[WIP] Hydrophone Tracker/Quick Plotter

[gutted]

Originally Posted by Nico71

He he! It's about dinner time, so one more try before I'm gone! Let's assume that your boat is creeping along with 2 kts while you track that lone merchant far away. You collect bearings, then make the final triangulation, do the drawing stuff on the in-game map, etc. Then you're done!

you still dont get it

read the part in bold again.

If you are moving, you need to figure out where you will be on the in-game map in the future and draw the predicted bearing from THAT location. otherwise your triangulation will be wrong.


Anyway, the pdf pisces posted is the direciton im probbly gonna head with this.

input your 1st course and speed... grab three bearings.
Input your second course and speed... grab another three bearings.

It then tells you the targets true course and speed.. .and the distance along the last bearing he was.

So all you'll need to do is draw out the last bearing from you to the distance it gives you... then move the other end of it in the direction the target is going and you're good to go.

[Nico71]

Originally Posted by gutted

you still dont get it

read the part in bold again.

If you are moving, you need to figure out where you will be on the in-game map in the future and draw the predicted bearing from THAT location. otherwise your triangulation will be wrong.


Anyway, the pdf pisces posted is the direciton im probbly gonna head with this.

input your 1st course and speed... grab three bearings.
Input your second course and speed... grab another three bearings.

It then tells you the targets true course and speed.. .and the distance along the last bearing he was.

So all you'll need to do is draw out the last bearing from you to the distance it gives you... then move the other end of it in the direction the target is going and you're good to go.

And while I wrote this I've burned my dinner! YIKES!!!

Hey don't get me wrong, I'm not saying that you are wrong and I am right! I'm just trying to get a point across, which turned out to be more difficult than I have expected!

To make a long story short: you say this doesn't work as this will introduce a new error (position), while I say it's all a matter of conversion! Can we at least agree on that?

Be that as it may, Pisces has made this debate obsolete anyway.

Man, I had a good laugh!

[gutted]

It's all good man, it's just we're looking at two different things. If you can tell me how you're going to triangulate it without a hassle im all ears.

[Nico71]

Originally Posted by gutted

It's all good man, it's just we're looking at two different things. If you can tell me how you're going to triangulate it without a hassle im all ears.

Dunno! I was so busy typing that I couldn't test it!

[gutted]

Ok i figured out a way to do step (11) in that pdf mathematically without using scales to figure out how far along to draw A line Z-Y.


Instead of drawing A-B where it shows on that moboard... you instead draw it with A centered at P2.


then you take the vector length of Z-P2 divided by the vector length of Ownship-B.... to get a ratio to multiply the relative distance to get the actual distance.

[Pisces]

Originally Posted by gutted

So do you think im wasing my time adding support for more than 3 bearings and averaging them?

in other words... we should just use bigger intervals between sightings intead?

Yes, small bearing changes/short intervals only make bad results. And many bad results do not average to a good one.

It's difficult to explain why. Your measurement accuracy cannot be better than 1 degree if you let your crew call out the degrees. Maybe better than 1 degree if you use the periscope bearing scale. But much worse if you have to listen for the sounds on the hydrophone yourself. Anyway, 1 degree off in for example a 4 degrees inerval means the target track during it is upto about 25% wrong in length. But 1 degree after a change of 10 degrees is only 10% wrong. 1 in 50 is 2%. The issue is called 'granularity'. Alot of bearing measurements over short intervals only means more work. The data is then all of the same bad quality. Longer intervals means less measurements and drawing work, and better accuracy. Trust me, patience pays of in reliability.

You could reuse old bearings by doubling the time interval. You keep bearing 1 as the baseline. Old bearing 2 is discarded. Old bearing 3 becomes new bearing 2, while the new bearing 3 is measured after the 4th short intervals. This way you get an early estimate on course and you may decide to let it go if if is moving away. But still get the luxury of better data.

If possible try to choose the interval so that the final bearing (3rd) to be when the target would show an AOB of 90 degrees, or when he would be the closest range from the listening location. That way the 4th bearing will get the best crossing when you move parallel to the target.

[Pisces]

Originally Posted by gutted

you still dont get it

read the part in bold again.

If you are moving, you need to figure out where you will be on the in-game map in the future and draw the predicted bearing from THAT location. otherwise your triangulation will be wrong.

This is a bit of a problem indeed if you need to submerge to listen and be on the move surfaced. It's tricky to adhere to that average speed between taking the bearings. Best is to just flank it
on the surface to somewhere along a course some certain distance away. Then wait there wasting time to make the average speed match up. When the 2nd bearing has been measures, surface like a madman and flank/sprint another such distance ahead to be there early. Again wait there submerged to make the average speed match to whatever you have chosen to use in the plot.

Yeah, who said a navigator's life is easy.

This 2-leg technique is better suited for modern subgames like Subcommand/Dangerous Waters/688i since they are submerged the whole time and can go faster than WW2 subs. The ranges in those are usually alot further though, meaning slower bearing changes. However the crappy drawing tools on the map prevent me from doing it in those games. And the TMA station there spoils alot.

In SH3/4 (5 if I ever buy it) I'll do the 3-bearing-then-sprint technique as shown in the images in the beginning of this thread. Sometimes I do the 2-leg technique using the periscope at extreme visual range. Still trying to perfect my skill in these methods though. Knowing is something else than doing in the heat of battle.

