TMA by least square method - Page 3

XLjedi · 11-08-07, 01:10 PM

Quote:

Originally Posted by Dr.Sid

Anyway in most languages the code should work with more or less copying it as it is, it's really that simple.

If it's VB-based that'll work fine.

Lemme know when you think you have something...

Dr.Sid · 11-08-07, 01:17 PM

Quote:

Originally Posted by aaronblood

Quote:

Originally Posted by Dr.Sid

Anyway in most languages the code should work with more or less copying it as it is, it's really that simple.

If it's VB-based that'll work fine.

Lemme know when you think you have something...

I've sent you the source codes on your work email, so check if it didn't end up in the spam.

XLjedi · 11-08-07, 01:26 PM

Hasn't made it yet... but I'll keep an eye out for it.
Sometimes our system is a tad slow with outside email.

Hey, what would you like to call your tool?

"DRT TMA""
"Pos/Brg TMA"
"Least Square TMA"
"Dr.Sid's Magic Medicine"

...or do you have some other preference?

This isn't by chance an implementation of "Ekelund Ranging" is it?

Dr.Sid · 11-08-07, 01:44 PM

I preffer Least square TMA. While Dr.Sid's magic medicine is not bad at all it does not have that hi-tek sound. :rotfl:

jmr · 11-08-07, 02:20 PM

Quote:

Originally Posted by aaronblood

Now the only question is... Should I continue bothering with this 1.1 beta or just focus my attention on jumping ahead to 2.0? :hmm:

How much was left for 1.1? 2.0 sounds pretty juicy so if you'd rather jump on that then go for it.

XLjedi · 11-08-07, 02:51 PM

Quote:

Originally Posted by Dr.Sid

In my implementation any times. It can even be more LOBs from same time. Hm .. I should better try this :rotfl:

Edit: tried .. and it works like charm. At each time there are 2 LOBs now. Now this can be used to limit target distance. You simply add 1 fake LOB to cross the actual LOB at the point of correct distance. Speed does not seem to be so easy anyway.

Yeah... but if I have two units taking bearings at the same times I don't need the tool. The crossing LOBs indicate the location, etc. I guess it's good that the logic properly falls on the LOB intersections, but I could've used my Mk1 eyeball to figure that out.

I wanted to see a solution with non-uniform LOB points of origin (ie. variable LOB time intervals). I guess I can test it myself when it arrives...

BTW, I'm not seeing it at work. The system may have blocked it.
Can you send a copy to my home email?

Dr.Sid · 11-08-07, 04:15 PM

Here you are. Ownship moves along a circle, times are random (not even in correct order, but that can't be seen). Note that the green circles at computed at times of the LOBs, to check that they fits them. Both ownship and target move at constant speed on this picture.

Dr.Sid · 11-08-07, 04:28 PM

Now I tried to add some position error to LOBs. Now it can be seen that near-parallel LOBs gives poor solution. I've also checked that stronger leg (bigger heading change) gives better solution, and with more then 90 degrees turn it gives good solutions even with large artificial errors.

XLjedi · 11-08-07, 06:00 PM

Looks good!

I think in MoBo this:

Might be implemented like this:

Where we would plot a series of LOBs and then run your gadget to plot the course line and put a contact with motion vector on the last LOB.

Dr.Sid · 11-08-07, 06:22 PM

There should be also some kind of dot stack, and possibility to adjust the solution manually and see what dots does. There should also be NUMERIC average error. My method then would only give you a estimate which you could work on manually.
But it looks great ! Have you got my mail ?

XLjedi · 11-09-07, 10:44 AM

No... I haven't seen any mail come thru.

If you like you can just copy/paste the code to a PM and I can piece it together on my end.

More than likely, I'll encapsulate it as a class.

jmr · 11-10-07, 09:35 PM

How quickly can this be integrated into Mobo? I'm anxious to try it out.

XLjedi · 11-11-07, 09:17 AM

It's gonna take a little time...

I think maybe before Thanksgiving is possible.

Zachstar · 05-01-08, 01:37 AM

It has been quite some time... Any progress?

XLjedi · 05-01-08, 06:52 AM

Yeah, you're right, it has been a while...

