TWS-Auto

[gyrovague]

Track While Scan - Auto

 

As most know, we had a TWS-Auto implementation for the F-14 at launch, but it was quite unreliable and ended up causing more trouble than it was worth, so it was disabled. After launch of course we were kept quite busy, and it was some time before we managed to revisit and eventually rewrite the TWS-Auto system. We’re finally ready to re-release it now, and this post will summarize the functionality.

 

As you may remember, both TWS modes require the scan frame to be 2 seconds, and only support 2-bar ±40° or 4-bar ±20° scan volume modes (the radar gimbal moves at 80°/s, so you can see where the 2 second frame comes from in these two scan volume options). Very simply put, TWS-Auto attempts to keep the TWS pattern pointing where it is most useful in azimuth and elevation, while also controlling the optimum scan volume. TWS (and all the PD radar modes in F-14) uses ground stabilized (a.k.a. Earth fixed) patterns, so the azimuth and elevation angles are relative to a tangential plane on the surface of the earth at your present position, with azimuth further being relative to the F-14 longitudinal axis too (i.e. 0° is directly ahead on the imaginary plane, -20° is to the left, etc.).

 

Function of centroid

 

The azimuth and elevation angles are determined by a weighted centroid of targets in the scan volume. What this means is that some targets contribute more to the centroid than others, depending on a few factors outlined below. Essentially the position of each target is multiplied by its weight factor divided by the total weight of all targets, and these weighted positions are added together to give a centroid. Only sensor targets (i.e. own radar) are considered, not datalink targets.

 

Two separate (but somewhat related) centroids are calculated as part of the TWS-Auto update procedure: a steering centroid, and an illumination centroid. The former facilitates steering cues (on HUD, VDI, TID, DDD) to help the pilot maintain optimum target coverage, and also displays a small X on the TID indicating the steering centroid position. The illumination centroid controls the azimuth and elevation of the scan pattern center by using the angles from the aircraft to the computed illumination centroid (in TWS-Manual, these are directly controlled by the radar azimuth and elevation knobs on the RIO sensor control panel). The velocities (change in position over time) of both centroids are also calculated, used to calculate steering, and dead reckoning for a short period when all targets are lost (in an attempt to re-acquire them at their expected positions).

 

Weighting targets

 

As mentioned, a few factors influence target weighting. These differ slightly between the calculations for the steering centroid and the illumination centroid, but largely consist of the presence of a launch zone (targets marked as friendly by RIO will never get a launch zone), whether a missile is already underway to a target, and some RIO selections on targets such as DO NOT ATTK (disregards from weights completely, shows a small vertical bar over a target) and MAND ATTK (forces evaluation and raises importance, only one target may be selected for this, and mutually exclusive with DO NOT ATTK. Shows a small horizontal bar over a target).

 

Targets that are not displayed on the TID screen in the currently selected TID mode and range are also disregarded completely. In addition, targets which are deemed to be leaving the scannable volume are also raised in importance, if they are already deemed important by the previous criteria. For the steering centroid, the radar gimbal limits (maximum extents, basically ±65° in both azimuth and elevation) are considered, while for the illumination centroid, the current scan volume limits are considered (i.e. either 2-bar ±40° or 4-bar ±20°). This causes the illumination centroid to adjust towards important targets leaving the current pattern, and allows the steering centroid to shift towards targets that are leaving the maximum radar gimbal limits (i.e. pilot would need to steer towards that to keep them illuminated).

 

Engaging TWS-Auto

 

When TWS-Auto is first engaged, there is a period of 8 seconds where the manual controls for azimuth, elevation and the scan volume pattern are still in effect and can still be controlled by the RIO. After this period, the computer takes over. The scan volume pattern (2-bar vs 4-bar) is re-evaluated every 4 seconds, while the centroids are re-evaluated multiple times per second. The weights are updated at the end of each 2 second scan frame. The scan volume algorithm considers the future positions of all targets, and selects between which of the two options would give a greater total illumination weight. If they are equal, 4-bar ±20° is selected.

