Underactuated Rotor for Simple Micro Air Vehicles

http://modlabupenn.org/underactuated-rotor/

by command_tab • 12 years ago

253 81 12 years ago

This thing is so nimble in the air yet has a fraction of the complexity of a regular helicopter. I wonder if this method will scale up to "full size"?

FrankenPC• 12 years ago

That's what I was thinking. That would require electrical motor control of the blades. Maybe turbine generator for power source. I would imagine this kind of pulsing would cause severe fatigue on the parts if they scale up too large.

TaylorAlexander14• 12 years ago

The parts will cycle back and forth exactly as much as they would in a traditional helicopter, and there are fewer comparatively small parts. It would likely be stronger and more fatigue resistant than current methods of controlling blade pitch. We can program the motor to change torque as smoothly as we want.

sliken• 12 years ago

So nimble? Looks barely flyable. Compare the video to a quadrotor for instance.

marvin• 12 years ago

Electric motors remove a lot of the complexity in powered-lift flying machines. As soon as your power/endurance requirements demand a combustion engine, you also have to manage the complexity of a piston engine or turbine (jet) engine.

This means either including a crankshaft and optionally a system of gears to route the rotational power along the axes you want the rotation in, or ducting the output of a jet turbine in some way that drives the rotors. In addition, there are different limitations with regards to the time required for a given change in RPM as compared to an electrical motor.

My impression is that a lot of the complexity in conventional helicopters (and flying machines in general) stems from the limitations of combustion engines. There's a reason that you very rarely see vectored thrust in conventional airplanes, for example.

Of course, if batteries and electric motors become sufficiently powerful, this changes the dynamic and it will become possible to design electric aircraft that re-evaluate the traditional design restrictions. This is one of the reasons that the rapid progress of electric cars is so exciting.

sebastianavina• 12 years ago

or you can generate electricity with a combustion engine a move two electric motors... like the railroads

wiredfool• 12 years ago

Probably not, for the same reason that a quadcopter doesn't really scale to full size. The inertia effects of larger rotors make changing the speed of the rotors within a rotation much harder.

With full sized helis, it takes a long time to spin up the rotors to speed before the pitch is changed to take off.

TaylorAlexander14• 12 years ago

Well the critical difference here is that changing blade speed is not needed or desired. They just need to change blade torque, which would apparently immediately cause a change in blade pitch.

I don't see a specific reason why this wouldn't scale. I do wonder what changes in load do to the system though. If it can't handle changing loads without messing up the blade dynamics, it would only work for fixed payload systems like camera platforms.

lotsofmangos3• 12 years ago

Sure it would work. It only needs a variable clutch on a standard engine and the motion on the blades is very slight and mostly balanced, so the torque difference should be pretty manageable.

spacefight• 12 years ago

Not quad exactly but still worth following is this man lifting multicopter from Germany:

http://www.e-volo.com/

https://www.youtube.com/watch?v=CzP0Zqxam7E#t=76

ajuc• 12 years ago

Big rotor = big moment of inertia = more energy lost on accelerating it and slowing it down each time. At some point it's better to rotate it at constant velocity and change pitch independently.

TaylorAlexander14• 12 years ago

You're not worried about speeding up the blades, just changing the torque on the drive system. The inertia of the blades actually helps you here.

foobarian• 12 years ago

It would probably work but I'm guessing the overhead of the extra actuators is much smaller at full size, so this may not be as helpful as at small sizes.

prbuckley3• 12 years ago

Depending on what you mean by scaling up. There have been single man vertical take off and landing (VTOL) vehicles that have two rotors and require no variable pitch for directional control so they are even simpler then the system in this video since the 1950's! The first such vehicle was the Hiller VZ-1, dubbed the "Flying Platform". If you are in the bay area you can go check a real one out at the Hiller Aviation museum next to the san carlos airbort (30 minutes south of sf). Or enjoy watching this video...

http://www.smithsonianmag.com/videos/category/history/the-se...

Directional control is achieved by leaning, the gyroscopic forces from the dual rotating blades creates inherent stability without any feedback system. They are like segways of the sky. Reports say it only took 20 minutes for a non pilot to learn how to fly it. These machines are incrdible and I don't understand why more hasn't been done with them.

