Wired: Entire U.S. Stealth Fighter Fleet Grounded

[mardi gras skin]

The original reason given to use them, was to find out where the bad guys have put their SAMs. (So that you can then kill the SAMs.) The reasoning was that looking for SAMs is dangerous. But that if a drone gets shot down, then a) You haven't lost a really expensive fighter, and a pilot, and Even if it gets destroyed, you still learn something.

When we stationed in Bitburg, Germany, our sister airbase at nearby Spangdahlem was home to the f-4 wild weasels. They were the SAM hunter/killers. First in/last out. If the USSR ever decided to come across the border, the wild weasels were to make them regret it immediately. I looked up a quote I couldn't quite remember: “A Wild Weasel provides a Warsaw Pact SAM operator the maximum opportunity to give his life for his country.” There has never been another fighter that looked more bad ass, sounded more bad ass, with a more bad ass mission than the f-4 wild weasel. Pure intimidation. To your concern about loss of life, they would have said, "We don't lose fighters, F you very much."

Its good that all this new tech may help save lives but I'm sure the drones have cut their pride to the quick.

[PeterMP]

You are seriously clueless on this subject. You don't even understand what you are reading. :doh:

I just want to add... One day software may be advanced enough to handle the advanced combat maneuvers and *think* creatively enough to outsmart a human. But we are a LONG way away from that day.

If you don't think computers are good enough to outsmart humans NOW, you need to talk to some chess players. There is a HUGE advantage to being able to play the game millions of times in a pretty short period of time.

Not knowing anything about actually flying, I'd guess the issue is going to come down to control. Humans are going to want to mantain some control, but mantiaining control also means you have the possibility of losing control. Imagine what would happen if your fighter squandran was "hacked" and turned around and attacked you.

It seems to me in a real war against about an equal opponent, you'd want like weekly software/code variants, but I think that would become cost and proceduerly prohibitive.

[Mad Mike]

If you don't think computers are good enough to outsmart humans NOW, you need to talk to some chess players. There is a HUGE advantage to being able to play the game millions of times in a pretty short period of time.

Not knowing anything about actually flying, I'd guess the issue is going to come down to control. Humans are going to want to mantain some control, but mantiaining control also means you have the possibility of losing control. Imagine what would happen if your fighter squandran was "hacked" and turned around and attacked you.

It seems to me in a real war against about an equal opponent, you'd want like weekly software/code variants, but I think that would become cost and proceduerly prohibitive.

Chess is actually a simple mathematical game for a computer. Each piece has a limited number of possible moves. The challenge is looking ahead far enough to see all of the possible outcomes of each move. If you can do that, you can determine the winning moves with complete certainty.

The difference between chess and the real world is that in the real world there are literally MILLIONS of possible moves per second and there is no move that guarantees anything. And that's just to make the thing fly. Our most advanced autopilots today can basically fly a straight line and land with the proper guidance data provided from the ground. That's about it. The aerial maneuvers required for combat alone are thousands of times more complex and that does not even take into account the creative thinking required to outsmart your enemy. Real, live combat, requires creativity that is far beyond the level of artificial intelligence as we know it today.

Hell, no one has yet figured out how to make a car drive itself. And for all of the variables involved in that, a car is only moving in two dimensions and trying to avoid obstacles and other vehicles which are also trying to avoid an accident. A fighter aircraft must operate at extreme limits in three dimensions with an enemy that is trying to shoot it down.

This is Google’s first self-driving car crash

http://jalopnik.com/5828101/this-is-googles-first-self+driving-car-crash

This photo of what looks like a minor case of Prius-on-Prius vehicular violence may actually be a piece of automotive history: the first accident caused by Google's self-driving car.

[DC9]

Needless to say....only the Air Force would rift 40% of their enlisted personnel to pay for something that didn't work. Oye.

[mardi gras skin]

Needless to say....only the Air Force would rift 40% of their enlisted personnel to pay for something that didn't work. Oye.

