visual detection rules

[CV32]

Hmmm... I wonder if this is the explanation for the occasional "self destruction" of pairs of my F-14s, etc. when none of my units ever detect anything in the area to account for it? In at least one of these instances, there were Eurofighters in the scenario, which could have been the culprits. Even so, I am still puzzled about how the Eurofighters (or whoever) even knew to look for the F-14s that were way out there in the ether... not to mention wondering how they had enough gas to get way out there to knock 'em off.

 

For the umpteenth time, JoeK, enough with the vague generalizations and suppositions already. If you have a specific issue or question, with specific details to flesh them out, then please share but otherwise this isn't helpful at all. Okay?

[broncepulido]

Some examples of values/signatures obtained using the aformentioned formula:

Very Small: WWI fighters, WWII fighters, Il-2, Me-262, Ar-234, Gripen, Tiger ARH, Mangusta, UH-1, MiG-15, MiG-17, MiG-19, MiG-21, AV-8B, Ka-50, J-10, AH-64, Mi-28, Lyn, Sea Hawk, Hawk, A-4 Skyhawk, Mitsubishi F-1, Jaguar, OV-10, F-5A/E, A-37, Yak-38, SH-2 Seasprite, F-104, Mirage F1, Vampire, Strikemaster, Ajeet, JF-17, Bearcat, Sikorsky 55, Ka-15 Hen, Magister, Alouette II/III, F-82, F-94, Ouragan, T-6 Texan (WWII), Sea Fury, F2H-2 Banshee, F9F-2 Panther, F11F Tiger, OH-23 Raven, H-13 Sioux, Sycamore, Po-2 Mule, Yak-15, Yak-23, FJ-1 Fury, F-84D/G Thunderjet, Jastreb, Galeb, Alpha Jet, L-39 Albatross, SF.260W Warrior, Super Galeb.

 

Small: Me-110, Mosquito, He-219, Ju-87, F/A-18A/E, Rafale, F-16, UH-60, EH-101, Su-7/17/22, Mi-8, SH-3 Sea King, A-10, Mi-25, Super Frelon, CH-53, Mirage III/5, Mirage 2000, Etendard, Super Etendard, MiG-23/27, MiG-29, Su-15 Flagon, A-7 Corsair II, F-8 Crusader, Draken, Tornado, Yak-25M, Yak-28P, Su-25, F-101, Alize, King Air, RQ-4, S-2 Tracker, Hunter, Marut, F-86 Sabre, A-1 Skyraider, F-82, Skynight, Su-11, F-100 Super Sabre, Mi-17, Skyray, Vatour, F-105, Gannet, Sikorsky 58, U-2A, Scimitar, Ar-196, Me-410, Fw-189, Mystere IV, Super Mystere, Il-10 Beast, F2H-3 Banshee, F9F-8 Cougar, Fairey Firefly, FJ-3 Fury, R/F-84F Thunderstreak/Thunderflash, TBF-3 Avenger, IAI Westwind, Skyvan, Tejas.

 

Large: Ju-88, B-25, Ju-52, F-102, A-6 Intruder, CH-47, Yak-25RV, Viggen, F-4 Phantom, Mirage IV, S-3 Viking, Su-24 Fencer, Su-27/30/34/35/37, F-15 Eagle, Il-28 Beagle, RB-66C, A-3 Skywarrior, F-111, Mig-25, F-3 Demon, Savage, Gotha (WWI), Constellation, F-106 Delta Dart, E-2 Hawkeye, A-12/SR-71, U-2R, CH-54 Tarhe, Javelin, Sea Vixen, Lancaster, B-17, B-52, B-58, Tu-16, Tu-22 Blinder/Backfire, Fw-200 Condor, C-82, C-119, C-123, B-36, Il-14 Crate, Il-18 Coot, F7U Cutlass, B-47 Stratojet, YS-11, Yak-24 Horse, AF Guardian, HU-16 Albatross, Tu-14 Bosun, F-89 Scorpion, CF-100 Canuck, B-26 Invader, C-47 Dakota, Privateer, Buccaneer, B-29, Tu-4, Eurofighter.

[CV32]

Fortunately, Harpoon4 provides an assessment of aircraft size that provides for consistency with the paper rules and is easily adapted to HCE.

[VictorInThePacific]

I use this formula:

Aircraft Height (in meters) X Aircraft Lenght (in meters) X Aircraft Wing Area (in meters, in helicopters Rotor Area/10)= Y

If Y <= 45 (if Y equal or smaller than 45), signature Very Small.

If Y between 45 and 70, signature Small.

If Y > 70 (if Y greater than 70), signature Large.

 

Interesting. Effectively, considering the aircraft to be rectangle-shaped, this is the same as

 

Y = (side area) x (wing area)

 

Note that wing area has units of square meters, not meters.

