fixed Rapier Guidance Logic/AI

[Northstar98]

Hi Everyone,

 

I was doing some tests with the Rapier and it seems that when firing at low altitude targets, the Rapier missile will do an extreme yoyo, going high, and then rapidly diving, only barely managing to pull-up and avoid a collision with the ground.

 

In the first track I have a Mi-24V flying near enough perpendicular (but not notching), a launcher and an optical tracker, with no blindfire. The Mi-24V is @200m AGL (ΔH ~ 195m), flying at 200kph - well within the engagement parameters for Rapier (according to the Rapier_FSA_Launcher.lua the minimum altitude is 50m). The first missile climbs, then performs a steep dive, only barely managing to pull-up to avoid a collision with the sea, and eventually hits the target.

 

The target descends, another Rapier is fired, this time crashing into the sea. 

 

The third missile repeats what the first missile did, and barely manages to avoid colliding with the sea.

 

In the second track I have everything set-up exactly the same, only now the helicopter is set to an altitude of 100m (ΔH ~95m), which should still be inside the Rapiers engagement altitude; this time every missile ends up flying into the water.

 

It should be noted that the missile behaves the same when paired with Blindfire or not.

 

Spoiler

AFAIK When Rapier is only used with the optical tracker, the missile uses a visual SACLOS guidance method; on the optical tracker there is essentially a periscope (though only the 'head' moves), and the operator uses hand-controllers to slew the 'head' around; and the missile attempts to stay centred in the reticule, with steering commands provided by the launcher via an uplink antenna (which is the small parabolic dish on the front of the launcher). It works in pretty much the exact same fashion as the Tunguska's 9M311 missiles.

 

When paired with Blindfire, the missile can use a purely RADAR ACLOS guidance method, whereby Blindfire tracks the target and the missile (AFAIK) via RADAR (unsure if Rapier uses a RADAR beacon - though in certain shots I can see what look to be antennae on the rear of the fins, those these could just be the uplink receivers). Here it works in a similar way to the SA-3, but I'm unsure of the exact guidance logic used.

 

Blindfire also has a TV camera co-located, but I'm not sure what its function is, it could be that it's used for optical tracking of the missile, via the flares on the missile's rear, in the exact same way the optical tracker does.

 

The launcher itself contains a small, doppler acquisition RADAR with IFF, though one battery component missing from Rapier is the SEZ or Sector Engagement Zone box, which is the operator console for the search RADAR. It consists of a ring of 32 red lights, with a light illuminating when a contact is detected in the corresponding direction. Each light has a switch, allowing for an operator to 'switch off' returns from certain directions. I suspect that the RADAR is essentially azimuth only, and doesn't provide range information; similar to the Tomcat's PD search mode.

 

The launcher and Blindfire have a small generator that can be mounted on them for transport; when deployed the generator is detached and placed a short distance away (we have the generator already, but it is non-functional, doesn't make any noise, and can't be moved independently to the launcher/Blindfire). 

 

One final component missing is the MST or Missile Supply Trailer, which is a small, 1-ton trailer that can carry several missile reloads (wiki says 10, but 6 seems more realistic, given the size of the reload containers). The containers for the reloads can actually be clearly seen already in DCS, when you select Open/Loaded on the Land Rover 101 FC. While not the aim of this thread, it would be great to see these as a separate unit, either as a cargo or a static object, preferably integrated with the warehouse system, so that 1 of these reload containers can reload one Rapier missile, it would be useful to set-up something like this, whereby the 4 containers carrying the initial 4 missiles, are covered with a camo net around the optical tracker, in a similar way sandbags would be used.

 

The entire system, including Blindfire is towed by 2x Land Rover 101 FCs and 1 Land Rover 109 S3.

 

One of the 101 FCs tows the launcher, with the optical tracker folded up inside the vehicle, along with the initial 4 missiles. The second 101 FC tows Blindfire, and the 109 S3 tows the MST.

 

Now this could be caused by probably 1 of 4 factors:

 

1. The missiles are initially elevated too high - unlikely, the missiles should elevate based on range and altitude.
2. The missiles are too aggressive in trying to centre themselves in the operators reticule (this is broadly what an optical SACLOS guidance scheme tries to do) - unlikely as this might play a part into reducing the minimum range, which is probably a good thing for a short ranged system.
3. The missiles aren't responsive/manoeuvrable enough - not sure, I can only base on my expectations (which may well be off). If you draw a line between the optical tracker and the target I'd expect the missile to follow that line more closely (like the Tunguska), not loft way high above it and then make an excessive dive past it.
4. The AI operator - here I think lies the issue, the AI operator most likely keeps the reticule as centred as possible on the target, this makes sense, however it is most likely (from my reckoning at least) causing this issue, if not for the above 3; a possible remedy is to have the operator aim high (at a similar elevation to the one the missiles are fired at), and then slowly bring the reticule down (so that it is on the target after about 2 seconds). Hopefully, this will result in a more gradual descend and the missile won't perform such an aggressive dive that it can barely climb away from. This should also increase the amount of energy the missile has, which is beneficial for achieving a hit.