I guess I didn't think there was a whole lot of interest so I wasn't exactly in a hurry. The 1.1 beta does have a few new features, including plot save and unit data entry.

Lemme scan through it again and I'll try to see where it stands.

11-08-07, 01:26 PM	#33
XLjedi Ace of the Deep Join Date: Jun 2005 Location: Palm Beach, Florida Posts: 1,243 Downloads: 53 Uploads: 8	Hasn't made it yet... but I'll keep an eye out for it. Sometimes our system is a tad slow with outside email. Hey, what would you like to call your tool? "DRT TMA"" "Pos/Brg TMA" "Least Square TMA" "Dr.Sid's Magic Medicine" ...or do you have some other preference? This isn't by chance an implementation of "Ekelund Ranging" is it? __________________

11-08-07, 06:00 PM	#39
XLjedi Ace of the Deep Join Date: Jun 2005 Location: Palm Beach, Florida Posts: 1,243 Downloads: 53 Uploads: 8	Looks good! I think in MoBo this: Might be implemented like this: Where we would plot a series of LOBs and then run your gadget to plot the course line and put a contact with motion vector on the last LOB. __________________

11-09-07, 10:44 AM	#41
XLjedi Ace of the Deep Join Date: Jun 2005 Location: Palm Beach, Florida Posts: 1,243 Downloads: 53 Uploads: 8	No... I haven't seen any mail come thru. If you like you can just copy/paste the code to a PM and I can piece it together on my end. More than likely, I'll encapsulate it as a class. __________________

11-11-07, 09:17 AM	#43
XLjedi Ace of the Deep Join Date: Jun 2005 Location: Palm Beach, Florida Posts: 1,243 Downloads: 53 Uploads: 8	It's gonna take a little time... I think maybe before Thanksgiving is possible. __________________

05-01-08, 01:37 AM	#44
Zachstar Sea Lord Join Date: Feb 2005 Location: Shreveport, Louisiana Posts: 1,956 Downloads: 13 Uploads: 0	It has been quite some time... Any progress? __________________

11-08-07, 01:44 PM	#34
Dr.Sid The Old Man Join Date: May 2005 Location: Czech Republic Posts: 1,458 Downloads: 6 Uploads: 0	I preffer Least square TMA. While Dr.Sid's magic medicine is not bad at all it does not have that hi-tek sound. :rotfl:

11-08-07, 04:15 PM	#37
Dr.Sid The Old Man Join Date: May 2005 Location: Czech Republic Posts: 1,458 Downloads: 6 Uploads: 0	Here you are. Ownship moves along a circle, times are random (not even in correct order, but that can't be seen). Note that the green circles at computed at times of the LOBs, to check that they fits them. Both ownship and target move at constant speed on this picture.

11-08-07, 04:28 PM	#38
Dr.Sid The Old Man Join Date: May 2005 Location: Czech Republic Posts: 1,458 Downloads: 6 Uploads: 0	Now I tried to add some position error to LOBs. Now it can be seen that near-parallel LOBs gives poor solution. I've also checked that stronger leg (bigger heading change) gives better solution, and with more then 90 degrees turn it gives good solutions even with large artificial errors.

11-08-07, 06:22 PM	#40
Dr.Sid The Old Man Join Date: May 2005 Location: Czech Republic Posts: 1,458 Downloads: 6 Uploads: 0	There should be also some kind of dot stack, and possibility to adjust the solution manually and see what dots does. There should also be NUMERIC average error. My method then would only give you a estimate which you could work on manually. But it looks great ! Have you got my mail ?

11-10-07, 09:35 PM	#42
jmr Commander Join Date: Sep 2004 Posts: 462 Downloads: 62 Uploads: 0	How quickly can this be integrated into Mobo? I'm anxious to try it out.

05-01-08, 06:52 AM	#45
XLjedi Ace of the Deep Join Date: Jun 2005 Location: Palm Beach, Florida Posts: 1,243 Downloads: 53 Uploads: 8	Yeah, you're right, it has been a while... I guess I didn't think there was a whole lot of interest so I wasn't exactly in a hurry. The 1.1 beta does have a few new features, including plot save and unit data entry. Lemme scan through it again and I'll try to see where it stands. __________________