Its also the only branch of the military where enlisted men send their officers off to die. Their personnel decisions to achieve mission readiness are necessarily different than the other branches of the military, current stealth setback aside.

[HailGreen28]

That is likely to change as well because of some of their hardheadedness

btw...they already have changed...and more are coming

http://www.airforcetimes.com/news/2009/03/airforce_uav_audit_030109/

In September,(09) Chief of Staff Gen. Norton Schwartz unveiled plans to create a UAV-specific career field and to train officers with no flying experience to control Predators and Reapers.

added

interesting read for those interested

http://www.strategypage.com/htmw/htlead/20101015.aspx

Thanks for the link. I'm surprised the new dedicated UAV operators aren't pilots.

[twa]

Thanks for the link. I'm surprised the new dedicated UAV operators aren't pilots.

The biggest change will come from breaking the AF's officer only barrier

on the grounding issue....ground testing has resumed on the F-35,so perhaps progress is coming.

[PeterMP]

Chess is actually a simple mathematical game for a computer. Each piece has a limited number of possible moves. The challenge is looking ahead far enough to see all of the possible outcomes of each move. If you can do that, you can determine the winning moves with complete certainty.

The difference between chess and the real world is that in the real world there are literally MILLIONS of possible moves per second and there is no move that guarantees anything. And that's just to make the thing fly. Our most advanced autopilots today can basically fly a straight line and land with the proper guidance data provided from the ground. That's about it. The aerial maneuvers required for combat alone are thousands of times more complex and that does not even take into account the creative thinking required to outsmart your enemy. Real, live combat, requires creativity that is far beyond the level of artificial intelligence as we know it today.

I just want to point out that much of the above simply can't be true. As you've already stated, the very nature of the air craft and the laws of physics greatly constrain the system, and the actions taken. Some moves are simply NOT possible, just like in chess I can't move a queen like I do a knight.

A computer doesn't even have to consider the moves that aren't possible, just like it doesn't have to consider moving its queen like a knight.

Essentially, a 3D chess board is EXACTLY what you have. Now, if you have the inferior air craft, then there might be no combination of winning moves, but that wouldn't be the computers fault.

You're assuming a human would NOT have had the accident. My wife once rear ended somebody that stopped in short on a light. It was raining, and she was going down hill, and the light just turned red, and the driver in the other car didn't even stop at the light.

She slammed on the breaks as soon she realized what was happening. Given the physical conditions, the car didn't/couldn't stop in time. It wouldn't have mattered if a computer was driving it or not.

Also, let's be realistic, the technology that military has and likely always will have exceeds what google is using to power its self-driving car.

**EDIT**

What do you think about the guidance system on things like cruise missiles?

[twa]

cruise missiles and terrain following radar are old school, we done went optical, as well as intuitive and swarm intelligence is advancing....with these advances even the plane itself is a weapon instead of a platform.

Simply slaving some UAV's to a manned fighter alone extends reach and firepower,but we can go well beyond that to reach out and touch someone:evilg:

[sacase]

I just want to point out that much of the above simply can't be true. As you've already stated, the very nature of the air craft and the laws of physics greatly constrain the system, and the actions taken. Some moves are simply NOT possible, just like in chess I can't move a queen like I do a knight.

A computer doesn't even have to consider the moves that aren't possible, just like it doesn't have to consider moving its queen like a knight.

Essentially, a 3D chess board is EXACTLY what you have. Now, if you have the inferior air craft, then there might be no combination of winning moves, but that wouldn't be the computers fault.

You're assuming a human would NOT have had the accident. My wife once rear ended somebody that stopped in short on a light. It was raining, and she was going down hill, and the light just turned red, and the driver in the other car didn't even stop at the light.

She slammed on the breaks as soon she realized what was happening. Given the physical conditions, the car didn't/couldn't stop in time. It wouldn't have mattered if a computer was driving it or not.

Also, let's be realistic, the technology that military has and likely always will have exceeds what google is using to power its self-driving car.

**EDIT**

What do you think about the guidance system on things like cruise missiles?

