Just want to mention that hobbyists into fixed wing RC aerobatics have been doing crazy stuff with variable pitch propellers (VPP) for several years. RC motors with integrated variable pitch propellers exist in small sizes, so no need to DIY the propeller for that application. Here's an example of the 4D aerobatics they enable: https://www.youtube.com/watch?v=j5EsvZNvVRs.
This is similar to but different than what the creator of the one motor drone is doing since he's spinning at higher RPM with the pitch varying under CPU control much faster than fixed wing aerobatic pilots who are controlling everything manually with insane skill. I assume the one motor drone guy already knows fixed wing RC VPP exist, just mentioning for others who might not.
I don't know the exact config on this plane as it's probably a custom design since Donatas is a well-known pilot in F3P indoor aerobatic competition. I fly small scale outdoor 3D aero and based on what I see it doing, I think it's traditional control surfaces, rudder, elevator and ailerons plus two additional elements: a variable pitch propeller and a thrust vectored motor.
VPP and thrust vectoring are two separate elements which do different things. A variable pitch propeller can reverse the propeller from pulling the plane forward to pushing it backward. That's what enables diving toward the ground and almost instantly reversing to start going back up. Thrust vectoring dynamically changes the angle of the motor allowing the plane to turn in an even tighter radius than a large rudder can alone. The turn can be so tight it's less than the length of the plane, technically becoming a spin.
In fixed wing RC aerobatics, thrust vectoring (TV) isn't a feature of the motor itself, it's actually a feature of the mount the motor is attached to, which can pivot left/right ~20 deg. A rudder changes the direction of travel by pushing the tail to the left from the rear indirectly causing the nose to turn toward the right. A thrust vectored motor mount increases the rate of turn by also angling the motor to directly 'pull' the nose toward the new direction. In RC the angle is usually changed mechanically by extending the rudder servo control rod forward to the pivoting motor mount and reversing the connection point so the motor pulls the nose toward the opposite direction the rudder is pushing the tail.
As a 3D pilot looking at the video, Donatas is doing things which are wildly impressive and, if you don't know about VPP and thrust vectoring, look impossible. Before we even get into VPP and TV, the first amazing thing about the video is simply the airspeed. I fly very lightweight 3D planes and the video looks slo-mo even to me. That size of plane shouldn't be able to stay in the air moving that slowly. F3P indoor-only competition planes are hand-crafted insanely light from esoteric materials using clever engineering tricks to sustain torsional rigidity. They sweat over every gram to maximize the cubic wing loading.
Second, is the insane piloting skill on display. Carving those perfectly straight lines and precisely uniform arcs so slowly in such a small volume takes years of practice. That perfection is what F3P competitions judge. Last is leveraging the VPP stopping the the plane in midair and moving back upward. Keeping the plane perfectly vertical on that down line is already like balancing an ice pick on your fingertip. Stopping and moving up is next level. The video is probably from an exhibition as I don't think they fly VPP props in F3P competition since they're heavier and these guys even test the strength vs weight of dried glues to save tenths of grams.
Yes, you should be (besides the flying skills), and yes, it is a hard technical problem.
Normally the plane would use control surfaces, this one uses thrust vectoring with only one item providing thrust. I do think the rudder moves though, but technically that is still redirecting the thrust (because otherwise they'd have to move the motor around).
Most variable pitch propellers use a swashplate (helicopters) or just a collective yoke (also helicopters, and acrobatic planes).
If I understand what op has done is to print a propeller with NO FIXED LINKAGE to control the propeller pitch, but instead lead or follow the inertia of the propeller by rapidly-adjusted motor rpm to induce variation in the angle of attack of the propeller.
I now realize I should have explained more in my original post, since there's some confusion between the what's unique about the experimental "One Motor Drone" (OMD) and more 'normal' VPP and thrust vectoring (TV), which are fairly unusual in RC but still common enough there are existing components and designs on the market which all work essentially the same way.
The OMD technically fits the definition of VPP and TV in that the propeller pitch varies and the thrust vectors but it accomplishes this in different ways and with very different trade-offs than the existing RC components, which I mentioned because they're still new-ish in RC aerobatics, unusual enough many RC pilots aren't familiar with them and provide interesting capabilities.
For the life of me I can't find the actuators, they must be absolutely tiny or camouflaged really well but there definitely is some kind of wire attached to the rudder. I can't find anything to the other control surfaces nor do I see any deflection there in the video, but it could of course be very subtle.
The flying is absolutely off the scale, that is unbelievable skill. The OMD is also impressive but it is very much a work in progress and I think it will be a while before we see fully control of the drone. But at least he has it stabilized.
F3P indoor planes typically use de-cased linear servos which are flat and tiny. Their control rods are made out of piano wire so thin, there's no way it would show up on video. F3P plane designers go to extremes of miniaturization and weight reduction. F3P isn't the corner of RC 3D aerobatics I play in but it's fun to look at photos of crazy F3P planes and the videos just look... wrong :-). https://www.youtube.com/watch?v=lXD6vbGVROw
These days, quads get all the attention (and blowback) but fixed wing is the OG RC. In recent years, meta-materials + SDR + open source have changed fixed wing dramatically, unlocking breakthroughs leading to some crazy extreme flight envelopes and many sub-specialties. There's dynamic soaring gliders which reach sustained speeds over 400mph from hand launch with no motor, cross country motorless gliders can go a hundred miles from launch, and long-range powered FPV (long range = >25 miles), though lately the first rule of LR-FPV Club is 'Don't talk about LR FPV Club' :-)
Personally, I find outdoor 3D aerobatics and slope soaring to both be extremely fun. Although my 3D aerobats are a bit larger than F3P (~ .8 - 1.2M), they are still very light and quite unlike most RC planes in unexpected ways. My planes can fly low, slow and so close I can just launch and catch them by hand. They are quiet, don't need much space and bother no one. Yet, unlike F3P planes, mine can fly outdoors whether in a large backyard, local park or even the beach. Heading into a hint of breeze, they can slow down until they're just stopped, hanging in the air (which looks odd), then zoom off in a blink at over 30 mph, and they keep flying in high winds that ground most RC planes (in fact, it gets even more fun if you have sufficient skill). Because they all have a thrust-to-weight ratio far above 1:1 they also have pretty insane flight envelopes enabling high-torque and high-G aerobatics far beyond the reach of any full-sized planes at airshows. Yet they only cost ~$50 to $200, take maybe 90 mins of my time to build, and can survive high-speed crashes that would turn most RC planes into confetti (the old joke in RC is they tell newbies who insist on starting with too fast and fragile of a plane, "bring a trash bag for all the pieces").
That all combines to make my corner of RC: fun, cheap, low stress and easy to start yet with endless depth to learn ever-cooler aerobatics. I love doing crazy aerobatics only 3 ft off the ground and 20 ft away yet if I 'dumb-thumb' the plane into the ground, it just bounces - or worst case it's a 5 min repair I can do on the spot while changing batteries. All thanks to cutting-edge materials, new motors, HV Lipos, and tiny servos with crazy high torque. It's also engineering-centric and nerd-friendly - so very unlike the vibe of "Your Dad's RC Plane Club." No balsa wood, gas engines or special flying field needed. My radio transmitter looks like an XBox controller yet is modular, runs open source firmware, BT connects to my phone, has SDR features and can be scripted with Lua. I learn new aerobatics with an open source 3D simulator on my laptop or in VR so I can slow down real-time as I learn.
Underactuated robotics are so cool. In the hobby BattleBots community, there is a design where the whole robot only has a single motor, but it's able to move in any direction and damage the opponent, all just by controlling the single motor's speed.
Most robots have at least three motors, two drive sides and a weapon motor. The trick to getting away with one is integrating an impactor into the chassis of your bot, using the single wheel to spin up to ~3k RPM, then by tracking the rotation angle of the bot as it spins, you speed up a little while the wheel is pointing where you want to go, and slow it down a little when it's pointing the other direction.
