When Drones Fall From The Sky | Hacker News Reader

When Drones Fall From The Sky

washingtonpost.com

61 pointsilamont45 comments

Only somewhat related to the article, the broad use of the word "drone" to include both military drones (usually gutted aircraft, to my understanding) and civilian drones (usually smaller-scale robotics, like quadcopters) is a little confusing. I kinda wish there was some better term to differentiate them, because while I can see a military drone falling from the sky being a great hazard, a smaller quadcopter falling from the sky isn't going to be that bad. (Bad enough, I guess, but not to this degree.)

Like, for instance, some town in Colorado recently passed (or attempted to pass) an anti-drone ordinance giving citizens the right to shoot down drones. In my mind, this means military drones -- not quadcopters -- but to the average person in this town, it probably means any automated and/or remote-controlled aircraft, including quadcopter-drones.

Amazon was recently talking about using drones to deliver stuff. Delivery quadcopter-drones are way, way, way different from military RC-planes-with-missiles drones. Due to the common label "drone," however, I'm not sure the average American really knows the difference.

Anyway, just thinking out loud.

The difference between unmanned airplanes and quadcopters is especially important, IMO.

Only to the extent that there is a scale issue. If it can fall from the sky and only give someone it lands on a bruise then it isn't much of a safety issue. A big quadrocoptor could do some damage.

Even a big quadrotor is a few orders of magnitude smaller than an unmanned aircraft.

I think that when regulation becomes more mature, weight class and speed will be central issues.

If the only difference is the scale of risk, then you should include many other things, like paper airplanes, kites, and baseballs.

I really appreciate you saying this. As a person starting a business involving UAS (UAV, RPV, "drone") I am really facing a PR problem due to articles like this.

They are not all the same.

Thats the very reason why they use the word drone for everything tho. Thats how news, etc. sell.

The reason they use the word "drone" for all unmanned aircraft is because the word "drone" has been used to refer to unmanned aircraft at least as far back as World War II.

Oh yeah im sure thats why. News are serious business with good information.

I have some experience with building large (2-3 foot rotors) quadrocopters. They are _heavy_. With batteries and motors and no payload, our copter weighed ~10-30 pounds depending on the battery loadout. I wouldn't want a 10 pound object falling at whatever the copter's terminal velocity is, smacking me on the head.

Luckily, in general the larger a quadcopter is, the more expensive it is, and thus the less likely it is to be used by an irresponsible person. This obviously isn't a perfect solution; there could be some rich person willing to spend big bucks on a quadcopter while having no clue what he or she is doing, or in the future there could be a $500 off-the-shelf quadcopter with 2 foot rotors (unlikely for quite a while, but possible). So these things self-regulate to a certain extent. That's why you hear of far fewer problems with 700 sized RC helicopters than with DJI Phantoms, despite the former being far more dangerous.

The issue I think is the cost to build redundant systems. We found that many times, external factors like weather and malfunctions were responsible for the vast majority of our critical failures.

If for example, some cloud cover shifts and we lose connection with our copter for just a few seconds, a lot can happen. Even though we built redundancies into our software (if no connection, hold position), those few seconds of lost control could cause some catastrophic problems. I think this is really the biggest issue surrounding drone flight; it's not the question of "Can we build autonomous drones?", and more of "Can we constantly guarantee control over our drones?"

DJI Phantoms and the more common clones by Walkera and other manufacturers are well known for going on "journeys of self discovery." When connection with remote drops, they are supposed to come back home, but some just tend to fly away, never to be seen again. One of those dropping on your head from 200 feet in the air will kill you!

I have never seen evidence that fly-aways area used by anything other than basic misuse, namely, not verifying that there is a GPS signal before flying. Granted, I think software changes could mitigate this, and I think I remember hearing that DJI released an update along these lines).

Still, misuse of plenty of things can have a probability of killing someone. Automobiles are the most obvious example. Bicycles are a great example, since they can be had for similar prices and similar restrictions (i.e. basically none) as quadcopters. Even a baseball propelled from a bat can kill someone at least as easily (and I would argue much more easily) as a DJI Phantom.

Then you should love this: The university in my town is allowing students to check drones out of the library.

http://www.wtsp.com/story/news/local/2014/06/20/students-tam...

Luckily, there are considerable safety measures so far.

Do you remember when civilian drones were just called RC planes and RC helicopters?

