DARPA Flies a Black Hawk Helicopter Without a Pilot For 30 Minutes (cnet.com) 81
The Defense Advanced Research Projects Agency has flown a UH-60A Black Hawk helicopter without a pilot for the first time ever. CNET reports: DARPA's Aircrew Labor In-Cockpit Automation System program was used to fly the helicopter on autopilot over Fort Campbell, Kentucky, on Saturday. The Black Hawk was kitted out with Sikorsky Matrix autonomous flying technology, and DARPA says it repeated the "uninhabited flight" on Monday. "Pilots can focus on mission management instead of the mechanics," Stuart Young, program manager in DARPA's Tactical Technology Office, said in a statement. "ALIAS ... includes the ability to operate aircraft at all times of the day or night, with and without pilots, and in a variety of difficult conditions, such as contested, congested, and degraded visual environments."
Not a big deal anymore (Score:1)
Re: (Score:3, Informative)
It's a bit more than a regular autopilot: As an automation system, ALIAS aims to support execution of an entire mission from takeoff to landing, even in the face of contingency events such as aircraft system failures. ALIAS system attributes, such as persistent-state monitoring and rapid recall of flight procedures, would further enhance flight safety. Easy-to-use touch and voice interfaces would facilitate supervisor-ALIAS interaction. ALIAS would also provide a platform for integrating additional automati
but combat and damage? (Score:2)
but combat and damage?
Re: (Score:3)
30 minutes - why is this news? Autopilots for helicopters are old hat.
Would it help if DARPA changed their name to Skynet?
You know, in case you might still be wondering why this is suddenly a story...
Re: (Score:1)
Most people overestimate what autopilot typically means, in most aircraft it's little more than things like altitude or attitude hold. Some support transition between steer points and so forth, and a few, but not all military jets have automatic landing systems like the F-18s ACLS, but anyone that thinks your typical autopilot just does everything from startup and takeoff, to flight and landing, parking, and shutdown has no idea what the fuck they're talking about.
Hell, if you're in a 737 Max then even alti
Re: (Score:2)
Re:Not a big deal anymore (Score:5, Insightful)
30 minutes - why is this news? Autopilots for helicopters are old hat.
There's a big difference between "keep her steady for a while" and flying an entire mission.
Re: Not a big deal anymore (Score:2, Flamebait)
Keeping her steady is the hardest part. An Arduino with an off the shelf IMU and GPS can do the whole mission, though. No part of this is impressive by modern standards.
Re: (Score:2)
An Arduino with an off the shelf IMU and GPS can do the whole mission, though.
A frick'n Arduino ain't gonna do target acquisition or skyhook a LRRP from a hot LZ.
Re: (Score:3)
Sorry, I meant fly the whole mission. You might need a raspi to do your fancy pantsy target acquisition bullshit. A fucking raspi zero has enough balls to recognize thousands of objects from an onboard database, and aim a weapon. Frankly you don't even need realtime processing for that.
We have literally had self-stabilizing heli tech since the Comanche project. The whole idea that we should be impressed by a self-flying helicopter today is fucking stupid. Again, fucking cheap off the shelf hardware with emb
Re: (Score:2)
I've seen one too many DJI drones fly away for no reason to trust in any off the shelf hardware. I'm sure the military has the forethought to design a bulletproof system (sorry for the pun). Systems that don't compete for resources that don't exist. Systems that don't crash (literally) and have multiple redundant failure modes while identifying software and system faults from the logs.
We haven't even come to the part where the different system sensors with different voltages and different communication i
Re: (Score:1)
An Arduino with an off the shelf IMU and GPS can do the whole mission, though.
A frick'n Arduino ain't gonna do target acquisition or skyhook a LRRP from a hot LZ.
The US Army has phased out all LRRPs and LRS units.
Re: (Score:3)
Obstacle avoidance, especially when landing is much more difficult.
Re: (Score:2)
Obstacle avoidance, especially when landing is much more difficult.
It takes more sensors or enough computer to keep track of what's where. But again, you can do that with a very trivial computer as well. However, if you program in waypoints well enough ahead of time, it's not that big a deal; just don't land it where there's obstacles. Since we always have high-res maps with elevation of mission areas, since these exist for the whole planet already and also because we have overflight, satellites etc. the only real danger is that someone will know where you're going and put
Re: (Score:1)
Re: (Score:2)
Focus on mission management my ass (Score:5, Insightful)
What they're developing is unmanned killing machine. Just tell it like it is already. This is DARPA: it won't be used to dust crops and improve agriculture.