This design is generally called Meltybrain, and it has been in professional robot fighting too. Nuts 2 was in Robot Wars, and got surprisingly far into the competition for being viewed as somewhat of a joke entry (all the roboteers were a bit mad and had silly clothes). It's success wasn't always its fault, but it was still very impressive
Well not just battlebots, real life 6th gen fighters now all lack the vertical stabilizer so they have to rely on very sensitive real-time software control of the elevator/spoilerons to do banks and yaws. Not sure if it counts as under-actuated but they are tremendously difficult to steer without computer control.
Generally these aircraft still have three+ actuators, they just replace the rudder with differential spoiler or split elevons. Three actuators, three axes. Same number, so it's not under-actuated.
You're referencing the fact that it's dynamically unstable, and thus needs to be stabilized with a computer. That's a different thing.
You can also have aircraft which are under-actuated yet are dynamically stable. One example is a parafoil, which achieves controlled flight using only two actuators (pull lines).
I have wondered how far you can get without an engine and with computer stabilization. A plastic bag flies in the wind, and sailships can sail upwind to some degree, so surely someone is experimenting with solely wind powered drones?
The easiest version of that is a balloon, they go where the wind goes but they can go very far and if the winds are favorable then you can do 800 km on a couple of AA batteries while powering up your controller every 10 minutes or so.
I'm not super familiar with heavy equipment like dozers and such, but some of them may also have a single motor to pressurize hydraulic fluid and valves to articulate the various pistons. So may be able to make the argument there too
That is common and you can't run two different cylinders at the same time. Depends on how the pump and valves are plumbed of course. Sometimes you can run more than one but you split power.
You can run your cylinders in parallel, however that means whichever cylinder has its easiest time moving is going to capture a proportion more share of the power. If things are not perfectly balanced, which in reality they're never going to be, that means the cylinder that needs to work the hardest is going to work the least. And so it's going to take a huge amount of effort to control.
Alternatively, you can run your cylinders in series, however that means that if you're trying to feather one to move slow, all the others are going to have to move slower too.
You end up needing multiple pumps in order to make this work well, but pumps are rather expensive and so that's going to bring your costs up. Plus, multiple pumps mean that you're going to potentially want more power than your motor can deliver.
All the above can be solved with careful engineering. However, engineering is expensive too in order to get this right and it's going to take a lot of work.
I was thinking in terms of pressure regulating valves on each output. That would solve the problem of different actuators needing different pressures, at the cost of efficiency as the pressure drop across each valve is converted to heat. Or flow control rather than pressure control maybe?
I presume aircraft hydraulics must do something like that, as they have multiple moving parts under servo control (which must involve some clever valves).
I have a layman's understanding of hydraulics. I can report my own experience on machines I have with hydraulics. However, you're getting into areas where a mechanical engineer with expertise in hydraulics is needed to comment further. I would not be surprised if the people with that knowledge are not reading Hacker News.
Oh yeah, you can even get involved! There's probably a group near you and it's a great community. The highest production value events are on the east coast, https://www.nhrl.io/, but there are events all over, check out https://www.robotcombatevents.com/event_map
This is well worth the entire watch. It feels like what would any hacker do in trying to figure out how to make something they want to exist. Also very well presented and good pace.
Only to figure out that the vibration in his drone was due to vibration and it was easily fixed by swapping to rotors that he should've used in the first place. Oh well! Live and learn.
These things are not always as obvious as you make them out to be.
I was fighting a lot of similar instability when making my drone only to realize that a software regression many versions prior had introduced a bug that made flying the drone next to impossible. After fixing that bug a whole pile of issues went away. Debugging that on a drone that does not fly is super hard because you never get to the point where you can properly observe what is going on. So improving the observability by doing that camera trick was just what was needed here.
A really neat concept. It uses a special hinge design, a magnetic encoder and torque modulation to control the lift of the rotor as it spins. Completely negates the need for a conventional helicopter's swash plate.
The single motor does have the pretty major downside that it spins the moment it's in the air. I suspect a 2 motor drone with a tail motor would be a lot more practical. But I really like how mechanically simple the rotor hinge design is - it pushes all the complexity into the control algorithm.
Perhaps there are specific applications where the payload spinning could be an advantage, imaging/scanning perhaps (which in the video is going to be in the next iteration for the builder).
I don't think it would be all that useful, give the very way it works (modulating motor speed) - your rotation would not be consistent in speed during even a single turn, which makes exposure times different each frame, and possibly too low.
GP is right and the exposure time difference can be dealt with in software because you can model it. The degree to which you can modulate a modern CCD's exposure time dwarfs the speed of rotation.
Side note, I used those same moteus controllers for a decidedly less energetic use case (two axis gimbal for RF satellite tracking) and found them to be super high quality and the lead developer always super helpful. Been a few years but def worth a look if you want to get into motor control.
Nice work! It's good to see things more from first principles also (although incremental improvements to mainstream things are also good). Also there's lots of historical things that could be revisited with modern technology. For example this is slightly similar: https://en.wikipedia.org/wiki/Dornier_Aerodyne
Very cool. I also built a drone with only one motor and only one rotor, it took about three months and I learned more about drone firmware than I ever wanted to. I can't fly it but one of my sons is an absolute genius with this kind of thing and he can make it do tricks.
This kind of project is extremely rewarding. What's interesting here is how many things we did different. I used four servos and vanes in the airstream so that the whole thing would be mechanically as simple as possible, which makes this solution even more impressive to me (though I think mine is a lot more stable in flight judging by the video).
I think I went through 10 prototypes before the first one actually flew and #13 flies very well.
Yes. Mine does actually fly though, you can run it for as long as you have juice and make it go where you want it to. This one is not controllable, it is stable but it doesn't really go where it is told to go. Though in the longer term he may be able to fix that, I doubt it will work without at least one additional actuator.
It is possible to get full control with just one motor and one control surface but not with offset payloads.
Yes, servos are motors too. But you'll notice that this drone here is not quite controllable. It is still unstable in flight. It can take off but it is seconds before it runs into something and it is yawing all the time.
It means my eyesight and control/reflexes are shot to the point where I can't fly a drone but my kid is super sharp with this stuff and can do it easily. Drones are fast.
Indeed. I've some furniture with a couple of bites taken out of it from flying a much larger drone indoors. Lessons learned and all that (it seemed perfectly safe until it was not...).
Very clever concept! Curious if you can now scale it to carry some payload. Reminds me of a story where drones deliver medicines / food to remote locations. And a drone with single motor is cost effective that can be used in low income provinces with no connectivity.
Payload attachment is hard on these. The COG is crucial with respect to maintaining control, you want it to be roughly between the battery and the rotor plane and as high as possible on the drone. So if you add a payload then you are immediately shifting that COG outside of the range were you have good control.
I don't think it would. This design has a trade-off: you lose some yaw control in favor of pitch/roll/altitude control. If you look at the design a bit more closely you'll see that the body has statically offset vanes on it that counter the bulk of the momentum of the prop. But that's an average and if you end up doing a bit more or a bit less you'll end up with some spin that you won't be able to compensate for without giving up something else that you want as well.
It is amazing but this one won't be practical the way it is right now when scaled up, but there is a unique element in there that probably could be re-used.
As others have said it's mostly just consistency of practice. There's a wealth of resources available freely online for these topics and you learn by doing. I would caution setting goals/expectations based on what others have achieved in youtube videos though - this guy was/is a computer science undergrad at Cambridge, so he certainly has a high capacity for learning hard stuff and has been exposed to some of the best schooling on offer.
You learn along the way by finding something you want done, never giving up.
I want to build a robot that does one simple thing - follow my voice and deliver something to it's original location. That means I need to learn batteries and power loads, electronics, audio processing, directional acoustics, more programming, soldering, and a half dozen other things.
