Chinese UCAV/CCA/flying wing drones (ISR, A2A, A2G) thread

[latenlazy]

[

My opinion is that if the Chinese MIC can support different specializations without financial strain, go ahead. I fear a jack of all trades situation if they try cramming too much capabilities in each individual drone, which has small internal space to start with.

It would actually defeat the point to try to make your drones too multirole. Piling on complexity into single platforms slows down the pace of iterative development and multiplies program risk and also dimensions of compromise. Every time you fly for a specific mission there are capabilities you’re not using. Multirole became the norm because it was better for cost management in a manned fleet context. If you can tackle the cost management side orthogonally, which is what unmanned should allow you to do, specialization combined with modularity should deliver much better and more tailored mission packages.

 

[Temstar]

So, they're developed enough to be flying (and upsetting PLAAF pilots with their ability) but not refined enough for deployment. I think that would put the program conservatively 4-5 years before deployment readiness?

I figured it's something like that because either last week or the week before when Guancha gang was talking about what it means to be "in service" as far as parade goes they raised the idea that if something is in PLA hands and they play with it on the regular, can that not be said to be "in service" by some definition? I don't think they specified aircraft but the language sounded like they were implying aircraft.

I even posted in the parade thread at the time after hearing that suggesting that maybe due to the unique nature of CCA and UCAV meant that they can be displayed in parade earlier since they're not allowed to fly in the aerial parade and if they're going to be shown on the back of a truck you can display earlier in their life cycle. I had in mind GJ-11 been shown in 2019 where maybe the aircraft isn't quite ready to fly. But by similar logic you can also display UADF earlier while they're still training their AI since they're doing it under PLAAF care.

How is that comforting? These things are literally gunning for their jobs.

Yankee and Shilao humorously suggested that AI needs a separate parallel reward program where instead of getting medals and commendations they get something else for scoring simulated combat kill. You see people making jokes in the scrolling text on screen of how the AI should be rewarded with another RTX 5090 installed if it becomes a simulated air combat ace.

They also raised some interesting prospect about AI pilots. Every time a PLAAF pilot figures out a way to beat one the knowledge of that event gets added to the AI's data structure, and with a data sync all the other instance of the AI also learns about it unlike human pilots who have to do classes and presentation to peers to dissimulate the knowledge. Depending on who trained the AI you might get different AI "personalities" and quirks. They suggested maybe future export model UADF could have a "Peshawar trained AI" as opposed to PLAAF trained.

I was having a discussion with someone about UASF and basically my point is that they could very well be more effective used in MUMT combat scenario than F-35s. Essentially, they are stealthier from frontal and side for sure and they have better sensors and probably has better supersonic performance. So if you have a few controller manned aircraft sitting further back that is conducting EW and synthesizing sensor data, then they can control the nearest UASF drone to do A direct B shoot kind of scenario and then getting closer to get full lock on to a target F-35 with weapon grade tracking (or maybe it doesn't need to if the quality of AAM seeker keeps improving) and then several nearby drones + manned aircraft all perform electronic attack on the target F-35 and leave it unable to track incoming missiles.

So, i'm not sure about 6). That seems quite unlikely.

Being networked missile carrier sounds like that should be a CCA Gen 1 kind of thing instead of CCA Gen 2/UADF?

 

[plawolf]

I was having a discussion with someone about UASF and basically my point is that they could very well be more effective used in MUMT combat scenario than F-35s. Essentially, they are stealthier from frontal and side for sure and they have better sensors and probably has better supersonic performance. So if you have a few controller manned aircraft sitting further back that is conducting EW and synthesizing sensor data, then they can control the nearest UASF drone to do A direct B shoot kind of scenario and then getting closer to get full lock on to a target F-35 with weapon grade tracking (or maybe it doesn't need to if the quality of AAM seeker keeps improving) and then several nearby drones + manned aircraft all perform electronic attack on the target F-35 and leave it unable to track incoming missiles.

So, i'm not sure about 6). That seems quite unlikely.

I think there will be many layers to how the UADFs will work with manned 5th gens and later 6th gens.

