From Ars Technica
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 hostile fire or terrible flight conditions, a significant percentage of drone crashes is caused by human error. A December 2004 Federal Aviation Administration (FAA) study of Defense Department drone crashes found human factors to be a causal factor in about a third of the cases the researchers examined.
But as four human factors engineering researchers have found, sometimes the accidents are by design. That is, the design of the systems that operators use to fly the drones are so bad that they invite accidents. A recent Ergonomics in Design article reported that a small but significant number of crashes could be directly attributed to bad ergonomics on ground control station hardware. These factors may have played a major part in crashes that were attributed to other causes.
Take, for example, one drone crash in 2006. As the operator brought the drone in for a landing, he meant to flip the landing gear button on the control joystick but accidentally hit the nearby ignition switch instead—shutting off the engine in mid-flight. The $1.5 million drone plummeted to the ground, a total loss. On another occasion, glare on a screen was so bad that a drone operator couldn't read an alert and mistook it for a landing signal—again killing the engines before the drone had landed.
Read more here
Comments
Good point Gary. And home brew stuff by definition means the pilot is the creator or has a vested interest in true understanding of the system. In contrast, G.I. Joe has been trained just enough to get the job done. His superiors also make sure the defense contractors don't blame him for the failure. Not wanting to lose a contract, the defense contractors blame the cause of crash on the obligatory yet nebulous "human factors". Reading the report, G.I. Joe's superiors instruct the defense contractor to remove the button that says "kill engine" because clearly, it was the cause of the crash. By the way, I've just described why defense contractor drones cost so much money.
Allegedly the photo shows a US Marine covering a roof-crashed UAVin 2006.
There is definitely fluid leaking from it. HERE is a more detailed explanation of the pic.
Regarding control layout, there is much legacy design being used. Translating our thoughts into commands via an interface is a challenge. Witness the variation in game controllers over the years.I can use my Spektrum DX6i well enough but when picking up a PS3 controller... I have to be retrained.
-=Doug
Sorry John, I can't agree. I can operate my 9C "blindly", but it doesn't mean I never hit the wrong switch, especially in a panic. The science, and the data in this case, would argue that not designing a controller ergonomically and with a logical layout will eventually lead to an accident.
Like, based on one incident they mentioned... the "ignition" switch right beside the "landing gear" switch on the joystick? What are those two things doing beside eachother, and why are they on a joystick in the first place? The ignition switch is a "wings fall off" button, it should be on a panel, probably in a section of switches related to startup, and have a cover over it.
On my 9C, if I had an ignition button, and a landing gear button, they likely would be right beside eachother. Actually, I can just check my old Spitfire setup... Yep, Gear is on B, Idle Down is on E and Throttle Cut is on F. Not quite *right* beside eachother, but close enough. And the Idle Down Throttle Cut button isn't quite as significant Ignition, neither would kill the engine unless it was already at an idle.
Most of these incidents have EGO stamped all over them. Something that is not in short supply here. To claim better incident rate for homebrew gear is plain nonsense. We don't log hours or equipment type accurately enough to make comparisons.
The sooner CRM and aviation knowledge courses start the better.
Once again some of us are ahead of the mil stuff. I just got a voice activation board that i plan on installing in my transmitter box along with an old apm1 .4 board between the easyVR and the 3dr radio, as well as buttons and a display. The vr module would be great for mode changes and requests for data ect. And I have an mp3 module to listen to. Just talk to your drone and it talks back. I have all the hardware, now i just need to get to work on the hard stuff, integration and code.
Thumbs up! Andy! wings stay on!
Thanks for posting this. Though the success of the R/C radio is self-evident, I believe it breaks down as taskings become more challenging. The tiny displays, the inability to add new features without significant time lag, and the tie to a particular manufacturer, among many others, are too limiting for me. And the controls aren't intuitive, and don't become intuitive, without a substantial investment of time (as John alluded to). I did have a LOL moment when reading the comments from the ars technica link (and I promise not to incorporate such a feature in my project):
Because then the operator says "what's it doing now?" Which isn't always a good thing.
When the operator makes a bad move can't the $$$$$$$ drone figure it out and RTL on it's own?Aren't they getting smarter all the time ?
OMG! I didnt see the guy crouched down next to the fuse to start with, I assumed it was a little 2M wingspan type setup at first glance... DAMN!
-
1
-
2
of 2 Next