Humanity is disgusting...
Re: (Score:2)
Re: (Score:1)
"SWITCH TO MANUAL!"
"...err
what could go wrong
ah , but it must be organic blockchain tribes who are synergetically designed !
ofcourse
Re: Focus on mission management my ass (Score:1)
Your slashdot chatbot is as unimpressive as it is unintelligible.
Re: (Score:1)
Re: (Score:2)
What they're developing is unmanned killing machine. Just tell it like it is already. This is DARPA: it won't be used to dust crops and improve agriculture.
Humanity is disgusting...
What do you mean. They could be working on this [youtube.com]. Or maybe you were thinking of this [youtube.com].
Re:Focus on mission management my ass (Score:5, Insightful)
Actually the primary use case for this is medical evac and resupply. For resupply it's a no brainer. For medivac it reduces a crew in danger by 2, presumably you still need the medics.
This is largely funded by Sikorsky, not Boeing who would be doing it with Apaches. It's not so much about making a killing machine as it is for Sikorsky to have a drop in product that keeps their products in high demand.
Re: (Score:3)
If you're going to have it fly itself, and not even include a flight crew, there's no reason to waste space on a cockpit. It makes more sense to cook up a new design to put your fancy tech into in order to maximize your cargo volume capacity.
Re: (Score:3)
This was sold as a retrofit of existing aircraft. You don't develop and try to sell a new aircraft to the army that they hasn't asked for. This allows them to more cheaply develop the technology, get their foot in the door and put them in a better position to pitch and win a contract for brand new aircraft.
For the military the advantage is leveraging this technology to cheaply make boneyards full of retired F16s, F4s, and other aircraft back into viability. Of course this goes back to the original comment
Re: (Score:3)
Actually the primary use case for this is medical evac and resupply. For resupply it's a no brainer. For medivac it reduces a crew in danger by 2, presumably you still need the medics.
This is largely funded by Sikorsky, not Boeing who would be doing it with Apaches. It's not so much about making a killing machine as it is for Sikorsky to have a drop in product that keeps their products in high demand.
Don't ruin the OP's emo/nihilistic self-lamenting narrative with facts and logic, bro. Facts and logic aren't as cathartic as the good old "omg we sux, why wai wai." This is /. after all.
Oh, how Shakespearian. (Score:3)
What they're developing is unmanned killing machine. Just tell it like it is already. This is DARPA: it won't be used to dust crops and improve agriculture.
Humanity is disgusting...
GPS and microwaves want to have a word with you. Sure, this has specific military applications, but most military applications have always had an spillover into civilian space.
We have had unnamed killing machines for a while. Drones, and before that missiles, and before that artillery... all the way back in the day when Grook back in the Neolithic discovered he could hurl a spear point or set up a trap with a snare or "punji sticks" to defend his village against instruders.
Yes, humanity is disgusting...
What's more important? (Score:3, Insightful)
Why are you more worried? (Score:3)
It is highly probable that driverless cars will kill far few people than those driven by people. Of course there will be press stories about where that isn't the case, but given the present number of deaths on the roads, a substantial reduction in the toll would be a GOOD THING(tm).
Re:Why are you more worried? (Score:4, Interesting)
The press is already salivating, waiting for the next Tesla crash.
Meanwhile the human drivers are killing 38000-odd people per year in the USA alone.
(and the number is rising, despite all the advances in vehicle safety)
Re: (Score:3)
No doubt caused by a lot of hubris coming from Tesla. Especially since the Tesla driver assistance systems are so poorly done they cause accidents in themselves. (Running into emergency vehicles is not a good look, as is many other Tesla faux pas like gaming while the car is in motion and refusal to implement proper driver monitoring systems when using assistance systems, instead opting for trivially bypassable ones).
Excluding those, about t
Re: (Score:2)
Running into emergency vehicles is not a good look
Sure, but humans do that too. Some of them get flustered by the sirens and lights and we tend to drive towards what we're looking at. That's why when you apex a turn you look at the apex until you get there, then look at the exit. I personally don't think Tesla has enough hardware onboard to do a substantially human-superior job of driving in all conditions, but the criteria isn't "is this embarrassing" but "is this safer than a human driver". Humans kill people while driving, so if self-driving cars simply
Re: (Score:2)
Humans manage it with two, medium-resolution imaging units that only work in the visible light spectrum. They can't even see through fog.
