Twin Seat F-35 Parent / Drone???

[blindpilot]

Re: the video

Is 'autonomous' formation flight with the AV-8B the same as being controlled and tasked by the single-seat AV-8B? The video was somewhat vague as to how much actual 'control' was exercised by the Harrier compared to the ground facilities that were referenced...

All the above.
http://ir.kratosdefense.com/news-releas ... ccessfully
We have to remember 5th Gen is not "who does what." It's the network, the blob, doing all.

FWIW,
BP

[outlaw162]

5th Gen is not "who does what."

From the news release:

Throughout the mission multiple UTAP-22 were controlled by a single operator.

I just don't get the feeling that the 'single-operator' is the guy in the single-seat AV-8B at any point.

But, old school, I've never flown with a blob, unless you count a WSO.

[blindpilot]

5th Gen is not "who does what."

From the news release:

Throughout the mission multiple UTAP-22 were controlled by a single operator.

I just don't get the feeling that the 'single-operator' is the guy in the single-seat AV-8B at any point.

But, old school, I've never flown with a blob, unless you count a WSO.

The difficulty in discussing 5th Gen approaches is that the framework is not 4th Gen.

But I'll give it a try with some simple points.

If you look at the specs/conops for the Tiger II pod, some interesting factors are apparent. Keep in mind, this is NOT a network pod, it's an "EW" pod. But that's 4th gen talk. The pod has capabilities that "can be controlled" from the cockpit or by a ground operator. That's a point worth putting in our back pocket for now. Whatever the capabilities (classified) of the ALQ 231, it is open architecture as to the point of control ... whatever control is in this (loyal wingman) environment.

The video showed the "App" switching control from one ground operator to another ... that's very TTNT ish (like F-35 controlling an SM-6). Which of course is no surprise since the demo also showed full TTNT functionality. ... so let's see ..."switching controller" ... "cockpit or ground" ...

Just a few things to think about. Now to my favorite subject, "What is a cell phone?"

Well it is not a wireless flip phone duct taped to a Garmin GPS glued to a battery operated WiFi router, with a keyboard and computer screen connected by bluetooth .. connected by paper clips to ... etc. etc.
It is/can do all those things. But that is not what a iPhone is. It is me asking Siri, "Take me home," and
she replies, "turn right, we'll avoid the traffic on the interstate, and I assume you want me to order a pizza delivery for when we get there? ... oops go straight, there was just an accident on 1st street ... so Pizza? yes? No? .. and should I text your wife with our ETA?"
And whatever else my App is set up to do. Waze controls the accident report from someone else's phone, the Pizza Hut computer controls the pizza order. The voice network controls all the phone calls needed, and ...

what was your question? Who controls what? Yeah, like I said "all the above."

And to the point, Kratos already did all of this in 2015, and you can actually buy a UTAP-22 now called "Mako," off the shelf today and connect it to your existing TTNT systems. (payload is a bit light(500lbs), but you can buy one) The Air Force has in inventory 37 BQM-167's (skeeter targets) operated by the 82nd ATS.

FWIW,
BP

[outlaw162]

Guys this was already demonstrated and proven by Kratos back in 2015 in the UTAP 22 project using a SINGLE SEAT AV-8B Harrier II. Yeah I'd say it could be done without stretch and second seat machinations.

Thank you for the explanation in the previous post.

I guess my confusion was with what was claimed to be proven by the video in 2015 and with what the video actually demonstrated. Show me the video of the single-seat AV-8B controlling and tasking the drones when (if) it becomes available, and that it demonstrates that a second crewmember is not only not necessary, but not even preferable.

(I don't use any geolocation features on any personal device, anywhere.....and I actually order in person at the pizza parlor )

[Corsair1963]

Another possibility is using F-35 as network node to relay control and information to and from drones. The controllers could be in some large aircraft or even on the ground well back. This would be easier to implement but has some serious drawbacks like latency time and reduced data transfer rate especially at longer ranges.

Possibly both of these should be combined with controllers doing drone tasking and F-35 sensor fusion engine actually controlling the drones according to what is wanted from them.

I was thinking the same thing....

[spazsinbad]

Why does an US University have to copy a well-known Iranian Design for a Stealth Aircraft - size included!?

TACE lays down foundation for future UAV test safety
29 Jul 2019 Giancarlo Casem, 412th Test Wing Public Affairs

"EDWARDS AIR FORCE BASE, Calif. -- The 412th Test Wing’s Emerging Technologies Combined Test Force conducted an autonomous test flight at Edwards Air Force Base, California, July 25. The flight’s mission was to test a software suite designed to make unmanned aerial vehicle flight safer. “What we have here today is a sub-scaled jet-powered aircraft; it’s about 12 feet long and do about 250 knots,” said Jeff Jessen, Chief Engineer, Emerging Technology Combined Test Force. “Our goal is to verify an autonomy safety net that was developed by John Hopkins Applied Physics Lab, called TACE, Testing Autonomy in a Complex Environment."