Indeed, I think there will be a lot of overlap and similarities between how 5th gens and 6th gens will operate alongside UADFs, with the biggest practical differences being you will need multiple 5th gens working together to get a similar effect as what a single 6th gen can do by itself.

At the most basic and furthest level, the UADFs will operate similar to manned 5th gens and prefer BVR over dogfights despite its great focus on dogfighting capabilities.

I think they used the F22 as an example because that’s one of the few opponents that might have any meaningful chance of weathering or bypassing the BVR layer to even get within WVR combat. At which point, with BVRAAMs exhausted, the UADFs will boost ahead to engage the surviving Raptors in WVR combat, with the J20s turning for home, or following up to mop up at their leisure and discretion.

At that point, it’s already a win for the PLAAF irrespective of how the Raptors fare against the UADFs, since the Americans at that point are essentially trading their top of the line fighters for China’s bottom tier air combat assets in terms of value.

My ranking is manned 5th gens top, manned legacy mid, and unmanned bottom. This is based on replacement price rather than pure combat effectiveness, since I expect the UADFs to rank higher than legacy manned jets in AA combat capabilities. But on the grand scheme of things, assuming no ridiculously lopsided combat exchange rates, it’s going to be force replacement and expansion capabilities that will win the war.

Also worth noting is that in WVR, they seem to be only releasing stories about 1 v 1 fights between manned fighters and UADFs. Whereas I think the biggest difference between manned and unmanned fighters in WVR combat will be how they perform as the number of combatants increase.

With UADF, especially with drone swarm tech, I would expect the combined combat capabilities of UADFs to grow disproportionately greater to increases in numbers of combatants, potentially even to an exponential rate. This is because with drones, you can programme them to fight as a team much more effectively and efficiently than human pilots can, where UADFs can set up the bogy they are chasing for easy kill shots by other UADFs coming in from a totally unexpected vector as the human pilot is forced to keep focus on the original UADF engaging him. And as the number of combatants increase in a given airspace, the human pilot will become quickly overwhelmed with the number of aircraft flying to the point that just basic safe flying might become a challenge. OTOH, if they employ drone swarm tech and design principles, UADFs could become more effective as the number of combatants increase and the UADFs share processing power as well.

 

[totenchan]

I think there will be many layers to how the UADFs will work with manned 5th gens and later 6th gens.

Indeed, I think there will be a lot of overlap and similarities between how 5th gens and 6th gens will operate alongside UADFs, with the biggest practical differences being you will need multiple 5th gens working together to get a similar effect as what a single 6th gen can do by itself.

At the most basic and furthest level, the UADFs will operate similar to manned 5th gens and prefer BVR over dogfights despite its great focus on dogfighting capabilities.

I think they used the F22 as an example because that’s one of the few opponents that might have any meaningful chance of weathering or bypassing the BVR layer to even get within WVR combat. At which point, with BVRAAMs exhausted, the UADFs will boost ahead to engage the surviving Raptors in WVR combat, with the J20s turning for home, or following up to mop up at their leisure and discretion.

At that point, it’s already a win for the PLAAF irrespective of how the Raptors fare against the UADFs, since the Americans at that point are essentially trading their top of the line fighters for China’s bottom tier air combat assets in terms of value.

My ranking is manned 5th gens top, manned legacy mid, and unmanned bottom. This is based on replacement price rather than pure combat effectiveness, since I expect the UADFs to rank higher than legacy manned jets in AA combat capabilities. But on the grand scheme of things, assuming no ridiculously lopsided combat exchange rates, it’s going to be force replacement and expansion capabilities that will win the war.

Also worth noting is that in WVR, they seem to be only releasing stories about 1 v 1 fights between manned fighters and UADFs. Whereas I think the biggest difference between manned and unmanned fighters in WVR combat will be how they perform as the number of combatants increase.

