Slashdot Mirror
Hit the Wrong Button, Drone Goes Boom
ios and web coder writes "An article at Ars notes, 'Unmanned aircraft crash. In fact, they crash a lot—though there's no recent specific data, the Congressional Research Service reported last year that despite improvements, "the accident rate for unmanned aircraft is still far above that of manned aircraft.' And while many of those accidents can be attributed to being exposed to hostile fire or operating in conditions when aircraft normally wouldn't, a significant percentage of drone crashes is caused by human error. A December 2004 FAA study of Defense Department drone crashes found human factors to be a causal factor in about a third of the cases they examined (PDF).' Drones are un-cheap. As yesterday's Super Hornet story noted, they are cheaper than manned planes... but not that much cheaper. Expect them to get more expensive. Also, as they get armed, the price paid for a bad UX decision could become quite tragic."
It's who they have flying these things. You would think they could do better.
The preceding post was not a Slashvertisement.
"the accident rate for unmanned aircraft is still far above that of manned aircraft.'
In addition to being cheaper, unmanned aircraft have no people on them. So much less of an incentive to worry about safety.
Until the fall on someone's head, that is.
It's "Death from Above due to incompetence" too? That makes me feel so much better. Overall, the people who're advocating these back in the States are missing on little detail: Yes, ostensibly, you will feel like you're actually doing something to curb crime, make the neighborhood safer [insert some other thinly veiled justification for surveillance] etc... But these are not foolproof devices. There are always better fools.
If computers were people, I'd be a misanthrope.
I think that's the right button.
We'll only use those drones abroad where nobody cares whether we drop them on some brown skinned folks. The world already doesn't give half a shit about it when we blow up a few houses 'cause someone heard someone consider pondering that there might be a terrorist somewhere in the general area. And if those things fall out of the sky, it will make our friends happy who sell them. Think of the jobs we create that way!
Also, think how we protect our valuable fighters. You know how long it takes to train a fighter pilot? And how expensive that is? Every single one of them that gets shot down or even killed has to be replaced and that person first has to be trained to the same level of experience. This ain't some idiot grunt that we use in the ground troops which pretty much serves as a way to hide unemployment, those are people who actually have to know something! They could technically actually have a decent job, those don't come in endless supply!
We used to have a Bill of Rights. Now, with the rights gone, all we have left is the bill.
"Un-cheap" is not a word. TFS should say "not cheap". Can we please have some minimal editing for language in future?
const int one = 65536; (Silvermoon, Texture.cs)
SJW, n: "Someone I don't like, and by the way I'm a fuckwit" - AC
just put another quarter in to start a new game.
Is that data even valid anymore?
You think some armchair-pilot cares, that's cute. It's not like he/she is sitting in the cockpit. The reason why these are cheap is beacause the human element has been taken out (to train a pilot is around $2M last time I checked, and it's been a decade.) So yeah, they're gonna crash more than manned flights. Who cares? You want the pilot to care? Sitting his/her ass in a manned jet is the only sure way to make that happen, and yes, that's expensive. It's easier/cheaper to train a mechanic to unbox another drone and screw the wings on...
One third is a surprisingly low percentage. The number of manned small plane crashes caused by human error is probably close to two thirds.
So while I'm sure lot could be done to improve the ergonomics of the pilot, it sounds like the drones' mechanical failure rate is a more worrying problem.
How can I believe you when you tell me what I don't want to hear?
Make the drones totally autonomous.
Now, excuse me while I go hide under a tree.
“He’s not deformed, he’s just drunk!”
Or a large wooden badger.
While I recognize this is not military quality, but for under a grand I can easily build a self navigating aircraft capable of carrying all sorts of surveillance and bombs such as Molotov cocktails. I may not get days of flight time, but I bet I can come pretty close on my first try. I don't even have to do any real work, just combine my existing R/C modeling skills with some parts from diydrones.com. The hardware is cheap, comes with software thats pretty solid from the start and easy as hell to extend.
I can not possibly imagine there is a real reason for ridiculous expense other than government waste. The only time any of my aircraft have crashed in the last 10 years is because I literally flew them into the ground (or lake as I've been doing seaplanes recently), not because of some sort of 'mishap'. In every case had I not been trying extreme maneuvers it never would have happened.
Persistent Volume manager for Kubernetes - https://github.com/dwimsey/openshift-pvmanager
Reminded me of this...
"Slashdot - News and Chat Sites Deviant". (Click "homepage" link above for details).
The risk of crashing a manned plane is your death...
Flying a drone is more like a video game, you don't have any fear of personal injury so you expend less effort to avoid crashing.
