US Navy robot stealth fighter in first unmanned carrier landings The last of the great barriers has fallen: an unmanned aircraft has successfully made autonomous arrested landings on the deck of an aircraft carrier. It's now plain that robots are not just as good as human pilots - they are as good as the best human pilots. As most regular Reg readers will be aware, the X-47B project …
".....why not just remove the pilot all together and use it as a drone?" Because a human pilot can face an infinite number of possible scenarios, assess each one, and make a decision on the spot. It would be impossible to program a drone with every possible occurence, so if the mission parameters wander too far from the ones thought likely the drone will need a remote operator to tell it what to do. In such situations, the on-the-spot pilot will shoot down the waiting-for-instructions-down-a-high-latency-satellite-link drone.
>a human pilot can face an infinite number of possible scenarios
Sorry. No go. It's not so much that you are wrong as this is just the canned glib answer BS that we can expect over the next 20 years or so to justify the F35 against hell and high water. I don't want my tax dollars wasted, or my country's servicemen put in danger, with no more reasoning than this argument. Wars have been lost, often, by focusing on the wrong weapons and tactics for reasons of tradition.
If you are talking US vs. China air to air, the scenarios are fairly limited - if it flies and the IFF says it's not yours, shoot it down.
And, if you are talking US vs. China airstrikes, I expect that a whole lot of non-scenario-aware missiles are going to be called for before any pilots fly over hot zones.
IF, and that's a big if, the F35 is as capable as its price tag makes it out to be, I suspect the Chinese will bypass trying to field their own manned jet fighters against it and opt for drones or other approaches themselves. Don't fight to your enemy's strengths.
Now, if you are talking avoiding civilian casualties in minor wars, that's indeed another story. Some experts have been advocating a return to slower aircraft like A10s or even prop-driven planes to allow human pilots _time_ to assess your scenarios and not shoot up weddings or the like. F35s in all that?
Footnote: Western military planners are correct to focus high-cost items on a possible war with China. But this is also a very dangerous game to play - arms races can be easier to get into than to get out of. Procuring masses of very expensive hardware with no other credible opponent sends a signal Chinese planners would be daft to ignore.
...the Chinese will bypass trying to field their own manned jet fighters against it and opt for drones or other approaches themselves. Don't fight to your enemy's strengths.
I think an apt comparison from the tank world might be the Panzer versus the Sherman. How impressive would it be for 100s of highly trained pilots flying very expensive birds to wipe out 1000s of cheap UAVs, but still go down to the rest after all the crewed vehicles ran out of munitions? I guess it would take a tragedy of this magnitude to get some minds changed.
"Spending billions". Oh, impressed I am, Matt. How _many_ billions is that, on the J31? Good of you if you can share with us, since defense analysts struggle to estimate PLA purchasing power from its official budget.
7 F35s, at a unit cost of 150M$ for the A variant, work out to 1.0B$. What's intended procurement, just for the US? 1500 F35s, down from 2400+ earlier on? What does that work out to - 250B$? You can fit a lot of "Chinese billions" in that kitty. Besides, if the Chinese want us to mis-allocate spending, they do need to prime their threats with a bit of pertinent kit.
Another example of not playing to enemy strength, the Chinese have also spent billions buying the Liaoning, their new aircraft carrier. Are the US carrier battle groups quaking in their boots? Not really, because they still outclass the Chinese.
They are however rather more worried about the DF-21D, a supposedly anti-ship capable ballistic missile. Given the cost of a carrier, it is really worrying that they might (not proven) be able to take out one of those and neutralize the huge sunken cost of carrier battle groups. That's exactly what the Chinese should aim to do - play on our hugely lengthy and costly development cycles to bankrupt us on committing to acquire weapons superbly fitted to fight the last wars but not proven to work for the next one. Their 2007 anti-satellite test served the same purpose.
There're plenty of unsolved issues that preclude an air-to-air autonomous drone today. And in 10 yrs too, most likely. Wanna bet they won't be solved in 20 yrs? 30 yrs? That's how long fighters typically fly for.