AI + deep thinking about the problem.
The goal is not the objective but to learn everything along the way.
I can't answer for the OP, but for me it was mostly trying, reading, failing, discussions with smart friends and then trying again.
Even with the amount of open source software for this sort of application out there it will still take you at least a month but more likely two or three to get something that can be controlled.
If you want a nice place to start with building then I think Dhremflight's videos are an amazing resource as is liftoff on Steam because you can get some flying skills before you actually control something irl.
I wanted to learn more about photography, I started asking an LLM and it helped me immensively. Definitely not the same realm but if you wanna pick up new skills, LLMs have tons of beginner friendly ways to explain things!
Yes, I’ve spent a couple hours lately asking Gemini about various home improvement topics in fundamental detail. I’ve learned a lot about:
- How to correctly hang a pre-framed door (shim it, tack it in place with 16 gauge nails, and use one wood screw in the top 2 hinges to prevent the door from sagging)
- How to properly adhere MDF trim to the ceiling to hide a closet door track (screw a furring strip or a header board to the studs above the drywall, shim it so it’s level in case the ceiling is uneven, and then nail through the face of the trim into the strip)
- What the difference is between the 27 DAP putties and spackles in the paint aisle (turns out Painters Putty 53 is typically used to hide finish nail holes, because you tend to make a lot of them and can just smoosh it in and level it off with a palette knife!)
The best part about using an LLM for learning new things like that is not having to deal with snarky and snobby comments from the community members involved in them. Thanks to an LLM, I learned how to install, maintain, and winterize an above-ground pool. It's not rocket science, but it was not easy either. I peeked into the forums surrounding the pool communities for help first and they were always so hostile towards newcomers. It's very demotivating when you're excited about something new and half the people that are supposed to be helping you are just rude elitists; completely ruins the whole experience.
It definitely depends on the community, I personally have had mostly good experiences, but I also spend a couple hours searching & reading before asking anything. And to be fair to some communities: when you are constantly bombarded with noobs asking the same questions that could have been answered with a simple search (not insinuating you did this), it can be frustrating. It's also the reason why I stopped helping in a lot of forums I used to frequent.
But I do agree about using LLMs for learning new things - it's great to just get those answers, hopefully the LLM doesn't hallucinate too much and cause you to destroy your project in the process ;) On the flip side, it removes the useful skill of learning to communicate with fellow human beings, and atrophies the skill of how to ask questions, which often can lead to answers just by phrasing the question correctly. It also removes the chance to join a community of like-minded individuals. It's a completely soulless interaction. Maybe it's a boon for anti-social types, but I think it might also be creating more anti-social behaviors in humans, in a time when IMHO we should be learning to become more social and accepting of each other.
For optics through photography (not CCD operation or post processing which are vast subjects themselves) I highly recommend View Camera Technique 7th Edition and Scientific America Image, Object and Illusion as general audience applicable introductions.
Serious answer - find something that you are so interested in that the problem is you can't stop. Become obsessed.
I went down a similar rabbit hole soon after covid started to build a small, eh, 'hand sanitizer factory' (that's my story and I'm sticking to it). By the time it was done I had a fully automated distillation setup complete with load cells, six different temperature readouts, solid state current control, flow/temp control for liebig condenser recycle, hierarchy of pid loops largely managing output flow and an attempt at live specific gravity testing. That with a nifty little dashboard to show live progress and compare metrics from various runs. Spent a couple grand on it in total and learned a LOT in the process.
Now with LLMs? Super easy. Just ask for step by step instructions to build a standard drone. You will learn about micro controllers, physics, pcbs, 3d printing etc.
If you have some time available definitely worth it.
A good engineering course would cover most (if not all) of the skills necessary to do this. This is notable because the person is young and wielding a broad slice of the discipline with maturity.
gyroscopes drift over time, that's why typically you need a kalman filter to combine the strengths of various sensors, magnetometers, where available, are bad at resolution, but don't drift.
MMs have their own set of issues though and calibrating them and knowing how the environment can influence them is a requirement for being able to use them productively.
true, sensor fusion of accel/gyro/mm only gets you so far, though of the 3, the mm is the only one that doesn't require integration, so while they are terrible at finding magnetic north in a given environment, the environment's big chunks of metal tend to not move, so with gravity, is the only external referential. That's why to do this stably you need something more, lidar/cameras+slam, the IR trick here, gnss (ideally with rtk/ppk/ppp).
> while they are terrible at finding magnetic north in a given environment
That's actually quite doable. The problem is that when the environment changes you will drift and then it becomes very hard to distinguish between actual heading changes and external influences. But if you can make a full spin you can (usually!) obtain magnetic North again. For drones this is a problem because they will drift and spin all the time. When you're on the ground and you have encoders on your axis you can do a lot better than with a drone.
RTK GPS is very nice to have but may not always be available (GPS itself may not always be available). I have some stuff in the works just for that. It does well in the lab but I still have (serious) problems in the field.
even with the calibration procedure, i've had nearly no success without going into industrial grade components, and as you point out, it's impractical.
gps also won't give you orientation really anyway, and indeed, is mostly outdoors, and rtk is needed for this resolution but also rather inaccessible to amateurs.
If you're working on indoor navigation, that's very cool, when I used to do it, in my context it was easier because we controlled the environment of the robots (markers on walls of a known colour type stuff). We also explored passive radar type arrangements for picking up wifi repeaters, but it became complicated to eg. ignore bluetooth.
Though in the case of this video, the wii sensor bar style approach seems fine, assuming you control reflections
Both indoor and outdoor. I have the relative positioning and heading working fine, it is the absolute positioning that is giving me a lot of trouble. I'll get there though, sooner or later or I'll die trying :)
It's insanely complex stuff, I've had to do a lot of re-studying math and this whole project has given me more respect than I already had for the engineers behind modern positioning systems.
Sure, get a 'tinywhoop' and learn to fly by hand. They're cheap, can be flown indoors and won't do damage so you can practice to your heart's content. The batteries don't last long, you'll get a minute, maybe two if you're lucky out of a single charge.
Even cheaper:lift-off on Steam. It's a simulator, you can then just get a controller (I'd recommend a TX10 to start with, a TX16 if you think you're going to be serious about this stuff) and hook it up to the computer.
It's a pretty cool concept, however, fails to mention coaxial rotor design that has a single motor and has much simpler control systems. I am also curious on how this affects frame vibration.
This is a newer trend in mechanical design that leverages more complex control software to overcome mechanical issues. A good example is https://turntide.com/
He leads the video with claims of being the first, and it’s reasonable to expect him to have done sufficient research on that claim. It’s a different claim than what GP thought.
If I understand correctly, those are single rotor but NOT single motor. This is a single motor design, which is why it's neat. There are no additional control servos/motors.
I strongly doubt adding a ton of extra weight in aerodynamic surfaces, and drag, and complex mechanical parts, to avoid a few motors/servos is worth it in that context.
Every gram of weight you remove is another gram of explosive you can add. Also consider the difference in supply lines for aerodynamic surfaces and mechanic parts and motors and servos. If you have four hundred servos, you can make one hundred quadcopters, or four hundred of these things. If you only have one hundred control boards powerful enough to control the single motor design though, you can still only make one hundred. (Tho the quadcopter can lift 4x the weight)
What's interesting is that the Ukrainians are building drones near their front lines because if they build them too far away and ship them in, the game has evolved and those drones are already yesterday's news.
A fixed wing drone only needs one motor, the control is easier, and the physics are more favorable for carrying high ordnance. I'm skeptical that this is going to change much.
It's cool in that it can hover and do quadcopter type flight with only a single motor, and it's that second part that is interesting, because a fixed wing can also do a prop hang, but it can't do quadrotor-like flight from there.