A late 1970's version called the williams x jet WASP added a cruise missle turbojet for more horsepower, and was coined the "flying pulpit".

https://www.youtube.com/watch?v=U-3Ql7G7qRc

prbuckley3• 12 years ago

As an aside I think a vehicle like this could be developed today and fall under the FAA ultralight powered vehicle classification (weighing less then 254 pounds and carrying less then 5 gallons of fuel). This classification doesn't require you to have a pilots license!

People here might also be interested in the Mosquito Air helicopter as well. It falls into this ultralight classification and is a kit you can build in 200-300 hours. It costs about $30,000...

http://www.innovator.mosquito.net.nz/mbbs2/mosquito.asp

usrusr• 12 years ago

The complexity of the conventional helicopter method is not only providing control, but also a considerable amount of safety by enabling autogyro capability. This lack of autogyro is already restricting the quadrocopter setup to "let it crash" dimensions and the same well be true for this admittedly very clever duocopter.

DickingAround• 12 years ago

I wonder if this will cause some parts of the motor to heat up more; since the blade is 1:1 rotations with the motor then driving more torque (power) during certain phases of the cycle will put more watts on certain coils.

It's a cool idea. Human size helicopters would love to avoid all that blade-pitch complexity. :)

theoh• 12 years ago

Couldn't that be addressed (if a large enough effect to be a problem) by just having the body of the vehicle spin at a low rate, like less that 1Hz. With appropriate sensors this wouldn't be a problem[1]. More to the point, there shouldn't be any need for sustained constant "cyclic" input anyway, in normal maneuvering.

[1] an axially-symmetric imaging system could compensate easily, e.g. a panoramic system.

alimoeeny• 12 years ago

The whole thing looks great and all. But I specially liked their last sentence that said "for civilian needs".

zan2434• 12 years ago

This is so clever! Turning mechanical control problems into informational control problems is critical to the ubiquity of micro air vehicles.

fasteddie31003• 12 years ago

There is a lot of potential in solving mechanical problems with information systems, rather than using complicated mechanical solutions. For instance, hydraulic automatic transmissions and differential.

lumpypua• 12 years ago

The mechanical solutions are always really interesting and clever, but digital systems are much more flexible.

Really cool video on the mechanical ignition timing control in a delorean to control emissions. Explanation of ignition advance at the end: https://www.youtube.com/watch?v=ge1GwepqtK0

swah3• 12 years ago

I have no idea what you're talking about.

Cushman• 12 years ago

"The linkage to change the pitch of this rotating blade is way too complex! Can we simplify it somehow?"

"How about we just add a simple device that associates the pitch of the blade with the torque, and let a computer figure out how to spin the motor to get the pitch we want? No linkage!"

Yeah, that is dang clever.

jahnu• 12 years ago

How do they determine blade position?

phkahler• 12 years ago

> How do they determine blade position?

I assume you mean the motor rotation angle. It's probably not a brush motor, but a 3-phase synchronous motor. That means instead of having a commutator to apply the voltage to the coils, a computer switches some FETs to apply the voltage. In the simplest case, the rotor is assumed to follow the applied voltage waveforms. In a more typical system, the rotor position is estimated from the applied phase voltages and measured currents (usually via PLL or a sliding mode observer).

If you mean "how do they determine the blade pitch?" then the answer is that they don't need to. The controller - weather that's a computer with inertial sensors, or a person watching it - will just manipulate the amount of torque variation until it gets the response that it wants from the copter. Much like you don't actually need feedback of your cars gas pedal position so long as you have vehicle speed feedback.

joosters• 12 years ago

Very clever! I hadn't fully understood the mechanical complexities involved in a standard single-rotor helicopter. I just thought that they were hard to fly manually (which made me very confused to see all the computer-controlled quadrocopters - why didn't they use just one rotor?)

TaylorAlexander14• 12 years ago

Well the issue isn't just that they're hard to fly or that they're mechanically complex - the issue is that they're both. So you crash often (when flying models) and crashes are expensive. This makes it a very expensive hobby.

The issue when flying is that they balance as if they are on a ball. You need to constantly adjust the controls to stop from "falling off the ball". Go the wrong direction and you speed up the rate at which it falls. When it is facing away from you, that's not very hard. When it is facing towards you everything is backwards. When you are turning, the correct direction to stay balanced is constantly changing.

Automated systems can do this automatically for you though, so a drone-like autopilot in a standard helicopter could make them as easy to fly as quadcopters. Traditional helicopters are still mechanically complex though. This system looks to fix that. And this would be more efficient than a multirotor.

aosmith• 12 years ago

Now if we could just start printing these parts at home...

thisjepisje• 12 years ago

What's stopping you?

