This is very impressive! I researched fault-tolerant octorotor control using RL in grad school for a NASA project. Perhaps this may be helpful[1, see section 8.3]! The field is moving fast, so there may be better or more suitable approaches out there now.
For folks who are interested in UAV physics, I wrote up an explainer[2].
You know you're doing a great job, OP, when the peanut gallery here has nothing more substantial to add than to critique your em-dashes; greek-latin root word mix-ups despite the common vernacular having moved on from that; and lack of title brevity.
Congratulations --- this is a super cool project. I wonder if you've considered using ultralight filaments and 3dprinting the frame? PLA is stiff but brittle, and I know Bambu and a few others sell specialised versions that supposedly weigh less than normal.
Most filament based printed frames end up with really nasty resonance; it’s possible to engineer damping around the issue with some clever 3D design if the parameters of the prints are measured, but overall 3D printing copter frames doesn’t tend to be a straightforward solution.
The milled fiberglass the author used is a much better UAS frame material than anything from a filament 3d printer due to stiffness and related considerations.
Something I've wondered for octocopters - could using a ring instead of arms be beneficial for weight? 6.28r < 8r, but then again the arm radius is usually less than the full circle, and some components want to be centrally located, etc. I could imagine holding the central components in tension via light filaments (carbon fiber, nylon, etc) in tension, vs having to have rigid structure, but the small factor between 6.28 and 8 and maybe makes it not worth it.
large schedule 40 or 80 tubing sliced into rings would be pretty quick source material, starting with duct tape and zipties until you find a good arrangment then get into the glue and screws.
Low end and most open source stuff will use a PID inner loop for “fast path” control (stabilization) and either a second PID loop or something a little better (Kalman filter etc) for the slow path (position / path hold).
Higher end stuff will use a ton of inputs (visual odometry, binocular vision, lidar, range finding, etc) fused into some kind of proprietary blended algorithm that you could probably call an MPC.
RL is pretty cutting edge, especially for fast path motor control; there are a lot of university competitions for drone control that lead to a lot of papers and projects in the space (some promising) but most commercial stuff has not adopted this yet, certainly not at the low end.
Phrased in a slightly different way, assuming the standard is Betaflight/Ardupilot/PX4. (This article uses BF): Inner PID rate loop; this compares IMU-measured rates vs that commanded by manual controls. This might run at 1-4kHz.
On top of this (Maybe at a few hundred hz), you can add outer controls to set attitude. This could be an autopilot, or having the controls command attitude instead of rate. Betaflight pilots usually don't both with this, and have the simple setup of control maps to rate.
I've programmed firmware using a weird hybrid where the controls command a change in the target attitude. So it flies like rate, but has the forced attitude stability of an attitude-based control system. Non-standard, but makes it so you don't need to worry as much about tuning the PID loop. In practice, you can do full aerobatic flight with this like you'd do with a rate-only setup. (Basically, there is a commanded attitude quaternion; controls nudge it; the PIDs update motor power to maintain this commanded quaternion.)
PID is more than enough to keep level. FPV relies on manual flight, but you can get Ardupilot for autonomous missions. There's no need for RL, nothing to gain here; level flight and following waypoints is a solved issue already.
And frankly as a pilot, I'd rather not see any completely autonomous drones with no oversight in the sky - that's one incident away in which blame cannot be put solely on the operator from getting the hobby completely banned.
As a researcher in the autonomous robotics space, there is a lot to gain in RL over PIDs and manual flight.
The delta between what is possible with current autonomous flight missions and manual FPV style flight is by having a brain on board that can dynamically adapt to a changing environment. There are a finite amount of PID profiles for each steadystate solution that a researcher can preprepare for. But RL allows an overarching heuristic to transiently alter the PIDs depending on the changing environment.
We use PIDs because analyzing robotics platforms as seeking a steadystate dramatically simplifies the math needed to where its computationally possible for us to solve for a situation.
We use RL in systems that have continuously changing environments with transient solution spaces that are easier to model in hyperspace with a RL model.
Take for example platforms that have tiltrotors. They ideally have a minimum of 3 PID profiles for flying. One when it best fits a multirotor profile. A second when it is transitioning from multirotor to fixed wing flight, and a third for when fixed wing flight is established. What happens when the researcher has a need to fly in the transition state, or subconfigurations of the states?
How many PID profiles are you looking to think of and train for?
This is where RL has dividends.
