Question about lead angle [Archive] - STEELBEASTS.COM FORUM

Lansfa

09-28-2006, 01:49 AM

There're different procedures to take lead angle on a modern tank, for example, on M60A3 the lead angle is taken by chasing a moving target to measure the angular velocity of the turret, and after pressing the lase/lead button the reticle would move according the angular velocity. Some tanks use the laser rangefinder to measure the lead angle by illumanating a moving target for a while. I guess this is an application of Doppler effect but I'm not so confident, since the Doppler effect can tell if the target is approaching or leaving, not its traverse movement.

So would someone please give me a hint about the principle of taking lead angle by lasing a target? And which method do those tanks in SB Pro take, laser measurement, angular velocity measurement, or relative motion of sight mirror with an index point? Thanks a lot.

3Star

09-28-2006, 02:54 AM

I'm not sure what tanks would use a continuous illumination to determine target motion: It seems to me that a continuous burst would be the sort of thing to burn out a tank-mounted laser pretty quickly.

Pretty much all the tanks in SB use the angular turret speed to determine lead, all but the Leopard 1AS automatically apply lead as soon as a range is entered, you can pull the trigger immediately. The Leopard 1AS has a slightly older system which doesn't apply lead to the gun, but reverse lead to the sight: The sight crosshair will jump behind the target, you then must 'catch up' to the target in order to have the correct lead, then pull the trigger.

NTM

NEpi

09-28-2006, 07:20 AM

Dopler effect occurs when a wave bounces of a surface that moves in the direction of the origin of the wave (or away of the origin). movement perpendicular to the origin won't result in a Dopler effect. so it can help lead calculation.
usually, the lasing only signals the FCS that you've found a target, and it should start calculating lead. the FCS does so by calculating average rotation speed over a short time and by the range set by the LRF.

Lansfa

10-09-2006, 07:29 PM

Thanks for your answers.

I know there's a sensor on M60A3 which is used to measure the rotation speed of turret, then the lead angle is given by moving the reticle backward solely according to it. So, is there a corresponding sensor on Abrams and Leo 2A4/A5?

112TYR

10-10-2006, 02:57 PM

Pretty much all the tanks in SB use the angular turret speed to determine lead, all but the Leopard 1AS automatically apply lead as soon as a range is entered, you can pull the trigger immediately.
NTM

The Leo1A5 and Leo2 needs 1,5 seconds to calculate correct lead. It does use angular turret speed to determine correct lead, but the system needs 1,5 secs from you start tracking the target until it has calculated a correct lead value, regardless of when you lase. Pull the trigger immideately, and your round will fall behind the target.

Hase

10-10-2006, 05:43 PM

The Leo1A5 and Leo2 needs 1,5 seconds to calculate correct lead. It does use angular turret speed to determine correct lead, but the system needs 1,5 secs from you start tracking the target until it has calculated a correct lead value, regardless of when you lase. Pull the trigger immideately, and your round will fall behind the target.

Hi all,

the 1,5 seconds thing is controversial.
In 1993 on an Leo2A4 we were taught to track for at least 1,5 second afaik. A calculation of an average turret traverse speed during the last second was said to happen internally, thus increasing accuracy. But we were taught a lot of stuff of doubtful validity... Might be BS.

In SB1 and ProPE it seems to work different: No gliding average over 1,5s, just the momentary turret traverse speed is used.

IMHO the Leos in SB clearly show that this approach is sufficient and faster. No need for averaging.

In both cases the turret traverse rate is used together with the average speed of the projectile for the range (measured or entered manually) taken from a look-up table to calculate the lead angle.

Different story:
From an engineers point of view I speculate (!) that no tank in the world uses it´s LRF with or without Doppler effect for calculating the lead angle.

There is a sensor for turret traverse speed in the Leo´s turret drive.

Has someone TRUE knowledge about the 1,5s theory in Leos?

Regards
Hase

RecceDG

10-10-2006, 06:41 PM

I don't know *specifically* about the Leo1, but I know that generally, when you need to know an angular velocity on something that needs to freewheel, you use some sort of pulsed sensor, like a magnetic reluctor or an optical interruptor.

You see these on the wheel speed sensors used with ABS brakes, vehicle speed sensors, crank and cam angle sensors in engines, etc.

