I'm currently building a quadcopter from parts and I'm writing my own control software.
I wasn't really familiar with RC components before starting this project, but I must say I'm impressed with the performance and reliability you get from very low cost motors, ESCs, propellers, gyroscopes and accelerometers.
However, I think Li-Po batteries are a huge safety issue. There are multiple factors:
* The battery chemistry is very unforgiving. You overcharge, it gets damaged, you over-discharge, it gets damaged. Same for charge and discharge current. The failure mode? Flammable gasses are released and eventually the whole thing sets itself on fire. And that's a fire you can't extinguish using typical means like water or regular fire extinguishers.
* All RC Li-Po batteries use a soft shell. If your model crashes the batteries can and will get crushed, which leads to an internal short-circuit, which eventually starts a fire.
* No RC Li-Po batteries incorporate Smart Battery controllers (like in laptop batteries) which could prevent overcharge/over-discharge and over-current conditions.
There have been plenty of incidents involving Li-Po fires, including whole houses burning down.
And still, the batteries don't change. Li-Po makes perfect sense for RC applications because it has great energy density and allows high discharge current. But to me it looks like some simple safety features could make this technology safe enough that you wouldn't have to treat your batteries like small bombs.
As a side-note, I've initially tried to use laptop Li-Ion batteries for my quadcopter. Each motor + propeller unit requires almost 12A @ 11V when at maximum speed - quite typical power requirements for a medium-sized quadcopter. I couldn't get more than 16A out of an older 6 cell battery (voltage was abruptly dropping if trying to draw more, so I guess this was the limit of the cells). I've then tried a new 9 cell battery (which was too heavy anyway). The controller on this battery considers 20A discharge an over-current condition and shuts off the output. So I'm stuck with Li-Po batteries and hoping that they won't set anything on fire.
I certainly wouldn't feel comfortable with current-technology Li-Po powered vehicles flying around a city without being watched.
I've built a few high performance rc cars and lipo batteries terrify me for all of the reasons you said. It seems like they are always just a hair away from bursting into uncontrollable flames. I quit using them because I was afraid to leave them at home unattended. I wouldn't want a swarm of these flying over my city.
No technical reason really, it's about cost, demand and volume.
Cost is going to be a problem because of the high current involved in normal operation and even more so in case of a short-circuit.
Let's take a popular battery size, let's say 2200mAh 3S 25C. This battery should be able to safely support a discharge current up to 55A. For reference, a 10KW power shower draws less current than this from the mains network. Our smart battery controller needs to be able to measure the discharge current, and needs to be able to switch the output on or off.
You can just buy a hall effect based current sensor and a power transistor that will handle 55A. But these components need to survive in case of a short-circuit for long enough for a micro-controller to read the current at its regular polling rate and then turn off the transistor. Expect the short-circuit current to be a number of times higher than 55A.
So you end up with a parts list that costs you about $10 per unit for small quantities (500). Have I mentioned the price of the battery I was building this for? It's $13.
I've thought about building some for myself, but 1) they really should be mounted inside a hard case together with the battery and 2) I don't have a space where it would be safe to prototype this, assuming that a failure will happen.
Okay HN developers -- when you get done reading the linked article and are finished laughing at the thought of a little helicopter delivering a taco or a bottle of beer, start thinking:
* Little helicopters can now lift a substantial weight.
* They aren't very expensive.
* They're easily controlled, more so than a full-sized helicopter (primarily because of computer-aided controls and GPS guidance). So you don't have to be Chuck Yeager to fly one.
* All you need to do is mate the helicopter with a decent camera that can simultaneously beam a picture to the ground for guidance and preview, and take high-resolution pictures on command by way of the radio link.
* Uses: real estate (who desperately need a way to take high-quality pictures of houses from above), surveillance, art, video productions, etc..
This is an opportunity waiting for someone willing to take it on.
Could you elaborate a little bit about what kind of stuff he's run into? I've been considering doing some arial chase scene footage of BMX events but wondering about the legalities of it? If you want, email works as well (In profile).
We already have plenty of photos of houses from above. Surveilance is still hard because these things are noisy and don't have a long linger time. The tacocopter is a real use case; I'd pay for it, particularly if it can deliver to my second-floor balcony rather than having to go down to the door.
My big fear: when the payload is hand grenades [1], i.e. a "multicopter bomber".
You can launch the multicopters from places where you're not seen. You can even do this in nighttime and have them land and wait undetected on some rooftop until you want them to continue their mission.
A handfull of such multicopters will make more terror than 9/11 ever could do because 9/11 was (almost) concentrated to a single place. 9/11 had passenger lists so it was easy to find out who took control of the planes. Multicopter bombers have no passenger lists, no sender address. You don't know how many more planes are in the air and when they will strike.