And there is the essence of your problem, there will be some moves that aren't possible that the computer will disregard, the human however will not disregard the impossible and actually attempt to do it and possible even succeed in doing it. I don't know how many times I have seen soldiers do the "impossible".

Computers are not perfect, they never will be, there are to many random variables for them to calculate. Flying in strait lines is easy, taking off an landing is easy. Combat is pure chaos. Not even close to a 3D chess board.

---------- Post added August-11th-2011 at 12:53 PM ----------

That is likely to change as well because of some of their hardheadedness

btw...they already have changed...and more are coming

http://www.airforcetimes.com/news/2009/03/airforce_uav_audit_030109/

In September,(09) Chief of Staff Gen. Norton Schwartz unveiled plans to create a UAV-specific career field and to train officers with no flying experience to control Predators and Reapers.

added

interesting read for those interested

http://www.strategypage.com/htmw/htlead/20101015.aspx

Most army UAV operators are enlisted troops, while all air force ones are officers. The Raven training only lasts 80 hours, but this tiny UAV was designed for ease of use. It takes about five times longer to train operators for larger UAVs like Shadow and Predator. General Schwartz made this point, in explaining that the largest UAVs, like the Global Hawk, can cross oceans, and requires a high degree of training and skill. But it's much more dangerous to fly a Raven within rifle range of enemy troops, and keep the little bird alive long enough to get the video feed needed to win the battle. Many of these army Raven operators are very, very good, mainly because they have hundreds of hours experience operating their UAVs while under fire. Few air force UAV drivers can claim this kind of experience.

General Schwartz was also getting pressed from above about how the air force operates its UAV fleet. The U.S. Department of Defense has been putting pressure on the air force to automate their UAV operations. The air force initially responded with complaints that they were overloaded. But then it became widely known that, while army UAVs have software that enables automatic take off and landing, similar air force UAVs do not, and this has led to higher UAV losses for the air force. At the same time, the navy, and the British, are developing flight control software that allows pilots to control two or more UAVs while flying their own aircraft. In effect, these pilots would fly into combat with two or more UAVs under their control. The U.S. Air Force is under a lot of pressure to make this happen, so one team of UAV operators can control two or more aircraft.

It's not that the air force can't make this happen, it's just that there's a lot of resistance in the air force to replacing pilots with a lot of UAVs.

Ok, let's address the article first. They will TRAIN no pilots to operate UAV's, which means, you will get your pilots license to finish your training. Have you actually seen a real predator up close? They are HUGE, they are not model airplanes. I wouldn't want someone flying them that isn't a pilot, these are still multi million dollar aircraft.

Now let's talk about the raven...the raven is nothing more than a RC plane, for gods sake, you throw it like a paper airplane. As far as them getting shot down, I doubt that happens alot, they are small and quick, not exactly something that is easy to shoot down. Their size helps them. The predator is the size of a private jet, much bigger target.

[twa]

you can be a UAV pilot w/o a pilots license or pilots qualification,even a Predator(and they DO have to train EVEN pilots to do so)

different skills....all of which still ignores the next level of autonomy

[nonniey]

The thing is, you're using still using pilots. (and still need the same basic training for the same missions.)

The Air Force does in fact require UAV Pilots to have the same training as normal aircraft pilots and be officers. We in the Army allow our NCOs to pilot UAVs and with much cheaper training (So far as I know there has been no discernable performance difference between the two (Larry got the Air Forces justification on resistance to UAVs right though)

[Mad Mike]

I just want to point out that much of the above simply can't be true. As you've already stated, the very nature of the air craft and the laws of physics greatly constrain the system, and the actions taken. Some moves are simply NOT possible, just like in chess I can't move a queen like I do a knight.

A computer doesn't even have to consider the moves that aren't possible, just like it doesn't have to consider moving its queen like a knight.

Essentially, a 3D chess board is EXACTLY what you have. Now, if you have the inferior air craft, then there might be no combination of winning moves, but that wouldn't be the computers fault.