Re: (Score:3)
Humans manage it with two, medium-resolution imaging units that only work in the visible light spectrum.
You're forgetting that those imaging units include cabling which does processing before information even reaches the brain, and then a connection to a big squishy brain that does a lot of spooky stuff that computers aren't good at yet. Don't get me wrong, they also do a lot of repetitive stuff that we aren't good at repeating exactly, so they have their uses, but they don't do everything we do as well as we do it either.
They can't even see through fog.
Teslas can't do a good job of seeing through fog with the hardware they have now, either
Re: (Score:3)
On the plus side: The car won't be checking social media or nodding off while driving.
Re: (Score:3)
Yes, that kind of focus and dedication is a massive plus. I'm not against self-driving cars, I'm not even against Tesla, I'm just in the camp that believes more kinds of sensor data is better than just cameras — especially when it includes explicit rather than derived depth information.
Re: (Score:2)
I'd imagine the Tesla's stereo depth perception is quite good.
Humans manage depth perception with sensors that are only about three inches apart. The Tesla cameras are a lot further apart than that.
Re: (Score:2)
They only have one camera which can see through fog, and not very well, so in the fog they have no depth perception. Which is why they need RADAR.
Everyone else uses RADAR for this purpose, and with good reason.
Re: (Score:3, Insightful)
Of those 38k fatal accidents every year, some will be down to recklessness by the driver. Those drivers are punished if they survive. Jail time, fines. Tesla cars on autopilot have killed people, but Elon Musk isn't being prosecuted. It's up to individuals to sue Tesla.
Corporations are good at diluting blame to the point where nobody can be held accountable.
Re: (Score:1)
Re:Why are you more worried? (Score:4, Insightful)
The real issue is that only the cases a decent driver would have avoided will be published. ...) could avoid 1000 deaths from possible driver error (sleeping, driving at summer speed in winter conditions, one second of inattentiveness, going too fast in a curve where he drove a thousand times, rubber necking, ...), kill 10 people in cases that couldn't be avoided by the best driver in the world, and people will only hear about the one case (due to one reason or another) when a decent driver would have fared better.
As such, Tesla (or whatever driverless car, Waymo, Mercedes,
Kind of a "survivorship bias" - only the catastrophic enough news survive the "20 articles per day" or whatever quota of the news agencies.
Re: (Score:2)
I'm more worried about pilot-less Tesla cars
Yes, because endless human warmongering, never harmed anyone.
And eventually AI on a bad day can probably do better than the social media addict behind the wheel who can't put their fucking phone for more than 5 minutes if their life depended on it.
Literally.
It won't be hard for AI to drive safer than 40,000 deaths a year. Hacking will become your primary concern. Oh, and that goes for pilot-less attack helicoptors too.
Instead of passenger drones and flying car nonsens (Score:3)
... wouldn't it make more sense to make normal helicopters much easier to fly using this type of technology? Not necessarily full autopilot but have fly by wire whereby the computer does the much harder than fixed wing flying stuff such as maintaining hover position, take off, landing etc that make learning to fly a chopper a very steep learning curve that some people never climb?
Re: (Score:2)
... wouldn't it make more sense to make normal helicopters much easier to fly using this type of technology? Not necessarily full autopilot but have fly by wire whereby the computer does the much harder than fixed wing flying stuff such as maintaining hover position, take off, landing etc that make learning to fly a chopper a very steep learning curve that some people never climb?
Well, I would say that you still need to ensure that pilots are training fairly evenly on all aspects of aircraft movement, for one main reason; Shit Happens. That includes when autopilot systems fail at the worst possible time, and a pilot isn't (muscle-memory) trained well enough to react quickly enough.
But to your point, is there a massive shortage of helo pilots in the world, and training complexity is the reason? Do they have an unusually higher track record of accidents and deaths due to this, in tr
Re: (Score:3)
But to your point, is there a massive shortage of helo pilots in the world, and training complexity is the reason?
I'd say that training cost is the reason, especially when people have to pay for it themselves just to get their first license.
I also think that helicopters could trivially do the takeoff/landing parts automatically, and they're the most dangerous parts.
('trivially' for the people who work on that sort of thing)
By extension: If a fly-by-wire helicopter can do takeoffs and landings, it could also have a more intuitive control interface where moving the stick does what you think it should without having to
Re: (Score:2)
Indeed. It should be a no brainer to have a control interface very similar to a fixed wing except for VTOL.
Re: (Score:2)
But to your point, is there a massive shortage of helo pilots in the world, and training complexity is the reason?
I'd say that training cost is the reason, especially when people have to pay for it themselves just to get their first license.
Compared to the cost of say, a college degree? It's a drop in the bucket. And that's a career you're creating, not something you merely polish and put up on your hobby mantle. I'd be willing to bet that fear of heights keeps the overwhelming majority of humans out of the cockpit. That includes millionaires.
Re: Instead of passenger drones and flying car non (Score:2)
Itâ(TM)s a career⦠until the automated technology discussed in this article eliminates that field of work altogether.
Re: (Score:2)
Except most chopper crashes are pilot error, not mechanical failure. If there was computer-says-no in the loop it might reduce them somewhat.
Re: (Score:2)
Re: (Score:1)
Re: (Score:3)
And yet, actual helos generally do not, beyond basics like hover hold. Here's a stab at an answer:
https://aviation.stackexchange... [stackexchange.com]
Re: (Score:2)
In the late 80's USSR was preparing the Buran space shuttle to be able to fly without a crew. This is hardly impressive...
Did they do it? No. Therefore, this is impressive.
Anyone can say they're planning to do something, but until they do it it's only words.
Re:The USSR could fly a space shuttle with no pilo (Score:4, Informative)
Yes, they did do it... https://www.buran-energia.com/... [buran-energia.com]
ever flown a helicopter? (Score:2)
The only way to hover is without thinking about it.
All the action is from your ass to your hands and feet. If you try to think about staying in one spot you'll be 2 or 3 seconds behind. Like learning to ride a bicycle.
My instructor had me trying to push a big paint can around with the skid. As soon as you concentrate on something else the "can't keep it in a 20 acre field" effect goes away.
Re: (Score:2)
That's because humans have crappy sensors, and also a pretty lousy system for filtering sensor input. It's actually very powerful, but what makes it bad is that it's extremely uneven and thus unpredictable. You can't know what the actual jitter is going to be. This is why it takes a well-trained human to hover a heli, but a $4 arduino nano knockoff with a $10 IMU can easily hover a helicopter or multicopter.
Regards to Captain Dunsel (Score:1)
"I think I'm a drone now" - autopilot computer parodying Weird Al
Good (Score:2)
My understanding is heli's aren't all that reliable relative to other aircraft. If we can get our soldiers out of harms way, that's always a good thing. I wouldn't go up in one unless my life depended on it.
Re: (Score:2)
If you just flew a helicopter from airport to airport like a fixed wing they would have accident rates very much lower than small fixed wing airplanes.
It's the jobs their asked to do that makes them more dangerous.
An engine failure in a helicopter cruising along is nothing more than picking out a spot a little bigger than the aircraft and landing. the same problem in a fixed wing means finding something like a runway.
Re: (Score:2)
An engine failure in a helicopter cruising along is nothing more than picking out a spot a little bigger than the aircraft and landing.
Uh, it's a lot more complicated than that. However, you still might be right about safety. Over 95% of autorotation attempts are successful. Sadly, I don't have a statistic on dead-stick landings to begin to make a comparison with, but then you'd have to also compare passenger capacity and so on...
Re: (Score:1)
I've either done or been in a helicopter for hundreds, if not thousands of practice autos. Also 3 actual engine failure autos.
Helicopter pilots used to practice an autorotation just about every chance they got. We had a 206 that did pole pickup that was so light even I could do a hovering auto and pick it back up and do another one with the engine at idle.
Those days of hovering next to a telephone sized pole snagging a bag full of checks above a bank was the wild west of helicopters.
Re: (Score:2)
Re: (Score:2)
30 years ago even turbine engines weren't as reliable as today. but they were better than recips!
Every pilot back then did so many practice autos that a real engine out was usually just go to the spot that you'd seen 20 seconds ago.
Re: (Score:2)
Nah, the statistically safest way to fly is for superman to carry you--the writers can't let you die, as it would tarnish the hero! :_)
Re: (Score:1)
Christopher Reeve or George Reeve ?
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Ha! I wasn't a licensed pilot. i was a mechanic for 45 years. But the pilots always wanted to be friends with us so we got to fly when no pax were aboard.
First time flying one was a CH-47. it was the biggest and easiest to fly!
How to remove the human factor from... (Score:3)
warmongering, destruction, and killing.
The ultimate goal of the lawless Military Industrial Complex that both Eisenhower and John F. Kennedy warned us about.
Nothing to brag about (Score:2)
Zeppelins have been doing for 100 years.