TR UAV/UCAV Programs | Anka - series | Kızılelma | TB - series

[DBdev]

in terms of a dog fight didn't tests already reveal that AI is more capable then humans?

Maybe a combination of human pilot completely assisted by AI will be better then either alone.

It did but humans were not allowed to think outside the box. Pilot was just there to promote the project.

For example I have heard a story from TEI CEO Mahmut Akşit yesterday.
He was talking how they created Şimşek target drone turbojet engines and delivered 12 of them but never received new orders.

Because our smart soldiers were tying a foam decoy behind the target drone and by painting it with radar reflective paint instead they were saving even the seemingly cheap target drone. Good luck to idiotic Bard or Bing AI outsmarting humans in real world.

 

[Afif]

I don't think it's really clear what "6th gen" is. US doesn't know it,

With all due respect, I like to disagree. Given an NGAD demonstrator has been flown already, I think USA in principle has a conceptual understanding of what they want out of NGAD and the whole system of systems.

 

[Siper>MMU]

Have you seen this plane before? Whats is that ?
View attachment 63603

Don't get hyped. Just a model which already doesn't even look like Kızılelma. I highly doubt they will change the design this much. Also this model has a single vertical stabiliser which is a big no no for stealth.

 

[Samba]

It did but humans were not allowed to think outside the box. Pilot was just there to promote the project.

For example I have heard a story from TEI CEO Mahmut Akşit yesterday.
He was talking how they created Şimşek target drone turbojet engines and delivered 12 of them but never received new orders.

Because our smart soldiers were tying a foam decoy behind the target drone and by painting it with radar reflective paint instead they were saving even the seemingly cheap target drone. Good luck to idiotic Bard or Bing AI outsmarting humans in real world.

WHAT!???

 

[uçuyorum]

WHAT!???

Towed decoy to a target drone now that's pod racing

 

[Strong AI]

For example I have heard a story from TEI CEO Mahmut Akşit yesterday.
He was talking how they created Şimşek target drone turbojet engines and delivered 12 of them but never received new orders.

Because our smart soldiers were tying a foam decoy behind the target drone and by painting it with radar reflective paint instead they were saving even the seemingly cheap target drone.

Starts talking about it at approx. 54:45 min

 

[Rooxbar]

The upper limit of level of intelligence and insight provided by natural language models is the data it's trained with, and that data comes from human minds, so that's that for those models. AI can outdo humans if there's any brute force calculation involved or when it's about doing one simple thing extremely efficiently. If there's decision-making involved in any complex environment, I'm yet to see any advancement, or the theoretical basis for any such claims. The guided trial and error way of ML will eventually be trained to find the best strategies involved even in those complex environments (like the success of DeepMind's AlphaGo, AlphaZero, AlphaFold and AlphaStar) but, to date, that requires unimaginable amounts of data and processing power and training times. There will be improvements in all these fronts over time, but how long that will take is up for debate.

 

[Sanchez]

With all due respect, I like to disagree. Given an NGAD demonstrator has been flown already, I think USA in principle has a conceptual understanding of what they want out of NGAD and the whole system of systems.

Certainly. What is a 6th gen tho, what it entails? Can whatever comes out of NGAD be called 6th gen? 5th gen was networked warfare, stealth and supercruise. 4th gen was fly by wire and higher maneuverability, and first real instance of a multi role fighter. I'm not seeing any great leaps from 5th to 6th for now. Nothing game changer. "6th gen" has become a buzzword, I think that was made clear when Northrop dubbed B-21 as a "6th generation aircraft". That's bs, there have been at least 7-8 generations of aircraft since Wright brothers. It's semantics, yes, but I love semantics.
Edit, ooh, my 1000th message.

edit 2.
I think using fighter gens for drones and future unmanned fighters takes away from their potential, of what could really be possible. We are now in a century of war, costs will have to lower, or we can't have wars. Attritability is the name of the game.
This was a good recent piece.