You're assuming a human would NOT have had the accident. My wife once rear ended somebody that stopped in short on a light. It was raining, and she was going down hill, and the light just turned red, and the driver in the other car didn't even stop at the light.

She slammed on the breaks as soon she realized what was happening. Given the physical conditions, the car didn't/couldn't stop in time. It wouldn't have mattered if a computer was driving it or not.

Also, let's be realistic, the technology that military has and likely always will have exceeds what google is using to power its self-driving car.

**EDIT**

What do you think about the guidance system on things like cruise missiles?

Look, I'm getting tired of trying to explain this but you are simply wrong.

There is a reason there are no self driving cars on the road. There is a reason there are no unmanned fighter aircraft. It simply is not possible at this time. PERIOD. It will be possible one day, but that day is still a long way off. Commercial, self driving cars that you can buy are at least 20 years away. AI that can manage combat maneuvers, at least as long. AI that can be creative enough to out wit a human are MUCH farther off.

The real limitation is creativity. Combat is a creative endeavor. It is fluid and completely unpredictable. Victory goes to the person who can adapt to chaos, and overcome unexpected difficulties with creative solutions. Computers suck at being creative.

Again. I'm getting tired of trying to educate you on this subject. It's like trying to debate a conspiracy theorist. You have your mind made up and no matter how wrong you are you will not change. I'm done.

But just to answer your last question before I go. the flight path of a cruise missile is a SIMPLE THING. It flies in a straight line according to a pre-programmed flightpath at pre-programmed altitudes. THEY ARE NOTHING LIKE FIGHTER AIRCRAFT.

_______

Edit for reference here is the Air Force UCAV program that is in development now. Note that a human pilot is required.

http://www.fas.org/programs/ssp/man/uswpns/air/attack/x-45_ucav.html

As a member of a tightly coupled system of systems, the UCAV will work cooperatively with manned systems and exploit the emerging command, control, communications, computer, intelligence, surveillance and reconnaissance (C4ISR) architecture to enable successful achievement of campaign and mission level objectives. Intelligence preparation of the battlefield will provide an initial mission/threat database for mission controllers. Controllers will exploit real-time data sources from the theater information architecture to plan for, and respond to, the dynamically changing battlefield.

This is our next generation UCAV. Any type of completely autonomous fighter is AT LEAST two generations away. In other words, not in our life time.

[HailGreen28]

The Air Force does in fact require UAV Pilots to have the same training as normal aircraft pilots and be officers. We in the Army allow our NCOs to pilot UAVs and with much cheaper training (So far as I know there has been no discernable performance difference between the two (Larry got the Air Forces justification on resistance to UAVs right though)

With around 400 hours needed for the bigger drones, I guess you're giving them a good basic training (about 10 weeks in the real world). I still think some kind of flight training (on props?) would give them an edge in flying the UAVs, even though it's a different beast.

[PeterMP]

But just to answer your last question before I go. the flight path of a cruise missile is a SIMPLE THING. It flies in a straight line according to a pre-programmed flightpath at pre-programmed altitudes. THEY ARE NOTHING LIKE FIGHTER AIRCRAFT.

This isn't true. Advanced cruise missiles are able to detect changes in topography and adjust their flights accordingly independent of humans.

http://www.wisegeek.com/what-is-tercom.htm

They have built in sophisticated guidance software that allows them to deviate significantly from their pre-programed flight plan. I know somebody that works on problems related to their software.

[Mad Mike]

This isn't true. Advanced cruise missiles are able to detect changes in topography and adjust their flights accordingly independent of humans.

Once again you failed to understand what you read. From your link. :doh:

The system uses a predefined contour map of the flight path which acts as a comparison master image. The missile is equipped with a sophisticated radar altimeter which constantly reads the terrain it is crossing and compares the readings to the master image. When deviations are detected, the missile's guidance system makes the necessary corrections to its flight path. This makes extremely accurate navigation and collision avoidance possible, thereby allowing the missile to fly closer to the ground and avoiding detection by radar.