 

Instead of wing area, you should probably use top area, which is not exactly the same.

 

But mainly, I think you should not be multiplying these two areas, but rather averaging them. This would be important for planes that are significantly larger in one view than the other. Or maybe it would have no practical effect at all.

[broncepulido]

I use this formula:

Aircraft Height (in meters) X Aircraft Lenght (in meters) X Aircraft Wing Area (in meters, in helicopters Rotor Area/10)= Y

If Y <= 45 (if Y equal or smaller than 45), signature Very Small.

If Y between 45 and 70, signature Small.

If Y > 70 (if Y greater than 70), signature Large.

 

Interesting. Effectively, considering the aircraft to be rectangle-shaped, this is the same as

 

Y = (side area) x (wing area)

 

Note that wing area has units of square meters, not meters.

 

Instead of wing area, you should probably use top area, which is not exactly the same.

 

But mainly, I think you should not be multiplying these two areas, but rather averaging them. This would be important for planes that are significantly larger in one view than the other. Or maybe it would have no practical effect at all.

Clearly you're right, in my original formula:

- The area size is in square meters.

- The values lenght+height+wing area are ADDED, not multiplied.

For me is clear issue ... the problem was to explain it (for sake of that, I've send the examples...)

[donaldseadog]

I use this formula:

Aircraft Height (in meters) X Aircraft Lenght (in meters) X Aircraft Wing Area (in meters, in helicopters Rotor Area/10)= Y

If Y <= 45 (if Y equal or smaller than 45), signature Very Small.

If Y between 45 and 70, signature Small.

If Y > 70 (if Y greater than 70), signature Large.

 

Interesting. Effectively, considering the aircraft to be rectangle-shaped, this is the same as

 

Y = (side area) x (wing area)

 

Note that wing area has units of square meters, not meters.

 

Instead of wing area, you should probably use top area, which is not exactly the same.

 

But mainly, I think you should not be multiplying these two areas, but rather averaging them. This would be important for planes that are significantly larger in one view than the other. Or maybe it would have no practical effect at all.

Clearly you're right, in my original formula:

- The area size is in square meters.

- The values lenght+height+wing area are ADDED, not multiplied.

For me is clear issue ... the problem was to explain it (for sake of that, I've send the examples...)

It seems to me that simulating visual detection is pretty hard as there are so many factors and some aren't modelled (eg sound). Things like

detecting a glider head on is difficult, but from above or below it is much easier. Also detecting a plane that is right behind the searcher as against one right in front, I'm sure there will be some things used in determining detection that are found to work but the explanation is not clear.

Don

[broncepulido]

I use this formula:

Aircraft Height (in meters) X Aircraft Lenght (in meters) X Aircraft Wing Area (in meters, in helicopters Rotor Area/10)= Y

If Y <= 45 (if Y equal or smaller than 45), signature Very Small.

If Y between 45 and 70, signature Small.

If Y > 70 (if Y greater than 70), signature Large.

 

Interesting. Effectively, considering the aircraft to be rectangle-shaped, this is the same as

 

Y = (side area) x (wing area)

 

Note that wing area has units of square meters, not meters.

 

Instead of wing area, you should probably use top area, which is not exactly the same.

 

But mainly, I think you should not be multiplying these two areas, but rather averaging them. This would be important for planes that are significantly larger in one view than the other. Or maybe it would have no practical effect at all.

Clearly you're right, in my original formula:

- The area size is in square meters.

- The values lenght+height+wing area are ADDED, not multiplied.

For me is clear issue ... the problem was to explain it (for sake of that, I've send the examples...)

It seems to me that simulating visual detection is pretty hard as there are so many factors and some aren't modelled (eg sound). Things like

detecting a glider head on is difficult, but from above or below it is much easier. Also detecting a plane that is right behind the searcher as against one right in front, I'm sure there will be some things used in determining detection that are found to work but the explanation is not clear.

Don

Yes, but counting with the limitations of the Game Engine, as it's not discrimination in the relative "facing" of any the platforms (the sound issue is a different question), we can considerer a middle "visual impact/radar cross section" value in all the platforms, as all the platforms are in the same relative case (i.e. none of the platforms are discriminated by a differente "facing" in any case).

Mmmm, I see my explanation is a little confuse ...

[TonyE]

I think this is a valid case to bring out the paper rules since this portion of the code is modelled directly to the paper rules (to the best of my current understanding).

 

See specifically 5.4.2, 5.4.3, and 5.4.4

 

You all appear to be doing good work here, if you end up with a new visual model, great! When chatting with Brad long ago about database maintenance I think he said he thought entering 4 signature values per frequency band would be bearable (correct me if I'm wrong Brad). Those would be:

 

Top/Bottom aspect

Front aspect

Rear aspect

Side aspect

 

In that case you need to agree upon both formulas for the raw detection (VitP is doing well there with raw physics with others helping bring in other aspects) as well as formulas for 'mixing' the signature values to give a value for any vector to the platform.

 

We've already seen the other inputs available to you, 3 states for time of day (night, day, dawn/dusk) and 4 weather modifiers ( 0)no precip, 1)light rain/fog, 2)moderate rain/fog, 3)heavy rain/fog ).

Visual.pdf

[donaldseadog]

Yes, but counting with the limitations of the Game Engine, as it's not discrimination in the relative "facing" of any the platforms (the sound issue is a different question), we can considerer a middle "visual impact/radar cross section" value in all the platforms, as all the platforms are in the same relative case (i.e. none of the platforms are discriminated by a differente "facing" in any case).

Mmmm, I see my explanation is a little confuse ...

I think I follow, but I'm not convinced it is always correct.

Consider:

two planes about the same size are at a bit less than visual limit, the plane's respective heading are about 90 deg apart with one plane approachinging the other and facing directly at it and the second plane heading 'across' the front of the first, ie an ideal situation for the first plane to cruise in behind the second on its tail and shoot. The approaching plane is seeing the second plane side on and its right in front, but for the second plane they are looking at the first front on and having to look right abeam (with maybe a wing in the line of site). The approaching plane is (in reality) in a very advantageous position.

[broncepulido]

Yes, but counting with the limitations of the Game Engine, as it's not discrimination in the relative "facing" of any the platforms (the sound issue is a different question), we can considerer a middle "visual impact/radar cross section" value in all the platforms, as all the platforms are in the same relative case (i.e. none of the platforms are discriminated by a differente "facing" in any case).

Mmmm, I see my explanation is a little confuse ...

I think I follow, but I'm not convinced it is always correct.

Consider:

two planes about the same size are at a bit less than visual limit, the plane's respective heading are about 90 deg apart with one plane approachinging the other and facing directly at it and the second plane heading 'across' the front of the first, ie an ideal situation for the first plane to cruise in behind the second on its tail and shoot. The approaching plane is seeing the second plane side on and its right in front, but for the second plane they are looking at the first front on and having to look right abeam (with maybe a wing in the line of site). The approaching plane is (in reality) in a very advantageous position.

For me clearly your right, Don, but if the GE don't give us "faceting" of the platforms, I should think all the platforms are "flying saucer" shaped (i.e, they are shapened with radial simetry).

[TonyE]

For me clearly your right, Don, but if the GE don't give us "faceting" of the platforms, I should think all the platforms are "flying saucer" shaped (i.e, they are shapened with radial simetry).

 

Hehe, yes, the platforms are all spheres to the game sensors.

[donaldseadog]

For me clearly your right, Don, but if the GE don't give us "faceting" of the platforms, I should think all the platforms are "flying saucer" shaped (i.e, they are shapened with radial simetry).

 

Hehe, yes, the platforms are all spheres to the game sensors.

Ideal if planes are finally replaced by flying saucers, eh

I suppose it comes down to how often is visual detection important in the game?

Certainly there are cases, especially with more and more stealthy fighters and the ability to rely on remote platforms for doing the detection for the shooter.

If we were to go to including faceting of the visual target then I think we'd need to also be looking at the viewing direction from the observer (in the case of an aircraft observer) as that may be more critical. If we could take into account the visual angles then maybe we only two facet values for the target and they are balanced taking into account its direction of flight relative to the observer... ( I think it is my turn to claim a somewhat less than clear explanation )

[CV32]

I think the term you folks are looking for is 'target aspect'.

[Kavik Kang]

I think the term you folks are looking for is 'target aspect'.

 

Haha, I was thinking the same thing reading this thread and then got your post at the end.

 

What you guys are talking about is actually very complex, as CV32 said it is "target aspect". In simple terms this means "radar cross section relative to facing". Star Trek writers used the concept, there are a couple episodes where you will hear Picard order "present minimal aspect to target". The ship/plane is a lot easier to see from the top down than if it is coming right at you. This would be silly to attempt to represent this acurately within a game. You'd essentially have a different radar cross section depending on the facing of the unit.

 

Harpoon does the only thing you really can do within a game, just pick an "average" value to represent this regardless of facing.

[TonyE]

Harpoon does the only thing you really can do within a game, just pick an "average" value to represent this regardless of facing.

 

Let's try not to get too pessimistic. We have the data to determine target aspect (emitter and target altitudes, headings, range) so we can determine the target aspect from as seen from each emitter.

 

Nor are we forever restricted to a single average value, in my previous post you hopefully got the idea that we can modify the database structure (we just prefer not to without good reason since it is a pain) to hold multiple reflectivity values for a particular sensor band.