Amazing: I'm watching lots of homemade builds atm.
I've got a question: why CNC milling and not just FDM 3D printed parts? TFA doesn't talk much about it except saying she went to a machine shop.
> 2. CNC milling forms out of G-10 fiberglass (arms) and 5mm carbon fiber (body)
TFA also says this:
> The solution for this is to 3D print a 0-tolerance assembly jig to hold the arms in perfect position while the center of the drone is superglued together.
Why not 3D print it all?
There's this guy who built a drone that can fly for 3 hours and cover hundreds of miles, 3D printed at home on a $250 printer:
Then there was the $200 K quote for the body for a car that just did Pike's Peak with a four times Pike's Peak champion and instead the team... 3D-printed the car's body at home (something like 40 parts, assembled together), which cost them less than $2 K to make (1/100th of the quoted price for the car's body). Here's the vid where they print all the parts (on a $1500 consumer printer):
Basically: why CNC milling and not 3D printing at home when many drones enthusiasts (and now too people building race cars) simply print parts at home on a consumer-grade 3D printer?
For "copter" style drones, 3D printing is basically a "noob" meme at this point; everyone thinks it's a good idea, tries it, and realizes that it's actually really hard and doesn't work.
Copter-style drones are exposed to vibration across a huge frequency range in every axis, and it's almost impossible to avoid really nasty resonance issues using generally-printable FDM filaments and "standard" design techniques; it's a lot easier to just use super-stiff carbon fiber and CNC it.
For planes, like what you linked, 3D printing is more "plausible" than for copters but also not really practical; you can 3D print a good plane, but plastic lacks the durability and favorable weight characteristics of foam - plastic planes tend to be "one time crashed" while foam is easy to repair, restore, and rebuild.
Love it, great to explore and learn, and I like the mixed background too of cybersec and robotics (all the way to CNC), just like myself, I think these two new fields will make a new industry, similar to OT cyber but more niche.
Which also means great people can go beyond what’s their school was about, so a CS major doing CNC isn’t “weird” or different, I remember when applying for jobs in systems in aerospace industry and get rejected despite having a systems background too, with feedback of “they are looking for people with education only in aerospace”, which is idiotic thing to consider.
So good luck OP, start exploring hacking mavlink or similar protocols which is what im working on.
Might be intended to preemptively deflect criticism of "reinventing the wheel"/solving subproblems in a non-standard/convoluted way.
I'd expect an engineering project with "no prior experience" to take weird/experimental approaches more often compared to a "from scratch" project (where I would expect proven minimalism instead).
Hm making an AI assisted page and replacing the emdashes with double dashes seems like more work than to just rewrite the text yourself. Not sure why you would do that.
What? That’s a simple find and replace vs rewriting the whole thing. If someone had the savvy to write the thing, they probably wouldn't have been using the the assistant in the first place. Either way, comparing a find/replace to rewriting is farcical
The name "octocopter" does not make sense. "Helicopter" is a compound word made of "helico-" and "pter", which means "screw-wings". "Octo-" means eight, "-co-" means nothing.
"Octopter" would be a correct compound word meaning "8-wings", but that would be ambiguous, so the object discussed in TFA is better named just "8-propeller drone".
That ship has long sailed. You’re correct, but the author isn’t the one who “named the thing” in this case, they are just using the name commonly used to describe it.
Multi-rotor drones have been called tricopters, quadcopters, hexacopters, octocopters based on their propeller counts conversationally for as long as I can remember.
There are plenty of commercial vendors who use the exact term for their expensive industrial drones.
Update: I see that in the four minutes it took for me to validate my initial inclination and post that plenty of others also had the same thought :) No need to me to belabor the point!
This is quite a common linguistic phonomenon, where a word is rebracketed to form a new suffix, even if it doesn't make sense with the original etymology. See also -holic (alcoholic -> workaholic), -thon (marathon -> danceathon) or -gate (Watergate -> partygate). Termed a "libfix" from liberated affix
Nit pick: "nit pick" means to remove tiny bugs from hair, which this is not.
Oh, language changes and now "nit pick" means "to make trivial criticisms" even though neither "nit" nor "pick" etymologically has anything to do with criticisms? How very self-serving of you ;)
Blame language evolving over time rather than OP, octocopter is a widely-used term for '8 propellor drones'.
A nit pick with your post - you use the word 'ambiguous' but really this is from the latin root 'ambiguus' so we don't need the supurflous 'o' in between the two u's.
