Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

One motor suddenly failed

Maybe drones need to start to incorporate parachutes to slow down the fall in case of problems like these, or maybe even airbags to ease the crash. Not that it would make it 99.9999% reliable, but it should be better than nothing. And also won't completely thrash your drone when it crashes.

edit seems nine_k already commented with a similar idea



Quadcopters can also be saved if only one motor fails using only the 3 other motors and an algo: http://diydrones.com/profiles/blogs/new-algorithm-can-save-a...


Recent research has had controlled landing on 2 motors in real robots (and one motor in simulation only). The main trick is to spin the whole vehicle around the z axis (up-down) very fast, and modulate the speed of the remaining rotors to apply forces at different places around the centre of gravity. IIRC the one-rotor version didn't work on a real robot due to limits on the frequency response of the real-world motor controller.


Spinning the drone so fast that the modulation is usable doesn't itself sound very safe to me.

It's an upgrade from "expensive rock with sharp blades falling out of the sky" to "expensive rock with sharp blades falling out of the sky spinning at X0 RPM, and the sharp blades are also spinning at X00 RPM"


Sounds plausible but I haven't read about this before and would like to know more, do you have a reference handy?


In regards to turning on the z-axis to be able to land with engine failure, I believe User is referring to Autorotation: http://en.wikipedia.org/wiki/Autorotation


No, it's not autorotation. Here's a lovely demo:

https://www.youtube.com/watch?v=ek0FrCaogcs&feature=youtu.be


I thought autorotation was only useful if you're moving ahead very fast (high ground speed) while also falling. Is that right? Most drones are essentially hovering much of the time.


Patent-pending.


Yes!! Much better than a parachute, that needs altitude to work.


Drones of this size are cheap enough that the extra weight and cost of engineering of safety devices generally isn't worth it. It's easier to fix or replace parts.


Are you factoring in your personal liability for killing someone in that equation?

The comment that mentioned parachutes was talking about how a drone crashed at an event and almost injured people. My personal homeowners liability policy covers things like hitting someone with a shopping cart in the grocery store and hurting them, it probably would also cover a drone, but if you don't have insurance it could expensive real quick.


Notice I said "drones of this size". While the propellers could still cause some damage, these kind of things aren't going to kill anybody, so let's not be overdramatic. As for potential dangers to other people, the solution to that is not a technical one, it is simply to not fly over or near to people. Just like we don't drive our cars on the sidewalk.


The "follow me" feature of newer drones seems to encourage people to fly near themselves and presumably others.


Are you willing to take a propeller to the face?

The cool thing about sidewalks is that they are elevated and it takes some effort to hop on to one and run over pedestrians (especially when there is a protective lane of parked cars). Drones, on the other hand, are poised to fall on people, with no barriers.


I've taken a propeller to the face from a motor similar in strength to the consumer "drones", and it was spinning full throttle. I really don't recommend this. I got lucky it didn't hit my eye (came close). End result was some 4 scratches across the cheek on the left side but no permanent scarring.

In my case it didn't occur when I was flying the quadcopter but rather when I was bench testing the ESCs (yes, I knew abstractly I should remove the props before doing this but I got lazy after a while of testing and eventually it bit me).

Having said all of this I mostly agree with the grandparent post (though I do recognize that a 2 lbs drone could, theoretically, kill someone) in that the solution to this as a safety issue is primarily just don't fly a quadcopter over anything you wouldn't feel comfortable crashing it into (basically don't fly over people or valuable property it could potentially damage). As long as you stick by this rule of not flying over anything you're not afraid to crash it into you already have a good (if not inexpensive) failsafe, which is just cut the throttle completely.


Of course not, where did you get that idea? I don't think drones should ever be operated in close vicinity to people, especially if those people are bystanders uninvolved in the operation of said drone. Well yes, but the slight elevation of the sidewalk isn't the sole reason drivers don't run over pedestrians. The barriers are hopefully in the discipline of the pilot.


You don't think a device that weighs over 2lbs falling from 40' could kill someone?

And that seems like it's being pretty generous. In most "quadcopter crash" videos I've seen on YouTube the drone was much higher. Sometimes even hundreds of meters.


Cars kill 40k people per year in the US. Yes a falling drone "could kill someone" however falling coconuts actually kill a lot of people each year. [1] If you want 100% safety in life, build a bunker and never go outside.

[1] http://diaryofnumbers.blogspot.com/2010/11/death-by-coconut....


Even if it's possible, it's incredibly unlikely compared to an automobile accident.