US needs attritable systems, strategic logistics analysis to win wars

The key to victory will rely on the adoption of weapon systems simplistic in design, rapidly reproducible and highly lethal.

www.defensenews.com

"The idea of securing every inch of a complex weapon system’s supply chain — while an imperative — is perhaps a near-impossible task. A forward-thinking way to solve this problem is to design and produce war machines that are as simplistic as a paper airplane and as easily replaced as a toy tank, but highly lethal."

Maybe 6th gen fighters won't cost 100 million apiece but 20 million apiece. Maybe 6th gen is unmanned and NGAD is just an F-22 replacement, but not a 6th gen, although it will be dubbed that way. If KE is a 50-60 million dolar drone fighter, losing it in combat is still akin to losing an F-16 cost wise. How many 50 million dollar planes can we lose and still call it a victory? 15-20?

 

[Afif]

Certainly. What is a 6th gen tho, what it entails? Can whatever comes out of NGAD be called 6th gen? 5th gen was networked warfare, stealth and supercruise. 4th gen was fly by wire and higher maneuverability, and first real instance of a multi role fighter. I'm not seeing any great leaps from 5th to 6th for now. Nothing game changer. "6th gen" has become a buzzword, I think that was made clear when Northrop dubbed B-21 as a "6th generation aircraft". That's bs, there have been at least 7-8 generations of aircraft since Wright brothers.

It's semantics, yes, but I love semantics.

Edit, ooh, my 1000th message.

Well, yes. I can certainly agree with that. The leap from 5th gen to 6th gen seems nowhere as big as it was from 4th gen to 5th gen. Who know? perhaps something will come out later, like NGAD carrying HCM in its weapon bay or shooting 300kw laser?

And btw congratulation on your 1000th massage.

 

[Rodeo]

Well, yes. I can certainly agree with that. The leap from 5th gen to 6th gen seems nowhere as big as it was from 4th gen to 5th gen. Who know? perhaps something will come out later, like NGAD carrying HCM in its weapon bay or shooting 300kw laser?

I think @Windchime explained the issue well in one of his posts.

So, which feature could warrant a fighter jet "6th gen" nameplate? A generational difference indicates a difference that fundamentally cannot be achieved by previous generations. Well, unless you significantly alter the origonal design, which at that point I'll rather call it a different plane.

For example a 1st gen fighter could never achieve supersonuc speeds on level flight or doing combat maneuvers. 2nd gen can't BVR since most of them don't have the provisons for the necessary avionics, nor the generator capacity. 3rd gen distributed avionics cannot be replaced with a federated one. 4th gen isn't stealth. All of those are what I would call a generational difference.

In that sense it is very hard to distinguish a feature that cannot be implemented onto the current 5th generational figthers, or more to be exact the F-35 and J-20 and those which could be. DEW or man-unmanned teaming often gets brought up, but those are things that the USAF are considering for future F-35 blocks. AI is already implemented to a certain degree, and a higher sophistication could be achieved with a better computer and software, which aren't irreplaceable for an integrated modular design of F-35. VCE is also being considered for block-4 with existing prototypes. Almost every other features brought up in this thread are already being implemented to 5th gen or could be implemented.

One thing I could think of is a wide-band stealth technology, or in simpler terms removing, or at least reducing vertical planes of the aircraft. Then again this was already explored with the likes of YF-23.

My suggestion would be to just wait for NGAD to arrive. US has been the pioneer in this regard for decades now and will almost certainly suggest us what the definition of 6th gen might or should be.

One thing's for certain though, and that's that the MMU cannot be 6th gen for obvious reasons. So are F-35, J-20 or J-35. I think there's a high chance that the GCAP and NGF are rather 5.5th ven fighters, not 6th gen when they arrive. Same thing happened in the 80s and 90s when the US was busy developing ATF while Europe was still stuck with EAP. No one can say for certain that same thing won't happen again in 2020s and 30s.

And btw congratulation on your 1000th massage.

He beat me to it.

 

[GoatsMilk]

It did but humans were not allowed to think outside the box. Pilot was just there to promote the project.

For example I have heard a story from TEI CEO Mahmut Akşit yesterday.
He was talking how they created Şimşek target drone turbojet engines and delivered 12 of them but never received new orders.