Well I was confused by it! I was expecting an article on amateur semiconductor fabrication. Granted, that was due to my misreading it as 'optocoupler'.
Octocopter makes perfect sense. Everyone understands immediately what it means, and that's the only purpose of language: to convey ideas. It should be clear, which this is, and concise, which this is.
Fidelity to ancient Greek is not, and should not, be a goal for English.
This is very impressive! I researched fault-tolerant octorotor control using RL in grad school for a NASA project. Perhaps this may be helpful[1, see section 8.3]! The field is moving fast, so there may be better or more suitable approaches out there now.
For folks who are interested in UAV physics, I wrote up an explainer[2].
[1]: https://drive.google.com/file/d/1RTEVRd0XCWLuDXY2nkbmYuOaa5x...
[2]: https://iahmed.me/post/drone-physics/
You know you're doing a great job, OP, when the peanut gallery here has nothing more substantial to add than to critique your em-dashes; greek-latin root word mix-ups despite the common vernacular having moved on from that; and lack of title brevity.
Congratulations --- this is a super cool project. I wonder if you've considered using ultralight filaments and 3dprinting the frame? PLA is stiff but brittle, and I know Bambu and a few others sell specialised versions that supposedly weigh less than normal.
Most filament based printed frames end up with really nasty resonance; it’s possible to engineer damping around the issue with some clever 3D design if the parameters of the prints are measured, but overall 3D printing copter frames doesn’t tend to be a straightforward solution.
I wish we had some standard filament testing that most manufacturers were willing to provide results for.
Until that happens, this guy here is probably the next best thing: https://www.youtube.com/@MyTechFun
Plot twist: many of the "special" filaments aren't special at all or at least very exaggerated.
Yes. And when your design is simply beautiful as this is:
https://karolina.mgdubiel.com/drone/drone-img/05-30-26/cnc_c...
I agree with your first point.
The milled fiberglass the author used is a much better UAS frame material than anything from a filament 3d printer due to stiffness and related considerations.
Oh no doubt. I'm no drone expert!
Something I've wondered for octocopters - could using a ring instead of arms be beneficial for weight? 6.28r < 8r, but then again the arm radius is usually less than the full circle, and some components want to be centrally located, etc. I could imagine holding the central components in tension via light filaments (carbon fiber, nylon, etc) in tension, vs having to have rigid structure, but the small factor between 6.28 and 8 and maybe makes it not worth it.
consider "tiled" hexagons, like honycomb.
large schedule 40 or 80 tubing sliced into rings would be pretty quick source material, starting with duct tape and zipties until you find a good arrangment then get into the glue and screws.
People are so jealous. This is cool as hell.
Something happened recently that attracted a whole bunch of quite ignorant and frustrated people to this site.
So many activists of different kinds.
It's cool, but I could have done it 2.45 weeks!
This is cool as sugar! I have to ask though, how many end mills did you go through milling G10 fiberglass and carbon fiber‽
I've heard the dust from carbon fiber is second to asbestos for inhaling.
Kudos for such a great learning journey!
I love everything about this. Well done! but missed opportunity to name it R_of_L-copter.
Will follow a fellow Polish inventor! :)
I do not notice that the time of posts is reversed haha. I am confused whether you had build it.
Thank you, it's cool!
Super cooool
If I were to get a dirt cheap Chinese drone, would that be more likely to use RL or MCP? What’s the “standard”?
Low end and most open source stuff will use a PID inner loop for “fast path” control (stabilization) and either a second PID loop or something a little better (Kalman filter etc) for the slow path (position / path hold).
Higher end stuff will use a ton of inputs (visual odometry, binocular vision, lidar, range finding, etc) fused into some kind of proprietary blended algorithm that you could probably call an MPC.
RL is pretty cutting edge, especially for fast path motor control; there are a lot of university competitions for drone control that lead to a lot of papers and projects in the space (some promising) but most commercial stuff has not adopted this yet, certainly not at the low end.
Phrased in a slightly different way, assuming the standard is Betaflight/Ardupilot/PX4. (This article uses BF): Inner PID rate loop; this compares IMU-measured rates vs that commanded by manual controls. This might run at 1-4kHz.
On top of this (Maybe at a few hundred hz), you can add outer controls to set attitude. This could be an autopilot, or having the controls command attitude instead of rate. Betaflight pilots usually don't both with this, and have the simple setup of control maps to rate.