It's usefulness is likely to be in other areas -- mainly applications where that flight mode is important.
This design still requires an additional servo for rotor actuation, so the minimum number of actuators is the same as for a fixed wing (roll and yaw are coupled, so the bare minimum is a single rudder servo).
My bad. I skimmed the early part of the video with the sound down on the train this morning and thought I saw a servo in there. That's clever, but it doesn't really change much. There's not really any advantage in terms of payload or performance over existing designs.
Drone tech is dual use, as is most other tech. It's hard to have the one without the other. And yes, drones killing civilians is very demented, such war crimes (because that is what they are) should not go unpunished.
Likewise. Even companies like Saildrone where ~100% of the marketing before 2025 was about mapping the oceans and autonomous storm monitoring with their bright orange craft are now releasing videos with missile launching pods attached to navy gray versions: https://www.saildrone.com/news/us-navy-rimpac-lockheed-marti...
Do you want superpowers to be able to wage wars, foreign or domestically, with hordes of non-human combatants that can't refuse orders or defect on moral grounds?
To be fair, a lot of it is moving to drones killing drones.
My prediction is that future wars will actually involve increasingly fewer humans and will be more about tech+manufacturing vs. tech+manufacturing. And of course insurgents + lowcost tech vs. increasingly nonlethal hightech.
This would allow wars to continue without as much of the growing political backlash.
There's also been a few people building monocopters with a fixed, non tilting motor and prop, and four servo motor actuated fins at the bottom that can be used for thrust vectoring and to counter rotational forces.
That's pretty close to what I built only I'm using a far more efficient motor/prop setup so much longer flying times as well as the option to transition to level flight.
Neat. I wonder if you could use the air it generates by having directional ports louvers that open and allow the air to flow for more control.
Would need 2 more servo, left/right port louvers.
Generally those small diameter motor/prop combinations aren't very efficient, in grams/watt of thrust. They look really bad if you compare them to like a 15" size carbon fiber prop from t-motor and one of the tmotor "antigravity" series motors on a thrust stand.
Quadcopters that are designed for 45 min to 1 hr loiter time carrying a useful payload, while not being too big for a person to move around seem to be converting on 15 to 18 inch size prop and motor and esc combinations carefully chosen for optimal grams/watt thrust.
I've built a whole slew of drones over the last couple of years, quadcopters are hard to beat on many axis. But I have some proof that it may be possible to do substantially better with a monocopter in terms of efficiency. The hard problems you will then have to solve are payload attachment because unbalance is very tricky to deal with (whereas that is trivial in a quadcopter) as well as transitioning to level flight so that you can make good forward speed.
I think the recent winner of the DARPA life challenge was a 'traditional' single rotor helicopter with tailrotor. The very large disc area of the traditional heli design made it more efficient than quad/hex/octo designs.
Yes, but their forward speed is limited because they can't translate into level flight and that also means the range is limited. I was trying to get the best of both worlds (but that's a hard problem to solve).
I know very little of the field, but aren't there applications where the ability to carry ad hoc payloads doesn't matter? Integrating a fixed sensor suite into the ring and balancing it ahead of time is possible then, right?
And what can be learned from full-scale classical helicopters that frequently carry underhanging payloads?
(I don't at all wanna be the typical HN person who thinks their questions about a field they know nothing about are in any way insightful. I'm just trying to learn.)
That's all true. But (1) that's not the application that I have in mind and (2) forward speed of helicopters is not fantastic (3) neither is their fuel efficiency.
For a fixed payload it is indeed possible to set up the balance beforehand, even then you will probably end up with a drone that will translate slower or faster depending on the inbalance instead of hovering in place.
It's incredible that changes in velocity still work to adjust the angle of the blades even at the insane speeds they're moving at. I would never have thought that was possible.
I have to wonder why this guy didn't give his drone a leash while testing so it couldn't go flying off. It's the first thing that popped into my head, so I'm sure it's something he thought of as well and didn't use for a good reason. Maybe a string would get caught in the blades?
Maybe the pull from a string/leash would affect how the thing flies, kind of like kites.
I like how he built a test stand near 7:22, which solves the motor flying off problem and is probably easier to iterate on. It reminds me of how rockets are tested.
I’m sure glad we are making immense progress towards the dystopian hell where massive swarms of mini drones are hunting down obsolete and marked for termination flesh bots, aka humans.
Just want to mention that hobbyists into fixed wing RC aerobatics have been doing crazy stuff with variable pitch propellers (VPP) for several years. RC motors with integrated variable pitch propellers exist in small sizes, so no need to DIY the propeller for that application. Here's an example of the 4D aerobatics they enable: https://www.youtube.com/watch?v=j5EsvZNvVRs.
This is similar to but different than what the creator of the one motor drone is doing since he's spinning at higher RPM with the pitch varying under CPU control much faster than fixed wing aerobatic pilots who are controlling everything manually with insane skill. I assume the one motor drone guy already knows fixed wing RC VPP exist, just mentioning for others who might not.
How is this working? I can't tell if they are using flaps as well??
the rudder seems to have a wire attached, barely visible in the opening shot. that seems to be it though.
I don't know the exact config on this plane as it's probably a custom design since Donatas is a well-known pilot in F3P indoor aerobatic competition. I fly small scale outdoor 3D aero and based on what I see it doing, I think it's traditional control surfaces, rudder, elevator and ailerons plus two additional elements: a variable pitch propeller and a thrust vectored motor.
VPP and thrust vectoring are two separate elements which do different things. A variable pitch propeller can reverse the propeller from pulling the plane forward to pushing it backward. That's what enables diving toward the ground and almost instantly reversing to start going back up. Thrust vectoring dynamically changes the angle of the motor allowing the plane to turn in an even tighter radius than a large rudder can alone. The turn can be so tight it's less than the length of the plane, technically becoming a spin.
In fixed wing RC aerobatics, thrust vectoring (TV) isn't a feature of the motor itself, it's actually a feature of the mount the motor is attached to, which can pivot left/right ~20 deg. A rudder changes the direction of travel by pushing the tail to the left from the rear indirectly causing the nose to turn toward the right. A thrust vectored motor mount increases the rate of turn by also angling the motor to directly 'pull' the nose toward the new direction. In RC the angle is usually changed mechanically by extending the rudder servo control rod forward to the pivoting motor mount and reversing the connection point so the motor pulls the nose toward the opposite direction the rudder is pushing the tail.
As a 3D pilot looking at the video, Donatas is doing things which are wildly impressive and, if you don't know about VPP and thrust vectoring, look impossible. Before we even get into VPP and TV, the first amazing thing about the video is simply the airspeed. I fly very lightweight 3D planes and the video looks slo-mo even to me. That size of plane shouldn't be able to stay in the air moving that slowly. F3P indoor-only competition planes are hand-crafted insanely light from esoteric materials using clever engineering tricks to sustain torsional rigidity. They sweat over every gram to maximize the cubic wing loading.
Second, is the insane piloting skill on display. Carving those perfectly straight lines and precisely uniform arcs so slowly in such a small volume takes years of practice. That perfection is what F3P competitions judge. Last is leveraging the VPP stopping the the plane in midair and moving back upward. Keeping the plane perfectly vertical on that down line is already like balancing an ice pick on your fingertip. Stopping and moving up is next level. The video is probably from an exhibition as I don't think they fly VPP props in F3P competition since they're heavier and these guys even test the strength vs weight of dried glues to save tenths of grams.
Bloody hell, I was expecting something interesting, but that level of control from a single propeller is a lot more than I had imagined.
Too many different things I want to play with and insufficient time in the day -.-
I watched the video. Should I be impressed, no emotions here? Is this a hard technical problem overcomed and demo-d?
Yes, you should be (besides the flying skills), and yes, it is a hard technical problem.