TaylorAlexander14• 12 years ago

I did printed helicopter blades three years ago: https://www.youtube.com/watch?v=qXlUSWrVzys

These other components look printable too. Right now I am making 3D printable robots that don't need any non-printed parts aside from motors, bearings, and drive belts (and batteries and electronics).

So these definitely should be printable at home.

lucaspiller• 12 years ago

That sounds cool, do you have any links to your projects?

brk• 12 years ago

Seriously, why? Injection molding has far lower cost, higher quality, and you're talking about things that are very cheap and easy to ship. What does 3D printing bring to this?

lotsofmangos3• 12 years ago

1m square deskspace requirements and fast tooling in software, rather than thousands of pounds and new metal every time you want to change a detail.

Injection moulding is only lower cost when you want to buy something that is already mass manufactured or millions of identical things. Up to about 10,000 units, 3d printing is cheaper.

TaylorAlexander14• 12 years ago

Injection molding is on the order of 10,000x more expensive if you only need a single part.

I'm currently on a quest to design highly functional robots that can be made with just a 3D printer and a minimum of external parts - so far only bearings, motors, drive belts, batteries and electronics are the non-printed parts needed. I make everything so that it fits together by interlocking or with minimal use of some coarse printed fasteners.

Since nothing needs to be bought from the store that won't be useful if the design changes, it is trivial to iterate and build a newer version.

My hope is that with developers worldwide working on improvements, injection molded versions of the parts would become obsolete before the mold was even cut.

When it comes to manufacturing, 3D printing is a whole different ballgame. No other manufacturing technique can make so many complex parts without human intervention. This means it opens up all new possibilities for how we manufacture things - like continuous iteration of shipping hardware.

fit2rule• 12 years ago

" .. at home".

Experiment and explore without the wait for Mr. FedEx.

vdm• 12 years ago

"The entire mechanical system can be assembled from 3D printed and off-the-shelf components." http://youtu.be/aEPf0QHVuMM?t=1m18s

lotsofmangos3• 12 years ago

This is really cool. Would be interesting to see a configuration with two of them front and back, or a conventional tailrotor version.

paulftw• 12 years ago

Curious to see whether this system is robust enough for unpredictable outdoor breezes.

ww520• 12 years ago

At first viewing of the video I didn't understand why the body doesn't spin around with one rotor and a pair of blades. That was some magic! Then looked at the picture again and saw another rotor underneath. Pretty neat to have two rotors counteracting each other.

vutekst• 12 years ago

This always bothered me about Babylon 5 - how/why did the spine of the station remain stationary while the bulk of the station rotated?

mhandley• 12 years ago

This is such a neat idea, but I wonder if it suffers from vibration problems. Because the blades are mounted on pivots, whenever you've commanding differential pitch, the high angle-of-attack blade will incur more drag and so lag slightly more than the low angle-of-attack blade. Thus the blades won't be exactly opposite each other anymore, creating vibration. Perhaps the blades are spinning so fast this is not a big deal? Probably would be if you scaled up though.

logfromblammo2• 12 years ago

I don't think the force on the rotor would change direction fast enough to cause a vibration. It would be more like a constant push relative to the intended tilt, or a slowly spiraling wobble.

I'm not a mechanical engineer, though.

MadManE• 12 years ago

Vibration only becomes a problem when the scales of the drag/lag and rotational frequency are similar. My gut feeling is that this will only happen when your lift-to-drag ratio is close to 1, so not in any meaningful case.

emmanueloga_• 12 years ago

Reminds me of PWM [1] on old PC speakers. Many old computers (e.g. of the 286 era) had a speaker that was only able to generate square tones. By controlling the start and duration of the pulses, programmers could generate sounds that were a lot more rich than what you would expect from "plain" square waves [2].

In a way this is similar: they start with two actuators that seem very limited in what they can accomplish: you can only speed them up and down. By modulating the speed of both motors in sync, they are able to control that cheap rotor mechanism they came up with, achieving 6DoF [3]

1: http://en.wikipedia.org/wiki/Pulse-width_modulation

2: https://www.youtube.com/watch?v=FSe3ysBrXq4

3: http://en.wikipedia.org/wiki/Six_degrees_of_freedom

Underactuated Rotor for Simple Micro Air Vehicles

Related Stories