'Phantom Ray' robot warjet to ride atop NASA shuttle-carrier 747 The "Phantom Ray" robot Stealth combat jet under development by US aerospace mammoth Boeing is ready to begin trials, according to the company. The machine will now be flown to the military test centre, Edwards airforce base in California, mounted atop one of NASA's space-shuttle-carrying jumbo jets. The Phantom Ray UCAV …
It's not like the empire is broke and washed ashore.
Still, I can't wait for the headlines this will make when a fully autonomous attack mistakenly blows up a few hundred rich Pakistani in a well-populated inner city because of an off-by-one error. That's gonna be The Lulz.
Beat me to it! Do these boffins not watch movies? I know we are past 1997 but be somewhere else on august 29th in case the machines have a sense of humour...
"It becomes self-aware at 2:14 a.m. Eastern time, August 29th. In a panic, they try to pull the plug."
I for one welcome our new stealth flying robo overlords..
Or is there not already a wealth of AI piloting programs contained within just about every flight/space combat sim, that already doesn't crash into the ground, flies evasively, follows bombing routes etc etc?
I realise you can't just take Tie Fighter and put it onto "Phantom Ray" but these games must be a good place to start? I mean the various militaries already use these games for pilot training, what's wrong with taking the next step?
"As for the Phantom Ray, Boeing intends to cough up for a six-month programme of flight testing at Edwards. If no government customer comes forward after that, the machine may find itself back in mothballs." ... Doesn't that smack of the way the usual Hollywood plots/ploys are (re-)imagined for us? Here's the trailer...
"The Phantom Ray - Truely the lone-ranger of the airways had grown tired of 'mothballs'.....
.... It was a time for decision.
.....It was a time for action.
.....It was time to....
Come out the closet!"
"....call yourself a warrior? Unclip that hanger, get out there and show 'em..."
"Don't wear it grey - If you're goin' to fight, do it bright and Clash!!"
"....your airframe's writing checks you're AI can't cash..."
Phantom Ray - The Dreaming's Over
www.thephantomraythedreamingsover.com available soon!
Human Brain has about 100*10^9 nerve cells. Each cell is connected to 10000 other cells on average. That's 100*10^9 nonlinear elements connected in highly funny ways no scientist or engineer understands. That's what delivers human reasoning, decision-making, improvisation and also fighting capability. It has been honed by evolution over hundreds of million of years.
This fly-robot-thing might have the equivalent number of RAM cells, but certainly not this level of interconnectedness, non-linearity and irregularity. That means it will be easy to outfox to whoever acquires this flything optically, acoustically, chemically or with some sophisticated HF/UHF device.
A competent guy in his Mig21 will down it as soon as the thingy is acquired by his eyes with 90% success rate.
The second part of this lecture is homework. Research "history of automatic air-to-air missiles and fighter cannons in the vietnam war. Include analysis of the F4's performance in your report".
First of all, your meatball won't be able to lock his eyes befare his antique mig 21 is down.
Second, you don't have to replicate a human brain. In fact, you really don't want to, as human brains are terribly inefficient, and prone to errors, not to mention how slow they are or have very little "memory" they have.
Third. It is understood how nerves and individual cells interconnect, and it CAN be replicated, only that it is very different to replicate 10.000 cells or millions. What we don't really get is the globally detailed image (the forest and exactly how it works). We do mostly understand the tree
4. As for human evolution, whe did not evolve to fly planes, drive cars, etc. we are good at learning from experience and clasifying objects, as well as taking decisions and making analogies.
5. Autonomous robotic planes are the future, and vietnam is well into the past, I could also point out that cloth biplanes can't destroy my fleet of dreadnoughts..
There exist other "killing machines" with quite simple logic and lethal Effect Systems. Think of a Tiger. Stupid brain, powerful feet, powerful body, extremely powerful teeth.
Guess what ?
BRAIN is wiping Tiger off the face of the planet. BRAIN knows more about Tigers than the Tiger itself. BRAIN knows it's instincts and it will simply trap Tiger, before Tiger can use the claws and teeth. BRAIN needs Tiger territory and resources and kills off Tiger slowly but surely.
This flything is much less intelligent than a Tiger. Probably less than an ant.
BRAIN will cobble together some serious trap for ant-flything and kill one after the other. BRAIN already defeated the mighty F4, the mighty B52 and the mighty Mi24. It will also defeat this one.
BRAIN defeated Sidewinder.
And your Fairey Swordfish is my Mig21.
http://en.wikipedia.org/wiki/Fairey_Swordfish
My guess is that you could deploy a couple of hundred of these things against a specific target knowing most will return. If you call in an airstrike you risk dead pilots on fox news and that is not polically expedient. Either that or artillery and which is better Autonomous slightly intelligent drones delivering a lethal payload or autonomous non intelligent artillery.
After all, sending in meatbags equipped with full human intelligence has been know to go wrong in the past so where is the downside?
- the plane will be commanded to kill by a real person, and the decisions on the way to target will be about 'abort or continue' - so the robot is actually deciding not to kill, otherwise it's continuing along it's programmed flight profile and making no decisions.
I agree that it's splitting hairs, but it can be argued that people have already made the decision before the mission goes on automatic.
Simple, have it fly most of the way out to enemy territory, then assume radio silence as a human pilot would. And just like human strategic bomber pilots of the past, once you enter radio silence there is no going back, no abort.
Of course the difference here being that there could be a sufficiently heavily encrypted abort. The decision to bomb is made when the robot plane receives its instructions to attack and that decision is made by a human.
1) It can carry a variety of weapons; cruise missiles are limited to one.
2) It can attack several targets far apart; cruise missiles are limited to one (OK, you can kill multiple targets but they have to be co-located).
3) It can be re-used; cruise missiles are designed for one-way operations.
4) Reusability MIGHT make the UCAV cheaper than cruise missiles in the long run, but it depends on a lot of factors.
As for script kiddies, sorry, no clue. However, to accomplish such a feat, the intercept would have to successfully hide amongst the real flying schedule to convince maintainers and weapons loaders to prepare it for flight (as well as the armament guys to build the munitions and deliver to the aircraft). It might be slightly more credible if the intercept were to occur in flight, though then you take a gamble of what its loaded with might not fit your plans.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