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
I imagine that training with an UAV simulator is much closer to reality than training with the simulator of a manned aircraft.
from TFA: "Analysis shows that general-purpose computer workstations and UAS GCSes are up to 98% similar. "
There's your problem, right there. Flying a desk is not the same thing as flying a computer.
There's a reason that certain cockpit controls have different shapes. For example, tactile feedback, as long as you're trained to pay attention to it, can spare you the embarrassment of mistaking your flaps from your landing gear. Just in case you've never flown, retracting the gear when you're "going around" is a good idea (it reduces drag and increases your rate of climb) but retracting your flaps at the wrong time can kill you during the same procedure.
And where are the software designers in all of this? Flying without any sort of contextually accurate or appropriate sensory feedback creates a deficit. As a pilot, if you can't hear the engine or feel the effects of flight control inputs, you're at a disadvantage. Pilots are taught to pay attention to more than flight gauges and readouts. Software designers have the luxury of setting up AI algorithms that could cross check parameters, provide feedback and require verification before allowing a desk jockey to shut down an engine while an aircraft is at-or-below some critical elevation.
It's nice to know that there are standards for ergonomics and all, but there are reasons that the pilots of commercial aviation operations are still in the plane. Yes, one of them is so the passengers feel better about the experience, but the rest of them have more to do with controlling safety than anything else. It amazes me that to know that software exists target and intercept an incoming ICBM, but the military isn't using software to control its drones (the aircraft, not the 'pilots') more effectively? Me thinks someone's having too much very expensive fun.
CFIT was identified as a cause of 25% of USAF Class A Mishaps between 1993 and 2002.
http://en.wikipedia.org/wiki/CFIT
Controlled (should be "uncontrolled?") Flight Into Terrain. Simply put, "the aircraft was working fine until someone drove it into the ground".
It's easy to do, especially with high workload in a fast jet, in a combat situation.
I guess that the UAV technology is still immature.
The current generation of drones are *much* cheaper than their manned counterparts. When the DoD reports "per unit cost" of drones, most of the time that cost includes multiple aircraft. Per aircraft cost of a Predator B (MQ-9) is ~$10mil. The full reported unit cost of ~$50 million includes *four* aircraft and all the hardware to operate them. Not to mention reduced training, maintenance, and pretty much every other sort of cost. The Air Force hates using drones, and they resisted widespread adoption for so long, but even they couldn't deny the economics of the platform.
I'd imagine that a drone could kill people on the ground if it hit them? But they'd probably be foreign, so that's okay.
Perhaps to incentivise the drone pilots to not crash, they should sit them in ejector seats? You know, inside their office!
You've got to be shitting me.
A news story based on a decade-old study?
In other news, a 2004 study shows that your iPad does not exist!
You are welcome on my lawn.
The reality is that drones are much easier to fly than regular aircraft. Flying a drone is a point and click affair; most accidents are caused by people trying to fly it like a conventional aircraft. IMO, the Air Force's insistence on re-purposing fighter pilots as drone pilots is responsible for a great deal of the high accident rate. The Army, in contrast, doesn't use "pilots" to operate their drones and yet enjoy a lower accident rate.
You push the wrong button down
The drone careens round and round
And slashmissions come out
Actually, UAV pilots are MUCH cheaper. The Army uses enlisted soldiers to fly its drones. The Air Force used to use "real" pilots, but finally caved and now trains cadets as drone operators without putting them in a plane, ever.
Get humans out of the loop.
-- SkyNet
Have gnu, will travel.
An artist at my University has worked towards trying to raise awareness of these sorts of risks - particularly topical for the San Diego region where we have a confluence of lots of defense companies and high-tech university research. His art piece generated a lot of attention for wanting to stimulate the conversations we'd have when a crash occurs in a residential or otherwise inopportune area, before the event actually happens. http://uccenterfordrones.wordpress.com/regarding-recent-drone-malfunction/ was his piece on it, and one of the many articles explaining the 'hoax' http://www.nbcsandiego.com/news/local/-Source-of-Mystery-Drone-Crash-Revealed-182407811.html
i don't know karate, but i know ca-razy
Also, since it's remote, and not where the 'pilot' is, they have a level of psychological disconnection much greater than that of the pilot of a manned vehicle.
No matter what they do, it will always be a kind of video game.
The theory is that drones should fly better than manned aircraft--makes logical sense since drones are essentially robots/computers.
The experience is that drone crash more often and mainly caused by human error. Since drones are remotely controlled exclusively, not in a supervisory manner.
The practice is that drone control, equipment, telemetry and pilot training, though not classified as manned operation, were based on manned operation principles (e.g. rules of engagement for instance).