These sensors - depending on the pulse frequency - tend to be inaccurate until they've measured a couple of pulses. One pulse doesn't really tell you anything; you need a couple of pulses to get an accurate measurement.

Then, depending on the pulse frequency per degree of traverse, there is a limit as to how much of a rate of change you can measure.

Ideally, you want to see the item being measured moving at a constant velocity for a few pulses to ensure you get an accurate measurement.

Then there is a computational lag between when the data goes into the computer and when a result is returned, and then there's a mechanical lag while either the sight mirror or the gun tube is slewed into the correct position. It's a good idea to try and keep everything stable for a length of time while all the various subsystems settle out.

If you fire before everything has stabilized, the result will depend on how close the current values/positions are to ideal, and how large/close the target is. A fast moving target at extreme range is going to be very sensitive to small errors, where a close target point-blank won't matter.

So I don't think it's a case of where you *must* wait 1.5 seconds, but rather, that a delay of 1.5 seconds while you hold everything as stable as possible maximizes the chance that the firing solution is correct and the gun tube is on line with the firing solution.

As far as the actual calculations go, you need an angular rate of traverse, a distance to target, and the ballistics for the round currently loaded, and you've got enough information for a firing solution. As such, the laser is only needed to get range; you aren't trying to use dopplar or anything to get a motion vector.

For an even better chance to hit, you need a muzzle reference sensor (to better fix which way the gun is pointed; it's more accurate than an angular sensor at the trunnions) a wind sensor (to compensate for windage) and possibly a tube temp sensor or maybe a muzzle velocity sensor to try and fine-tune the ballistics of the round - or maybe you programatically degrade expected tube performance based on empirical or statistical data and an internal counter.

DG

112TYR

10-10-2006, 10:00 PM

Has someone TRUE knowledge about the 1,5s theory in Leos?

Regards
Hase

I tried this in a live Leo1A5. Integrator on, Stab Ein, traverse continuous speed, engage dynamic lead. Read the lead-value from the computer. The value gradually increased over 1,5 secs, then stabilised. I talked to a danish tanker about this on this forum and via mail, he tried the same thing, same results.

3Star

10-10-2006, 10:32 PM

112TYR

10-10-2006, 10:44 PM

I seem to recall that last time this came up, we had people trying it out on Danish and German Leopards, and the results were inconclusive. (different).

I'll see if I can track down the thread, but it was some time ago.

Got to admit, there aren't many sims when people will hop into the real thing to test something.

NTM

As I've said, I tried it in the Leo1a5, and the results were pretty obvious; the system DID need 1,5 sec to provide a stable lead value. IIRC, Johnny said his danish tanks did the same thing.

By the way, this is the original thread... (http://www.steelbeasts.com/sbforums/showthread.php?t=5504&page=6)

3Star

10-10-2006, 10:52 PM

112TYR

10-10-2006, 11:04 PM

Found it.

http://www.steelbeasts.com/sbforums/showthread.php?t=5504&highlight=Lead+Calculatoin

I guess the definitive answer is the very last post in the thread, when they asked Rheinmetall.

A 'good enough' lead is applied almost instantly, anything beyond that is just 'fine tuning' and takes about a second and a half.

[edit: I see that I was beaten on the thread link]

NTM

Ok, I know I'm nitpicking here, but the way I read that last post was that the system had calculated a "good enough" lead after the first second, the following 0,7 seconds are fine tuning - giving you a "perfect" solution after 1,7 secs.

This is apparantly because:

The computer will compute the lead in fractions of second - but there is a programmed "gateway" of 1 sec, to make sure you have a steady lead and thereby the best possible solution.

This corresponds well with my own observations, the difference in lead values was definitely bigger than "fractions of a mil" during the first second...enough to miss a tank-sized target, at least.

3Star

10-11-2006, 12:48 AM

112TYR

10-11-2006, 08:20 AM

Here's a different one, then: As the poster indicates that the gunner spends about that one second tracking and lasing, just where in the process does that 1.7 seconds start? i.e. if you were to have your computer showing inputs as you lase, track, and then hit 'dynamic lead', at what point would the numbers start changing? After the 'lase' or after the 'lead'? If the former, i.e. if the range input is the trigger to start calculating (which seems reasonable: Rucksterung in the Leo 2 at least requires calculation without the use of the Dynamic Lead), then the practical difference is as Johnny puts it: Very little at all, you can pull the trigger pretty quickly after hitting 'dynamic lead' unless you hit 'dynamic lead' particularly quickly after entering the range.