The TACE test is part of the Air Force Research Laboratory’s Skyborg program. The Skyborg program is a developing software tool that allows engineers and researchers to develop autonomous capabilities. AFRL plans to have Skyborg as an Early Operational Capability as early as 2023. To help meet that requirement, engineers are testing autonomous elements like TACE to provide safety foundations for future tests.

The TACE safety net programming sits between a vehicle’s safety critical control system and its mission system. Proven algorithms keep the plane within safe bounds defined before takeoff. Jessen said, the idea is to let the autonomous test aircraft fly itself out of the safety bounds; the TACE program would then take over control of the airplane and take it to a safe point.

Skyborg’s TACE complements AFRL’s previous work with artificial intelligence used with programs such as the Automatic Ground and Air Collision Avoidance Systems, which have been proven to have saved lives.

Earlier this year, ET-CTF conducted a TACE test on a smaller, lower-performance aircraft. That earlier test proved that the program works on a slower aircraft, one that flew at 30 knots. After today’s test, ET-CTF will eventually equip TACE on even larger, more powerful, full-size aircraft. “Ultimately this is going to be our safety net for future autonomous testing, we hope to be testing autonomy that does air combat-type maneuvers and TACE is fundamental to being able to allow us to do that safely and effectively,” Jessen said."

Photo: "An unmanned jet-powered aircraft takes off from dry lake bed at Edwards Air Force Base, Calif., July 25. The flight tested a software suite called TACE, Testing Autonomy in a Complex Environment, developed by John Hopkins Applied Physics Lab. (U.S. Air Force photo by Giancarlo Casem)" https://media.defense.gov/2019/Jul/29/2 ... 1-1017.JPG

Source: https://www.edwards.af.mil/News/Article ... st-safety/

https://cdn.theatlantic.com/media/old_w ... ghter_.jpg

[blindpilot]

... Show me the video of the single-seat AV-8B controlling and tasking the drones when (if) it becomes available, and that it demonstrates that a second crewmember is not only not necessary, but not even preferable.
...

That's a 4th gen statement and that's the problem trying to answer. The ALQ 231 with TTNT code/comm can/did/whatever execute such a concept in the demos, but it wasn't "demonstrating controlling the drone." I suppose it could have at some level, to satisfy your 4th gen question, but no one was asking that question and that isn't how it works. It demo'd how it works. This is not a model airplane in the park with a hand held XBox controller. It is a system of systems with mission objectives integrated in the overall network. In the 5th gen world, pilots may not - actually already today WILL NOT know where the sensor data comes from. It could be from one of the drones, but no mission operator will know that, and they don't care to know it.

So now we need to address the shooter. Already in the F-35/22 world systems are not even stove piping shooters. But because there needs to be a "human in the loop," people may know who the shooter is, but that isn't necessary to actually do the mission - shoot the weapon, but simply a check in the actual kill chain, artificially placed for keeping HAL from going crazy on us.

The integrated network shoots .. fires the weapon. The AI resides at the network level, not in the individual node units per se. So there may be one or more ground unit controllers launching 2 drones from a ship, who hands off control to a AWACS style air node controller, perhaps sending 1 drone left to air controller one, and the other right to air node controller two. TTNT supports all of this, and they did demo this exact stuff in 2015.

So as seen in the video App screen, the pickup controller tells the drone launched by another controller to "Follow Him>" Him is an AV-8B, who with his node connection knows what is happening and watches his loyal wingman join up. He might "take control" but at this point, it is probably autonomous under previous instruction. With F-35's this gets even more robust, with the MADL/Link 16/Sensor fusion capabilities. But the AV8B / F-35 doesn't care yet what the Mako is doing .. except it is "following him" (see the video) They - the system of systems - are now off to the targets, known or not yet detected.

Someone, some sensor, .. and remember no one knows or cares who it is ... the system .. detects and marks a target.
https://www.unmannedsystemstechnology.c ... perations/
"Additionally, the Kratos UTAP-22 successfully coordinated semi-autonomous payload deployment, breaking formation to perform independently with a subsequent rejoin, and loyal/trusted wingman flight with one UTAP being flown as if it were a manned aircraft and the second UTAP-22 joining and sustaining autonomous formation."

There is no reason the "manned aircraft simulating" UTAP could not have been an actual manned AV8B. The comm, the situation layout, the overall integration is there in the ALQ 231 interface. Certainly, in this demo if an air boss (AWACS type) controller was the one "commanding" the semi autonomous strike, the AV8B could have flown as the flight lead in place of the UTAP sim leader for the demo. That lead was just drilling holes in the sky "leading" the "follow him" elements.

Keep in mind that all that is needed beyond this described demo is the AV8B pilot to do the [sarc on] very taxing cockpit action[sarc off] of touching a screen with his finger one time, after mouse/thumb selecting (or having the network already having selected prior) the shooter (ie. second attacking drone).