Normally the plane would use control surfaces, this one uses thrust vectoring with only one item providing thrust. I do think the rudder moves though, but technically that is still redirecting the thrust (because otherwise they'd have to move the motor around).
As near as i can tell this is very cool.
Most variable pitch propellers use a swashplate (helicopters) or just a collective yoke (also helicopters, and acrobatic planes).
If I understand what op has done is to print a propeller with NO FIXED LINKAGE to control the propeller pitch, but instead lead or follow the inertia of the propeller by rapidly-adjusted motor rpm to induce variation in the angle of attack of the propeller.
Yes, that is exactly what he's doing. I posted a link elsewhere in this thread where someone does the same thing but with a fixed pitch prop.
GP here. I'm an outdoor 3D RC aero pilot and explained what I think the config of the plane in the video is in this sister comment: https://news.ycombinator.com/item?id=50169548.
I now realize I should have explained more in my original post, since there's some confusion between the what's unique about the experimental "One Motor Drone" (OMD) and more 'normal' VPP and thrust vectoring (TV), which are fairly unusual in RC but still common enough there are existing components and designs on the market which all work essentially the same way.
The OMD technically fits the definition of VPP and TV in that the propeller pitch varies and the thrust vectors but it accomplishes this in different ways and with very different trade-offs than the existing RC components, which I mentioned because they're still new-ish in RC aerobatics, unusual enough many RC pilots aren't familiar with them and provide interesting capabilities.
For the life of me I can't find the actuators, they must be absolutely tiny or camouflaged really well but there definitely is some kind of wire attached to the rudder. I can't find anything to the other control surfaces nor do I see any deflection there in the video, but it could of course be very subtle.
The flying is absolutely off the scale, that is unbelievable skill. The OMD is also impressive but it is very much a work in progress and I think it will be a while before we see fully control of the drone. But at least he has it stabilized.
> For the life of me I can't find the actuators
F3P indoor planes typically use de-cased linear servos which are flat and tiny. Their control rods are made out of piano wire so thin, there's no way it would show up on video. F3P plane designers go to extremes of miniaturization and weight reduction. F3P isn't the corner of RC 3D aerobatics I play in but it's fun to look at photos of crazy F3P planes and the videos just look... wrong :-). https://www.youtube.com/watch?v=lXD6vbGVROw
Unlike the larger, more powerful rotary servos I typically use which need to be mounted near the plane's center of gravity to maintain balance, F3P esoteric, custom linears are so light they can mount them right next to surface they move. F3P seems largely Euro-centric, but there's a growing scene in NA (https://nsrca.us/index.php/flying/f3pindoorsequencesseparato...) and a sub on RCGroups https://www.rcgroups.com/indoor-pattern-f3p-373/.
These days, quads get all the attention (and blowback) but fixed wing is the OG RC. In recent years, meta-materials + SDR + open source have changed fixed wing dramatically, unlocking breakthroughs leading to some crazy extreme flight envelopes and many sub-specialties. There's dynamic soaring gliders which reach sustained speeds over 400mph from hand launch with no motor, cross country motorless gliders can go a hundred miles from launch, and long-range powered FPV (long range = >25 miles), though lately the first rule of LR-FPV Club is 'Don't talk about LR FPV Club' :-)
Personally, I find outdoor 3D aerobatics and slope soaring to both be extremely fun. Although my 3D aerobats are a bit larger than F3P (~ .8 - 1.2M), they are still very light and quite unlike most RC planes in unexpected ways. My planes can fly low, slow and so close I can just launch and catch them by hand. They are quiet, don't need much space and bother no one. Yet, unlike F3P planes, mine can fly outdoors whether in a large backyard, local park or even the beach. Heading into a hint of breeze, they can slow down until they're just stopped, hanging in the air (which looks odd), then zoom off in a blink at over 30 mph, and they keep flying in high winds that ground most RC planes (in fact, it gets even more fun if you have sufficient skill). Because they all have a thrust-to-weight ratio far above 1:1 they also have pretty insane flight envelopes enabling high-torque and high-G aerobatics far beyond the reach of any full-sized planes at airshows. Yet they only cost ~$50 to $200, take maybe 90 mins of my time to build, and can survive high-speed crashes that would turn most RC planes into confetti (the old joke in RC is they tell newbies who insist on starting with too fast and fragile of a plane, "bring a trash bag for all the pieces").
That all combines to make my corner of RC: fun, cheap, low stress and easy to start yet with endless depth to learn ever-cooler aerobatics. I love doing crazy aerobatics only 3 ft off the ground and 20 ft away yet if I 'dumb-thumb' the plane into the ground, it just bounces - or worst case it's a 5 min repair I can do on the spot while changing batteries. All thanks to cutting-edge materials, new motors, HV Lipos, and tiny servos with crazy high torque. It's also engineering-centric and nerd-friendly - so very unlike the vibe of "Your Dad's RC Plane Club." No balsa wood, gas engines or special flying field needed. My radio transmitter looks like an XBox controller yet is modular, runs open source firmware, BT connects to my phone, has SDR features and can be scripted with Lua. I learn new aerobatics with an open source 3D simulator on my laptop or in VR so I can slow down real-time as I learn.
Here's a 1M wingspan outdoor foamy mixing high-energy and low/slow 3D aerobatics: https://www.youtube.com/watch?v=vgkrVimII1E
Yeah but what if I want those RC plane to carry 100kg of metal (or more) ???
Underactuated robotics are so cool. In the hobby BattleBots community, there is a design where the whole robot only has a single motor, but it's able to move in any direction and damage the opponent, all just by controlling the single motor's speed.
Most robots have at least three motors, two drive sides and a weapon motor. The trick to getting away with one is integrating an impactor into the chassis of your bot, using the single wheel to spin up to ~3k RPM, then by tracking the rotation angle of the bot as it spins, you speed up a little while the wheel is pointing where you want to go, and slow it down a little when it's pointing the other direction.
Can highly recommend this course https://underactuated.csail.mit.edu/ Russ Tedrake is an amazing lecturer (and researcher)
At this rate, HN is responsible for my whole adult education.
That's one of the better uses of HN.
This design is generally called Meltybrain, and it has been in professional robot fighting too. Nuts 2 was in Robot Wars, and got surprisingly far into the competition for being viewed as somewhat of a joke entry (all the roboteers were a bit mad and had silly clothes). It's success wasn't always its fault, but it was still very impressive
Well not just battlebots, real life 6th gen fighters now all lack the vertical stabilizer so they have to rely on very sensitive real-time software control of the elevator/spoilerons to do banks and yaws. Not sure if it counts as under-actuated but they are tremendously difficult to steer without computer control.
You're referencing the fact that it's dynamically unstable, and thus needs to be stabilized with a computer. That's a different thing.
You can also have aircraft which are under-actuated yet are dynamically stable. One example is a parafoil, which achieves controlled flight using only two actuators (pull lines).
I have wondered how far you can get without an engine and with computer stabilization. A plastic bag flies in the wind, and sailships can sail upwind to some degree, so surely someone is experimenting with solely wind powered drones?
The easiest version of that is a balloon, they go where the wind goes but they can go very far and if the winds are favorable then you can do 800 km on a couple of AA batteries while powering up your controller every 10 minutes or so.
I'm not super familiar with heavy equipment like dozers and such, but some of them may also have a single motor to pressurize hydraulic fluid and valves to articulate the various pistons. So may be able to make the argument there too
In that case, the valves/pistons pairs are all independent actuators. The hydraulic motor is more a constant power source for all of them.
Yeah, the analogy for drones is hydraulic piston -> electric motor and hydraulic pump -> battery
That is common and you can't run two different cylinders at the same time. Depends on how the pump and valves are plumbed of course. Sometimes you can run more than one but you split power.
That sounds like it could be solved with more valves.
it's not that easy.