Redundancy and multi-target missions are even cheaper and more efficient with multicopters than in the 9/11 attacks. Multicopters are so cheap that the terrorists can launch a bunch of them and not care if half of them fail to get to their target. Multicopters bombers works for targeting e.g. Disney Parks and other places with a lot of easily frightened people out in the open.
With 9/11 it was easy to get away from the attacks (i.e. just leave that part of Manhatten). Multicopter bombers are like a swarm of supersized wasps. You don't know where to go, you don't know when they'll be back.
When you say "my big fear", I can't help but read "m irrational fear".
Do you _really_ fear "multicopter bombers" with a fleet "like a swarm of supersized wasps"?
To me, that's in the same class as people wo are afraid of flying, yet happily drive their car to the airport.
Realistically, terrorists don't skill up with brand new technologies to try carrying hand grenades around, they use cars/trucks/airplanes. If _I_ had a budget set to get explosives in amongst people, I'd be buying or renting cars/pickups/small trucks which can carry thousands of pounds of explosives around a city relatively inconspicuously. I wouldn't be training up teams of rc-pilots/hobby-UVA-geeks to be able to deliver a pound or three of explosive at a time.
a) buying a few hand grenades and premade multicopters? Hobbyists in Germany have already multicopters which are GPS pre-programmed so you don't need any pilots to control them during the mission.
b) creating thousands of pounds of explosives? Try read the detailed diary from the 2011 Norwegian terrorist to see how much time and trouble it took produce his bomb: way more time and effort than most would think.
The next 10 years will make multicopters much more accessible to non-geeks, but "thousands of pounds of explosives" won't be any easier than today. New tech gets old very quickly.
Why do you need multicopters? Terrorists already can (and some have) plant small timed bombs in crowded areas. I don't really see much difference other than that they can target more secure areas. This would be useful if they wanted to attack military targets, but since their usual goal is civilian casualties, they can by definition already get to places civilians congregate.
I have absolutely no idea where where I can purchase a live grenade. However, living in a rural area, I'd have no trouble finding people who know how to make explosives from common (and cheap) farm materials. Hell, I was making blackpowder when I was 11 years old.
Now add to that the fact that in the US I can buy a used car or pickup truck cheaper than I can buy a cheap GPS-equipped model helicopter and I definitely know which option I'd be more worried about.
Hobbyists in Germany have already multicopters which are GPS pre-programmed so you don't need any pilots to control them during the mission.
Hobbyists everywhere can have this - Arducopter (DIYDrones.com) and other projects have had this for years. Arduino + compass + GPS + accelerometer + altimeter + optional XBee radio (~2km+ range on some models) + autopilot software... preprogrammed routes, live telemetry and course changes via Xbee and ground control software.
This has been happening since 2000, over Afghanistan, Pakistan, Iraq, Yemen, Libya, and Somalia.
With guided rocket propelled "grenades" and the terrorists are the CIA and pals in USA government.
You should put your talent to good use, I bet you have it in you to write a first rate thriller.
(I'm not being viscicious - plenty of people have imagination, plenty of other people can rationalise about what could work in practice, drawing these together is a skill).
This is a bit too pessimistic, even for me. Technology is only going to get better, and the quadcopters will become quieter, and with less vibration, with longer battery lives.
As quadcopter technology and things like computer vision become more advanced, it really does seem like there's a lot of possibilities for these types of devices.
I've been daydreaming of the day when I have my own personal swarm of quadcopters that continuously circle my house with a wireless camera, and then land to recharge, and another one takes off. Not that I'm particularly worried about security (my barred windows, landmines and laser booby traps are more than enough protection), but it would be fun to see that.
The other thing that would be fun with quadcopters is if you could give it your gear, like your cellphone, wallet, etc, and have it fly above you everywhere, and whenever you needed something, you could just summon it. It's like the old school Dungeons and Dragons concept of having a familiar that you can summon whenever you needed it. You could also video yourself, from a WoW-like view, just for fun.
Except they're insanely loud when they get in close to you.
Which also is a (potential) feature. Consider a burglar alarm at your home, when it goes off you can just launch the quadcopter from wherever you are to investigate (and at the same time make the burglar aware of your presence). A lot of deterrents go out of their way to be visible. It gives you the opportunity to cancel the alarm or call in the police quickly.
And as for video productions and art, there quite a few products targeting this already, where you can preprogram movements etc. that just wasn't possible before.
> Aren't there already unmanned blimps built to solve exactly this problem?
This solution might be less expensive, more flexible, and easier to control on a windy day. Not that a model helicopter would be easy to control in a high wind, but compared to a blimp -- no contest.
Also I think the setup and takedown time for a model helicopter would be much less than for a blimp.