Basically the Human Machine Interface, which includes user interfaces, operational protocol, and vehicle capabilities (i.e. features) is based on manned experience and at this point, we can conclude it doesn't work. We are scratching the surface on proper drone oepration. I'm sure experts back in 1995 generals were thinking, "Oh, it's just like manned flight/operation w/o the physical person in the craft... viola! Done..."
BUT, it doesn't mean drones and the concept of drones are less worthy than manned aircraft. The HMI interface is just wrong. And a lot of the autonomous tech and what's being created on the hobby side is showing that there's a better HMI interface.
...is remove any potential human error by developing an artificial intelligence to completely computerize operations. We could call it Skynet.
Are we at the point now that every 'ios and web' coder must consider a default 'UX' for every touchable object on the planet? Or do they really think the 'UX' of killing is to be regarded a major sales point for modern military budgets.
Is it too hard to accept that users still create a large part of their own experience learning any device and is not something which needs to be self-built in, unlike a UI? Where exactly are we aiming to hurl our technological consciousness back towards?
Also, since it's remote, and not where the 'pilot' is, they have a level of psychological disconnection much greater than that of the pilot of a manned vehicle.
No matter what they do, it will always be a kind of video game.
This also makes it easier to kill people.
The militaries (especially in Western nations) are always looking for ways to make it easier for people to 'override' their instinct to not take lives. A lot of military training goes into removing the delay between seeing an enemy human being and pulling the trigger; for infantry its a matter of sheer psychological conditioning and drilling. For drone jockeys, maybe too easy...
In the free world the media isn't government run; the government is media run.
You think some armchair-pilot cares, that's cute. It's not like he/she is sitting in the cockpit. The reason why these are cheap is beacause the human element has been taken out (to train a pilot is around $2M last time I checked, and it's been a decade.) So yeah, they're gonna crash more than manned flights. Who cares? You want the pilot to care? Sitting his/her ass in a manned jet is the only sure way to make that happen, and yes, that's expensive.
It's easier/cheaper to train a mechanic to unbox another drone and screw the wings on...
I can think of two solutions;
Electric shock feedback for the drone jockies, maybe some heating elements wired to the soles of their feet stuff like that. So they feel pain when the drone 'feels pain'.
Alternatively program the drones so that they fear death and don't want to crash.
The first will probably end better than the second...
In the free world the media isn't government run; the government is media run.
Well it kind of makes sense. Drone pilots only need a portion of the training. They don't need to be physically conditioned to deal with g forces, how to survive an ejection and most of their communication (radio/etc) is probably handled by somebody else in the room.
Soon they'll be autonomous, solar powered, and small and smart enough
to track you down and crawl into your ear before they blow your mind out your nostrils.
Ban the earworm now, before it's too late!
We build and operate manned aircrafts to a very very high safety standard.
A simple software upgrade for a manned aircraft takes years to complete, because of the standards employed.
When operating manned aircrafts we have strict standards as well on how to do everything and many many small and large things we don't do.
These are all very limiting both in the cost they impose and in the ability to get the job done.
We use unmanned aircrafts so we operate more freely both when building and when flying them, this comes with a higher accident rate from all causes but that
is the whole point.
We need to remember unmanned aircrafts are not only cheaper then their manned counterparts they also do things the manned counterparts can't or won't do.
The relatives of somewhere around 3000 murdered Pakistanis would like a word with you about verb tenses.
=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Friends don't let friends enable ecmascript.
Once weather, gusty winds etc get into the picture that "under a grand" in a weekend or two isn't going to be enough.
Have you tried our new force-feedback controller? ;-)
It VIBRATES!!!
Down With Slashdot BETA!!! I've been around the corner and seen the oliphant; you can only abuse me from your perspecti
Actually, drone pilots do a basic flying course all the way up through solo flights and cross-country flights. They are just as real pilots as anybody that just got their private pilot's license. They don't have the extensive flight training of their fighter/bomber/transport brothers but they do have extensive training flying their drones.
Yes, the manes proved to induce extreme aeronautical drag while attempting flight.
So they had to remove the manes from the drones.
Thus, the mystery of why there are no known cases of flying lions, is now solved.
Down With Slashdot BETA!!! I've been around the corner and seen the oliphant; you can only abuse me from your perspecti
I think I feel sick now... :-(
Down With Slashdot BETA!!! I've been around the corner and seen the oliphant; you can only abuse me from your perspecti
It funny as the whole drone issue is made out to be something new. I was training horses at the Atlantic City Racetrack in the late 70's. The racetrack was actually 10 miles from Atlantic City proper and at this time the casino industry was just a thought. One afternoon a drone from a nearby military installation actually went off course and crashed in the infield of the racetrack. We thought we were being attacked and had no idea that these things were even in existence. The secrecy was the same then as it is now but in these times the government has a harder time of hiding things from the public.
Pilot's butt isn't on the line...