This might sound like a dumb question, but why are there no AI + IR sensitive camera guided missiles that could literally help track the aircraft and ignore chaff and flares?
Clouds wouldn’t be a problem since they would have an IR sensitive camera, size also wouldn’t really be a problem if we used Nvidia’s Jetson Nano. So my question is, Why are there still no missiles guided by AI + IR sensitive cameras?
P.S. My first time on the forum btw.
Modern IIR missiles do use pattern recognition to reject conventional flares iirc so they sort of do exist but it’s not really the same type of “AI” you are talking about
What about the accuracy? Wouldn’t my idea with more advanced AI be even more accurate?
AI, do you mean artificial intellegence?
Then it probably is in development, but I dont think the tech has matured enough to be any better than conventional IIR missiles like ASRAAM, Aim-9X, etc etc. Which are already essentially immune to most countermeasures. Id also imagine that the necessary kit for “AI” would be expensive and heavy and largely a negative attribute for what is normally a small weapon type.
I could see AI being developed and introduced in larger, longer ranged Cruise missiles though
They can already usually target the cockpit of an aircraft for a “hard-kill”. So… Already fairly accurate
Not that we need it, because we’ve already achieved 90% of the remotely positive outcomes that could be achieved with a black box of “probably working right, but who’s to say” without said blackbox of who-bloody-knows. This is before we remotely consider the logistics of it.
Short and long of it, consistency is key and there is no future outside of screwing around for the technology proposed here, unless suddenly they make a version of it that is A) computationally cheap, B) logistically sound to power over other systems, and C) easily corrected in the event it becomes flawed.
In the sense of cruise missiles. I see it being used for adaptive navigation to avoid threats
Issue is, we kinda already done that, at least for ASMs, and it wouldn’t be all too difficult to move that logic over to cruise missiles (assuming we’re splitting hairs and not counting ASMs as cruise missiles).
There is also just making it faster which is generally easier, and as we’ve seen, generally equally effective in defeating point defence systems. The technology sounds great in theory, but the inconsistency it introduces makes it may be worth it, and to my mind, “may be worth it” is not the most sound principle when there are others demonstrated that carry a bit more weight.
More response it does to threats, longer it is in the air, longer it is in the air, longer window of engaging it. This contrasts with making the weapon just really sodding fast, which decreases the window it is in the air, and the window to engage it.
Yeah. That is true, but we’ll have to wait and see what they come up with
There are infra-red missiles resistant to countermeasures in use in real world, adding artificial inteligence to the equation is just more work, initially, I can’t imagine the amount of trouble to make it differ from an enemy and ally aircraft if visual sensors weren’t stupidly expensive to send clear information for recognition, if it’s not broke, don’t fix it. But somewhere down in a bunker or a government office may be someone making this true.
And I might be fearing the future when I see that Helsing is part of this. Jokes apart by the way.
Modern IIR missiles (ASRAAM/AIM-9X/IRIS-T etc) already use algorithms for flare rejection and tracking gates. The task they have isn’t a complex one by modern standards as such dumping some crappy LLM running on a tiny Nvidia board is just a waste of time and effort.
Image recognition is actually used by many or all IIR missiles depending on how you look at it. Even the more basic and older contrast detection IIR method is, for most implementations, a CNN (Convolutional Neural Network, a simply kind of AI). Or on the more advanced end proper image recognition.
There’s actually multiple different kinds of image recognition algorithms used by IIR missiles. Some like the Python-5 use models trained with a list of targets to identify a target from the background clutter. Meanwhile others like the AAM-5B appear to use the original profile of the target as the base for comparing to for maintaining a lock.
In other words, the missile knows where it is.
And the missile knows who you are.
The missile knows where it is at all times. It knows this because it knows where it isn’t. By subtracting where it is from where it isn’t, or where it isn’t from where it is (whichever is greater), it obtains a difference, or deviation. The guidance subsystem uses deviations to generate corrective commands to drive the missile from a position where it is to a position where it isn’t, and arriving at a position where it wasn’t, it now is. Consequently, the position where it is, is now the position that it wasn’t, and it follows that the position that it was, is now the position that it isn’t.
In the event that the position that it is in is not the position that it wasn’t, the system has acquired a variation, the variation being the difference between where the missile is, and where it wasn’t. If variation is considered to be a significant factor, it too may be corrected by the GEA. However, the missile must also know where it was.
The missile guidance computer scenario works as follows. Because a variation has modified some of the information the missile has obtained, it is not sure just where it is. However, it is sure where it isn’t, within reason, and it knows where it was. It now subtracts where it should be from where it wasn’t, or vice-versa, and by differentiating this from the algebraic sum of where it shouldn’t be, and where it was, it is able to obtain the deviation and its variation, which is called error.
What if the AI missile decided that the plane/person/country who launched it, is actually the baddy?
AI is a term used for what is just a LLM. It doesn’t “think”, it’s just weighing of different answers/options/data in order to produce the most likely result per the models data/training.
AGI or Artificial general intelligence is what you would typically think of as a thinking or self aware type AI.
All the “AI” we currently interact with (ChatGP/Co-Pilot) etc are all just LLM’s and not in anyway an AGI.
Ai: Oh hey look, a civilian airliner!
while AI might(or probably will) give better ‘peak’ performance, it’s practically impossible to anticipate any sort of error that the system will produce. so while a system without AI may show a somewhat worse performance, the error that it has can be quantized and accounted for
Nurosama
The AI models you think of are not good at this task at all
There are already system that are much faster and a lot more accurate available on most modern IIR missiles that can recognize what aircraft they are tracking, what range it is at and filter out flares as well as IRCM like the one found on the Ka-52