They're quadcopters rather than single mainrotor + tailrotor because it greatly simplifies the mechanics.
Single rotors need a cyclic mechanism to vary the blade pitch depending on where the blade is in the rotation and the control inputs. This is VERY complicated with a lot of moving parts that have to maintained and adjusted. (See https://upload.wikimedia.org/wikipedia/commons/thumb/4/4f/Si... for one example)
Quads (and hexes, etc) can have full control authority via RPM only, using a simple injection molded constant pitch prop.
Basically with a quad if you lose any of the 4 engines you're going down.
This is much less of an issue with an unmanned aircraft.
I say by and large because there are exceptions, but they are complex, expensive, and compromised.
As for complex and expensive the first google result for "rc plane vtol" for me is this one [1] and it costs around $50. And then there are things like [2]...
[1] https://www.youtube.com/watch?v=xFISwEQbaz0 [2] https://www.youtube.com/watch?v=MDxHamYEQNU
On a quadcopter you have a horizontal plane on which you can mount cameras or other sensors. Suppose you want to do some mapping, which is a pretty common scenario for these kind of industrial drones. With a quad, you can easily mount a camera which is mostly pointing in the same direction, regardless of air speed. With a gimbal you can direct it pretty freely.
With a VTOL aircraft like your [1], the ground angle would shift 90+° degrees depending on airspeed (nose-up hover vs nose-forward flight). It would require a pretty aggressive gimbal to just keep the camera pointed towards the ground, and some angles would be impossible to get in hover mode.
But for some industries like AG/large scale mapping, range is so important and you are basically always just interested in looking straight down and fly in simple pattern, that fixed vings are a thing: https://wingtra.com
It is implied. Costumers shopping for $20,000+ industrial drones already know the tradeoffs between fixed wing and quadcopters.
There is no shortage of fixed wing drones and commercial operators are not idiots. They buy what they need for their application.
> I think it's mostly just a matter of power-to-weigh ratio, which small aircraft have in spades.
Yeah! These small aircraft parts actually provide a respectable amount of thrust. The low weight of drones and model aircraft usually means they have massive thrust-to-weight ratio. T/W > 1 means the engine alone is able to propel it straight up. The larger the ratio, the more it acts like a rocket.
Sort of like fly 20 minutes or take off vertically for 30 seconds.
For efficiency, I think the best bet might be to take off from the top of a downhill slope gaining momentum and flying. To land, return facing up the slope killing momentum.
sort of like this: https://youtu.be/vscvM8ysAwM
This means an outstanding resistance to wind, which fix-wings do not have.
maybe one big part involves the landing & takeoff:
for plane-like drones both are a lot more dangerous (therefore more difficult to be automated) and need a lot more space than quadcopters (and the likes). Saying this because I remember that, when I was a kid, my RC-plane (which took weeks and $$ to build) crashed 1.8 seconds after the first take-off attempt (damn, I got so angry & depressed...).
I guess that soon or later we'll see mixed designs that can take off & land like a quadcopter but which can increase their range/efficiency by "transforming" in mid-air into an airplane-like form? (something similar to https://en.wikipedia.org/wiki/Bell_Boeing_V-22_Osprey or maybe just a passive design that works by changing the fixed orientation of the whole drone from horizontal to almost vertical, or the opposite depending on the form of the drone?) In any case complexity would increase, therefore more specific demand for long-range capability is probably needed to go towards developing something in this area?
This made me wonder if a drone gyrocopter might be a useful compromise, since they're inherently very stable. They do seem to exist but are rare[0][1].
According to the few discussions I could find on the topic, gyros are also less efficient than helicopters, so I guess that doesn't really make it a useful compromise in the context of drone range[2].
[0] https://www.youtube.com/watch?v=TPbQlRm_lvI
[1] https://www.rc-airplane-world.com/rc-autogyros.html
[2] https://www.rotaryforum.com/threads/gyro-vs-helicopter-cruis...
I crashed my first Phantom drone as well since I didn't know you have to warm up the batteries in cold weather. Live and learn I guess :)
Most commercial versions opt for simplicity and use a separate set of propellers for quad and forward flight configurations.
Quadcopters are just much safer and easier to control. Compare "fall down at random without input" vs "stop without input".