I'll try one last time to give you some perspective. A cruise missiles flight programming *IS* sophisticated. BY TODAY'S STANDARDS. That's what I keep trying to tell you. Flying combat like a fighter pilot is thousands, maybe tens of thousands of times more difficult. That's why we cant do it yet.

I know somebody that works on problems related to their software.

Good. Why don't ask him to educate you on the subject. Ask him if he can program a computer to fly combat maneuvers that would beat a human pilot.

[HailGreen28]

I'll try one last time to give you some perspective. A cruise missiles flight programming *IS* sophisticated. BY TODAY'S STANDARDS. That's what I keep trying to tell you. Flying combat like a fighter pilot is thousands, maybe tens of thousands of times more difficult. That's why we cant do it yet.

Good. Why don't ask him to educate you on the subject. Ask him if he can program a computer to fly combat maneuvers that would beat a human pilot.

Dude, even I know that military hardware and software runs behind civilians. For example, the US spent billions on that "Land Warrior" or "Nett Warrior" or whatever they call all the expensive networking gear they were going to pile on a soldiers back pack, helmet, and gun, so he could be "networked" with everybody else, with all kind of cool features. The brass found out soldiers were doing more with their relatively cheap and reliable cell phones already.

However, you could have a computer controlled jet out-dogfight a human pilot. We've had that simple "kill" AI for awhile in video games. The AI would think quicker, and could withstand maneuvers that a pilot couldn't.

Right now various militaries (including ours), have some aircraft that can land automatically. F-18 carrier landings, for example. And drones programmed to carry out missions, and even return to base if something unforeseen happens. It's not farfetched to say the capability to program an unmanned fighter to out dogfight a human probably exists already.

[PeterMP]

Once again you failed to understand what you read. From your link:

"The system uses a predefined contour map of the flight path"

:doh:

They have built in sophisticated guidance software that allows them to deviate significantly from their pre-programed flight plan.

Why didn't you quote the whole post? Oh wow, I used the "pre-programmed", and they used predefined. That makes ALL the difference in the world.

Again. I'm getting tired of trying to educate you on this subject. It's like trying to debate a conspiracy theorist. You have your mind made up and no matter how wrong you are you will not change. I'm done.

It is like trying to debate with a conspiracy theorist. They argue against a point that nobody ever made (nobody has said we can do it NOW. My point about NOW dealt not with flying an airplane which has issues related to how to sense and transmit data on other rapidly moving objects in real time (which I know nothing about), but about the ability for computers to deal with human creativity with respect to competitive problem solving "games", and I'll stand by that statement in the face of the non-existent piles of real information you've supplied to back up your opinion), and when they say something that is plainly wrong and you correct them (w/ a link), they pretend like YOU are the one that is wrong.

I'm done.

(Seriously, this topic is somewhat interesting, but it isn't worth my time given that you clearly can't have a real conversation.)

[twa]

Peter have you seen some of the latest results from swarm intelligence and it's dependent cooperation?

Giving a objective and setting them loose had some interesting results.

hailgreen,what he overlooks is the concept is a pack mentality and application,not individual....mano a mano ain't in the game plan....the reduced costs and risks enable it as well

[Mad Mike]

Dude, even I know that military hardware and software runs behind civilians. For example, the US spent billions on that "Land Warrior" or "Nett Warrior" or whatever they call all the expensive networking gear they were going to pile on a soldiers back pack, helmet, and gun, so he could be "networked" with everybody else, with all kind of cool features. The brass found out soldiers were doing more with their relatively cheap and reliable cell phones already.

However, you could have a computer controlled jet out-dogfight a human pilot. We've had that simple "kill" AI for awhile in video games. The AI would think quicker, and could withstand maneuvers that a pilot couldn't.

Right now various militaries (including ours), have some aircraft that can land automatically. F-18 carrier landings, for example. And drones programmed to carry out missions, and even return to base if something unforeseen happens. It's not farfetched to say the capability to program an unmanned fighter to out dogfight a human probably exists already.