 

Firing an AIM-54 Phoenix while in TWS-Manual will result in automatic selection of TWS-Auto. Furthermore, when any AIM-54 missiles are deemed to still be in flight (up to 16s beyond their expected time to impact), TWS-Manual cannot be entered (the button press is ignored). Target tracks that have a missile launched at them will also behave as if Track Hold (the button to the left of the TID fishbowl) has been selected, i.e. they will continue to be extrapolated for up to 2 minutes if their radar contact is lost.

 

Symbology

 

The steering cues on HUD, VDI, TID and DDD will direct the pilot horizontally only, based on the steering centroid. By default, lead collision steering is employed. If the RIO uses the CLSN button next to the TID fishbowl, pure collision steering is instead calculated. Since both the pilot and RIO can see steering cues on their respective displays, coordination of required maneuvering is made easier. The steering cue on HUD and VDI is an upside-down T (hence called a steering tee), while on the TID and DDD the steering cue is a small square. These four displays have different steering sensitivities: TID is 40°/inch, DDD is 128.5°/inch, VDI is 25°/inch and HUD is 26.5°/inch.

 

In conclusion

 

TWS-Auto can be a handy tool for maintaining illumination on targets under attack, even while maneuvering about 60° away from the direct path to the targets. However, as pointed out by one of our RIO SMEs, good RIOs did not make exclusive use of TWS-Auto while hunting for new contacts, as it limits your overall situational awareness. RWS is still useful for scanning huge volumes, and TWS-M is still useful while maintaining the ability to control the scan volume angles. The Track Hold function (button to the left of the TID fishbowl) is also quite important, to maintain tracking (by extrapolation) on targets that would otherwise be lost in the PD filters.

 

Take a first look at TWS-A together with Jabbers here:

 

 

 

Edited November 9, 2022 by IronMike

[TLTeo]

Hype! Really looking forward to this :)

[DD_Fenrir]

Amazing work Gyro - and (as with many things in avionics) a great deal more complex than the concept would initially appear to the layman. I can well appreciate that simulating this feature was no simple task and that this took some time to perfect. Thank you for your efforts to get it to such a faithful representation.

[Sam]

Incredible work Gyro !

 

One question : If only sensored targets are considered, is it possible to use TWS-M to point the antenna on a target that we only see on datalink and then switch back to TWS-A to track it automatically ?

Edited April 4, 2020 by Panther 976

[TOViper]

I can't wait looking deeper into the F-14 ... what an amazing bird.

Think the next weeks will be very busy for me ... as they were/are/will be for HB!

What an amazing digital representation of a machine, I love your work. :thumbup:

 

Congrats guys, you really rock!

 

No questions from my side, but I really hope that Jester will then know what he has to do with all this ... since finding a real RIO might become a tricky part because the systems become more and more complex ... :noexpression:

[Kula66]

Looks great ... how much of this and how will it be controllable for us single seat flyers?

Edited April 4, 2020 by Kula66

[Alphabet_Ghost]

Any news about manual update? Thanks!

great work!

[QuiGon]

Amazing details! I'm really curious how TWS-A will perform against my own TWS-M skills, that I developed while flying as a RIO for more than a year now. :joystick:

[MBot]

Fantastic update. The first thing I am going to try with TWS-Auto is intercepting multiple incoming AS-4 Kitchen at 80'000 ft. So far this has been extremely difficult with manual elevation control.

 

Having read the piece multiple times now, I still have troubles understanding the concept of the steering centroid. Let's analyze the following picture:

 

 

Why is the steering centroid to the left and what would be the suggested action?

 

I think I understand the weighting of the tracks. The friendly track has no weight. The unknown track is marked as Do Not Attack and has no weight either. Of the two hostile tracks, the left-hand one is marked as mandatory attack and is about to leave the scan volume limits, therefore it is weighted higher.

 

Now here is what I assume will happen: The scan volume is going to shift left in order to keep the higher weighted track within its limits, until right-hand track falls outside the scan volume. When the right hand track is no longer updated and is dropped, the left hostile remains the only track with any weight and the scan volume will move further left to center it (illumination centroid).

 

But why is the steering centroid offset to the left? No maneuvering is going to prevent the loss of a track when two tracks are more than 40° from each other and no maneuvering is required to keep tracking the left-hand hostile, as it is well within the radar gimbal limits.