NTM

The process starts after you hit dynamic lead, and will measure the AVERAGE speed over 1 second. Rucksteurung doesn't enter into it, that's 2 different equations. The unknown factor with moving targets are target speed and range. With rucksteurung, the speed is known(since the speed is the speed of your own vehicle), so you don't need to track the target. Dynamic lead use 1 second to determine target speed(or turret angular speed, to be more presise), and dynamic lead is based on this. So it won't matter when you lase, the system will still have an average turret angular speed measured over 1 second from which to measure target speed. Any tracking errors you make during this second will inflict the shot.

During my testing, all other factors were eliminated, as I used the TC integrator switch and had a manual range entry right from the start. I could probably try to film the results again, if you're interested. (It can only be done this accurate on the Leo1A5, as the Leo2 lacks dynamic lead for the TC...)

3Star

10-11-2006, 10:29 PM

While Rucksteurung and D.L. are different concepts, I'm having a little difficulty getting my head around the concept that Rstg doesn't require input from the turret. Any time either the shooter or target is moving, lead is required, and for that you need a relative velocity vector and a distance, not just speed and range. Put another way: if you're doing 30mph right angles to a stationary target, you need to apply a different amount of lead than if you're doing 30mph at a 30 degree angle to that same stationary target.

For Rstg (or whatever the abbreviation is), the LRF will tell you how far away the target is, the speedometer (or whatever it is) will tell you how fast you are going. Surely the only thing that gives a vector component to the Rstg calculation is the direction the gun is facing relative to the hull: The same sensors that are used to determine angular lead when dynamic lead is pressed.

NTM

112TYR

10-11-2006, 10:45 PM

For Rstg (or whatever the abbreviation is), the LRF will tell you how far away the target is, the speedometer (or whatever it is) will tell you how fast you are going. Surely the only thing that gives a vector component to the Rstg calculation is the direction the gun is facing relative to the hull: The same sensors that are used to determine angular lead when dynamic lead is pressed.

NTM

None of these factors needs to be calculated as an "average" value. When you use dynamic lead, target speed is calculated as an average value over 1 second, to get a correct estimate. With rucksteurung, you know the relative speed, since that speed is the speed of your own vehicle. All of the other factors can be read directly from components within your own tank. That's why you don't get the same delay with Rucksteurung. All factors are calculated instantly by the computer.

3Star

10-11-2006, 11:03 PM

112TYR

10-11-2006, 11:15 PM

What if you're using Rstg and the driver is performing a turn? The angle / vector portion of the calculation is constantly changing. For the turret sensors not to get a continuous and average sensing would result in Rstg being inaccurate/non-functional as long as the hull is rotating in relation to the turret.

NTM

Well, if the system is smart enough to calculate turret angular movement from the gunners handles, and not just the sensors, that shouldn't be a problem. Regardless, dynamic lead needs to measure "average" speed, whereas rucksteurung knows the exact speed all along. That's the essential difference that causes rucksteurung to be instant, while DL has a delay

3Star

10-12-2006, 12:40 AM

No, I'm not letting go of this one. It's not enough for me to know that something works the way it does, I need to know why it works the way it does. (It's why I hate cars with automatic transmissions)

Right now, I'm still stuck on the fact that the only mechanism that tells the Rucksteurung computer where the target is in relation to the hull are the exact same sensors that tell the dynamic lead calculator what the target is doing compared to the hull, and that as the Leopard moves and turns, the effects on the fire control solution are exactly the same as if the Leopard was stationary and the target was moving at the same speed in the opposite direction.