With UADF, especially with drone swarm tech, I would expect the combined combat capabilities of UADFs to grow disproportionately greater to increases in numbers of combatants, potentially even to an exponential rate. This is because with drones, you can programme them to fight as a team much more effectively and efficiently than human pilots can, where UADFs can set up the bogy they are chasing for easy kill shots by other UADFs coming in from a totally unexpected vector as the human pilot is forced to keep focus on the original UADF engaging him. And as the number of combatants increase in a given airspace, the human pilot will become quickly overwhelmed with the number of aircraft flying to the point that just basic safe flying might become a challenge. OTOH, if they employ drone swarm tech and design principles, UADFs could become more effective as the number of combatants increase and the UADFs share processing power as well.

I see no reason that the larger CCAs would necessarily have a lower replacement price than the manned legacy aircraft, at least for the foreseeable future. Things might change if you try to price in the pilot of course, but I'm not sure how you'd do that.

 

[dingyibvs]

Still a lot cheaper and more scalable than a manned multirole fighter. You’re trading some flexibility for modularity. That’s the value proposition.


Keep in mind that active contestation of the electromagnetic magnetic environment makes this kind of tactical employment concept far from a sure fire bet. Manned contingent will likely need to be an active part of the picture as redundancy and countermeasure mitigation for quite a while longer.

I would think that for aerial combat laser based communications might be better. Atmosphere is thinner and you might get hundreds of kms of range.

 

[latenlazy]

I would think that for aerial combat laser based communications might be better. Atmosphere is thinner and you might get hundreds of kms of range.

Tight beam might be a decent countermeasure comms method against EW interference but probably restricts employment flexibility.

 

[dingyibvs]

I see no reason that the larger CCAs would necessarily have a lower replacement price than the manned legacy aircraft, at least for the foreseeable future. Things might change if you try to price in the pilot of course, but I'm not sure how you'd do that.

It may not even be a matter of price. Near the end of WWII Japan had more aircrafts than capable pilots. You just might not be able to train pilots fast enough for replacement especially given how complex modern aircrafts are to operate.

Tight beam might be a decent countermeasure comms method against EW interference but probably restricts employment flexibility.

Autonomy will be key here IMO as it would limit the amount and timing of communications needed.

 

[discspinner]

unmanned ADF by definition would prove most irreplaceable in situations where hitting a high value target at max range (tankers) that would require a one way trip, but would in cost benefit analysis be highly desirable, not only due to the ratio eg 1 UADF : 1 tanker, but any collateral losses from F35/F22s ditching for lack of fuel; though this would not be as critical if the CJ-1000 can really hit slow strategic airborne targets at 5000km

 

[ChineseToTheBone]

I see no reason that the larger CCAs would necessarily have a lower replacement price than the manned legacy aircraft, at least for the foreseeable future. Things might change if you try to price in the pilot of course, but I'm not sure how you'd do that.

Yeah pricing in the pilot is something that decisionmakers are certainly taking into account here, given how training each combat pilot incurs new costs and how they eventually retire after some years in the service.

Recently the think tank Mitchell Institute for Aerospace Studies estimated that Chinese fighter jet pilots were getting approximately 200 flight hours per year in contrast to roughly 120 flight hours per year for Americans, so we can use that for a rough baseline to extrapolate some assumptions on training costs. Let us first take your assumption that unmanned air dominance fighters have similar unit costs as legacy fighter jets and have similar maintenance costs of maybe ¥100,000 per hour to randomly use here for my examples.

Two squadrons of loyal wingmen of 40 in total sparring against each other at the side of manned fighter jets could be a minimum for basic collaborative autonomous training.
¥100,000 maintenance costs per flight hour × 200 flight hours × 40 unmanned fighters = ¥800,000,000 per year spent developing limited scale air combat strategies
¥800,000,000 would be the minimum annual fixed cost here and every new loyal wingmen from that point on benefits from the strategies developed over every single year of collaborative autonomous training.

May 7th had involved over 70 manned fighter jets on the Indian side and can be used as the example of a more intense training regiment for how major conflicts would play out for a higher number like 150 unmanned air dominance fighters total being used for training.
¥100,000 maintenance costs per flight hour × 200 flight hours × 150 unmanned fighters = ¥3,000,000,000 per year spent developing new large scale air combat strategies
¥3,000,000,000 would be my higher annual fixed cost here in this hypothetical future to develop institutional knowledge for utilizing loyal wingmen in major conflicts.