I've programmed firmware using a weird hybrid where the controls command a change in the target attitude. So it flies like rate, but has the forced attitude stability of an attitude-based control system. Non-standard, but makes it so you don't need to worry as much about tuning the PID loop. In practice, you can do full aerobatic flight with this like you'd do with a rate-only setup. (Basically, there is a commanded attitude quaternion; controls nudge it; the PIDs update motor power to maintain this commanded quaternion.)
PID is more than enough to keep level. FPV relies on manual flight, but you can get Ardupilot for autonomous missions. There's no need for RL, nothing to gain here; level flight and following waypoints is a solved issue already.
And frankly as a pilot, I'd rather not see any completely autonomous drones with no oversight in the sky - that's one incident away in which blame cannot be put solely on the operator from getting the hobby completely banned.
As a researcher in the autonomous robotics space, there is a lot to gain in RL over PIDs and manual flight.
The delta between what is possible with current autonomous flight missions and manual FPV style flight is by having a brain on board that can dynamically adapt to a changing environment. There are a finite amount of PID profiles for each steadystate solution that a researcher can preprepare for. But RL allows an overarching heuristic to transiently alter the PIDs depending on the changing environment.
We use PIDs because analyzing robotics platforms as seeking a steadystate dramatically simplifies the math needed to where its computationally possible for us to solve for a situation.
We use RL in systems that have continuously changing environments with transient solution spaces that are easier to model in hyperspace with a RL model.
Take for example platforms that have tiltrotors. They ideally have a minimum of 3 PID profiles for flying. One when it best fits a multirotor profile. A second when it is transitioning from multirotor to fixed wing flight, and a third for when fixed wing flight is established. What happens when the researcher has a need to fly in the transition state, or subconfigurations of the states? How many PID profiles are you looking to think of and train for? This is where RL has dividends.
Interesting - thanks! OP's drone IS using RL and that's what jumped out at me - it felt a bit overkill for the usecase.
Amazing: I'm watching lots of homemade builds atm.
I've got a question: why CNC milling and not just FDM 3D printed parts? TFA doesn't talk much about it except saying she went to a machine shop.
> 2. CNC milling forms out of G-10 fiberglass (arms) and 5mm carbon fiber (body)
TFA also says this:
> The solution for this is to 3D print a 0-tolerance assembly jig to hold the arms in perfect position while the center of the drone is superglued together.
Why not 3D print it all?
There's this guy who built a drone that can fly for 3 hours and cover hundreds of miles, 3D printed at home on a $250 printer:
https://youtu.be/e7AIKGDrlgs
Then there was the $200 K quote for the body for a car that just did Pike's Peak with a four times Pike's Peak champion and instead the team... 3D-printed the car's body at home (something like 40 parts, assembled together), which cost them less than $2 K to make (1/100th of the quoted price for the car's body). Here's the vid where they print all the parts (on a $1500 consumer printer):
https://youtu.be/nt85nTMnY1w
Basically: why CNC milling and not 3D printing at home when many drones enthusiasts (and now too people building race cars) simply print parts at home on a consumer-grade 3D printer?
For "copter" style drones, 3D printing is basically a "noob" meme at this point; everyone thinks it's a good idea, tries it, and realizes that it's actually really hard and doesn't work.
Copter-style drones are exposed to vibration across a huge frequency range in every axis, and it's almost impossible to avoid really nasty resonance issues using generally-printable FDM filaments and "standard" design techniques; it's a lot easier to just use super-stiff carbon fiber and CNC it.
For planes, like what you linked, 3D printing is more "plausible" than for copters but also not really practical; you can 3D print a good plane, but plastic lacks the durability and favorable weight characteristics of foam - plastic planes tend to be "one time crashed" while foam is easy to repair, restore, and rebuild.
Stiffness. Important for how the frame behaves under load and buffetting etc.
Love it, great to explore and learn, and I like the mixed background too of cybersec and robotics (all the way to CNC), just like myself, I think these two new fields will make a new industry, similar to OT cyber but more niche.
Which also means great people can go beyond what’s their school was about, so a CS major doing CNC isn’t “weird” or different, I remember when applying for jobs in systems in aerospace industry and get rejected despite having a systems background too, with feedback of “they are looking for people with education only in aerospace”, which is idiotic thing to consider.
So good luck OP, start exploring hacking mavlink or similar protocols which is what im working on.