NTM

112TYR

10-12-2006, 08:38 AM

Ok, I'll try to explain it in a different way. The system is fully capable of calculating and applying dynamic lead within fractions of a second, just as it does with rucksteurung. It just doesn't do that, but that's not because of limitations within the system:

In a stationary tank - moving target scenario, the system needs to know:

- Range(given by LRF)
- Target speed(measured by average turret movement)

in order to calculate the correct amount of lead.(I'll exlude factors such as indexing ammo, temperatures, wind, and so on)

In a moving tank - stationary target scenario, the system needs to know:

- Range(given by LRF)
- Own speed(provided by the hull sensor)
- Turret posistion(provided by TPS)

The difference between these two methods, is that in the last, the system provides all the variables, whereas the gunner is the "sensor" calculating speed in the first one, and the system has no way of knowing how accurate/correct his inputs are. In order to get the input as correct as possible, the system uses an average lead value measured over 1 second, instead of just one, near instant, measuring. This reduces the impact of any small errors the gunner might make during the tracking.(Such as a slight handle movement as he pulls the trigger) After one second, the system applies the average lead, which is fine tuned over the next 0,7 seconds. After that, the average turret angular movement over the last 1,7 seconds will determine the lead - regardless of how long you have been pressing the dynamic lead-button. (At least this is how I THINK it works.)

NEpi

10-12-2006, 01:25 PM

If I understand both of you correctly, you're talking abuot two different systems.
dynamic lead is what you should use when firing, since most of your targets will move. that's the american philosophy, anyway. you can reduce problems with tracking by using more sophisticated "stabilization": your gun will point to your (stationary) target whatever your driver does. how does this work?
the FCS (or stab. computer, whatever) computes your angular movement from the last spot you lased. it's done by simple trigonometry.
so if your target is stationary, you will stay on target, unless you intentionally move. it might seem like a useless feature, but it is not, since the gunner doesn't need to see the target at all. that means that if you're going behind cover and reemerge on the other side, your sights are still spot-on. very cool feature indeed.
it can go a little off-target if:
1. target moves (obviously)
2. tank fires (recoil moves the tank a little, but enough to take you off target).
3. tank being hit (impact does that)
4. tank slips. if terrain isn't holding well enough, the tank will wrongfully gauge speed, unless it is using gyros for that, which I doubt.

Any tracking errors you make during this second will inflict the shot.

not exactly. since your correcting motion counters your mistake in the first place, the average is still good. the only problem is what if the gunner blunders right before firing the shot. in that case, he should get a boot in the head, since he fired poorly.

112TYR

10-12-2006, 01:52 PM

dynamic lead is what you should use when firing, since most of your targets will move.

Eh...you press dynamic lead when you engage a moving target. If the target doesn't move, you don't.

not exactly. since your correcting motion counters your mistake in the first place, the average is still good. the only problem is what if the gunner blunders right before firing the shot. in that case, he should get a boot in the head, since he fired poorly.

Well, that's only assuming the error is corrected through your own tracking.
If you make a tracking error, such as "catching up" to the target within that second, the system will generate too much lead, and you will miss. If you catch up to the target, then stay on for 1,5-1,7 seconds, the lead will be correct, and you'll hit.

NEpi

10-12-2006, 02:33 PM

you misread me on the first quote. I said it is one way to look at it. afaik, in M1 the dynamic lead is calculated once you lase, hence always.

If you make a tracking error, such as "catching up" to the target within that second, the system will generate too much lead, and you will miss. If you catch up to the target, then stay on for 1,5-1,7 seconds, the lead will be correct, and you'll hit.

let's have an extreme case: I pressed DL and not tracked, then managed to catch up in 0.1sec and fired. I traversed the same angle I would have if I have done a better tracking, which is the angular distance the target moved. since we're talking average angular velocity, it's angle/time. the ratio is still correct.
so, theoretically, catching up is just as good, if it is accurate.
the problem is catching up is not so accurate in reality, and since we assumed I was aiming well in the first place, which is hard to determine when I have to catch up, the lead might be erroneous.
that's why any gunner should track well first, and then press the DL, and not vice versa. having some catching up won't make you miss, if you're good at that, and if you're taking the few more fractions of a second making sure you haven't overcompensated.

112TYR

10-12-2006, 02:43 PM

let's have an extreme case: I pressed DL and not tracked, then managed to catch up in 0.1sec and fired. I traversed the same angle I would have if I have done a better tracking, which is the angular distance the target moved. since we're talking average angular velocity, it's angle/time. the ratio is still correct.
so, theoretically, catching up is just as good, if it is accurate.
the problem is catching up is not so accurate in reality, and since we assumed I was aiming well in the first place, which is hard to determine when I have to catch up, the lead might be erroneous.
that's why any gunner should track well first, and then press the DL, and not vice versa. having some catching up won't make you miss, if you're good at that, and if you're taking the few more fractions of a second making sure you haven't overcompensated.