I think with increased use of drones both for attack and surveillance there would also an increase in anti-drone technology.

Is there any easy way to detect and fight drones? Lasers? Jammers? Iran presumably jammed the drone they captured. Can that be done with a transmitter one can build at home?

Drones will be used more and more by law enforcement agencies, and I think eventually they will come with some (at least at first) non-lethal weapons -- flash grenades, rubber bullets, tazers, spikes for your car tires. Ran the red light? -- better watch out, a drone will be dropping tire spikes soon in front of your car.

In that respect I bet counter-drone technology will become interesting. You know harming a helicopter with real police officers or even harming a K-9 officer is a serious offense. Attempted murder, very threatening. Disabling a flying camera hovering over your head is not there yet. So there is less threat of a harsh punishment. Of course laws will catch up and there will probably be special provisions and penalties for disabling or tampering with drones that go beyond run of the mill vandalism.

Assuming you know the frequency a drone is flying on, it is trivial to jam. However a lot of the more advanced flight control systems (even on hobbyist drones) will have a failsafe that causes it to return to home or similar in the case of loss of radio contact. I suppose even in that situation you could use a GPS jammer to stop that from happening. Of course this would all be highly illegal.

Is a directed EMP beam possible to fry some its electronics. Also presumably if it uses GPS and not inertial systems a GPS jammer might be useful.

[This is just theoretical, don't try any of this it is probably highly illegal]

Hardware for autonomous flight will have to include both GPS and inertial sensors. I suppose an EMP beam could theoretically be used to damage the electronics, but it probably isn't very practical and the electronics could easily be designed to shield against this if it became a legitimate threat.

Assuming you know the frequency a drone is flying on, it is trivial to jam.

Hence they make it hard to know the frequency: http://en.wikipedia.org/wiki/Frequency_hopping

Yep, the vast majority of 2.4 GHz hobbyist RC aircraft do this, but you're still subject to noise floor issues. It might take a heck of a lot of power, but there's no technical reason why this stuff still can't be jammed, and it's not that rare to have signal drop-outs caused by (presumably) unintentional interference.

Powering up a multi-megawatt 2.4 GHz jammer will get you arrested. You are not allowed to transmit in the 2.4GHz band over a few watts (and even then) without a license and the rules are very specific for what you can send (in Australia for example you're not allowed to encrypt that traffic).

And frankly, all the panicky idiots who keep salivating at the prospect of wrecking things makes me glad the rules for radio transmissions are the way they are. Heads off a heck of a lot of problems.

My favorite comment was the user who was so worried about the danger of drones falling on people in Yosemite that he/she would throw a rock to knock the drone out of the sky. Solid logic.

You don't need multiple megawatts. You just need to be louder than the original transmitter. And that's not all that hard, especially if you can use a directional antenna pointing in the direction of the drone.

For frequency hopping radio links, you would need to be louder than the original transmitter on every part of the spectrum the radio is jumping to.

Being 'dirty' is a lot easier than being clean. The bands are typically not all that wide and to mess them up is easier than to be nice. A sawtooth modulator with a suitably high frequency modulating a carrier will disrupt a significant chunk of the spectrum with reasonably little power input. After all you only need to disrupt part of each packet to disrupt the packet. Jamming is a lot easier than getting a nicely modulated signal through.

"In that respect I bet counter-drone technology will become interesting."

How about a net? http://commons.wikimedia.org/wiki/File:Pioneer_RPV_recovered...

I find it a bit odd they crashed one because the pilot didn't realize it was upside down, I would expect there would be some intelligence to the avionics, like - NO DAVE I WILL NOT DO THAT.

Let alone automating landings, I guess corners have to be cut in places.

Now as for civilian drones, liability laws will quickly sort this out. The desire to have them will come up against the costs to insure them leading to better designs and rapid improvements. The military can afford to lose its toys, private individuals don't usually have that luxury

> I find it a bit odd they crashed one because the pilot didn't realize it was upside down, I would expect there would be some intelligence to the avionics, like - NO DAVE I WILL NOT DO THAT.

There's an attitude indicator, I'm sure, just like there is in any aircraft. It seems like that ought to be enough.

maybe they just felt like throwing that in there. its actually possible for pilots of real airplanes to confuse ground/sky during disorientation periods after taking some Gs regardless of the avionics warnings.

its not really possible for drone pilots unless they're flying drunk.

its kinda easy to throw in random little things in news articles since nothing is checked. just look at any news item about something you know very well and you'll be upset how almost everything is wrong - why would news about anything else be accurate then?