[kylania]

Originally Posted by Pisces

Yes, small bearing changes/short intervals only make bad results. And many bad results do not average to a good one.

It's difficult to explain why. Your measurement accuracy cannot be better than 1 degree if you let your crew call out the degrees. Maybe better than 1 degree if you use the periscope bearing scale. But much worse if you have to listen for the sounds on the hydrophone yourself.

That's the problem I ran into. I made a test mission with a ship heading course 118 at 8kts. The first time I ran the numbers I used exactly 15 minute intervals and used the "Report nearest contact".

The numbers he gave me were 333, 340 and 347. This ended up with a target course of 70 degrees! No where near the true course. I asked gutted about it and he pointed out that using 334, 340, 347, a ONE degree difference in one of three bearings and you ended up with 104 degree course. Still wrong, but at least in the right general direction.

Not sure how to get around this really.

[Pisces]

Originally Posted by Pisces

Yes, small bearing changes/short intervals only make bad results. And many bad results do not average to a good one.

[EDIT] However, some people can't stand it sitting still and feel they must do something in that time. Something may be said for that. Might be good for them to achieve Zen-state.

[Nico71]

Okay, tested my method in SH3 sub school, torp training! This is what I did: 3 bearings, 10 minute marks, enter them into the Hydrophone tracker. From the last position, at 20 min, the point where I took the last regular reading, I entered the bearing given to me by HT. Then I sprinted to a new location for triangulation. Then I went on as usual, took the new bearing for triangulation, measured the distance that I entered into HT. Now I get some results.

I have tried this several times now! With and without moving. Without moving always were some degrees off the mark, speed was a bit off, too.

With moving the results I got from HT were strange at first glance. To be on the safe side, I entered them into Mobo, using the method I have explained in another post above.

What can I say? 3 bearings plus 1 triangulation, and the resulting target course and speed were always less than 1 degree off, speed also very close to the real thing! Much better than just standing still!

I have to admit that I was a bit surprised myself. I don't understand the math behind it. In fact, I'm a doofus at math. But for whatever reason it seems to work! The precision seems to be somewhat dependend on the direction I take for the triangulation.

This is all certainly worth further investigation!

Could anyone else please try this method and report the results?

[Pisces]

Originally Posted by kylania

That's the problem I ran into. I made a test mission with a ship heading course 118 at 8kts. The first time I ran the numbers I used exactly 15 minute intervals and used the "Report nearest contact".

The numbers he gave me were 333, 340 and 347. This ended up with a target course of 70 degrees! No where near the true course. I asked gutted about it and he pointed out that using 334, 340, 347, a ONE degree difference in one of three bearings and you ended up with 104 degree course. Still wrong, but at least in the right general direction.

Not sure how to get around this really.

The numbers 333, 340 and 347 indicate that bearing 2 is about the bearing that shows an AOB of 90 degrees. Since bearing 1 and bearing 3 are equal degrees from the 2nd one.

Perhaps it's easiest to see if you make this 3 bearing AOB tool (shameless plug ) It does the same thing as doing the 3 bearing drawing in Gutted's tool, ( has to be) except for calculating the angle of and distance to bearing 4. It's just an AOB tool.

http://www.subsim.com/radioroom/showthread.php?t=147719

Notice how the smaller bearing differences are wider apart(i.e. 5 to 6 on outside scale), compared to the larger bearing differences (i.e. 20 to 21 on the same outside scale). This means the top disk needs to turn less if you want to see how big an effect one degree error has on long interval bearing changes.

[gutted]

Ok after a bit of hacking on the test version of the tracker...


I get this when using the example in the PDF... cant figure out why it's off:


Are we sure the solution in the PDF is correct? Mine was done mathematically with floating point precision and not drawn by hand.

Note: the above UI will not be the final version.. its just the previous version hacked to work with 2 groups of bearings.

[gutted]

If anyone wants to try out the moving version:
http://www.subsim.com/radioroom/down...o=file&id=1704

Be warned its had very little testing, and i'm not yet certain if the results are correct.

It's a beta version that needs testing.

And yes i know the UI is a bit strict (.ie forcing you to start over if you screw up), but thats because this is a very early version and im forcing the user to do things correctly.

Once its working as intended i'll free up the UI and make it more friendly by accounting for changes in data.

[gutted]

Just tried the moving version on a target that was doing 90 degrees at 9kts.

I came up with 108 degrees at 8kts using 10 minute intervals. Not bad accuracy for such short time span.

Only problem? It took so long to freaking do two sets that he was already in view when i started doing the second set LOL. I started using the scope for the bearings.

Will have to not use the moving method when the target starts so close.

[gutted]

Now that i think about it.. im not even sure how practical the 2-leg moving method is given the short hydrophone ranges on a u-boat.

It's rather pointless to be listening that long.

If you look at that example in the pdf the target was like 68nm away.. thats like 125km. Much more practical.

Now scale that down to U-boat a hydrophone of like 20km listening distance. You run for 30 minutes listenging at 3kts for a target thats doing 11kts... and he's already travelled almost half the distance (8km), while you've traveled nearly 3km... and thats just the first leg !!!


It seems much more practical to sit stationary and listen for like 20-30 minutes and see what he's doing before you start moving. If he's closing it wont be long before you get contact anyway. You have about enough time to run to a triangulation point to get his true distance and speed then setup for an intercept.