AI in games has nothing to do with an AI that would control an aircraft.
Actually that is not true at all... the AI for just flying is already in use and is called auto pilot on most aircraft.
For realism in a computer game the enemy AI need to be dumbed down because they are part of the system and as such know what is happening everywhere else and what you are doing because it sees your inputs.
In a real life system it would be collecting information around it with passive sensors and receiving target information from aircraft it is operating with... it should be smart enough to hold formation and fly with the aircraft controlling it, but in combat could be sent ahead and climb up and just listen for threats... the aircraft it is operating with could then scan the airspace for targets and this drone could use its radar and direct communication with the scanning aircraft to form a synthetic radar antenna that is rather big.... the drone might get reflections from the target that would not reach the aircraft that scanned because stealth is all about redirecting radar signals away from the radar of origin... even 20 degrees off will make you less visible to that radar... but not other radars nearby.
The drone can pass information about radar returns and IR target it detects... if the Su-57 detects a target he can choose to accelerate and climb and launch a missile himself or direct his forward deployed drone to launch a missile... the jobs the drone will do are not complex and it can have built in algorithms ... like if a missile actively lights you up with ARH radar then start to jam, release chaff and manouver and warn other aircraft you are under attack and perform a manouver.... if it detects an incoming missile with no radar it can use DIRCM to try to defeat optical homing and try a steep dive or other manouver...
Even a drone like a TB2 or Orion that basically flies straight is a pain in the ass to produce.
Actually the biggest problems are decent engines of a suitable power and fuel consumption to get the endurance and range, but also decent datalinks to allow communication and transmission of live video feeds from multiple cameras and sensors, but most of all it is cameras... you average 1.3 megapixel phone camera from the 2000s is simply not good enough... you need fully stabilised high mag high quality video for it to be able to operate 5-10 km above the ground yet still be able to see targets on the ground clearly.
The autopilot and navigation stuff they already put in normal planes.
And fly by wire fighter planes already are fully digital flight commands anyway... whether you are in the cockpit or a van thousands of kms away... the aircraft does not know and most of the time it flys on autopilot anyway... the controller operates the cameras and makes changes to the waypoints on the autopilot.
There will likely be extra features like an orbit button you can press if you see something interesting, when you push it the aircraft starts flying a circle over that spot while you use the cameras to get a closer look and the thing of interest.
Push the button again to continue with the previous flight plan or change the flight plan on the fly.
You would have emergency climb buttons in case you start seeing AA fire that worries you, but most of the time on recon UAVs by the time you see a missile it is probably too late.
With a fighter UAV you need to be able to detect incoming missiles... being able to manouver hard or release decoys or jammers will mean you will be much more survivable than most drones, but essentially you will actually be a missile/bomb truck most of the time.
As LMFS mentioned... UCAVs don't have guns generally... though in some roles in COIN operations where the enemy don't have good AD a gun pod might be useful against a column of enemy light trucks etc...
The software would be like 1000 times more complicated than the one in the f-35. Those things of 6th generation super fighter controled by AI are a joke.
The AI for the S-70 will be more for a missile truck and support aircraft than a dog fighter which simplifies things... it will be able to do a range of very simple things and perform very specific roles as commanded by its lead aircraft.
G force impacts aircraft also. Go more than 12 Gs in modern aircraft for few secobd and it will break in pieces or be so damaged that it can't fly anymore. It's not like for missiles.
The fact that electronics in Missiles can handle much higher G forces is the hint.
Avionics in fighter planes with people in them would never need to endure more than 12g or so... simply because the human occupants couldn't survive that anyway. A missile might be rated for 60g or more because when you are moving at mach 4 plus then the speed massively adds g force in a turn... the faster you go the higher the g force.
Try it in your car... in a (empty) car park put the car in first or drive but just let it idle forward...don't push the accelerator... turn the wheel as far as it will go and you will find yourself turning very tight circles... but no noticable g force... because you are travelling so slow you have very little inertia... a WWI biplane didn't pull 9g because it was not fast enough to pull 9g.
On the road at higher speed you will notice when the road curves on a motorway you feel yourself wanting to continue straight as the car turns... it becomes more noticable the faster you go.
Don't speed and wear your seat belt.
The life you save could be your own.
Agree . That is why the UCAV needs the man in the loop. At least for foreseeable future . I think that even with modern 3D Radar and image recognition by computer , that AI can not compete with human operator yet.
So in the attack mode , the operator takes the lead , only using AI to fix the flight or firing parameters by drone itself . Such that can not be done from the ground or air , by operator. Since local visual detection may be the only reliable means of identification for low RCS planes.
