No, we have no evidence of the WCS being able to correlate new tracks to old tracks so it must be the same track. If it is lost it's lost and goes for the held track.

 

This is the major weakness against fighters afaik, the missile itself does absolutely fine against them as long as it finds the target.

I've yet to find any evidence of the missile itself being particularly bad against fighters even if that is an opinion held by some. The evidence I've seen points towards the AN/AWG-9 being the weaker link of the two against small maneuvering targets.

[THE KING]

39 minutes ago, Naquaii said:

 

A truer implementation would be to tell the missile exactly where to go and then activate the seeker if within the scan zone there but that is not currently possible.

 

 

I take it this is something ED plans on addressing eventually right? As you may or may not know, all fox three missiles right now go stupid if TWS track is lost before going active instead of homing onto the targets last known calculated position.

 

10 minutes ago, Naquaii said:

No, we have no evidence of the WCS being able to correlate new tracks to old tracks so it must be the same track. If it is lost it's lost and goes for the held track.

 

 

 

I’m a bit confused by this, right now, if a TWS track is lost on a target and then re-acquired after a Phoenix launch, the missile will guide back onto the target. Is this what is supposed to happen or is this a bug? 

[GGTharos]

I would suggest that the track 'sticking' for 3 frames before it is timed out (where did I read this?  I forget) is evidence that there is TWS correlation attempts for 3 frames.   This is pretty much TWS bread and butter stuff and is itself vulnerable to things like tracks crossing/merging and probably other obvious things.  Right now it seems that in-game missing a single hit drops the track and I don't believe that's quite right.

At the same time, I don't think the AWG-9 tells the AIM-54 where to go once in flight - only the pre-launch commands set up the initial steering and seeker look angles, from there on the commands are updating seeker looking angles, doppler gate and I think at least one other radar parameter (maybe range gating?) - the missile uses this information to fly some form of PN mid-course while receiving reflections from the targets as they are scanned by the AWG-9 in TWS (thus SD/A mode, as opposed to the more modern inertial with mid-course update)  ... it sounds like the missiles in game don't hold the ability to be 'instructed' to act like this.

6 minutes ago, THE KING said:

I’m a bit confused by this, right now, if a TWS track is lost on a target and then re-acquired after a Phoenix launch, the missile will guide back onto the target. Is this what is supposed to happen or is this a bug? 

 

I would guess more like a limitation ... what the game really does is say something like, 'if the tracks on object xyx is regained, send this information to the missile launched at xyz'.   hat should happen is that the channel for the dead track is now gone/abandoned in the sense that it will only send commands based on either the 2 minute hold OR if the track is deleted, that channel is also gone, no longer used.   Basically, commands will not be transmitted on that missile's data-link channel.

[Naquaii]

49 minutes ago, THE KING said:

I take it this is something ED plans on addressing eventually right? As you may or may not know, all fox three missiles right now go stupid if TWS track is lost before going active instead of homing onto the targets last known calculated position.

 

 

I’m a bit confused by this, right now, if a TWS track is lost on a target and then re-acquired after a Phoenix launch, the missile will guide back onto the target. Is this what is supposed to happen or is this a bug? 

 

I can't talk for ED in these matters and have no insights into their plans. What we asked for and eventually received was the ability to control seeker state and loft allowing us to at least make the seeker behavior in regards to active/passive correct and also make it loft when it should and shouldn't.

 

There is currently a bug that seems to sometime allow the missile to reaquire even if the track is deleted, we're working on that one. If the track is reaquired or the target remains inside/near where the WCS extrapolates it to be it should go active as long as it's within the AWG-9 scan zone (i.e. can see the commands). That is how it should be.

[Naquaii]

47 minutes ago, GGTharos said:

I would suggest that the track 'sticking' for 3 frames before it is timed out (where did I read this?  I forget) is evidence that there is TWS correlation attempts for 3 frames.   This is pretty much TWS bread and butter stuff and is itself vulnerable to things like tracks crossing/merging and probably other obvious things.  Right now it seems that in-game missing a single hit drops the track and I don't believe that's quite right.