Because our smart soldiers were tying a foam decoy behind the target drone and by painting it with radar reflective paint instead they were saving even the seemingly cheap target drone. Good luck to idiotic Bard or Bing AI outsmarting humans in real world.

when i think of AI for fighter jets in dog fight situations, all I'm thinking is there is number of possible variables to do with your plane and the enemies plane in terms of how it moves and when it can be locked onto. You create enough algorithms to anticipate and calculate all of these variables in real time collectively giving the plane the best possible course of action at every moment.

I don't think AI as something that is sentient, just a tool that allows precise "rapid" calculations in limited situations where most if not all variables can be defined prior to the engagement.

But personally i wouldn't know and you could be completely right, its not something I've ever studied.

 

[uçuyorum]

when i think of AI for fighter jets in dog fight situations, all I'm thinking is there is number of possible variables to do with your plane and the enemies plane in terms of how it moves and when it can be locked onto. You create enough algorithms to anticipate and calculate all of these variables in real time collectively giving the plane the best possible course of action at every moment.

I don't think AI as something that is sentient, just a tool that allows precise "rapid" calculations in limited situations where most if not all variables can be defined prior to the engagement.

But personally i wouldn't know and you could be completely right, its not something I've ever studied.

No you can't know and calculate all variables prior to engagement, that is the issue. That's why we have AI, is a different approach than regular algorithms. When the search space created by variables are so large that you can't possibly search for actual optimal actions, you use methods and tricks that make smart decisions. This comes in different forms, there are heuristic searches and there are machine learning algorithms that learn from data and can generalize from there.

Large language models are as we say, very large, they learn from a lot of data and define a lot of internal parameters that allows them to process anything that is coded to a language to really high level. At this point they start being able to come up with elaborate plans and ideas as well.

 

[dBSPL]

in terms of a dog fight didn't tests already reveal that AI is more capable then humans?

Maybe a combination of human pilot completely assisted by AI will be better then either alone.

This is truly inevitable. While your limit in a manned aircraft is the human factor, in an unmanned jet it is theoretically possible to increase this limit to the structural strength limits of the aircraft. The aircraft can have a high maneuverability up to the limit where it can remain in one piece. The other dimension of the issue is artificial intelligence, which I think means a lot, even though current advances have only demonstrated the capacity in this field in a ground simulation environment! I think 6th generation fighter jets will inevitably have to be unmanned machines, and for the same reason, we will start to see more and more unusual, ugly geometries.

 

[GoatsMilk]

This is truly inevitable. While your limit in a manned aircraft is the human factor, in an unmanned jet it is theoretically possible to increase this limit to the structural strength limits of the aircraft. The aircraft can have a high maneuverability up to the limit where it can remain in one piece. The other dimension of the issue is artificial intelligence, which I think means a lot, even though current advances have only demonstrated the capacity in this field in a ground simulation environment! I think 6th generation fighter jets will inevitably have to be unmanned machines, and for the same reason, we will start to see more and more unusual, ugly geometries.

When a striker is on the attack the moment he decides to dribble, strike the ball, pass the ball or cross the ball, even go back is instead decided by AI.

I would suspect that given the parameters of the playing field and the limitations of both the player and the limitations of the opposing players AI will be able to calculate all of the variables to a degree of accuracy that a player cannot.

So a player may think to shoot because he sees a chance on goal. However when we watch football we may see this player has a bad long distance shot, or had he taken simpler option and passed to a team mate he would have scored easily, or even taking the shot while the entire team is in attack once deflected or given back to the opposing team sets up a counter account leading to your own team conceding.

AI i would presume would be able to calculate all of these variables instantaneously and then take the best course of action. And the best course of action would be what you decide it to be. So AI would conclude the shot is unlikely to lead to a goal and would instead to lead to an opposing counter attack. The pass to the team mate may also be calculated as risky and instead the safest and best course of action is to play the ball back. And then once the ball is played back the AI calculates the options and tries to anticipate the options available moving forward. So what is likely to occur several passes ahead.