One possible approach is way more X47 type exploratory R&D, budget for training and retain manufacturing capability, rather than commit straight out to particular weapons platforms too long in advance. Shorten development cycles and aim to lower unit costs. Given some lead time, conventional weapons can be cranked out and nukes should keep a first-tier opponent from gambling too aggressively in the short term. Remain nimble at procurement in other words.
That's one view and it may very well be wrong. But our militaries do need a substantial debate and not just "we need a 5th gen at any cost and we'll lie to you about that cost if we have to".
".....How _many_ billions is that, on the J31?...." Out of their much smaller military budget, quite a chunk, apparently.
".....They are however rather more worried about the DF-21D...." <Sigh> You really missed the bit about the constant development of countermeasures. The DF-21 is not the first time someone has thought of using a nuke-tipped missile to blast a fleet, the Soviets having put a lot and time and thought into the idea. Hence the USN's adoption of anti-ICBM kit like the SM-3 missiles. There's also the measures the USN could use prior to moving into an area threatened by DF-21s, such as wiping out the Chinese GPS satellites that the DF-21 is reliant on and the Chinese long-range radar (the more sneaky-minded might suspect that one of the roles of the X-41 secret spaceplane is to launch fake Chinese and Russian GPS satellites to confuse and degrade their GPS systems). Kinda hard to even hit something as big as a carrier when the oceans are so massive and you don't know where to shoot, and just imagine the fun if you accidentally fry a Russian warship! And if worst came to worst, the Chinese only have about 80 units, so the US could swamp the launch sites with hundreds of air-launched cruise missiles - much cheaper than DF-21s.
"..... That's exactly what the Chinese should aim to do - play on our hugely lengthy and costly development cycles to bankrupt us....." Yeah, I think the USSR tried that, it didn't work out too well for them. You do remember that whole falling of the Berlin Wall thing, or were you not born then? The US military budget still dwarves that of even the Chinese and they are nowhere near the capability of the US in areas such as space, let alone ballistic missiles and missile defence measures.
>The DF-21 is not the first time someone has thought of using a nuke-tipped missile to blast a fleet
The whole point of this _new_ Chinese carrier-threatening approach is that is NOT using nukes.
Does it work? Can it be countered? Who knows, but how can you write this rebuttal and not be aware that the _primary_ characteristic of this weapon is that it does not rely on nukes?
If you don't even know that much, do you expect me to take any of your other arguments seriously? Really?
"Because a human pilot can face an infinite number of possible scenarios, assess each one, and make a decision on the spot."
Utter crock. No wait, it's Matt Bryant. Talking whatever.
Could you dig out any commentary of yours on chess computers before Big Blue, Matt? Just want to check.
I just rewatched the Star Trek Voyager episode in which B'elanna has to go aboard a Cardassian-built, reprogrammed-for-the-Maquis, smart missile that got pulled into the Delta Quadrant and is now on its way to wipe out an innocent populated planet, thinking it's the original Cardassian target. The smart missile's computer assumes that B'elanna is being coerced by the Cardassians when she tries to shut it down, and does everything in its power to thwart her efforts.
I'd have to laugh if something similar happened to one of these drones. I'd be holding my sides if one of these things got turned arse-about-face and set off to take out the White House, dismissing all attempts to stop it as the result of its controllers acting under Chinese duress!
an infinite number of choices? that's every possible choice at that moment, in the past and in the future...
surely an infinite number of decisions would take an infinite amount of time to compute (human or machine)
nuke - as that's what would happen to any brain faced with that sort of choice!
The tech going into making this work is amazing.
I know a few things about robotics, sensors, and more stuff thats needed and they achieved something extraordinary.
Question is if this technology will be used for the greater good or if its kept back for a tactical advantage, if its the latter i pity the wasted creative hours spend by engineers which could've been put to good use elsewhere.
The biggest achievements require one to have their own aircraft carrier on which to apply them. The automated landing bit is already commercially available, it is very possible the pilot commanded no input on landing during your last flight. R/C aircraft can land themselves too, but not as reliably.
But yes, I agree with your overall sentiment, more military technology should be in the public sphere. A very large portion of military tech doesn't have to translate to explody things, that's just how it is used.