[escaner]

These articles are very nice, I hope they eventually make it to the manual.

[Scrape]

Long awaited goodness inbound. Like smelling cookie baking in the oven. You know it'll be great and it's only a matter of time.

[QuiGon]

@MBot: Great questions (as always)! :thumbup:

 

These articles are very nice, I hope they eventually make it to the manual.

+1 :thumbup:

[Uxi]

Awesome. Can't wait!

[gyrovague]

Incredible work Gyro !

 

One question : If only sensored targets are considered, is it possible to use TWS-M to point the antenna on a target that we only see on datalink and then switch back to TWS-A to track it automatically ?

 

As long as you see it on own radar, it doesn't matter whether you were guided there by datalink or some other reason.

[LazyBoot]

Quick question...

You mention DO NOT ATTK and MAND ATTK, but I can only find "DO NOT ATTK" on the CAP... How do we set something as MAND ATTK?

[Naquaii]

Fantastic update. The first thing I am going to try with TWS-Auto is intercepting multiple incoming AS-4 Kitchen at 80'000 ft. So far this has been extremely difficult with manual elevation control.

 

Having read the piece multiple times now, I still have troubles understanding the concept of the steering centroid. Let's analyze the following picture:

 

 

Why is the steering centroid to the left and what would be the suggested action?

 

I think I understand the weighting of the tracks. The friendly track has no weight. The unknown track is marked as Do Not Attack and has no weight either. Of the two hostile tracks, the left-hand one is marked as mandatory attack and is about to leave the scan volume limits, therefore it is weighted higher.

 

Now here is what I assume will happen: The scan volume is going to shift left in order to keep the higher weighted track within its limits, until right-hand track falls outside the scan volume. When the right hand track is no longer updated and is dropped, the left hostile remains the only track with any weight and the scan volume will move further left to center it (illumination centroid).

 

But why is the steering centroid offset to the left? No maneuvering is going to prevent the loss of a track when two tracks are more than 40° from each other and no maneuvering is required to keep tracking the left-hand hostile, as it is well within the radar gimbal limits.

 

The visible steering centroid (the X) is as you might have realised the steering command for the pilot. And in this case it is to the left due to the fact that the left target has a greater weight due to it being further from the center than the other DLZ track. This might change soon after the instant in the image if the WCS has to choose between the two which would likely result in it chosing the higher prio track, the number 1.

 

This it will do even if it is likely that one of the tracks will have to be dropped due to being too far apart as it is still possible for them to converge again. In short, the WCS will try to track both until the moment it will have to drop one of them, not before.

 

And in any case, the purpose of the visible X (steering centroid) is to help the pilot point the aircraft in the most optimum heading while the other centrum (controlling scan volume) is only visible in how the scan volume is pointed and it's purpose is to point the radar antenna in the most optimum way possible.

 

Hope that helps, might have went of on a tangent there... :-)

 

Quick question...

You mention DO NOT ATTK and MAND ATTK, but I can only find "DO NOT ATTK" on the CAP... How do we set something as MAND ATTK?

 

MAND ATTK will be added to the CAP when TWS-A is released.

[Belphe]

Great work! Can't wait for this to be released!

[amalahama]

Next OB maybe?

[gyrovague]

Quick question...

You mention DO NOT ATTK and MAND ATTK, but I can only find "DO NOT ATTK" on the CAP... How do we set something as MAND ATTK?

 

Well spotted :detective: We replaced the GND MAP button above DO NOT ATTK with MAND ATTK. Those labels changed quite a bit across F-14 blocks over the years. IRL the GND MAP feature changed radar to a shorter pulse length (IIRC) and some other presets, we haven't used that button anyway.

[MBot]

Thanks for taking the time Naquaii to reply. I am sorry for being slow, but I still don't quite understand what the steering centroid is good for. Hopefully clarifying this now might help to find an easy to understand explanation for the manual.