You can run your cylinders in parallel, however that means whichever cylinder has its easiest time moving is going to capture a proportion more share of the power. If things are not perfectly balanced, which in reality they're never going to be, that means the cylinder that needs to work the hardest is going to work the least. And so it's going to take a huge amount of effort to control.
Alternatively, you can run your cylinders in series, however that means that if you're trying to feather one to move slow, all the others are going to have to move slower too.
You end up needing multiple pumps in order to make this work well, but pumps are rather expensive and so that's going to bring your costs up. Plus, multiple pumps mean that you're going to potentially want more power than your motor can deliver.
All the above can be solved with careful engineering. However, engineering is expensive too in order to get this right and it's going to take a lot of work.
I was thinking in terms of pressure regulating valves on each output. That would solve the problem of different actuators needing different pressures, at the cost of efficiency as the pressure drop across each valve is converted to heat. Or flow control rather than pressure control maybe?
I presume aircraft hydraulics must do something like that, as they have multiple moving parts under servo control (which must involve some clever valves).
I have a layman's understanding of hydraulics. I can report my own experience on machines I have with hydraulics. However, you're getting into areas where a mechanical engineer with expertise in hydraulics is needed to comment further. I would not be surprised if the people with that knowledge are not reading Hacker News.
Almost everything can be improved by making it cost more.
Or more power ;)
Is there a way to track what's going on in the battle bots community? It was pretty cliche, but battle bots was fun to watch in the 90s/2000s.
Oh yeah, you can even get involved! There's probably a group near you and it's a great community. The highest production value events are on the east coast, https://www.nhrl.io/, but there are events all over, check out https://www.robotcombatevents.com/event_map
It was brought back by Discovery and videos are also published on YouTube: https://www.youtube.com/channel/UCxNDM5ZhvDa_Z1qTUap7IRw
> but they are tremendously difficult to steer without computer control.
they are not humanly pilotable without computers. they are inherently unstable.
This is well worth the entire watch. It feels like what would any hacker do in trying to figure out how to make something they want to exist. Also very well presented and good pace.
Inspires me to try and start doing some robotics.
Love the little side project of building a high resolution high speed camera to investigate the flutter.
Only to figure out that the vibration in his drone was due to vibration and it was easily fixed by swapping to rotors that he should've used in the first place. Oh well! Live and learn.
These things are not always as obvious as you make them out to be.
I was fighting a lot of similar instability when making my drone only to realize that a software regression many versions prior had introduced a bug that made flying the drone next to impossible. After fixing that bug a whole pile of issues went away. Debugging that on a drone that does not fly is super hard because you never get to the point where you can properly observe what is going on. So improving the observability by doing that camera trick was just what was needed here.
And that's why mathematicians are so pissed about the OpenAI math paper drop.
And not a mention of like and subscribe for my next videos. Pure passion video blog post.
A really neat concept. It uses a special hinge design, a magnetic encoder and torque modulation to control the lift of the rotor as it spins. Completely negates the need for a conventional helicopter's swash plate.
The single motor does have the pretty major downside that it spins the moment it's in the air. I suspect a 2 motor drone with a tail motor would be a lot more practical. But I really like how mechanically simple the rotor hinge design is - it pushes all the complexity into the control algorithm.
Note that the concept itself has been done before in RC helicopters. The only difference here is using a one-propeller drone.
https://hackaday.com/2020/06/25/building-and-flying-a-helico...
https://www.youtube.com/watch?v=d80oXSCcHTk
Yep, he mentions the video you linked as well. (Mentioning because I watched the video you linked first.)
Perhaps there are specific applications where the payload spinning could be an advantage, imaging/scanning perhaps (which in the video is going to be in the next iteration for the builder).
I don't think it would be all that useful, give the very way it works (modulating motor speed) - your rotation would not be consistent in speed during even a single turn, which makes exposure times different each frame, and possibly too low.
GP is right and the exposure time difference can be dealt with in software because you can model it. The degree to which you can modulate a modern CCD's exposure time dwarfs the speed of rotation.
Couldn't you use one motor attached to two props spinning in opposite directions?
You absolutely could, and that's exactly how a lot of RC helicopters have been flying around for decades now.
https://www.scribd.com/document/840293339/Lama-V3
This is from the same guy (Joshua Bird) that built the 4 axis 3d printer (and special algo for it too) https://www.youtube.com/watch?v=VEgwnhLHy3g
The Picocolissimo also uses only one motor: https://www.youtube.com/watch?v=UE47bcCxT_I
Side note, I used those same moteus controllers for a decidedly less energetic use case (two axis gimbal for RF satellite tracking) and found them to be super high quality and the lead developer always super helpful. Been a few years but def worth a look if you want to get into motor control.
Nice work! It's good to see things more from first principles also (although incremental improvements to mainstream things are also good). Also there's lots of historical things that could be revisited with modern technology. For example this is slightly similar: https://en.wikipedia.org/wiki/Dornier_Aerodyne
Very cool. I also built a drone with only one motor and only one rotor, it took about three months and I learned more about drone firmware than I ever wanted to. I can't fly it but one of my sons is an absolute genius with this kind of thing and he can make it do tricks.
This kind of project is extremely rewarding. What's interesting here is how many things we did different. I used four servos and vanes in the airstream so that the whole thing would be mechanically as simple as possible, which makes this solution even more impressive to me (though I think mine is a lot more stable in flight judging by the video).
I think I went through 10 prototypes before the first one actually flew and #13 flies very well.
That's five motors. 500% as many motors as this guy.
Yes. Mine does actually fly though, you can run it for as long as you have juice and make it go where you want it to. This one is not controllable, it is stable but it doesn't really go where it is told to go. Though in the longer term he may be able to fix that, I doubt it will work without at least one additional actuator.
It is possible to get full control with just one motor and one control surface but not with offset payloads.
He meant one motor all up, including actuators.
Yes, servos are motors too. But you'll notice that this drone here is not quite controllable. It is still unstable in flight. It can take off but it is seconds before it runs into something and it is yawing all the time.
What do you mean "I can't fly it"? You built a drone and only someone else can fly it?
It means my eyesight and control/reflexes are shot to the point where I can't fly a drone but my kid is super sharp with this stuff and can do it easily. Drones are fast.
Reminded me of the Manhack: https://half-life.fandom.com/wiki/Manhack
manhackernews
I am surprised he is not wearing more protection. that thing looks like a hornet in bedroom test mode. Play safe.
Indeed. I've some furniture with a couple of bites taken out of it from flying a much larger drone indoors. Lessons learned and all that (it seemed perfectly safe until it was not...).
Very clever concept! Curious if you can now scale it to carry some payload. Reminds me of a story where drones deliver medicines / food to remote locations. And a drone with single motor is cost effective that can be used in low income provinces with no connectivity.
Payload attachment is hard on these. The COG is crucial with respect to maintaining control, you want it to be roughly between the battery and the rotor plane and as high as possible on the drone. So if you add a payload then you are immediately shifting that COG outside of the range were you have good control.
Yeah that's probably the coolest thing I've seen all month. How could you not like this?
Rather than slap patents on it, he's open sourced it. This guy gets it.
This is smart design and huge amount of work. Thanks the creator! Would adding a gyroscope help stablize the drone direction?
I imagine an alternative design with single motor mounted horizontally controls 4-wings like the dragonfly in Dune. It would eliminate the spinning.
I don't think it would. This design has a trade-off: you lose some yaw control in favor of pitch/roll/altitude control. If you look at the design a bit more closely you'll see that the body has statically offset vanes on it that counter the bulk of the momentum of the prop. But that's an average and if you end up doing a bit more or a bit less you'll end up with some spin that you won't be able to compensate for without giving up something else that you want as well.
That was quite inspiring. So much hope for the future in all this AI doom and gloom.
Amazing work this is really innovative and it could be applied to a larger drones provided that the margin of error is reduced.
It is amazing but this one won't be practical the way it is right now when scaled up, but there is a unique element in there that probably could be re-used.