Just think where motion control automation like this could be used.
We could have automated tube trains, in the 21st century! With even less technology, as the lefty-righty, uppy-downy, and twisty-turvy controls would be unnecessary, and just a faster-slower one needed. Well, might need additional tech to handle plebs blocking doors and the like, but more powerful actuators might suffice.
Why not? Bob twat-dangle Crow. That's why we can't have it.
I ride on fully automated Metro trains every working day. The Lille Metro has been fully automatic, with NO staff on the trains (not even a DLR-style conductor) since it was opened 30 years ago. It gave me the creepie-crawlies the first few times, but now it's just the way the Metro works. And it means I get trains once a minute at rush hour.
It seems that airport automatic trams have been around forever, but that isn't the case; one (I forget which) when first turned on for public use, kept stopping midway between terminals, because, as it turned out, its computer interpreted any uncommanded change, even burnt out pilot lamps, as safety-critical failures.
Not to mention they never get tired or lose focus, don't need food or drink, don't need safety equipment/ejector seats of any kind, or a cockpit and if they crash you've only lost the airframe and not a pilot that cost millions to train.
You can only resist progress for so long, there's just way too many advantages to these things.
"Aircraft carriers do not "bob up and down"
You are correct. But they sure as hell do pitch and roll. Most U.S. Navy surface ships have a design limit of 10 degrees pitch but can experience up to 11 degrees in heavy seas and maintain steerage. 1-2 degrees in a 'light' sea state is common. Pitch periods between perpendiculars of 8 seconds are experienced with carrier length ships in high seas.
So it isn't a bob up and down motion. It is an incredible pitching motion that feels like a HUGE roller coaster at either end of the vessel and never stops and isn't remotely regular. Roll in large ships is somewhat mitigated by the sheer size of the vessels but it is nowhere close to stable.
I have long ago forgotten the math, but you can plot pitch and roll of the vessel on a curve and do up a nifty sketch for yourself if the sheer scale movement isn't clear. The movies do a good job of making carriers and tankers look big, but absolutely nothing is stable in open water, even on a good day. Bad days are bad.
That's exactly what I mean Don, the period is far longer so it's not like you're landing on something which is jittering on every wave, the motion is regular and predictable and slow. The deck will not suddenly buck as you touch down.
I din't claim the ship wouldn't move, that would be daft.
...Now lets see you do it in a storm, with wet decks, and the boat bobbing up and down in a strong swell....
Do you really think that a robot won't beat a human hands down on a complex moving situation with gusting wind and shifting target? That's the kind of calculation they are built for...
Night won't affect it as it is not as dependent on visuals as a human would be. Looking at this thread i think most commentards do not seem to get that this is not meant to belittle naval pilots skill (its fscking hard to do even in simulators let alone when your neck is on the line) but more that the robotics/coders/aviation engineers have increased their skills.
Sure it might not be able to save itself if battle damaged beyond a certain point but who cares? Navy pilots that land severely damaged kit in barely flying condition are doing it to save themselves (the plane is often a write off at that point anyway). Ejection seats save your life but they often break legs and can cause your back to compress and end a military pilots career.
I can't think of any good reason not qualify the automatic pilots for deck landings, if that is what the software is supposed to do then every time there is a change in the software I think I'd want it validated as still being a competent pilot, before I would let it slap into the back of my ship filled with fuel and possibly other explosives.
Oddly enough if you're above landing weight you dump fuel rather than weapons. Rarely do you plan to launch with so many stores you'll have to jettison them before landing, i.e. anything over that you'd expect to deliver somewhere else, ballistically.
In fact Sea Harriers had an issue with repeated landings causing damage to AMRAAM missiles when they were carried for multiple flights.
I wouldn't really trust Lewis' articles on aviation, as a pilot he makes a great diver...
The launch crew's visual queue is not all for the pilot. Remember, the catapult operator (sitting in a semi-sunk control box), the tower (on the Island), the other folks around the deck all takes their queue from the launch crew and his visual signals. Launching a plane from the deck of a carrier is less about a pilot than everyone else around. Only after he's launched does he really take control.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