In a modern IADS you have different components... missile and gun batteries, sensors and radars, command and control etc etc... an unmanned drone is just something that will work with another manned weapon a bit like a Pantsir battery. Different targets require different tools and the different physical locations of the vehicles means one platform might be closer to the target and equipped with the correct weapon for the engagement.
Part of the IADS is management where you look at the threat and the danger and you select the most appropriate solution or weapon to deal with that threat.
If the target is a B-2 cruising at 10km altitude at 800km/h then the solution is not a Verba battery for instance, yet an S-70 armed with a dozen Verba missiles might approach the B-2s from the front at 14km altitude for instance and do a half roll so it is flying upside down as it flys past and then a nose pull up to position itself behind and slightly above the B-2s where it can launch Verba missiles at the bombers and any cruise missiles they might launch...
Of course R-73s would be more potent with a bigger warhead and of course the new WVR AAMs might be even better.
How do you think okhotnik will fair in the MUMT future vs the likes of Australia's Loyal Wingman built by Boeing?
I doubt they would ever meet.
My initial thought is that a flying wing might be less practical as an AA platform due to its seemingly lesser manoeuvrability
A missile is a missile... the S-70 is stealthy and being a flying wing would have a naturally small RCS. Being able to fly high would allow it to have good range performance and reasonable speed for a platform like that.
No plane is going to be dodging missiles like they do in hollywood movies.
With high offboresight WVR AAMs it is probably a case of both platforms getting shot down a lot of the time, so if your platform is a drone you are ahead of the game.
It seams to me that without a okhotnik equivalent, the USA plans to send in mostly f35s (since there are only 187 f22s) followed by f15ex, which will be AA missile trucks, launching at everything an f35 picks up as they go into contested territory. By comparison, Russia seems to be wanting to send in okhotnik followed by su57 to launch at everything okhotnik picks up. Is this wrong or a fair assessment?
It is a terrible over simplification of the actual situation.
Russia has a robust and comprehensive air defence network with missiles and guns and also aircraft, but does not have the numbers of aircraft that HATO has.
Having said that the purpose of an IADS network is to allow a much smaller force to over come a much bigger and more potent force... like the British in the Battle of Britain and HATO during the cold war we are led to believe.
The problem for HATO is that they didn't know where exactly they were going to fight, so their IADS is based on AWACS aircraft and JSTARS aircraft and inflight refuelling aircraft and a wide range of fighter aircraft.
Their obvious problem is that take out those key C4IR aircraft and it becomes much less powerful and organised... and the multi country organisation would likely collapse in terms of both attack and defence because their attack and defence is largely air based.
The Russians on the other had have a robust and capable air defence force but probably not enough aircraft, though with ground based missiles they are not something HATO could take on and beat any time soon.
The Air Force and the Navy are works in progress... the Navy being a lower priority and the Air Force needing a lot of rebuilding along with the Army.
Now that the F-35s are revealing themselves to be under armed pieces of crap they are scrambling to find shiny bits on the turd and the shiny bits are the advanced sensors and avionics so they plan to use them as forward operating recons... like manned UCAVs that pick out targets for following conventional aircraft will deal with.
The problem there is that forward deployed F-35s will be obliterated by S-350 and S-400 missile systems as well as BUK etc, and the aircraft following up will likely suffer the same even without a Russian AF.
When you add the Russian AF in it is not just Su-57s and S-70 drones... they are not going to make thousands of Su-57s but they could make 3-400 perhaps over the next 30 years... but they will also have Su-35s and Su-30s upgraded to be twin seat Su-35s, and of course MiG-31s and MiG-41s and MiG-35s which are all modern capable aircraft with modern capable missiles and radars and systems and those other fighters could also use S-70 support drones for their operations to extend their reach and view and capability.
The Americans thought they were leading the way, but they realised the Russian way actually made more sense, but their problem will always be the crippling cost of military equipment.
I had a lecturer who was in the Singapore Air Force, he worked with computers in the field of Human computer interaction... basically usability... and he said they wanted a keyboard with a kill switch... a covered key where you could flip up the cover and push a button and the computer would start erasing everything on the hard drive in case their position was overrun. It was going to cost $1,000 US dollars per keyboard for the switch... and this was in the 1980s so more like $10K now. Rather than spend that money like the Americans did, they got some red tape and put two pieces on the computer box precisely where the hard drive on that machine was located. Every soldier was issued with a 45 calibre automatic pistol. Four shots through the crossed tape and data on hard drive was "safe".
10c for the tape and they carry the pistols and ammo anyway.
Such an elegant efficient and cheap solution would not occur to a current engineer in the US MIC these days.