I'm not sure what you mean by the "extreme" example. The system will not have a correct lead-value until you have kept the lead-button pressed for 1 second.(and that's just a "good enough" value) The average movement during that second will determine the lead. That means if you track the target perfect for 0,1 second, the lead will be incorrect and you will miss. If you catch up to the target for 0,9 seconds, and track perfectly for 0,1 second, the system will calculate too much lead, and you will miss. If you track perfectly for 1 second w/o pressing DL, then press DL and fire immideately, the lead will be incorrect and you will miss. Bottom line: You need a steady, correct tracking for at least 1 second(preferably 1,7 seconds) with DL engaged in order to generate the correct amount of lead.

NEpi

10-12-2006, 02:56 PM

112TYR

10-12-2006, 03:01 PM

by an extreme example I meant pressing the DL, then 0.9sec not tracking at all and then 0.1sec catching up to the target.
I assumed I aimed perfectly well when I pressed the DL button, and that I managed to catch-up with the target perfectly.

That is a pretty extreme case. In most (basically all) cases where you catch up to the target while tracking, the system will generate too much lead.

3Star

10-12-2006, 03:59 PM

Eh...you press dynamic lead when you engage a moving target. If the target doesn't move, you don't.

Not that pressing dynamic lead and tracking the stationary target will make you miss either. For the lead calculation, the end result is the same.

NTM

RecceDG

10-12-2006, 04:11 PM

the system will generate too much lead, and you will miss.

The statement "the system will compute too much lead" is correct, but the statement "and you will miss" depends on the amount of angular error, the width of the target, the range to the target, and where (relative to the target's width) the sight was located when the trigger was pulled.

Consider a wide target moving broadside at 500m. The sight picture "catches up" with the target (with dynamic lead engaged) and the trigger is pulled as soon as the pip hits the tail edge of the target.

The computer (correctly) calculates too much lead, but because the trigger was pulled at the tail edge of the target, and because the target is close, the round still strikes the target - just not in the same place it would have had the gunnery procedure been followed correctly.

That's not to say you SHOULD do this, but it is possible to have less than ideal lead calculations and still put steel on target.

DG

112TYR

10-12-2006, 04:13 PM

In that case, you might as well skip DL and apply manual lead.

NEpi

10-12-2006, 05:22 PM

That is a pretty extreme case. In most (basically all) cases where you catch up to the target while tracking, the system will generate too much lead.

I used an extreme case to show that catching up that will take a very short time is ok. the problem with that example is catching up seldom takes so little time, unless your tracking is not bad to begin with. the reason most of the times catching up will ruin lead computation is because it will take more than 1sec to catch-up, and since it is a abrupt and relatively violent movement you're bound to over-compensate, and find yourself aiming in front of the target.

RecceDG, even a stopped clock shows the right time twice a day. if my gunner ever fire from the hip while using lead like that, he'll feel my boot in his head.

112TYR

10-12-2006, 05:37 PM

I used an extreme case to show that catching up that will take a very short time is ok. the problem with that example is catching up seldom takes so little time,

Another problem with the example was that you DIDN'T track for 0,9 seconds - that is, you kept aiming at the spot where you first aquired your target, then swiftly moved the reticle onto the target, all while pressing DL. That's a different equation. If you catch up to a target and end up with the reticle in centre of mass, you'll still get the wrong lead calculated, regardless of how smooth or perfect your motion was.

RecceDG

10-12-2006, 05:43 PM

RecceDG, even a stopped clock shows the right time twice a day. if my gunner ever fire from the hip while using lead like that, he'll feel my boot in his head.

"Perfect" is the enemy of "Good Enough".

While we strive for perfection in practice and training, and we do what we can to innoculate gunners against stresses, unforseen circumstances, and whatnot, in the real world, Shit Happens. A nervous/excited/frightened gunner may forget to track, or track poorly, or fire when the sight picture is less than optimum, or get bumped right before shooting, or any of a million other things may happen that preclude the ideal "smoothly track the centre of exposed mass for 1.5 seconds then pull trigger".