Serious question - how do you even learn all of the skills to do something like this? Where would you start?
As others have said it's mostly just consistency of practice. There's a wealth of resources available freely online for these topics and you learn by doing. I would caution setting goals/expectations based on what others have achieved in youtube videos though - this guy was/is a computer science undergrad at Cambridge, so he certainly has a high capacity for learning hard stuff and has been exposed to some of the best schooling on offer.
You learn along the way by finding something you want done, never giving up.
I want to build a robot that does one simple thing - follow my voice and deliver something to it's original location. That means I need to learn batteries and power loads, electronics, audio processing, directional acoustics, more programming, soldering, and a half dozen other things.
AI + deep thinking about the problem.
The goal is not the objective but to learn everything along the way.
I can't answer for the OP, but for me it was mostly trying, reading, failing, discussions with smart friends and then trying again.
Even with the amount of open source software for this sort of application out there it will still take you at least a month but more likely two or three to get something that can be controlled.
If you want a nice place to start with building then I think Dhremflight's videos are an amazing resource as is liftoff on Steam because you can get some flying skills before you actually control something irl.
I wanted to learn more about photography, I started asking an LLM and it helped me immensively. Definitely not the same realm but if you wanna pick up new skills, LLMs have tons of beginner friendly ways to explain things!
Yes, I’ve spent a couple hours lately asking Gemini about various home improvement topics in fundamental detail. I’ve learned a lot about:
- How to correctly hang a pre-framed door (shim it, tack it in place with 16 gauge nails, and use one wood screw in the top 2 hinges to prevent the door from sagging)
- How to properly adhere MDF trim to the ceiling to hide a closet door track (screw a furring strip or a header board to the studs above the drywall, shim it so it’s level in case the ceiling is uneven, and then nail through the face of the trim into the strip)
- What the difference is between the 27 DAP putties and spackles in the paint aisle (turns out Painters Putty 53 is typically used to hide finish nail holes, because you tend to make a lot of them and can just smoosh it in and level it off with a palette knife!)
The best part about using an LLM for learning new things like that is not having to deal with snarky and snobby comments from the community members involved in them. Thanks to an LLM, I learned how to install, maintain, and winterize an above-ground pool. It's not rocket science, but it was not easy either. I peeked into the forums surrounding the pool communities for help first and they were always so hostile towards newcomers. It's very demotivating when you're excited about something new and half the people that are supposed to be helping you are just rude elitists; completely ruins the whole experience.
It definitely depends on the community, I personally have had mostly good experiences, but I also spend a couple hours searching & reading before asking anything. And to be fair to some communities: when you are constantly bombarded with noobs asking the same questions that could have been answered with a simple search (not insinuating you did this), it can be frustrating. It's also the reason why I stopped helping in a lot of forums I used to frequent.
But I do agree about using LLMs for learning new things - it's great to just get those answers, hopefully the LLM doesn't hallucinate too much and cause you to destroy your project in the process ;) On the flip side, it removes the useful skill of learning to communicate with fellow human beings, and atrophies the skill of how to ask questions, which often can lead to answers just by phrasing the question correctly. It also removes the chance to join a community of like-minded individuals. It's a completely soulless interaction. Maybe it's a boon for anti-social types, but I think it might also be creating more anti-social behaviors in humans, in a time when IMHO we should be learning to become more social and accepting of each other.
For optics through photography (not CCD operation or post processing which are vast subjects themselves) I highly recommend View Camera Technique 7th Edition and Scientific America Image, Object and Illusion as general audience applicable introductions.
Serious answer - find something that you are so interested in that the problem is you can't stop. Become obsessed.
I went down a similar rabbit hole soon after covid started to build a small, eh, 'hand sanitizer factory' (that's my story and I'm sticking to it). By the time it was done I had a fully automated distillation setup complete with load cells, six different temperature readouts, solid state current control, flow/temp control for liebig condenser recycle, hierarchy of pid loops largely managing output flow and an attempt at live specific gravity testing. That with a nifty little dashboard to show live progress and compare metrics from various runs. Spent a couple grand on it in total and learned a LOT in the process.
Now with LLMs? Super easy. Just ask for step by step instructions to build a standard drone. You will learn about micro controllers, physics, pcbs, 3d printing etc.
If you have some time available definitely worth it.
A good engineering course would cover most (if not all) of the skills necessary to do this. This is notable because the person is young and wielding a broad slice of the discipline with maturity.
> Where would you start?
Exploration and tinkering! But you have to have passion and deep interest in that topic/problem
Slap a MEMS gyroscope on that bad boy. https://www.vectornav.com/resources/inertial-navigation-prim...
gyroscopes drift over time, that's why typically you need a kalman filter to combine the strengths of various sensors, magnetometers, where available, are bad at resolution, but don't drift.
MMs have their own set of issues though and calibrating them and knowing how the environment can influence them is a requirement for being able to use them productively.
true, sensor fusion of accel/gyro/mm only gets you so far, though of the 3, the mm is the only one that doesn't require integration, so while they are terrible at finding magnetic north in a given environment, the environment's big chunks of metal tend to not move, so with gravity, is the only external referential. That's why to do this stably you need something more, lidar/cameras+slam, the IR trick here, gnss (ideally with rtk/ppk/ppp).
> while they are terrible at finding magnetic north in a given environment
That's actually quite doable. The problem is that when the environment changes you will drift and then it becomes very hard to distinguish between actual heading changes and external influences. But if you can make a full spin you can (usually!) obtain magnetic North again. For drones this is a problem because they will drift and spin all the time. When you're on the ground and you have encoders on your axis you can do a lot better than with a drone.
RTK GPS is very nice to have but may not always be available (GPS itself may not always be available). I have some stuff in the works just for that. It does well in the lab but I still have (serious) problems in the field.
even with the calibration procedure, i've had nearly no success without going into industrial grade components, and as you point out, it's impractical.
gps also won't give you orientation really anyway, and indeed, is mostly outdoors, and rtk is needed for this resolution but also rather inaccessible to amateurs.
If you're working on indoor navigation, that's very cool, when I used to do it, in my context it was easier because we controlled the environment of the robots (markers on walls of a known colour type stuff). We also explored passive radar type arrangements for picking up wifi repeaters, but it became complicated to eg. ignore bluetooth.
Though in the case of this video, the wii sensor bar style approach seems fine, assuming you control reflections
Both indoor and outdoor. I have the relative positioning and heading working fine, it is the absolute positioning that is giving me a lot of trouble. I'll get there though, sooner or later or I'll die trying :)
It's insanely complex stuff, I've had to do a lot of re-studying math and this whole project has given me more respect than I already had for the engineers behind modern positioning systems.
I was wondering did he use the control mixer MCU (initially mentioned in the video) at all? The encoder should always feed into the MCU directly.
Any recommended "drone for dummies" for programmers to fly their first drone?
Sure, get a 'tinywhoop' and learn to fly by hand. They're cheap, can be flown indoors and won't do damage so you can practice to your heart's content. The batteries don't last long, you'll get a minute, maybe two if you're lucky out of a single charge.
Even cheaper:lift-off on Steam. It's a simulator, you can then just get a controller (I'd recommend a TX10 to start with, a TX16 if you think you're going to be serious about this stuff) and hook it up to the computer.
Good luck & have fun!
It's a pretty cool concept, however, fails to mention coaxial rotor design that has a single motor and has much simpler control systems. I am also curious on how this affects frame vibration.
This is a newer trend in mechanical design that leverages more complex control software to overcome mechanical issues. A good example is https://turntide.com/
It’s an entertainmening YouTube video not a phd thesis. Bizzare standard to hold him to
He leads the video with claims of being the first, and it’s reasonable to expect him to have done sufficient research on that claim. It’s a different claim than what GP thought.