Why not just say from scratch instead of no prior experience, is it to brag
Might be intended to preemptively deflect criticism of "reinventing the wheel"/solving subproblems in a non-standard/convoluted way.
I'd expect an engineering project with "no prior experience" to take weird/experimental approaches more often compared to a "from scratch" project (where I would expect proven minimalism instead).
"from scratch" and "no prior experience" have very little in common.
Hm making an AI assisted page and replacing the emdashes with double dashes seems like more work than to just rewrite the text yourself. Not sure why you would do that.
Tbh this is cooler than anything on your github so he (edit: she) can do what he wants IMO
The abstract certainly smells like 100% LLM-generated text.
What? That’s a simple find and replace vs rewriting the whole thing. If someone had the savvy to write the thing, they probably wouldn't have been using the the assistant in the first place. Either way, comparing a find/replace to rewriting is farcical
> more work than to just rewrite the text yourself
You think s/—/--/g is more work than rewriting a whole article? Is this what you're saying?
Nit pick:
The name "octocopter" does not make sense. "Helicopter" is a compound word made of "helico-" and "pter", which means "screw-wings". "Octo-" means eight, "-co-" means nothing.
"Octopter" would be a correct compound word meaning "8-wings", but that would be ambiguous, so the object discussed in TFA is better named just "8-propeller drone".
That ship has long sailed. You’re correct, but the author isn’t the one who “named the thing” in this case, they are just using the name commonly used to describe it.
Multi-rotor drones have been called tricopters, quadcopters, hexacopters, octocopters based on their propeller counts conversationally for as long as I can remember.
There are plenty of commercial vendors who use the exact term for their expensive industrial drones.
Update: I see that in the four minutes it took for me to validate my initial inclination and post that plenty of others also had the same thought :) No need to me to belabor the point!
This is quite a common linguistic phonomenon, where a word is rebracketed to form a new suffix, even if it doesn't make sense with the original etymology. See also -holic (alcoholic -> workaholic), -thon (marathon -> danceathon) or -gate (Watergate -> partygate). Termed a "libfix" from liberated affix
The -copter suffix is very common in the drone community.eg quadcopter is widely accepted https://en.wikipedia.org/wiki/Quadcopter
Similarly, "heli" is a commonly recognized clipping of helicopter: https://en.wiktionary.org/wiki/heli#English
See e.g. https://en.wikipedia.org/wiki/Helipad
Are you trying to say that it’s been co-opted? Did anyone consult the Egyptian Christian community about this?
Hence a nit.
Nit pick: "nit pick" means to remove tiny bugs from hair, which this is not.
Oh, language changes and now "nit pick" means "to make trivial criticisms" even though neither "nit" nor "pick" etymologically has anything to do with criticisms? How very self-serving of you ;)
It's a metaphor:)
Unrelated.
McDonald’s getting a strongly worded letter from the Mayor of Hamburg over their use of “cheeseburger”.
Blame language evolving over time rather than OP, octocopter is a widely-used term for '8 propellor drones'.
A nit pick with your post - you use the word 'ambiguous' but really this is from the latin root 'ambiguus' so we don't need the supurflous 'o' in between the two u's.
Well I was confused by it! I was expecting an article on amateur semiconductor fabrication. Granted, that was due to my misreading it as 'optocoupler'.
"Copter" is a known word, short for helicopter.
https://dictionary.cambridge.org/dictionary/english/copter
On a related note, pronunciation variance in "Helicopter" -> "Helacopter" -> "Helocopter" leads to a confusing abbreviation - "Helee" vs "Heelo"
In this fundamental paper, the authors argue for multirotor instead of -copter. In academia, this term seems to have stuck.
https://ieeexplore.ieee.org/document/6289431
Counterpoint: -copter is a perfectly cromulent suffix. https://en.wiktionary.org/wiki/-copter
gyrocopter, helicopter, quadcopter, hexacopter, octocopter, parcelcopter, and—most famously—
roflcopter, https://en.wiktionary.org/wiki/roflcopter#/media/File:Roflco...
They all have their own dictionary entries.
Octocopter makes perfect sense. Everyone understands immediately what it means, and that's the only purpose of language: to convey ideas. It should be clear, which this is, and concise, which this is.
Fidelity to ancient Greek is not, and should not, be a goal for English.
Great examples. The English lexicon is continuously embiggened by the adoption and expansion of terms.
I guess it's a play on the popular word "quadcopter", rather than "helicopter".