At the same time, I don't think the AWG-9 tells the AIM-54 where to go once in flight - only the pre-launch commands set up the initial steering and seeker look angles, from there on the commands are updating seeker looking angles, doppler gate and I think at least one other radar parameter (maybe range gating?) - the missile uses this information to fly some form of PN mid-course while receiving reflections from the targets as they are scanned by the AWG-9 in TWS (thus SD/A mode, as opposed to the more modern inertial with mid-course update)  ... it sounds like the missiles in game don't hold the ability to be 'instructed' to act like this.

 

I would guess more like a limitation ... what the game really does is say something like, 'if the tracks on object xyx is regained, send this information to the missile launched at xyz'.   hat should happen is that the channel for the dead track is now gone/abandoned in the sense that it will only send commands based on either the 2 minute hold OR if the track is deleted, that channel is also gone, no longer used.   Basically, commands will not be transmitted on that missile's data-link channel.

 

The current implementation that we have and believe is correct from our docs is that it marks a track as lost after 3 missed frames and drops it after 7. During that time it can still be correlated if it gets new radar returns fitting the parameters. We have seem evidence that this might be affected if playing with high latency as dcs aircraft tend to rubberband a lot unfortunately and have yet to figure out a good way to counteract that.

 

In regards to the missile guidance the only thing we have control over in DCS is wether we tell the missile that we're tracking it's target or not and the last few patches we added the ability to control seeker state and lofting when that was implemented. So we can't really tell the missile to fly out to a specific location, that's why the best solution we can do atm is to check wether the real location of the target related to a held track is close enough to the missile for it to be able to see it and then if so activate the seeker.

[Sideburns]

I was about to comment on player latency / connection quality as a contributing factor to this perceived issue, I suspect other modules with high fidelity radar modelling also suffer to some extent from player rubber banding / warping around. It would be handy if ED could introduce "connection quality" tools for server operators to allow warning and kicking of clients that drop many packets or have poor latency.

[near_blind]

49 minutes ago, Naquaii said:

The current implementation that we have and believe is correct from our docs is that it marks a track as lost after 3 missed frames and drops it after 7. During that time it can still be correlated if it gets new radar returns fitting the parameters. We have seem evidence that this might be affected if playing with high latency as dcs aircraft tend to rubberband a lot unfortunately and have yet to figure out a good way to counteract that.

 

Is this code live in the open beta? I have never seen the AWG-9 correlate a lost track back to a healthy state. 

[Naquaii]

6 minutes ago, near_blind said:

 

Is this code live in the open beta? I have never seen the AWG-9 correlate a lost track back to a healthy state. 

 

It is but for it to correlate back to the same track the target has to have basically flown more or less as how the WCS calculates the target to fly, otherwise it will be outside the parameters for the correlation and those are modelled to the best of our ability from real data. Realistically it will only happen after a few missed hits (like 1-3) and far from everytime. Like mentioned before this TWS is quite old and not really close to a more modern one like the one in the Viper or Hornet.

[near_blind]

1 minute ago, Naquaii said:

It is but for it to correlate back to the same track the target has to have basically flown more or less as how the WCS calculates the target to fly, otherwise it will be outside the parameters for the correlation and those are modelled to the best of our ability from real data. Realistically it will only happen after a few missed hits (like 1-3) and far from everytime. Like mentioned before this TWS is quite old and not really close to a more modern one like the one in the Viper or Hornet.

 

How does Track Hold affect the implementation of the correlation logic?

Reflecting back the radar is pretty resilient since the implementation of TWS-A in April when Tack Hold is off, but when it gets enabled the radar seems far less capable of correlating. The radar seems less prone to dropping the track since the November patch, but once its considered lost, it seems like it is totally gone, even if the radar creates a new track right on top of the old. 

 

I've described this to a former RIO who asked to remain anonymous, and their opinion was the radar should be capable of correlating in that situation. 