> Except they're insanely loud when they get in close to you.
Not an electric one -- they're surprisingly quiet. I was thinking of a battery-operated helicopter, to make the system easier to use as well as quieter.
It's an opportunity for sure, but lots of FAA red tape if you start doing this on a larger scale for commercial use. Many existing projects and products fly under rules for hobbyists, which prevent commercial use. Taking pictures commercially for real estate probably would fly "under the radar", delivering tacos might be harder to keep quiet. Probably could be tackled, but a definite question mark.
IIRC, Rotomotion was the first outfit to Open Source their autopilot, and that was almost 10 years ago. I think they've since gone closed source, but their initial market was powerline and pipeline monitoring from semi-autonomous helicopters carrying cameras. Whether related to their OSS efforts or not, the DIY UAV scene exploded soon after.
Fun stuff. How about a RC plane dropping off quadcopters loaded with the payload for the last mile? (the copters can meet the plane on the way back or fly home alone. You get the range + efficiency of the plane with the agility of the copter.
I'd really like to see a "you can fly anything commercially that weighs less than 5 pounds and has a terminal velocity below 30mph in freefall" rule. Bigger loads and speeds would be subject to some regulation and licensing.
This would cause a Cambrian explosion in development. It would be wild.
Imagine one of these payloads landing on someone's head. Wolfram Alpha offers up "2.4x impulse force acting on a baseball being hit by a bat" as a comparison when you plug in "5 lbs * 30 mph".
I like the idea, but there probably needs to be more development in safety systems before anyone could reasonably consider allowing that.
Umm, the point was more about the rule, not the specific numbers. Most of the little quad's like the parrot(1) can free-fall unpowered right onto your head and cause little injury, but you are absolutely not allowed to use the pictures your parrot sends back for any commercial purpose whatsoever. Even as a hobbyist, you are not allowed to fly your parrot outside of direct line of sight, even though its technology clearly begs to be used this way.
Also bear in mind that a 5 pound sponge that hits you at 30 might be a bit more pleasant than a 5 pound chunk of concrete. Perhaps I should have worked out a formula for average density per unit of speed modulated by compressibility,
but again, that's not the point. First, acknowledge we need common sense rules that allow a great deal of freedom at the bottom end. Worry about the specifics later.
The worst thing that can happen is that all unapproved drone flights remain off limits "for our safety" and exceptions are made only for those rich and patient enough to clear a bunch of bureaucratic hurdles (which will become higher each year as bigcorps dominate the space and lobby to "increase our safety" and keep the upstarts out).
(1) I am well aware that a parrot weighs less than 5 lbs.
> Imagine one of these payloads landing on someone's head. Wolfram Alpha offers up "2.4x impulse force acting on a baseball being hit by a bat" as a comparison when you plug in "5 lbs * 30 mph".
I'm not sure what value this comparison has. Impulse force doesn't tell you much of anything about how much damage would be done to a persons head.
You are working with the wrong concept, m * v is momentum, not impulse. Impulse would be equivalent to the change in momentum over time [1]. Actually 5 lbs falling at 30 mph would probably crush a human skull or at least crack it open. It is easy to see why from the equations:
F * delta t = m * delta V, therefore
F = m * delta V / delta t
This means that the force generated in the impact is inversely proportional to the time of the impact. Since the skull is pretty hard, we can say that the impact time will be very short, probably in the order of 1e-2 seconds. Such a sudden change in velocity, from 30mph to 0mph in 1e-2 seconds, generates a big force, and therefore a big impulse.
Indeed, however to say anything about damage you need to know the area where the force is applied and more about the materials. If the bottom of the copter is padded the duration of the impact coud be much longer and spread over a larger area.
Also, F = m * delta V / delta t is only valid if force is constant in this context. I would expect acceleration to vary during the course of the impact so this relation would likely not hold [EDIT] in terms of determining damage I mean.
Yes, in this case the equation refers to the average force over an interval of time: if at time instant A we had a speed v1 and at a later instant B we have some speed v2 then a force with a magnitude of at least m * (v2 - v1) / (tB - tA) must have acted, therefore the analysis should hold.
As you point out the area is very important in the analysis, since Pressure = F / A. For large areas the force "spreads out" and the damage is less.
The fatality of the accident can vary due to many circumstances however I think we can safely assume that for the purposes of our initial analysis, 5 lbs traveling at 30mph and crashing into my head is not something I am looking forward to ;)
Is this for real or is it the most cleverly disguised viral advertising campaign for Doritos? I don't see any mention of any other taco brand plus the use of a Doritos branded image...
I really like the idea of multirotors taking off, literally, to be part of a delivery team.