If there were as many drones (of that size & rotor size) overhead as there are vehicles on the road, I'd still expect to have zero to single-digit fatalities from drones every single year--at least from "falling out of the sky onto people's bodies."

Perhaps there are more likely cases where the drone is the indirect cause, like breaking someone's windshield and causing a fatal car accident.


We live in a society with motor vehicles. They are here to stay. It doesn't matter what their safety record may be or under which parameters they are operated. Drones are an entirely separate issue. We do not live in a society with drones. Period. Now if we would like to introduce them into our society it doesn't matter what laws or regulations are established for their safe operation without on-board fail safes their introduction is unlikely. How do I know? I'll just pose this question: Are you OK with any drone dropping out of the sky and landing on top of your new born infants soft little squishy skull? Find me a person that will say yes to that question.


Nobody is okay with that, it's an asinine question.

Am I okay with my new born infant's soft little squishy skull living in a world with drones, which occasionally drop out of the sky?

Yep.

And we do live in that world, actually. They've already been "introduced," as you put it.


Not even apples and oranges. I know that they can still fly safely with only two motors if they are opposing. They may be so safe that it won't be for years to come but eventually after enough mad mothers are up in arms about somebody losing an eye or it getting sucked into an airliner jet that they're going to start heaping on the restrictions. One of the concerns with using them for deliveries at least in a city like Los Angeles is that people will try and shoot them down.

Already been introduced(bad term)? So 1 in 3 households have a drone in the U.S.? OK 1 in 4? You can always count on this statistic- 1 in 10? Nah? Surely 5% of Americans own or use the services of a drone then right? Yeah then they haven't been introduced yet


I believe a jet engine can handle this with no issues. They're tested to pass birds through with no issues.


Really? I didn't know that. That flight out of New York where "Captain Sully"? made the first successful water ditch of a commercial airliner in the history of aviation had a flame out from geese, but that was an entire flock.

They're so small I doubt they'd have much of an impact on light aircraft either.


http://en.wikipedia.org/wiki/Foreign_object_damage#Jet_engin...

For the large bird test, the engine is allowed to shutdown.


It's very unlikely to kill someone, or even seriously injure them. You should be at least as worried about playing baseball or riding a bicycle around people.


The battery packs on those things have quite a bit of heft to them, hit anything at terminal velocity and it's gonna be pretty devastating.


Terminal velocity for these things will not be terribly high.


Probably on the order of 120 MPH or 55.8 meters per second. The drone weighs 1.2 Kilograms. The kinetic energy on impact would be 1800 Joules give or take. A 357 Magnum generates about 873 Joules.

The energy we are talking about is not insignificant. I'm actually in favor of allowing people to fly drones with as little restriction as possible but let's not pretend there aren't some risks.


How do you figure 120mph? It's ~1kg and I would guess has a higher surface area relative to mass than a person whose terminal velocity is about 120mph.

In any event, it takes a person 8s to get close to terminal velocity — that's a drop of 300m or so.

The energy involved is certainly non trivial but energy is not the whole story. The densest, heaviest pieces will be the battery and motors.


So, they go 25mph normally, i wonder if the terminal velocity is more or less than that.

Too lazy to do the calculations :P


Way more. Probably on the order of around 120 mph


depends where you fly. A 2-5+ kg drone, even if the props aren't spinning is gonna do serious damage if it drops from the sky and hits someone on the head. With spinning CF props you can easily slice someone up pretty bad


Not if the safety device is reusable.


But a drone with a parachute falling into a pack of racing bicyclists sounds like a really bad idea as well, doesn't it? The larger issue seems to be the possibility of drones crashing into people who are involved in an activity that requires 100% concentration.


You could make the same argument against driverless cars.


One parachute-for-drone is already being Kickstarted (at least).

But the issue is the effective minimum height for a parachute to be any use.


Perhaps an airbag underneath could deploy faster.


Or a self-destruct to blow the whole thing up before it hits the ground.


Yes, lets add explosives, that will make it safer.


Timing matters here, as the drone could easily land on its heavier side if all rotors fail.


http://www.digitaltrends.com/cool-tech/djis-dropsafe-parachu...

In order to make flying aerial drones safer, DJI, maker of popular quadcopters like the Phantom 2, is currently developing a parachute system called DropSafe, which can be deployed instantly in a case of emergeny


The problem is these drones often have prop guards that would make them a lot safer to fly, but you almost never see the prop guards on a drone do you? Any other safety gear would be stripped by operators just like that. It hurts their flight time and performance after all.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: