A-10 Suddenly Air to Air Powerhouse

[ttaylor0024]

 

I was trying to find a way to defend myself against an attack by a lone SU-25 when I thought, "Why not give the AGM a try and see if I could lock on to a moving aircraft." Turns out, you can by using ATG mode on your TGP, set SPI, slew the maverick and fire as normal.

 

Video is me vs 4 SU-25Ts. Very effective, although it's not the AGM's intended use...

[Bullitt]

There's a novel I read when a Mudhen driver in an exciting climax uses a Mav he still had after depleting the rest of his ammo to take out a Mig-23 (maybe a 27 it been a while). Based on that I've tried this myself a few times usually on helicopters with great success. I never though to try it with much else figuring it would be easy to spoof a Mav against a fixed wing target with some decent maneuverability. Maybe I should rethink that. Future MiG-21Bis drivers have been warned :)

[JB3DG]

I have used AGMs on numerous occasions as A2A weapons...even taking down some MiG-29s with em.

[Frogisis]

Ha, both Sidewinders miss but the Mavericks are right on target...

 

I suppose they wouldn't show up on an RWR until they were launched, either, would they? I know other people's Mavs show up on mine, so is that supposed to be a reaction to the rocket flame, or what?

I think I did read somewhere you could engage air targets with Mavs if you really wanted... If they can pick out and lock onto a tank in ground clutter but not a tree or a shed I don't see why a plane would be any different.

Slightly OT, but since the sim "knows" what's a target and what isn't, obviously we wouldn't have this problem, but I can't believe real Mavericks wouldn't regularly get stuck on some sufficiently high-contrast spot on the ground, or innocent sharp-edged object like a power transformer or a phone booth if you gave it enough leeway to slew around. Even if it had the resources to try matching everything it passes over against a database, you'd figure there'd be a lot of false positives...

[Echo38]

I've been killed by Mavericks a couple of times. It happened yesterday. Pretty embarrassing to be shot down by an air-to-ground weapon. ; )

[sobek]

Mavs obviously bleed energy far too slow when maneuvering as well as being able to pull too much gs. There's room for improvement, i guess. :)

[Eddie]

Not to mention being able to maintain lock in conditions that they simply shouldn't be able to.

 

Hopefully one day all weapon flight models, guidance and control will be updated to be as close to reality as can be obtained, but until then such problems just have to be lived with.

[159th_Viper]

I dare say we have stumbled upon a particularly large Wetapunga :D

[GGTharos]

Ha, both Sidewinders miss but the Mavericks are right on target...

 

The Su-25T's might be using their IR jammers, in which case all is fine with the 9 ... the maverick needs to be ... taken care of.

 

If they can pick out and lock onto a tank in ground clutter but not a tree or a shed I don't see why a plane would be any different.

 

It's not a problem with tracking the aircraft, it's a problem with hitting it after launch. This missile is designed to hit targets moving with a speed of 70kph at most (there's probably a safety margin there though). Neither the missile's control surfaces nor its seeker, or warhead are set up to attack aircraft. Not that it couldn't happen, but that should be the exception to the rule.

[Eddie]

I'd be happy if polarity switching and the "keyhole", and other launch limitations were implemented. That alone would more or less eliminate Maverick being used against aircraft.

[159th_Falcon]

Keyhole? could you elaborate a bit please? Got me curious.

[Eddie]

The Keyhole is the launch window of the missile, the pointing cross must be within the keyhole (among other things) in order for the lock to survive missile launch (see diagram below).

 

[159th_Falcon]

Ah, nicely explained.

What also means that its not always possible to launch MAV's from 20k feet up as a lot of us (me included) do now. (at least not without pointing the nose of the aircraft down.)

(hope i didn't get that wrong)

 

Would like to see this modeled.

[Eddie]

Ah, nicely explained.

What also means that its not always possible to launch MAV's from 20k feet up as a lot of us (me included) do now. (at least not without pointing the nose of the aircraft down.)

(hope i didn't get that wrong)

 

It does indeed. A Mav launched in such a manner would be almost sure to loose lock and go stupid off the rail.

 

There are other restrictions as well, such as bank angle, roll rate, altitude (minimum) and range (minimum).

 

Would like to see this modeled.

 

So would I. Along with polarity switching.

[159th_Falcon]

Can explain that one as well?

Assume it has to do whit Bhot vs Whot?

Can't switch once locked without losing said lock?

[marcos]

An F-15 pilot took out a helicopter with an LGB.

[Eddie]

Polarity for Mavericks doesn't refer to the display switching between BHOT or WHOT on IR models. If refers to the Maverick seeker tracking either light/dark (CCD Mav) or hot/cold (IIR Mav) "objects".