Happily, in the real world the target is not a tiny dot; it is a great big Jesus tank several metres long. At shorter ranges especially, a less-than-perfect shot may still hit and kill the target - that's good to know, and it is good for gunners to know exactly how big than envelope is should they need to get a snapshot off.

Play the shooting gallery long enough, attempting to get the time between targets as small as possible, and you'll see what I mean.

That's what I consider the real magic of dynamic lead; it lets a sloppy gunner still hit a moving target with a good chance of first-round-hit at reasonable ranges. Wiether that gunner is natually sloppy, or a good gunner momentarily distracted, if the end result is us getting them first, that's fine by me.

DG

NEpi

10-14-2006, 05:36 AM

Another problem with the example was that you DIDN'T track for 0,9 seconds

you misread me. I said:

I assumed I aimed perfectly well when I pressed the DL button, and that I managed to catch-up with the target perfectly.

I explicitly said I was pushing the DL button all that second.
I merely pointed out that your tracking can jitter like hell as long as it takes less than a second to re-adjust and manage to get back on the target.
it's bad practice only because you can't be sure it took you less than a second to get back on target.

While we strive for perfection in practice and training, and we do what we can to innoculate gunners against stresses, unforseen circumstances, and whatnot, in the real world, Shit Happens.

yeah, and gunners miss, and it takes a few more seconds till the next round is loaded, and counter-fire starts being a problem. that deserves a boot to the head, not a shrug.I agree that boot should wait till the target that fires back is dead, and by then if I'm a gunner, I'd manage to kick myself in the head for missing by poor gunnery.

Happily, in the real world the target is not a tiny dot; it is a great big Jesus tank several metres long. At shorter ranges especially, a less-than-perfect shot may still hit and kill the target - that's good to know, and it is good for gunners to know exactly how big than envelope is should they need to get a snapshot off.

when your target is big and close, STOP FUCKING TRACKING IT! don't even lase! battlesight and fire away! you'll hit anyway (if to quote Omar from "The Wire": "from this range? even if I miss I can't miss")
if not, then making sure your first round is a kill is very important. take the extra sec or two, so you won't waste 5.
sometime, target is just over battlesight range. then lase, and fire right away. no tracking needed.

That's what I consider the real magic of dynamic lead; it lets a sloppy gunner still hit a moving target with a good chance of first-round-hit at reasonable ranges.

dynamic lead doesn't help you if you track poorly. it helps you aim, since you don't have to aim ahead of the target. tracking as bad as you're describing will send the round way off. if your range is short enough it doesn't matter. on the other hand, why using DL at that range in the first place?

112TYR

10-14-2006, 09:51 AM

you misread me. I said:
I explicitly said I was pushing the DL button all that second.

Yes, but you also said that you did not track for 0,9 seconds, or as I read it, keeping the reticle still:

by an extreme example I meant pressing the DL, then 0.9sec not tracking at all and then 0.1sec catching up to the target.

This says "no tracking at all".

I merely pointed out that your tracking can jitter like hell as long as it takes less than a second to re-adjust and manage to get back on the target.
it's bad practice only because you can't be sure it took you less than a second to get back on target.

Apparantly, you make some assumptions here that you don't write. The tracking needs to average out. 0,9 seconds of erroneous tracking followed by a 0,1 second "catch-up" will result in a miss. You explicitly said that "catching up" would be just as accurate as proper tracking, if done properly. This either means that you a) are wrong or b) require the "catching up" to be a part of a tracking that started out correctly(iow, merely a corrective motion), which is not what I am talking about at all. I am talking about a situation where you
1) aquire your target,
2) traverse onto it from behind,
3) press DL for 0,9 seconds BEFORE you're properly onto the target,
4) then track properly for 0,1 seconds....
5) and then fire

This is what I define as "catching up", and this will result in a wrong lead value.

desertowl

10-14-2006, 11:50 AM

if i understand NEPI correctly, he meant that he can retain a speed avarage over 1 second, but instead of moving the turret around the target's avarage speed (following the target constantly), he can make it from the target's position at T=0[s] to the target location at T=1[s] in a peak. he can get the same avarage speed as constant tracking BUT the turret ACCELERATION will be different as read by the turret encoder (the device that measures angular position,velocity and acceleration).
the question is, will the FCS take acceleration into account (i think it will), or just dead angular velocity avarage.

i guess that in modern tanks, tracking is done automaticly, via image proccessing of the signal from the CCD/FLIR sight, thus eliminating human tracking errors (relevant for tanks that have target acquiring, don't know if such capaility exists in the leo 2a5).