[Naquaii]

The only thing the Track Hold really does is change the time from last return to dropping the track. If a new hit occurs that could correlate it should work as it does without track hold. The problem is that the new radar track/return has to be within a certain range and speed gate of the current trackfile to correlate and which each cycle that likelyhood becomes less and less. Against a maneuvering target like a fighter this is quite unlikely because if the track has changed direction it is likely to be outside of those parameters.

[Sonoda Umi]

 

 

10 hours ago, Naquaii said:

The only thing the Track Hold really does is change the time from last return to dropping the track. If a new hit occurs that could correlate it should work as it does without track hold. The problem is that the new radar track/return has to be within a certain range and speed gate of the current trackfile to correlate and which each cycle that likelyhood becomes less and less. Against a maneuvering target like a fighter this is quite unlikely because if the track has changed direction it is likely to be outside of those parameters.

That is to say, the effective way to against a maneuvering fighter is closer range attack such as go close within appx. 18nm and make a "missile directly go active attack"

 

On the other word, if it wants hit a fighter under TWS mode, it should make sure that the target is not aware of the AIM-54 launched at a long distance. However, it seems not handful to attack DCS AI fighters - they can detect the missile once it lanunched and go defence maneuver immediately.

[Naquaii]

2 hours ago, Sonoda Umi said:

 

 

That is to say, the effective way to against a maneuvering fighter is closer range attack such as go close within appx. 18nm and make a "missile directly go active attack"

 

On the other word, if it wants hit a fighter under TWS mode, it should make sure that the target is not aware of the AIM-54 launched at a long distance. However, it seems not handful to attack DCS AI fighters - they can detect the missile once it lanunched and go defence maneuver immediately.

 

Unfortunately DCS AI behavior is completely out of our hands, we're certainly not giving them any sort of warning in our code.

[Sonoda Umi]

15 minutes ago, Naquaii said:

 

Unfortunately DCS AI behavior is completely out of our hands, we're certainly not giving them any sort of warning in our code.

Above all, it is only useful manner that attack DCS AI fighters to get close within 18nm and then fire and forget.

[Naquaii]

7 minutes ago, Sonoda Umi said:

Above all, it is only useful manner that attack DCS AI fighters to get close within 18nm and then fire and forget.

 

Like I said, this is not due to us modelling our system wrong, that's due to the standard DCS AI behavior defending before they should know that a missile is incoming. Not much we can do about that unfortunately.

[Sonoda Umi]

5 minutes ago, Naquaii said:

 

Like I said, this is not due to us modelling our system wrong, that's due to the standard DCS AI behavior defending before they should know that a missile is incoming. Not much we can do about that unfortunately.

I agree.

[THE KING]

On 12/20/2020 at 5:48 PM, Naquaii said:

If the track is reaquired or the target remains inside/near where the WCS extrapolates it to be it should go active as long as it's within the AWG-9 scan zone (i.e. can see the commands). That is how it should be.

So does this mean that the missile should guide onto the “lost”, track and go active when it’s a certain distance away?

Edited December 26, 2020 by THE KING

[Naquaii]

4 hours ago, THE KING said:

So does this mean that the missile should guide onto the “lost”, track and go active when it’s a certain distance away?

 

 

As it is now the missile continues to fly at the lost track and when at TTI we check if the correct track is still close enough that the missile would find it when going active. If it is the missile goes active and starts homing and if it is not it will stay dead. This is due to the fact that the current missile modelling does not allow us to point the missile at a point in space and just activate the seeker to find whatever it is in front of it.

 

This way at least the missile functions closer to reality in that it will still need the AWG-9 to tell it to go active making it less of a fire and forget missile. The alternative would be to leave it functioning as other ARH missiles in DCS in that it would go active on its own which it couldn't do IRL.

Edited December 26, 2020 by Naquaii

[Kula66]

@Naquaii Is there any difference between the 54A & C in terms of seeker modelling???

 

[QuiGon]

17 minutes ago, Kula66 said:

@Naquaii Is there any difference between the 54A & C in terms of seeker modelling???

 

The C has a higher chaff resistance AFAIK.