If I understand correctly, those are single rotor but NOT single motor. This is a single motor design, which is why it's neat. There are no additional control servos/motors.
It can be single motor too.
Can you expand on this? How? Any examples?
I am humbled. Wow!
How will this new tech affect the current drone war between Russia and Ukraine?
Smaller, cheaper, more numerous offensive drones?
This tech is more than a decade old, originally from UPenn AFAIK[0].
[0] https://youtube.com/watch?v=aEPf0QHVuMM
I strongly doubt adding a ton of extra weight in aerodynamic surfaces, and drag, and complex mechanical parts, to avoid a few motors/servos is worth it in that context.
Every gram of weight you remove is another gram of explosive you can add. Also consider the difference in supply lines for aerodynamic surfaces and mechanic parts and motors and servos. If you have four hundred servos, you can make one hundred quadcopters, or four hundred of these things. If you only have one hundred control boards powerful enough to control the single motor design though, you can still only make one hundred. (Tho the quadcopter can lift 4x the weight)
What's interesting is that the Ukrainians are building drones near their front lines because if they build them too far away and ship them in, the game has evolved and those drones are already yesterday's news.
A fixed wing drone only needs one motor, the control is easier, and the physics are more favorable for carrying high ordnance. I'm skeptical that this is going to change much.
It's cool in that it can hover and do quadcopter type flight with only a single motor, and it's that second part that is interesting, because a fixed wing can also do a prop hang, but it can't do quadrotor-like flight from there.
It's usefulness is likely to be in other areas -- mainly applications where that flight mode is important.
Larger fixed wing drone could deliver many smaller quadcopters that don't have the long range.
A fixed wing drone needs at least one extra motor for the control surfaces. More likely two or three.
This design still requires an additional servo for rotor actuation, so the minimum number of actuators is the same as for a fixed wing (roll and yaw are coupled, so the bare minimum is a single rudder servo).
No, the rotors are tilted by timing the acceleration from the single motor - there's no servo in the design in the video.
My bad. I skimmed the early part of the video with the sound down on the train this morning and thought I saw a servo in there. That's clever, but it doesn't really change much. There's not really any advantage in terms of payload or performance over existing designs.
> the Ukrainians are building drones near their front lines
Both sides 3d print parts near front lines. Just to be clear.
Especially when it comes to drops.
It won't. This is not going to work with a payload and the degree of control is too low. It's a very cool POC though.
This type of rotors are not highly resistant to winds. That's why drones don't look like this already.
maybe new breakthroughs for the darpa project. kamikazes drones do not need so advanced prop tbh it is against the idea of being cheap.
I am torn about drone tech now.
I used to be super stoked about it.
But watching drones kill civilians kind of ruined it. [0]
Then again ... they do put a dent in superpowers powers apparently so ... like I said ... I am torn.
[0] https://www.bbc.com/news/videos/czdmrm6y7r0o
Drone tech is dual use, as is most other tech. It's hard to have the one without the other. And yes, drones killing civilians is very demented, such war crimes (because that is what they are) should not go unpunished.
Likewise. Even companies like Saildrone where ~100% of the marketing before 2025 was about mapping the oceans and autonomous storm monitoring with their bright orange craft are now releasing videos with missile launching pods attached to navy gray versions: https://www.saildrone.com/news/us-navy-rimpac-lockheed-marti...
Do you want superpowers to be able to wage wars, foreign or domestically, with hordes of non-human combatants that can't refuse orders or defect on moral grounds?
To be fair, a lot of it is moving to drones killing drones.
My prediction is that future wars will actually involve increasingly fewer humans and will be more about tech+manufacturing vs. tech+manufacturing. And of course insurgents + lowcost tech vs. increasingly nonlethal hightech.
This would allow wars to continue without as much of the growing political backlash.
The next iteration is the drones removing those pesky humans from the command chain too :/
You are overly hopeful, I suspect.
A significant portion of Russia's attacks are purely on civilians. Even Ukraine's attacks on RU targets have civilian side casualties.
Terrorism will probably always be a thing.
You could say the same about smelting iron. Now we can wage war without tin.
No arguments here.
I guess the most important thing is to spread the tech then... make it fair at least? (Like this guy is doing)
There's also been a few people building monocopters with a fixed, non tilting motor and prop, and four servo motor actuated fins at the bottom that can be used for thrust vectoring and to counter rotational forces.
https://www.youtube.com/watch?v=u2cETOyuJ20
That's pretty close to what I built only I'm using a far more efficient motor/prop setup so much longer flying times as well as the option to transition to level flight.
Neat. I wonder if you could use the air it generates by having directional ports louvers that open and allow the air to flow for more control. Would need 2 more servo, left/right port louvers.
I love this angle so much! Do you know if there's been more detailed analysis of performance?
Very good control, low flying time (those ducted fans use an insane amount of energy like this).
Generally those small diameter motor/prop combinations aren't very efficient, in grams/watt of thrust. They look really bad if you compare them to like a 15" size carbon fiber prop from t-motor and one of the tmotor "antigravity" series motors on a thrust stand.
Quadcopters that are designed for 45 min to 1 hr loiter time carrying a useful payload, while not being too big for a person to move around seem to be converting on 15 to 18 inch size prop and motor and esc combinations carefully chosen for optimal grams/watt thrust.
Yes, that's in line with everything I've seen.
I've built a whole slew of drones over the last couple of years, quadcopters are hard to beat on many axis. But I have some proof that it may be possible to do substantially better with a monocopter in terms of efficiency. The hard problems you will then have to solve are payload attachment because unbalance is very tricky to deal with (whereas that is trivial in a quadcopter) as well as transitioning to level flight so that you can make good forward speed.
I think the recent winner of the DARPA life challenge was a 'traditional' single rotor helicopter with tailrotor. The very large disc area of the traditional heli design made it more efficient than quad/hex/octo designs.
https://www.google.com/search?client=firefox-b-d&q=darpa+lif...
Yes, but their forward speed is limited because they can't translate into level flight and that also means the range is limited. I was trying to get the best of both worlds (but that's a hard problem to solve).
I know very little of the field, but aren't there applications where the ability to carry ad hoc payloads doesn't matter? Integrating a fixed sensor suite into the ring and balancing it ahead of time is possible then, right?
And what can be learned from full-scale classical helicopters that frequently carry underhanging payloads?
(I don't at all wanna be the typical HN person who thinks their questions about a field they know nothing about are in any way insightful. I'm just trying to learn.)
That's all true. But (1) that's not the application that I have in mind and (2) forward speed of helicopters is not fantastic (3) neither is their fuel efficiency.
For a fixed payload it is indeed possible to set up the balance beforehand, even then you will probably end up with a drone that will translate slower or faster depending on the inbalance instead of hovering in place.
It's incredible that changes in velocity still work to adjust the angle of the blades even at the insane speeds they're moving at. I would never have thought that was possible.
I have to wonder why this guy didn't give his drone a leash while testing so it couldn't go flying off. It's the first thing that popped into my head, so I'm sure it's something he thought of as well and didn't use for a good reason. Maybe a string would get caught in the blades?
This one is very impressive too:
https://www.youtube.com/watch?v=K0POjjIC5xs
It does use a bunch of external sensors but the control is all with just the prop and rapid speed changes, they don't even use a variable pitch prop.
Maybe the pull from a string/leash would affect how the thing flies, kind of like kites.
I like how he built a test stand near 7:22, which solves the motor flying off problem and is probably easier to iterate on. It reminds me of how rockets are tested.
Possibly the fact that the TPU fin body doesn't stop all counterrotation so the leash would get twisted into a cord and interfere with flight?
print a slipring
Open source ?
I’m sure glad we are making immense progress towards the dystopian hell where massive swarms of mini drones are hunting down obsolete and marked for termination flesh bots, aka humans.
oh yeah? leonardo davinci built one with zero motors