112TYR

10-14-2006, 12:12 PM

if i understand NEPI correctly, he meant that he can retain a speed avarage over 1 second, but instead of moving the turret around the target's avarage speed (following the target constantly), he can make it from the target's position at T=0[s] to the target location at T=1[s] in a peak. he can get the same avarage speed as constant tracking BUT the turret ACCELERATION will be different as read by the turret encoder (the device that measures angular position,velocity and acceleration).
the question is, will the FCS take acceleration into account (i think it will), or just dead angular velocity avarage.

Yes, that's how I understood it as well. However, that does not address the error of "catching up", where the average speed will NOT be the same.

Also, while we have been using 1 second as an example in this thread, lead is calculated over 1,7 seconds...

i guess that in modern tanks, tracking is done automaticly, via image proccessing of the signal from the CCD/FLIR sight, thus eliminating human tracking errors (relevant for tanks that have target acquiring, don't know if such capaility exists in the leo 2a5).

Never heard of such capabilities, they certainly do not exist on the Leos I've crewed. How would this work?

desertowl

10-14-2006, 12:38 PM

i am no tanker, but i am qualified with technical studies and military hardware, so what i say regarding tanks is more of an educated guess.

ATR (AUTOMATIC TARGET RECOGNITION) consists of the ability of the FCS image proccessing algorithm to discriminate the target from the surroundings, like done in modern anti tank missiles such as the javelin and the spike.
once the computer "knows" what the target is (using edge detection algorithms and a variety of filters) it can track target movement (a technology usually refered as VMD - VIDEO MOTION DETECTION) , by comparing one frame to the other, seeking differences.
now, via a closed loop control, it keeps the target in the center of the FOV, thus measuring the angular velocity without human interference.
all that the gunner needs to do, is to put the target on the center of the cross, to give the FCS a reference and from there, the tracking is fully automatic.
hope it made some sense.

those capabilities do exist in the merkava 4 and 3 BAZ:
http://www.army-technology.com/projects/merkava4/

"capability to acquire and lock onto moving targets, even airborne helicopters, while the tank itself is on the move."

112TYR

10-14-2006, 12:41 PM

Impressive. Does it actually work?

desertowl

10-14-2006, 01:03 PM

like a charm, if you reffer to the technology in general.
ofcourse i can't supply more data about the merkava other than publicly availiable material , however i do recall a video footage of merkava shooting down an aerial target drone, i'll try to look it up.

the F&F capabilities of the javelin and spike are derivatives of the same technology.

edit:
http://www.youtube.com/watch?v=_T-GKplzI0k
this clip of the merk 4 shows the AA capability.

NEpi

10-14-2006, 01:25 PM

Yes, that's how I understood it as well. However, that does not address the error of "catching up", where the average speed will NOT be the same.

your "catching up" scenario is not mine. why are you changing the basic data and then arguing about it? not aiming well enough at the beginning of the track will botch your equations, since the turret will have to move faster than the target right from the start. I specifically said I start ok, and then jitter.

Also, while we have been using 1 second as an example in this thread, lead is calculated over 1,7 seconds...

that's irrelevant nitpicking. if I press the DL for a second, it will average over a second.

i guess that in modern tanks, tracking is done automaticly

even if you have such a system, it won't be perfect every time (for example, if you drive by a tree it might recognise it as movement, and "lock-on" the tree instead of the target). also, never count on any kind of technology ither than the integrity of your gun in battle. systems crash.

desertowl, I'm sure it's angular velocity being calculated, not acceleration, since velocity is what matters to the ballistic equation (how fast your target is moving latterally), and not acceleration (how fast your target changes speed). why adding more error factors?

112TYR

10-14-2006, 01:39 PM

your "catching up" scenario is not mine. why are you changing the basic data and then arguing about it? not aiming well enough at the beginning of the track will botch your equations, since the turret will have to move faster than the target right from the start. I specifically said I start ok, and then jitter.