Incidentally, I'd really like to be able to use those rotors to recycle things - literally to shred, perhaps, my plastic trash - such that the remaining processed materials can be used in .. say .. a 3d printer.
Also, lets get the 3d printer and quadrocopters pretty much working smoothly together so that one provides resources to the other. Oh, no wait, lets just make the 3d printer print quadrocopters, and the quadro's feed the 3dprinters .. and .. well now lets just make a flying recycling 3d printer robot, and be done with it.
On Mars. Because if we do it here, the damn thing will take over our planet.
That could be useful for delivering mail in skyscrapers. But The biggest challenge to solve would be finding the way to prevent interceptions/stealing.
If the cost of copter + payload is low enough, there is little incentive to steal. It should be similar to shopping cart theft or most probably less. The system can be built to go up, travel horizontally above the reach of most people, and then drop down at destination.
You could buy a perch like they show here: http://www.youtube.com/watch?v=E7X0_6o9J10&feature=playe... - the perch has a GUID, knows its GPS location, and carries insurance. When you order from any vendor, you just give them your perch ID along with your credit card. If the order is lost before it reaches your perch, it's the vendor's problem. Once the copter touches down on your perch (and informs the vendor of the safe delivery), you press 'Go Home' to launch it back. If it doesn't reach the vendor, your perch insurance takes over. Too many lost copters in your neighborhood and vendors will be able to deny service based on risk.
I think if the copters are no louder than a car pulling up the driveway, this can work very well.
Things that fly are really, really dangerous, and the benefits really need to be worth the risks. I don't think a Taco is worth the risk that one of these things goes out of control over the freeway and causes a 100-car pileup. The idea of cheap and plentiful multi copters would make this sort of event inevitable.
Of course there may be a much simpler problem: hungry seagulls.
The article suggests a "modest" speed of 40mph for the quadrocopter, which seems pretty fast to me. Let's call it 30mph. Our car on the freeway is doing say 60mph. Presuming they're travelling in opposite directions, that's 5lbs travelling at 100mph.
So what you're saying is that if I fire a MacBook Pro at car at 100mph then nothing will happen. Unlikely...
And one of these smacking a pedestrian in the head at 40mph isn't going to be much better.
Depending on the design of the frame. Some are made of plastic and break apart really easy on impact. None of them, even the the aluminum frames, are even close to as dense as a MacBook.
When the drones come to US cities, and it seems they will within a decade, I'll be moving to a smaller town, however small or remote it takes to escape the noise. I'll accept that I'm an old fogey who can't handle "progress".
How about ornithopters? Naively I'd guess they'd come in between planes and quadcopters in both efficiency and agility. Since seagulls can land without a runway, a tacothopter ought to be able to, too.
The Osprey is incredibly complicated to design and operate. The only reason the Marine Corps even bothers with it anymore is because it makes the impossible possible in terms of long range vertical insertion. The cost-benefit just isn't there for tacos.
I'm currently building a quadcopter from parts and I'm writing my own control software.
I wasn't really familiar with RC components before starting this project, but I must say I'm impressed with the performance and reliability you get from very low cost motors, ESCs, propellers, gyroscopes and accelerometers.
However, I think Li-Po batteries are a huge safety issue. There are multiple factors:
* The battery chemistry is very unforgiving. You overcharge, it gets damaged, you over-discharge, it gets damaged. Same for charge and discharge current. The failure mode? Flammable gasses are released and eventually the whole thing sets itself on fire. And that's a fire you can't extinguish using typical means like water or regular fire extinguishers.
* All RC Li-Po batteries use a soft shell. If your model crashes the batteries can and will get crushed, which leads to an internal short-circuit, which eventually starts a fire.
* No RC Li-Po batteries incorporate Smart Battery controllers (like in laptop batteries) which could prevent overcharge/over-discharge and over-current conditions.
There have been plenty of incidents involving Li-Po fires, including whole houses burning down.
And still, the batteries don't change. Li-Po makes perfect sense for RC applications because it has great energy density and allows high discharge current. But to me it looks like some simple safety features could make this technology safe enough that you wouldn't have to treat your batteries like small bombs.
As a side-note, I've initially tried to use laptop Li-Ion batteries for my quadcopter. Each motor + propeller unit requires almost 12A @ 11V when at maximum speed - quite typical power requirements for a medium-sized quadcopter. I couldn't get more than 16A out of an older 6 cell battery (voltage was abruptly dropping if trying to draw more, so I guess this was the limit of the cells). I've then tried a new 9 cell battery (which was too heavy anyway). The controller on this battery considers 20A discharge an over-current condition and shuts off the output. So I'm stuck with Li-Po batteries and hoping that they won't set anything on fire.
I certainly wouldn't feel comfortable with current-technology Li-Po powered vehicles flying around a city without being watched.