As for why quadcopters aren't used at big scale - moment of inertia scales with cube of the propeller radius. And to control a quadcopter you're constantly speeding up and down each propeller. That's a huge loss of energy when your propellers are big enough.
So instead at big scales we use very complicated mechanically designs like helicopter. It uses a propeller that rotates at constant speed (so moment of inertia doesn't matter), and instead changes the angle of attack to control the aircraft.
BTW moment of inertia is also why nobody in their right mind would build huge mech robots like in sci-fi. Wheeled vehicles move forward without fighting inertia - wheels spin at constant speed and you just add the energy lost to friction. Mechs constantly swing their huge legs back and forth fighting the whole inertia twice with each step. Fine for toys, not fine for big vehicles.
Computer-controlling a fixed-wing plane requires a lot of sensors, and a lot of math, and a lot of space, and forward airspeed.
Computer-controlling a quadcopter requires a lot of sensors, and a lot of math, and a lot of power.
Arduplane exists alongside arducopter, and it's easy enough to build both, but I can't fly a fixed-wing model in my backyard. I can fly a quadcopter just fine.
https://en.wikipedia.org/wiki/V_speeds
Rotary wing aircraft use their engines to constantly move the wings through the air instead of moving the aircraft itself. So the wings are always generating lift and the aircraft is free to move in much more flexible patterns. A very useful maneuver is hovering in place: the aircraft is able to simply hold its position in the air. Since only the wings need to move to generate lift, the aircraft itself does not need to accelerate in order to take off and therefore a long runway is not necessary.
Civilian drones usually have rotary wings because this allows people to do useful things like having the drone stay in one place in order to film or photograph something. A fixed wing aircraft would have to establish a loitering pattern around an area in order to do the same thing. In other words, rotary wings can just hover in place while a fixed wing would have to fly around in circles.
Fixed wing aircraft are more fuel efficient. Rotary wing aircraft must constantly spend fuel in order to keep their wings spinning. They lose lift otherwise. A fixed wing aircraft would probably be able to glide great distances and even land safely even if it lost all engines in the middle of its flight.
Maybe it is possible to combine a "normal" quadrirotor (for vertical takeoff and hovering) with a rotating wing aircraft, the autogyro:
https://en.wikipedia.org/wiki/Autogyro
to save a lot of energy when traveling horizontally.
There is actually this thing here that is already a sort of cross-breed between a helicopter and an autogyro:
Since the rotary wings are unpowered, the autogyro loses many of the properties that make the helicopter so valuable. It requires forward motion in order to move air through the unpowered rotary wings and maintain lift. Looks like they have propellers providing forward thrust just like fixed wing aircraft. The article also states that they're not capable of vertical take off and landing. They need a runway to accelerate just like fixed wing aircraft.
>Helicopter mode for hovering and slow forward flight any time
>Gyrocopter mode for fixed wing fast flight and long distance
See also gyrocopters:
https://www.youtube.com/watch?v=a882megKhVM
Now, in scale for drones, it may also be possible to have an always unpowered rotor (rotating wing) AND 4 (or 5/6[1]) more conventional small rotors for vertical take off/landing and hovering.
[1] just thinking aloud, but it would probably be simpler to have four horizontal small rotors, like a "common" drone and an additional one or two vertical to provide trust forward like a traditional airplane or autogyro.
Several hours of flight time is trivial with a battery powered fixed wing model airplane.
The rotary-wing UAV is a power hungry gussler while fixed-wing UAV is not that versatile (e.g. for hovering, etc). The best compromise will be Gyrocopter or Autogyro since it is both energy efficient and flexible. It can also easily fly at sustain high speed wind of more than 50 knots.
From aerodynamic standpoint planes are much easier to pilot (all but military, same story with computers for better maneuverability) since they are designed to be inherently stable. So if all power is lost and you still have control of aerodynamic surfaces, there is not problem to land a plane after planning for some tens or even hundreds of kilometers (depends on initial altitude). This is obviously not the case with quadropters, loss of a single propeller leads to complete loss of control.
The aerodynamic stability is an interesting topic, but useless here. If you assume your computer is dead, you won't control neither a remote quadcopter nor a remote winged plane.