They did not demonstrate a linear "AV8B pilot" stove pipe controlling ... take handoff control, fly to the wing position, ready weapons, set target, go kill, now return to me, follow me, bank left, slow down, level off, get ready to climb, speed up, bank right ... BECAUSE THAT'S not how 5th gen architecture works. It has nothing to do with having the AV8B pilot showing he is controlling a model airplane. It demonstrated full range functionality in a loyal wingman system of systems, BECAUSE THAT'S how 5th gen works.

Now the 4 years ago test AV8B is probably a parts queen at some Marine base, and they will probably not be slapping EW Tiger pods on active AV8B's, although they could I guess. They have this cool thing now, called the F-35B and ...
It don't need no stinkin pods!
The Mako meanwhile has demonstrated it can do "Loyal Wingman" today. Of course it's old news also and the way cooler XQ-58 Valkyrie is what Kratos wants to sell now ...
https://en.wikipedia.org/wiki/Kratos_XQ-58_Valkyrie

but if you want a Skeeter(Mako) following a Harrier, they are still off the shelf, just say the word. Kratos could probably find a retiree to set it up for you. All the current workers are doing Valkyrie stuff.

FWIW,
BP

[quicksilver]

Similar henny penny stuff was common when the F-18 first showed up in the early 80s. Same when single seat FAC(A) was proposed, and when some folks at Fallon (iirc) actually data linked video off of a fighter to a station on the ground about 20 years ago.

‘Controlling drones’? Heresy I tell you...

[outlaw162]

I guess I'm doomed to be stuck in 4th gen thinking (or maybe even 3rd). But I get somewhat worried when there's a prevalent attitude of unlimited optimism about how this will work when push comes to shove. I hope you're right.

The integrated network shoots .. fires the weapon

(semi-sarc on) Is there a good 'ol' fashion cover-guarded pilot nuke consent switch in the F-35.....or does the 'integrated network' take care of that? (semi-sarc off)

edit: I just can't imagine that all the potential lurking failure cases have been uncovered....the aviation equifax equivalents

[blindpilot]

...

(semi-sarc on) Is there a good 'ol' fashion cover-guarded pilot nuke consent switch in the F-35.....or does the 'integrated network' take care of that? (semi-sarc off)

edit: I just can't imagine that all the potential lurking failure cases have been uncovered....the aviation equifax equivalents

Agreed there is some risk here, not lightly set aside. The same thing exists today in the car you drive. Very little that you "command" is actually physically linked to the wheels, engine, brakes etc. Your steering wheel nor brake pedal(normally) isn't. And certainly for over a decade, your accelerator pedal isn't even remotely linked. It's not like fly by wire aircraft stick and rudder pedals actually "control" the flight control surfaces. One reason software testing is so burdensome/difficult (see F-35 delays et al) is that such marginal areas need to be tested, and proven as "fool proof" as possible.

The Boeing 737 MAX certainly stands as a beacon to what can happen. But usually the code, lacking a cover-guard, will have a kill switch option. Most of the actual "shoot" code has a gatekeeper "man in the loop" routine that is required for a positive "shoot" result to occur. But even this is being examined closely, since much of what happens in war now is too fast for the human to deal with at the time. So you have things like the CWIS Phalanx system requiring a weapons free condition set before the engagement might happen (cover lifted up off the switch). And as with the Phalanx, sometimes the Japanese destroyer hits a neighboring ship ... "Ooops my bad .." We could put a man back on the pedestal mounted .50 cal. I suppose, but chances are pretty good, he will not take down that cruise missile coming in ... and as has happened before, even he will sometimes hit the neighboring ship .. "oops, my bad". .. just saying.

If you are driving a car made this century, you are already in that danger zone, driving to the grocery store. And as the news sometimes reports, ocassionaly people's cars kill them. Yep it happens. See the lawsuits.

And finally Siri has killed folks sending them into a blizzard blanketed dead end dirt road during a storm. That too has happened. But again you can come to the war with a .50 cal if you want. Others will be shooting at you with something else, so good luck with that.

MHO,
BP

[blindpilot]

... But I get somewhat worried when there's a prevalent attitude of unlimited optimism about how this will work when push comes to shove. I hope you're right.
...

Optimism. hope and pushing really have little to do with it. Smart systems, like rocket science, ... wait some of it is rocket science ... is hard.

But actually some of it is actually safer, see GCAS. My introduction into the realm was in the early 90's consulting R&D robotics stuff for Gates. They had experienced a robot car (think XBox controller simple) crashing into a wall and scaring some poor secretaries on the other side. The project was to create processes, simulations, activation of the robotic systems so you could be reasonably confident the lumbering robot cars weren't going to kill anyone when you hit the on switch. We got it to the point where if the cars were going to crash, then they would crash in the sim 100% of the time first, so we could adjust the code accordingly. I did similar consulting for the Army in an Urban Robotic live fire training set up. A great deal of the work actually makes it safer than just turning loose some teenagers with M-16's and telling them to be careful.

MHO,
BP