reported earlier RWR way to sensitive (or AI radar's just voodoo powered.)

[norman99]

On multiplayer servers, I always receive RWR lock & launch warnings from AI aircraft 70nm-100+nm away, as they engage other aircraft, so long as they are pointing roughly in my direction (+/- 45 degrees). Whilst I understand getting false warnings if I'm very close to the aircraft actually targeted, what currently happens seems very unrealistic.

Below is an example of receiving a lock RWR warning sitting on a carrier deck, whilst the source SU27 engages an F-14 70nm away.

On a multiplayer server so I don't have a track. I don't seem to get this in single player.

Not sure if it's a Hornet RWR problem, or a general DCS radar/AI issue.

Edited January 15, 2022 by norman99

[Bunny Clark]

This is a known issue. Right now the lock/launch warning cone for aircraft is very large in multiplayer servers. It's some kind of technical / net code optimization thing that'll hopefully get reworked eventually. 

[norman99]

Hopefully..... eventually..... Sounds promising. Haha. 

[norman99]

Wow, just found out this bug is over 3 years old!

Surely no one actually thinks that this is acceptable?

ED’s disregard for things that are just ‘too hard’, no matter how important, never ceases to astound me.

Edited January 15, 2022 by norman99

[WobblyFlops]

3 hours ago, norman99 said:

Surely no one actually thinks that this is acceptable?

 

Oh boy, you've stirred up the Hornet's nest. While ED accepts that this is a bug (since it's a netcode bug it is likely going to take a very long time until it's fixed) a lot of people very often argue that this is 'realistic' because the RWR would pick up the sidelobes from the radar. In terms of DCS, this is 100% an unintended bug since side lobes aren't even simulated, especially not for AI and when talking about realistic behaviour, we'd have to specifically address this on a case by case basis utilizing open source information about a specific RWR system. 

[norman99]

Oh ok, thanks for the background.

I understand the logic behind it being somewhat realistic, but it can’t be to the extent that is currently experienced. Aircraft basically facing 90° from me, 100nm away causing a launch warning in my aircraft that is parked on the carrier deck just doesn’t seem right.

If attempting to operate realistically, and therefore treat launch warnings as real (even when you kinda know it’s not), it just constantly ruins missions as you have to react accordingly. The other option is to view enemy aircraft on the F10 map an check it’s a false warning. Such an immersion killer.

Seems like it’s a challenge to fix, but I’m really surprised there isn’t more noise made about this issue.

Anyway, seems enough was said in the previous thread, so I’ll leave it be for now. Maybe if we come back in another 3 years things may have improved….

Edited January 15, 2022 by norman99

[Swift.]

6 hours ago, WobblyFlops said:

Oh boy, you've stirred up the Hornet's nest. While ED accepts that this is a bug (since it's a netcode bug it is likely going to take a very long time until it's fixed) a lot of people very often argue that this is 'realistic' because the RWR would pick up the sidelobes from the radar. In terms of DCS, this is 100% an unintended bug since side lobes aren't even simulated, especially not for AI and when talking about realistic behaviour, we'd have to specifically address this on a case by case basis utilizing open source information about a specific RWR system. 

Yeah it seems the Bug is that Player radars behave differently to AI radars. If either one goes either way you can excuse it as DCS not modelling real life perfectly or DCS modelling RWR ambiguity, just so long as players and AI behave the same way.

[WobblyFlops]

22 minutes ago, Swiftwin9s said:

modelling RWR ambiguity, just so long as players and AI behave the same way.

But the current way it's 'modelled' for AI points towards a bug since it can trigger even if there's no way to be hit by any sidelobes. I'm all for implementing inaccuracies like physical blanking of the RWR antennas, actual sidelobe detection (with simulated sidelobes from the radars themselves), mippling, worse azimuth resolution and so on. These are documented effects and it would greatly enhance the fidelity of the RWR simulation to not have them be essentially perfect sensors on par with a radar tactically in many situations. But unfortunately the way it behaves currently when it comes to these erroneous spikes makes it clear that we're dealing with a bug.

[Swift.]

41 minutes ago, WobblyFlops said:

But the current way it's 'modelled' for AI points towards a bug since it can trigger even if there's no way to be hit by any sidelobes. I'm all for implementing inaccuracies like physical blanking of the RWR antennas, actual sidelobe detection (with simulated sidelobes from the radars themselves), mippling, worse azimuth resolution and so on. These are documented effects and it would greatly enhance the fidelity of the RWR simulation to not have them be essentially perfect sensors on par with a radar tactically in many situations. But unfortunately the way it behaves currently when it comes to these erroneous spikes makes it clear that we're dealing with a bug.

I suppose you could see it that way, or just that the AI radars have a ridiculous beam width. A friend of mine tested it a while ago and I think there were some that had a 'beam width' of 90 degrees either side. 

But yeah, on the whole. ED Gib proper RWR modelling

[Northstar98]

1 hour ago, Swiftwin9s said:

I suppose you could see it that way, or just that the AI radars have a ridiculous beam width. A friend of mine tested it a while ago and I think there were some that had a 'beam width' of 90 degrees either side.

Yep, this is the likely culprit, I can't seem to find the value in the .lua datamine, but I remember the beamwidth being something ridiculous, like 90/180°, it should be a small fraction of that.

And yes, while sidelobes are a thing, you'll be receiving a small fraction of the received power compared to the mainlobe for our RADARs.

1 hour ago, Swiftwin9s said:

But yeah, on the whole. ED Gib proper RWR modelling

Make that better RWR modelling and better AI RADAR modelling.

Edited January 15, 2022 by Northstar98

[WobblyFlops]

15 minutes ago, Northstar98 said:

and better AI RADAR modelling.

And just in general better radar modelling. Proper channelization, dynamic RCS values based on loadout and aspect, SNR based performance in the notch for radar with digital receiver, channelization, atmospheric effects, PRF ambiguities, the aformentioned sidelobe modelling, every radar utilizing a probability of detection value for a contact like the Mirage, adjustable antenna gain/noise floor through the API, imperfect azimuth resolution and proper doppler proccessing, mutual interference and I could go on and on. 

 

And if we're at it, the Hornet's radar requires a substantial rework. The amount of bugged or imporperly simulated functions are really really high, even when we're talking about basics like trackfile correlation or radar memory logic. Having an API that makes it possible and easy for other developers (and ED themselves) to elevate the sensor modelling fidelity is another very important part of the future of DCS, considering it's full of aircraft that highly revolve around sensor usage. And I don't just mean radar or RWR but things like the upcominc FLIR rework (TV missiles would also require some tune ups when it comes to realistic behaviour anyway), more limited targeting pods, IFF and better jamming simulation.

[BIGNEWY]

this is already reported and is a core issue in DCS. 

thanks