Bla, bla, bla,...

Again. Landing an aircraft is SIMPLE *in comparison* to making an aircraft THINK. A landing aircraft is homing into guide signals using pre-programmed instructions. Combat requires creative thought which computers DO NOT DO WELL.

None of us here are experts. I certainly don't claim to be. But I happen to know a little about these things. You see, I had a good friend years ago how flew off carriers. I did graphics work for Langley AFB and NASA where I've talked to fighter pilots and engineers. I have followed this stuff for most of my life. My godson has a Masters degree in artificial intelligence. Regardless of what you assume. You are wrong. We just aren't there yet. Not by a long shot.

http://en.wikipedia.org/wiki/Artificial_intelligence

Deduction, reasoning, problem solving

Early AI researchers developed algorithms that imitated the step-by-step reasoning that humans were often assumed to use when they solve puzzles, play board games or make logical deductions.[39] By the late 1980s and '90s, AI research had also developed highly successful methods for dealing with uncertain or incomplete information, employing concepts from probability and economics.[40]

For difficult problems, most of these algorithms can require enormous computational resources — most experience a "combinatorial explosion": the amount of memory or computer time required becomes astronomical when the problem goes beyond a certain size. The search for more efficient problem solving algorithms is a high priority for AI research.[41]

Human beings solve most of their problems using fast, intuitive judgments rather than the conscious, step-by-step deduction that early AI research was able to model.[42] AI has made some progress at imitating this kind of "sub-symbolic" problem solving: embodied agent approaches emphasize the importance of sensorimotor skills to higher reasoning; neural net research attempts to simulate the structures inside human and animal brains that give rise to this skill.

As for real creative AI, it is still in it's infancy.

http://en.wikipedia.org/wiki/Computational_creativity

Creativity in problem solving

Creativity is also useful in allowing for unusual solutions in problem solving. In psychology and cognitive science, this research area is called creative problem solving. The Explicit-Implicit Interaction (EII) theory of creativity has recently been implemented using a CLARION-based computational model that allows for the simulation of incubation and insight in problem solving.[36] The emphasis of this computational creativity project is not on performance per se (as in artificial intelligence projects) but rather on the explanation of the psychological processes leading to human creativity and the reproduction of data collected in psychology experiments. So far, this project has been successful in providing an explanation for incubation effects in simple memory experiments, insight in problem solving, and reproducing the overshadowing effect in problem solving.

One day, computers *will* be able to think, adapt and be as creative as humans. Thank god we aren't there yet. Because on that day (and I'm not joking here), be afraid. Be Very afraid.

[gbear]

Reading this thread is like reading an intro to Enders Game.

[Burgold]

Reading this thread is like reading an intro to Enders Game.

I was thinking the same thing. The scary thing about unmanned aircraft... either piloted via remote control or piloted via computer AI is that it is one more step in sanitizing war, making it clean, making it easy. Right now, we grieve or get upset when our pilots die or when a helicopter crashes. When the day comes and it is coming, that more and more is done remotely without direct risk to our guys... it will be so much easier to kill. Killing with a gun is easier than killing with a knife. Killing with a missile is easier than killing with a gun. Killing with a robot will be the easiest of all.

[Koolblue13]

The last thing we need, is to make it easier to kill other people.

[DRSmith]

I was thinking the same thing. The scary thing about unmanned aircraft... either piloted via remote control or piloted via computer AI is that it is one more step in sanitizing war, making it clean, making it easy. Right now, we grieve or get upset when our pilots die or when a helicopter crashes. When the day comes and it is coming, that more and more is done remotely without direct risk to our guys... it will be so much easier to kill. Killing with a gun is easier than killing with a knife. Killing with a missile is easier than killing with a gun. Killing with a robot will be the easiest of all.

Gunshots by computer.

Human lif will become of no mater value than that which is shot in the first person shooter games

[twa]

The last thing we need, is to make it easier to kill other people.

unless they need killing