In the open air this is how i would imagine AI could help a dog fight. And in terms of on the fly learning this would occur to just like when i play football you soon pick up what others can and cannot do and you adjust accordingly. For example if a player is a great dribbler you make sure to be on him before he's received the ball so that you can intercept before he takes control. Or if a player likes to shoot and he's terrible you always give him just enough room to take a shot compared to a player who has high accuracy you close him and always try to black off his shot. In the air open air calculating the variables associated with it should be doable i would think and i would think AI could for the most part do this better then a human with his limited senses.

 

[Cabatli_TR]

Akinci with AESA

https://twitter.com/i/web/status/1732703560196837795

 

[DBdev]

For me, megawatt laser weapons that somehow don't bloom are a must for 6th gen.

If you have lasers you won't need fast speeds or AI team ups or any of the other not so new crap. We have heard NGAD broke SR71's records but I suspect it was for mere seconds. Considering how fuel hungry ramjets are.

AI will not outsmart humans anytime soon.

That leaves us with my fav, laser weapons! Both for air combat and ground attacks. Yes, yes, I know... No way a jet engine can possibly power a megawatt laser, or a fighter jet can possibly carry it etc. etc. But still until laser weapons on planes become real everything else will still be only 5+ gen IMHO.

Lasers are perfect for space wars since there are no pesky air molecules that cause them to weaken via bloom by always being in the way. In fact low energy laser communication is already being used between far apart satellites.

Now, how do we make science fictional laser weapons science and reach similar distances without blooming?
We somehow have to get all the air molecules out of the way momentarily. Apparently that is possible with plasma windows. Artificially between strong magnets in tokamaks etc., naturally when a lightning strikes. So why not use nature's MASSIVE existing energy to create a straight lightning which not only strikes the target with electricity, hot plasma, but also with our fighter jet's own laser which suddenly finds itself a plasma window without pesky air molecules.

Plasma window - Wikipedia

en.wikipedia.org

"the plasma window will allow radiation such as lasers and electron beams to pass. This property is the key to the plasma window's usefulness — the technology of the plasma window permits for radiation that can only be generated in a vacuum to be applied to objects in an atmosphere"

And this is how we can create a straight lightning and a plasma window for our laser.

Yes yes, I know a rocket and a wire? Seems redundant. But if this is possible better solutions without rockets or even lightning should be possible too. It is only impossible until someone invents a brand new way to make it possible.

 

[Zafer]

Lasers are perfect for space wars since there are no pesky air molecules that cause them to weaken via bloom by always being in the way. In fact low energy laser communication is already being used between far apart satellites.

Now, how do we make science fictional laser weapons science and reach similar distances without blooming?
We somehow have to get all the air molecules out of the way momentarily. Apparently that is possible with plasma windows. Artificially between strong magnets in tokamaks etc., naturally when a lightning strikes. So why not use nature's MASSIVE existing energy to create a straight lightning which not only strikes the target with electricity, hot plasma, but also with our fighter jet's own laser which suddenly finds itself a plasma window without pesky air molecules.

Plasma window - Wikipedia

en.wikipedia.org

"the plasma window will allow radiation such as lasers and electron beams to pass. This property is the key to the plasma window's usefulness — the technology of the plasma window permits for radiation that can only be generated in a vacuum to be applied to objects in an atmosphere"

And this is how we can create a straight lightning and a plasma window for our laser.

Yes yes, I know a rocket and a wire? Seems redundant. But if this is possible better solutions without rockets or even lightning should be possible too. It is only impossible until someone invents a brand new way to make it possible.

There was laser triggered lightning work in the past to initiate lightning and store the energy in the ground batteries. So laser can cause ionisation of the air which is similar to plasma but colder, afaik. This is not in Türkiye.

 

[DBdev]

Sadly our Tubitak. The entity that should be giving our engineers brand new ways to create their inventions is just a google translator service. They just translate the English language research papers. Or keep repeating their most useless experiments.

 

[Bogeyman]

https://twitter.com/i/web/status/1732751402198335498

 

[Afif]

https://twitter.com/i/web/status/1732751402198335498

@Yasar will like it.