This is why pilots absolutely need to be trained to very deliberately rely on instruments and ignore their "feelings" when they get into a tricky situation. There are tons of well-known sensory illusions in aviation: https://en.wikipedia.org/wiki/Sensory_Illusions_in_Aviation

> its not really possible for drone pilots

On the contrary, it sounds much EASIER for drone-pilots to make that mistake, because no matter how the drone is oriented they (the pilots) perceive themselves as right-side up.

Unlike direct pilots, you don't even need a heavy G-force situation, you simply roll 180 degrees.

I flew non-military drones for 10 years and I dont see how thats possible to lose orientation (unless you lost signal/video/etc.) honestly. If for any reason you were blind or stupid (I never, ever confused ground/sky on clear video), the avionics are extremely clear and do not depend on video quality.

"A $3.8 million Predator carrying a Hellfire missile cratered near Kandahar in January 2010 because the pilot did not realize she had been flying the aircraft upside-down." - how can this happen? Isn't the picture upside down on the screen?

Presumably at a high enough altitude or at night it would be hard to tell. Looking at that video is pretty low res.

The artificial horizon might not been working correctly as well.

Getting disoriented even on a real aircraft is not that hard. Sometimes even during constant turning the brain's accelerator resets and can interpret the turning as level flight.

At a "high enough altitude" it probably wouldn't have crashed ;-)

But it could have stalled or got into a spin didn't have enough time to recover.

A good point ... and wings are designed to provide lift while flying right-side-up. While upside down, you lower the elevators to promote level flight.

Apparently, USAF pilots aren't trained to watch out for mountains descending from above them.

The aircraft I flew in the '80s had an artificial horizon that showed whether you were straight and level. That aircraft wasn't capable of inverted flight (or maybe I wasn't) but the dial was vertically symmetrical and the only indication you were upside down would have been the (small) printed numbers showing you degrees. I saw the same dial in a flight simulator for an A-10 Warthog, so I'm assuming there were cues but no flashing red light.

Drone pilots obviously don't experience acceleration, so some of these shouldn't apply, but there are many well-understood sensory illusions that pilots are vulnerable to. Many of these will cause flight characteristics that will sound ridiculous to the layperson, like the graveyard spiral, where a pilot incorrectly perceives the craft to be flying level and thus reacts incorrectly to the altimeter's indication of a descent.

https://en.wikipedia.org/wiki/Sensory_Illusions_in_Aviation

https://en.wikipedia.org/wiki/Graveyard_spiral

Is there any kind of cheap and easy crash avoidance system for drones? If so then it should be mandatory in military drones and private drones that are flown over other people's private land or public/government land. If that kind of system doesn't exist then maybe someone should try to build it, "Open Crash Avoidance" or something similar.

http://gcn.com/articles/2013/07/12/drone-uav-sense-and-avoid...

The article focuses on military use of drones.

Crash avoidance systems are about saving the lives of the pilot more than saving the craft. Drones coming down in one piece will give the enemy more information than drones crashing so the cynic in me says that if you're willing to fly drones over other countries' territory protecting their population will take a back-seat to military objectives.

After all piloted craft come with ejection seats and parachutes for the pilots not with parachutes for the whole plane.

when i was a kid (90s), i saw these planes going between the hills/tree tops. they were too large to be a remote control plane, but too small to be a maned aircraft. when i was in the woods i saw the belly of one of them going over the tree tops 'us airforce', but didnt look like any plane that i knew of. then there was this one night on the news about an airplane crash everyone was calling in, but FAA and military was saying nothing crashed. however...

i go outside and there was about a dozen chinook helicopters coming my way. they were going over the ground in a grid pattern. one of them stopped, pointed lights, and landed. they were there for about 20-30 minutes and took off. I would have checked it out but it was night time and I didnt have a flash light.

this is the type of thing that makes people think UFO's exist. at the time i knew something was going on, that the air force lost something and now they found it, but I couldnt quite figure out what.

now that i know what i know now, i figured out they were testing drones in my area, and they crashed one of them at night. (specifically an early form of predator drones, propeller on back, some weird jet engine thingy on top, and a strange bubbled front is how i would have described it)i guess my area outside pittsburgh was ideal for testing, hills/trees, and they probably didnt think anyone was paying attention.

Reply on news.ycombinator.com