Ship Radar Signatures (Full Version)

All Forums >> [New Releases from Matrix Games] >> Command: Modern Operations series



Message


ARCNA442 -> Ship Radar Signatures (8/31/2019 3:46:19 AM)

I'm not sure if this belongs in the database updates thread, but I was digging through the database and found some seemingly odd numbers for ship signatures.

First the really weird ones (all signatures are broadside in the E-M bands):

The Indian Talwar and the Russian Grigorovich are basically the same ship - yet the former is 26.8db while the latter is 6.8db.

The Dutch De Zeven Provincien and the Danish Iver Huitfeld are also extremely similar designs - yet the former is 28.4db while the latter is 17.0db.

The German Sachsen is 28.3db while the larger and closely related and larger Baden-Wurttemberg is 17.8db.

Finally, the British Type 26 is 17.8db while the larger and older Type 45 is 9.0db.

But aside from these oddities, the many of the recent generation of European/American/Russian ships appears to have signatures that are far too low. For instance, Horizon in 9.0db, FREMM is 8.2db, and Gorshkov is 7.7db, while the Type 054D is 32.2db, Shivalik is 28.3db, and the Type 054A is 27.6db despite all of them being built around the same time and having similar levels of stealth shaping. Even other European designs are nowhere near as low, such as the above mentioned Sachsen at 28.3db or the Norwegian Nansen at 27.5db.

The American LCS designs also have incredibly low signatures at -0.7db and 0.1db for ships that have few stealth features beyond basic shaping - especially when the extremely stealthy and similarly-sized la Fayette is a massive 15.8db. The American-built Ambassador missile boats are another example at -0.7db when the smaller Chinese Type 022 are 4.0db.




c3k -> RE: Ship Radar Signatures (8/31/2019 12:01:35 PM)

Nice job pulling all this data out of the (massively sized) database and bringing them forward for discussion.

I'm curious to see what comes of this.




DavidRob0 -> RE: Ship Radar Signatures (9/24/2019 1:54:12 AM)

Any data storage system the size of CMANO's is going to have inconsistencies - not only in radar signatures but (probably at least as importantly) in ship and submarine sound signatures, including increased noise at speed.

Would make an interesting project to run a side project to search for inconsistencies and to verify the "reasonable" accuracy of the data already stored there.




KnightHawk75 -> RE: Ship Radar Signatures (9/24/2019 6:22:50 AM)

@DavidRob0

That would be one hell of a 'side' project. eek lol
Anyway if it helps toward that end, attached are tab delimited dumps for all ships and subs and their related signature and dimension profiles, can import it into whatever tool to play around\visualize with it more. I included the sqllite queries if you want to tweak the output or re-run. Beats the hell out of going page to page in the viewer at least.

I wouldn't even know where to start trying to programmatically spot comparative inconsistencies though[&:]. I guess for starters you could figure out some sort of typical value range per band,per direction, per ship type + given dimensions(assign ing for example your own 'HullDesignClass' classification for every entry) and then query who falls outside of said typical ranges for a given assigned HullDesignClass and then for each of those manually see if those make some sense or do not make sense based on research (deep or cursory) into each specific ship, or at least relative to similar ships in db. oof, gives me a headache just thinking about it.




DavidRob0 -> RE: Ship Radar Signatures (9/24/2019 6:42:01 AM)

It certainly would be a large job[:'(]

Thanks for that KnightHawk75.

Saved your files and will give it some thought[8|]




Luckschaden -> RE: Ship Radar Signatures (9/30/2019 6:44:39 AM)

Thanks for this data!

I only have a few hundred hours in this game and don't know much about naval stuff at all, but the data set was quite interesting to me.

I did some analysis using a simple Linear Regression to predict the expected signature for a given ship size, and then looked which entries have the largest deviation between real and predicted signature.

I've done it in a python notebook, I'll attach a html version which can be used without installing python / jupyter notebook and hopefully includes the plot outputs etc.

The interesting stuff is in cell 37, log scaling really improved the model fit.

This is a first stab for only A-D bands, but can easily be extended if it's useful at all.




KnightHawk75 -> RE: Ship Radar Signatures (9/30/2019 4:10:58 PM)

@Luckschaden interesting stuff and approach indeed. :)





Luckschaden -> RE: Ship Radar Signatures (9/30/2019 6:34:36 PM)

Let me know if it's useful at all, it'd be quite easy to do the same analysis for all the other detection types and output a list with outliers, for example.




Page: [1]

Valid CSS!




Forum Software © ASPPlayground.NET Advanced Edition 2.4.5 ANSI
1.796875