I stated in one of my posts that catching up would result in inaccurate lead, and that's where the whole thing started.
You replied that catching up would be just as accurate as ordinary tracking. Obviously, you misunderstood what I meant by catching up, since you did not address that at all, in stead you started talking about corrective tracking

that's irrelevant nitpicking. if I press the DL for a second, it will average over a second.

It is far from irrelevant, it determines how long you need to track to eliminate incorrect tracking. If you keep the DL-button pressed for 4 seconds, it will NOT average over 4 seconds, and it should be quite interesting to know how far back the memory goes. If the system needs 1,7 seconds, and the target makes a turn after 1 second, that will affect the lead calculated.

desertowl

10-14-2006, 01:42 PM

because tracking an accelerating target requires different transfer function than tracking a target that moves at a constant speed, BUT , and its a big BUT - if we are dealing with high velocity projectiles at reasonble ranges and target that is few meters long, it may be neglected (i have to run some calculations to find out the Pk for each fire solution).
as i said before - i am no tanker, but i am aware to the latest innovations in tankionics.

as for ATR - you can program the system to disregard trees and other interferences in your FOV , it's a function of good sensors and smart algorithms.
by no means did i point out that manual tracking is negligable and it is still a vital gunnery skill, however, emerging technologies offer us the ability to extend preformance, and the human factor becomes more and more a limiting bottleneck, as automation preformance increases.
you can find examples for this trend in every military branch,with emphasis on the more technical fields such as aviation , PGM's and in our case - tanks.

NEpi

10-14-2006, 06:00 PM

It is far from irrelevant

how does that relevant to our debate over the "catching up"?
why once you knew what I called "catching up" and that it is different from your definition didn't you volunteer your own definition?

you can program the system to disregard trees and other interferences in your FOV

I suppose so, but then you'll add more controls for the gunner, to be able to respond to every different situation he might encounter. that will add to the cognitive load on that stressed gunner. that might not be as good in reality as it is on paper.
keeping it simple is high-priority.

the human factor becomes more and more a limiting bottleneck
the current paradigm is humans ARE the bottlenecks. it's a little awkward, but it is efficient when trying to reduce technical stuff from the operator so he could keep focused on the decision making.

112TYR

10-14-2006, 06:20 PM

how does that relevant to our debate over the "catching up"?

This is not our personal thread, the term "relevant" means linked to the topic, not necessarily linked to YOUR statements within this topic. The post you refer to wasn't even a reply to one of your posts. One of the questions asked in this thread was how long time the system needed to calculate the correct amount of lead. Answering that question can hardly be called irrelevant. I can't even see why you would bother to mention it.

why once you knew what I called "catching up" and that it is different from your definition didn't you volunteer your own definition?

I did. To quote myself:

I am talking about a situation where you
1) aquire your target,
2) traverse onto it from behind,
3) press DL for 0,9 seconds BEFORE you're properly onto the target,
4) then track properly for 0,1 seconds....
5) and then fire

This is what I define as "catching up", and this will result in a wrong lead value.

desertowl

10-14-2006, 06:43 PM

no, you wont have to add controls - this is the idea behind ATR , that it's automatic.
by implicating pattern recognition the system can tell from the target to a tree or disturbance in the FOV (pixel count,formation and color values).
of course there are situations that a lock will be broken, but it's true for human operator as well, and as i said , this does not mean the gunner won't need to master manual control.

i find the whole discussion very interesting, it could be nice to hear from a tank system engineer the in depth method of operation of the lead calculation.

NEpi

10-15-2006, 02:02 AM

it should be very interesting to hear from someone who actually design those systems, but I don't believe we'll ever hear one here. noone likes to share such sensitive data on the internet.

I know just a little about pattern recognition in humans and in machine, and afaik, human are much better still. the problem of vegetation is particularly hard, since most camo paints are designed to blend into vegetation, and tanks can be the size of a small tree or a shrub when more distant than the said vegetation.we, as humans, have no problem differentiating tank from shrub due to object recognition, which is quite hard to implement in algorythm. we also get better at this at remarkable speed.

telling the computer what you are trying to track isn't easy some times. these times are not negligible.