Chinese fighter jet pilots of course have valuable institutional knowledge as well but that will one day be miniscule compared with the library of knowledge that unmanned air dominance fighters and loyal wingmen could accumulate through shared data. Every single new loyal wingmen comes programmed with all the experience gathered from the air combat exercises done by their predecessors and thus it has a flat cost not requiring additional training with just the need for keeping them airworthy not unlike maintaining a stockpile of missiles.

In the other hand we have yet to calculate for the annual fixed cost of roughly 2000 manned fighter jets operated by China.
¥100,000 maintenance costs per flight hour × 200 flight hours × 2000 combat pilots = ¥40,000,000,000 per year spent on providing training alone
¥40,000,000,000 is still money well spent to maintain what is the second largest air force in the world, but with just less than 10% more you could feasibly get real training for decently capable and ever improving unmanned air dominance fighters where one collaborative autonomous improvement means an improvement for all.

Please understand this calculation does ignore the need for combat pilots to all eventually familiarize themselves with flying alongside loyal wingmen and thus require more squadrons of them dispersed across the country, but my assumption of drilling multiple squadrons of unmanned air dominance fighters that hard is already quite maximalist for a near future hypothetical.

 

[latenlazy]

Yeah pricing in the pilot is something that decisionmakers are certainly taking into account here, given how training each combat pilot incurs new costs and how they eventually retire after some years in the service.

Recently the think tank Mitchell Institute for Aerospace Studies estimated that Chinese fighter jet pilots were getting approximately 200 flight hours per year in contrast to roughly 120 flight hours per year for Americans, so we can use that for a rough baseline to extrapolate some assumptions on training costs. Let us first take your assumption that unmanned air dominance fighters have similar unit costs as legacy fighter jets and have similar maintenance costs of maybe ¥100,000 per hour to randomly use here for my examples.

Two squadrons of loyal wingmen of 40 in total sparring against each other at the side of manned fighter jets could be a minimum for basic collaborative autonomous training.
¥100,000 maintenance costs per flight hour × 200 flight hours × 40 unmanned fighters = ¥800,000,000 per year spent developing limited scale air combat strategies
¥800,000,000 would be the minimum annual fixed cost here and every new loyal wingmen from that point on benefits from the strategies developed over every single year of collaborative autonomous training.

May 7th had involved over 70 manned fighter jets on the Indian side and can be used as the example of a more intense training regiment for how major conflicts would play out for a higher number like 150 unmanned air dominance fighters total being used for training.
¥100,000 maintenance costs per flight hour × 200 flight hours × 150 unmanned fighters = ¥3,000,000,000 per year spent developing new large scale air combat strategies
¥3,000,000,000 would be my higher annual fixed cost here in this hypothetical future to develop institutional knowledge for utilizing loyal wingmen in major conflicts.

Chinese fighter jet pilots of course have valuable institutional knowledge as well but that will one day be miniscule compared with the library of knowledge that unmanned air dominance fighters and loyal wingmen could accumulate through shared data. Every single new loyal wingmen comes programmed with all the experience gathered from the air combat exercises done by their predecessors and thus it has a flat cost not requiring additional training with just the need for keeping them airworthy not unlike maintaining a stockpile of missiles.

In the other hand we have yet to calculate for the annual fixed cost of roughly 2000 manned fighter jets operated by China.
¥100,000 maintenance costs per flight hour × 200 flight hours × 2000 combat pilots = ¥40,000,000,000 per year spent on providing training alone
¥40,000,000,000 is still money well spent to maintain what is the second largest air force in the world, but with just less than 10% more you could feasibly get real training for decently capable and ever improving unmanned air dominance fighters where one collaborative autonomous improvement means an improvement for all.

Please understand this calculation does ignore the need for combat pilots to all eventually familiarize themselves with flying alongside loyal wingmen and thus require more squadrons of them dispersed across the country, but my assumption of drilling multiple squadrons of unmanned air dominance fighters that hard is already quite maximalist for a near future hypothetical.

Pilots are not just additional institutional costs but scaling bottlenecks for force capacity.