 

There is this video of a Rapier launch failure (here the system is the much more modern FSC variant, but AFAIK the command-guidance scheme is the same, just instead target tracking is done automatically via EO/IR and/or RADAR). In DCS it seems to do something similar every time it's launched at a low-altitude target.

 

Can anyone confirm what's going on here, and how Rapier should behave? I've done some digging but can't find much, other than basics.

 

 

[sidenote: @Tippis  contrary to your excellent reference wiki Rapier in DCS will in fact launch with just the optical tracker and without Blindfire, under an optical SACLOS guidance scheme, as it should IRL]. 

 

 

Rapier_FSA_Guidance_(200m_AGL).trk Rapier_FSA_Guidance_(100m_AGL).trk

Edited January 22, 2021 by Northstar98

[Tippis]

Oh? Neat. How did you set it up because I could never get it to do that.

Is it just a matter of giving it enough time to lock on?

[Northstar98]

41 minutes ago, Tippis said:

Oh? Neat. How did you set it up because I could never get it to do that.

Is it just a matter of giving it enough time to lock on?

 

I just set it up with 1 launcher and 1 optical tracker, worked without me doing anything. I'll PM you the mission file.

 

The only thing I did was set the interception range to 75% and the alarm state to red.

Edited January 18, 2021 by Northstar98

[dorianR666]

23 hours ago, Northstar98 said:

I'll PM you the mission file.

The mission is already inside the trackfile btw.

[Northstar98]

6 minutes ago, dorianR666 said:

The mission is already inside the trackfile btw.

 

Is it? Never knew that.

[Flappie]

Yes. Transform a ".trk" into a ".miz" and you will be able to load the mission.

Edited January 19, 2021 by Flappie

[Raven (Elysian Angel)]

7 hours ago, Northstar98 said:

Is it? Never knew that.

Neither did I... One learns something new every day!

[VirusAM]

The mission is already inside the trackfile btw.

Exactly, if you open the trk file with a 7zip you will notice that there are all the mission files, and some more folders which are the track data

[Flappie]

100% reproducible when target is low enough (here, 100m ASL).

 

 

Edited January 20, 2021 by Flappie

[nighthawk2174]

Hmmm yeah that ain't, from what little footage i've seen it doesn't loft at all but flies right at the target with PN.

[Northstar98]

 

@Flappie The target doesn't have to be low either, even at 200m, the missiles barely manage to avoid a collision; they even do a similar dive when fired at a target at 1000m, though at least then it recovers from the subsequent dive sooner, thereby reducing the chance for a collision with the ground/sea.

 

@nighthawk2174 It's not so much lofting, rather it's just flying straight, at the angle the missile was elevated at on launch, presumably in this stage, the missile is unguided and will keep going until it receives uplink commands. These might be happening too late, meaning the missile climbs too much, I'm not sure - I don't know much about this system other than the basics. But the main problem is the excessive dive the missile does (presumably after capturing the uplink) after, and the fact that it recovers way too late.

 

If you were to attach a laser pointer to the optical tracker and boresight to the reticule; the missile would try to stay centred in the beam (SACLOS being like beam-riding, just without the beam, and the guidance commands being sent to the missile, instead of the missile computing them itself).

 

On capturing the uplink, I would expect the missile to dive to centre itself in the beam of our imaginary laser pointer, and then recover from the dive when intersecting the beam at the latest. Right now, at least initially, the missile pulls up well too late, far after crossing the imaginary beam, it's almost as if the command to dive is getting stuck, essentially telling the missile to keep diving when it should be recovering.

 

However, if its current behaviour is accurate, another solution would be for the AI operator to aim high and then slowly (say within a couple of seconds) bring the reticule onto the target, so that hopefully the missile is less aggressive in its dive, it would also cause the missile to expend its energy more efficiently, as it's not having to do such hard manoeuvring. 

 

And FWIW, I'm not sure Rapier uses PN, at least when only using the optical tracker, given that it doesn't know anything about the target, just where an operator wants it to go, though I wouldn't be surprised if it used it with Blindfire (bear in mind, near enough all Rapier footage is probably from the newer FSC/Rapier 2000 systems (identifiable mainly from the new launcher, which has an EO/IR and laser ranging (AFAIK) system in a spherical turret on top, 8 missiles, different uplink antenna and an integrated generator; the FSC Blindfire also has EO/IR instead of just the TV, the system also now has the search RADAR in its own separate trailer), you'll probably be hard pressed to find footage of the system, as it was before Blindfire, and just using the optical tracker).  

Edited January 20, 2021 by Northstar98

[Mad_Shell]

Bump! I can confirm that many of its missiles crash when targetting low altitude aircraft.

[BIGNEWY]

this has been reported

 

thanks