 

You mention:

the purpose of the visible X (steering centroid) is to help the pilot point the aircraft in the most optimum heading

 

And the article says:

Two separate (but somewhat related) centroids are calculated as part of the TWS-Auto update procedure: a steering centroid, and an illumination centroid. The former facilitates steering cues (on HUD, VDI, TID, DDD) to help the pilot maintain optimum target coverage, and also displays a small X on the TID indicating the steering centroid position.

 

What I don't understand is, as long as all tracks are within the +/-65° gimbal limits, my steering of the aircraft will have no influence on my radar's ability to maintain optimal target coverage. If I turn 30° right, the radar will simply gimbal 30° left. Target coverage by the radar will still be optimal as long as I don't reach the gimbal limits. So what is the criteria for the optimal heading?

 

Just to be clear on this, the steering centroid is NOT displaying the lead collision steering (average of all tracks), right? As I read the first post, lead collision steering (or pure if selected) is displayed on the TID by the steering cue (little square), so this is a separate thing from the X of the steering centroid, correct?

[TLTeo]

If I understand it correctly - the tracks also have a heading. If you're tracking something that's flying to your left and it's barely within gimbal limits, and then you crank right 30 degrees for example, you will lose that track.

 

 

In the screenshot, #2 is heading to the left while #1 is heading straight for the F14. As a result, the steering centroid also suggest pointing left slightly

[*Rage*]

Nice work, looking forward to flying with it.

[Naquaii]

Thanks for taking the time Naquaii to reply. I am sorry for being slow, but I still don't quite understand what the steering centroid is good for. Hopefully clarifying this now might help to find an easy to understand explanation for the manual.

 

You mention:

 

 

And the article says:

 

 

What I don't understand is, as long as all tracks are within the +/-65° gimbal limits, my steering of the aircraft will have no influence on my radar's ability to maintain optimal target coverage. If I turn 30° right, the radar will simply gimbal 30° left. Target coverage by the radar will still be optimal as long as I don't reach the gimbal limits. So what is the criteria for the optimal heading?

 

Just to be clear on this, the steering centroid is NOT displaying the lead collision steering (average of all tracks), right? As I read the first post, lead collision steering (or pure if selected) is displayed on the TID by the steering cue (little square), so this is a separate thing from the X of the steering centroid, correct?

 

If I understand it correctly - the tracks also have a heading. If you're tracking something that's flying to your left and it's barely within gimbal limits, and then you crank right 30 degrees for example, you will lose that track.

 

 

In the screenshot, #2 is heading to the left while #1 is heading straight for the F14. As a result, the steering centroid also suggest pointing left slightly

 

Nice work, looking forward to flying with it.

 

The second quote is about right.

 

You can think of the steering centroid as a way of helping the pilot maximising his chances of keeping the tracks within the gimbal limits in the future, if the tracks are heading left it will point left as the tracks will eventually move outside of the limits to the left. And thus by telling the pilot to turn left it will mitigate that.

 

It's not really a correlated steering cue for all tracks combined, it really is a steering cue towards helping the pilot keeping the tracks within the gimbal limits.

[MBot]

I am sorry for being an idiot, but I simply not get it. Am I the only one that doesn't understand it?

 

In this situation, steering left will do nothing to help track both targets. One will be lost when they get more than 40° from each other. Steering left will also not help to continue tracking just the left-hand target, as it is nowhere near the radar gimbal limit. The radar will track it just fine for quite some time to come without any steering being necessary. What is the useful information the X is transporting here?

 

[captain_dalan]

I am sorry for being an idiot, but I simply not get it. Am I the only one that doesn't understand it?

 

In this situation, steering left will do nothing to help track both targets. One will be lost when they get more than 40° from each other. Steering left will also not help to continue tracking just the left-hand target, as it is nowhere near the radar gimbal limit. The radar will track it just fine for quite some time to come without any steering being necessary. What is the useful information the X is transporting here?

 

 

My interpretation:

Both 1 and 2 are co-altitude (roughly). 2 is about to leave volume limits in azimuth. The FCC will switch to 40 degree 2 BAR to keep it inside the scan volume, but it will still move away from the F-14 faster then 1 will (it's on a direct intercept course). By pointing your nose slightly left, you will keep both in the scan zone longer.