 

So if you're trying to lock a target with a CCD Mav (B, H, K) and said target is lighter than the background you'd need the Boat switch aft. If the target is darker than the background you'd need to boat switch forward.

 

If you don't set the correct polarity you won't get a lock.

[159th_Falcon]

Ah, also on the wishlist please.

[GGTharos]

And? That helicopter was barely moving.

 

An F-15 pilot took out a helicopter with an LGB.

[Frogisis]

Thanks for the explanations, I had no idea that was a factor on the real Mavericks.

Would definitely like to see that modeled, too, since something about them does seem suspiciously too easy to me as-is. Once you slew it onto a target it's basically a "make this explode" button, and it'd be a nice challenge to add specific maneuvering on top of it.

...You could probably leave the boat switch forward most of the time - I can't think of an instance off the top of my head I wanted to use a Mav on something lighter than the background... A ship, maybe?

[Pyroflash]

Mav's are incredibly easy to dodge due to their inability to maneuver. Most of their ability comes from people's lack of ability to dodge missiles in general, and over-reliance on CM's, which are ineffective against Mavs.

 

I do agree however, that they need to be implemented in a more realistic fashion. I cringe every time I see a missile launch coming from an A-10C flying straight and level at 25,000'.

[SmokeyTheLung]

If you don't set the correct polarity you won't get a lock.

I'd really love to see this simulated as well. Would add a wonderful challenge to the weapon

 

I'd imagine that with the real deal you could lock up a vehicles shadow (assuming the vehicle appears light in the display) if you had the wrong polarity set!

 

Fingers crossed but I'm not holding my breath... Also have to acknowledge that the smulated maverick that ED came up with is amazing, despite it being a bit of smoke and mirrors its definitely the most convincing simulation of a maverick that I've ever seen on a home PC (close enough for rock and roll as they say)

[ralfidude]

In AA, i only use AGMs.

 

 

Sidewinders against a S25T has about a 5% hit chance. They either go stupid on launch, or lose and bleed energy after the very first little tiny manouver they make. They are shit.

 

But somehow the Su-25Ts missiles are capable of making 13 high G turns and still hit their target, at longer range may I add.

 

 

So AGMs it is.

Edited September 28, 2012 by ralfidude

[msalama]

capable of making 13 high G turns and still hit their target

 

So is this actually a whine of a) biased modelling b) commies being evil c) sidewinders being shyte? Enquiring punters want to know ;)

Edited September 28, 2012 by msalama

[Bushmanni]

In real life there's plenty of situations where you would need polarity switching but I can't figure any where it would be useful in DCS. If a tank is hiding in a shadowy place and then drives out of the shadow it's going to be most likely cooler than the ground until it warms up and hence you would need to switch polarity to get a lock on it. In DCS this isn't modeled but the tank is always hotter than the ground. There's also no camouflage or things like that that could necessitate polarity switching or prevent getting a lock altogether.

 

I don't know what kind of algorithm is actually used in real life for auto image tracking, but it certainly isn't proper pattern recognition. Most likely it's simple edge detection. Here's how edge detections works. You scan a line of pixels in an image and calculate the rate of change in luminance (or some other variable that works). If the rate of change is higher than a preset threshold you declare there to be an edge. Edge detection can be two way or one way, ie. it only detects transitions from black to white or white to black or both at the same time. If you detect edges in both ways you get a signal for both upper and lower or left and right edges so you can't distinguish them. When you you detect only in one way you get only one edge. If do this for up, down, left and right directions (four times total) you get a bunch of pixels for every edge. You can calculate how much the pixels in one group of edge pixels are spread around and if they are spread less than a preset amount you can decide that they represent an edge in a image and their outwardly weighted average position is the position of the edge. This way you get a screen position for all four edges of a bounding box surrounding the target. If the preset sensitivities for image edge detection and edge pixel group spread are too high the calculated target position will jump around the screen too much because of noise and if the presets are too insensitive the algorithm will require more contrast to get a lock. Because of the way the algorithm works you need to tell it a rough position for where it should be looking for the edge hence you have the tracking gate size controls for IAT so it will pick the correct edge from the many there likely are in the image. And because it will detect edges in only one way you will need the polarity switch to lock targets that have reversed polarity.

 

The image edge detection is computationally easy to do and runs relatively fast, especially if the processed image has relatively low number of pixels. Reducing resolution and cropping the to be processed image to contain only the area where the edge needs to be found you can have pretty small images to be processed. The question is if this algorithm can be tweaked fast enough to handle multiple pods and missiles tracking targets at the same time and leaving enough resources for the rest of the sim to run. With multiple cores I'd think it shouldn't be too optimistic. Maybe the GPU could be used for this also. But of course there's also other things to use these cores for so it's more of a design question.