A nice example of a different approach is a variable angle of attack propellers which are driven by a single motor [1]. See after 0:52 and 7:13. In case battery is dead since all the propellers are synchronized they auto-rotate and stabilize the drone (near the earth the pilot activates the motor just to land softly, but likely similar effect can be reached by properly adjusting the angle of propellers in proper time, like in case of auto-rotation landing for helicopters).
[1] https://youtu.be/TnGhEInTXYc?t=52 (this video was posted below by @jcims)
(Note also that doing this outside of a mocap arena would probably require a pretty high-quality IMU since you need very accurate yaw and you're constantly integrating your gyro to get it, with only the magnetometer and maybe the camera providing any kind of absolute reference.)
Perhaps the air is so turbulent around the wings due to the propellers that wings just don't get any airflow to create lift, or would need to be much longer to be out of the way of the propellers, that it's not worth it?
The closest thing in the real world to the hybrid that you describe would probably be the tiltrotor, where the same engines which supply lift can turn 90 degrees to become forward thrusters on wings: https://en.wikipedia.org/wiki/Tiltrotor
Amazon have made at least two designs, one that flies “tilted” and uses the boundary ring as wings, another with more separate wings and forward propulsion
Version 1: https://techcrunch.com/2020/08/31/amazons-prime-air-drone-de...
Version 2: https://www.amazon.jobs/en/teams/prime-air
The cost (in terms of both price and structural integrity) of gratuitously dynamic wings (think turning the whole wing 45 degrees to horizontal, bending the wing to 90 degrees at it's half way point) are much smaller for small aircraft.
Such gratuitously dynamic wings give more degrees of freedom for automated piloting software to work with when it comes to optimizing near vertical take-off and landing when compared to a fixed wing aircraft, while retaining fixed wing efficiency advantages during elevated flight.
I think its a super interesting piece of kit so I hope you find success with it!
Because somebody, somewhere published the first opensource autopilot code, and that first code was for the quadcopter.
But before quadcopters we used large-wingspan EPO foam / balsa / plastic planes.
It's still common place.
Same here.
> Why then a quadcopter is the default shape for smaller drones?
Large fixed-wing planes are expensive. Piloting them requires jumping through many high-hanging and heavily regulated hoops. The most dangerous parts of the flight near the airports happen in well-known space, according to pre-defined paths known to all parties.
In many places of the airspace, pilots are guided by data from the ground. Optical from visual approach slope indicator (these weird light arrays around airports which are visible as 0-4 red circles depending on whether you’re too high or too low). Radio from beacons. Most importantly humans, the job title is “air traffic controller”.
For the scale of a conventional airplanes, accurate map data is available. GPS data+SRTM height map is accurate enough for most cases. Also, real-time weather data is pretty accurate at that scale.
None of the above applies to small drones. They’re cheap, most are piloted by amateurs and no licensing is required. They can take off and land whenever. No ground control is available. At their scale, no offline map data is available, and ground-induced wind turbulence is borderline unpredictable.
This might change in a few years.
We might get chips smart enough to reliably do right things based on the limited input data despite the unpredictable factors.
We might get sensors sensitive enough to generate much better data for these chips. E.g. if you have spatial data for 1x1x1km space around the drone, with 1mm spatial precision, and 20ms refresh rate consistent over the whole volume, today’s mobile chips will do mostly OK in good weather.
We might get good enough sensors to cover all airspace with them, and make robotic equivalent of air traffic controllers, a software that detects future collisions and tells the drones to do something about it. Or broadcast current wind conditions at 1m^3 resolution.
Least likely but still, we might get good enough actuators & power sources to combine these two. Birds do OK, many of them can hover, glide, and arbitrarily combine hovering with gliding. When you can stop flying and hover, or land almost everywhere, collision avoidance becomes way simpler. That’s one reason how birds and modern-day small drones are functional.
They have a stable and basically stationary anywhere in 3-space mode (until the juice runs out). Human and automated control is much simpler as a result.
Besides, the mechanics are simpler and therefore more reliable. And the propellers are smaller than a helicopter rotor, which is safer. Multirotors are also much simpler to operate than a helicopter. Most industrial customers do not want to operate unmanned helicopters because of their complexity.