@@NicholasRehm - I'd love to see you do something with plasma actuators and ducted fans. There are plenty of RC "jet" models out there that use ducted fans in place of turbines. But I'd always heard that ducts create increased drag problems. Meanwhile, I'd also read that plasma actuators which ionize the air can improve the flow characteristics and reduce drag. I'd like to know if it's possible to put the plasma actuators around the duct inlets to reduce the drag and improve the performance of the aircraft.
Now that is spectacular engineering 👏👏 I do fishing drones, for years we have known an encouraged people to use longer pendulums, (suspending the loads further down) basically just to slow the pendulum down enough for the standard off-the-shelf GPS position guided drone to cope better with the swinging mass. The two masses in motion can cause some serious problems especially when the suspended mass is on a short/fast pendulum and it’s mass big in relation to the quad’s own mass. Seriously well done on your projects 👏👏
@@NicholasRehm give it a try, it has been gaining mass popularity since inception in 2013/4, I currently have around 30k guys around the globe fishing with Gannets. Applications will keep growing, smart guys like you have endless opportunities to contribute to the advancement in this new world in flight. Drones are now where mobile phones were around the 2000’s, look what we can now do with a cellphone, then think what drones will be able to do in 2040?
Wow, how could anyone not like this video!?! So interesting, and I'm jealous of your engineering skills. Next up, a dual pendulum drone...?(lol) Seriously, great work, thanks!
I decide to take a quick break from my control systems lab assignment and end up watching a video on an amazing control system. I can't get away from this stuff man.
This is a challenge I would like to try. I've seen examples of using a vision system to control it, but this allows for much more freedom of movement. I enjoyed the video and I have a desire to try it someday.
I'd love to see you do something with plasma actuators and ducted fans. There are plenty of RC "jet" models out there that use ducted fans in place of turbines. But I'd always heard that ducts create increased drag problems. Meanwhile, I'd also read that plasma actuators which ionize the air can improve the flow characteristics and reduce drag. I'd like to know if it's possible to put the plasma actuators around the duct inlets to reduce the drag and improve the performance of the aircraft.
@@NicholasRehm - awesome! FYI, there's a real light sport aircraft that's recently come out called the *PJ-2 Dreamer.* It's styled to look like a real jet fighter, but it uses a piston engine to power twin ducted fans, in order to make it more affordable to operate, and also to comply with LSA restrictions on engines. But its speed tops out at 200mph due to the drag on the flow inside the ducts: ua-cam.com/video/OkyUr-3ZtQ8/v-deo.html
Rafaello d'andrea (cofounder of Amazon Robotics, Professor at ETH Zurich) showed similar acrobatic experiments with drone in a ted talk around 10 years ago
@Nicholas Rehm, I had a challenge but I am now realizing the double pendulum pid stabilized videos I've been seeing are not three dimensional. I was going to challenge you to get double pendy with it... but that's a lot of maths when the three dimensional double pendulum cannot freely swing through itself in reality (i cannot immediately think of a way); if we did have access to such a device that would be my challenge. Double Pendulum in all the axis's flying around scaring folk and crashing into flowerpots after a totally working for a good while. Hey, thanks for sharing; this was quality content.
As long as it stays relatively vertical and doesn't swing down like you mentioned, the double pendulum on a drone is absolutely do-able. My ability to get a controller working well enough is a different story haha
@@NicholasRehm are signal IO'ing with pwm, serial, other? There are some protocols that could provide far faster than pwm/serial bus signaling. Thank you for replying to my double pendy suggestion.
@@NicholasRehm as in the "tune" of the quad's flight controller? I... dare I say, I specialize in betaflight tuning from 80 to 800 grams without blackbox and obviously benefit from the non-heuristic approach, using black box logs... A good pattern for pure logists, have someone you trust throw your rig like a frisbee and you pretend your in Space Camp the movie and do your best (it's easier done than articulated) to regain control and then do normal flight patterns and the grocery list of isolated movements by way of your partner signal audibly for each change to the next movement based on timing; these would be reviewed as a time series set to evaluate for the various (scripted) movement event times. I prefer the heuristic approach personally. As an engineer, I will argue that floating point maths are best codified and abstracted for quick changes with unknown variations on the over all theme, the tune, so to speak. Am I way off?
@@slowvoltage more so a problem of compounding error and reduced responsiveness as you go down the line of controlled variables from drone rate -> drone angle -> lower pendulum angle -> upper pendulum angle
Another question: How do you tune your control algorithm for the inverted pendulum? Are you using Simulink, after modeling the drone? Do you maybe have some hints in this direction, and maybe some code?
Trial and error & making sure each term in the PID controller is correcting in the right direction and at an approximately right magnitude that lets me go for a test flight. Then more refining PID gains to try to tease out more controller tracking performance
Great channel, great projects. Did you consider using the Arduino Nano RP 2040? It has a fast processor and the IMU included on board ... looks like a perfect platform for these projects. And a GPS module too, to top it up...
Hello, I'd like to build a drone and make the same experiments you did in this video. Could you please post the drone specs? I'd like to buy the same motors and controller. Are these the same you post on your site in the "store" section?
@@NicholasRehm The drone aperture is 250mm? Do you have a name/reference? The motors are iFlight XING-E Pro 2207 2450KV 4S ? Would a 6ch radio like the fs-i6 be enough for this experiment (the price differential is quite significant). For the remaining, I hope I can figure it out.
Have you tried porting the flight controller to less powerful HW, such as the f405? Is it too limited (almost 4 times lower clock frequency than the Teensy v4, and less powerful architecture - Cortex M4 vs M7). I'm pretty interested in having a hackable FC where I would implement my algorithms. In your presentation "How the Flight Controller Code Works" you say the control loop runs at 2kHz. Do you know how much free time you have per loop iteration? That is, how much function loopRate takes of the 500usec loop iteration? Finally: I assume the Arduino interface brings a lot of comfort. Do you know how much it penalizes performance? I see that Betaflight runs its PID at 4k on such HW (8k may be too demanding), but I assume input integration is run at a lower frequency, whereas here everything runs at the same frequecy.
There’s about 150 out of 500us overhead free in the 2khz loop. The whole thing about dRehmFlight with the teensy is that it’s just a simple arduino project, so yea there’s some performance & feature sacrifices compared to betaflight, but it’s still “good enough” and offers the benefits of easy integration of new hacky stuff. Hardcore drone racing guys would hate how ‘sloppy’ dRehmFlight feels. But for fun little VTOL projects where performance doesn’t matter as much, it’s more than enough
Hi Nicholas, I was just getting involved in model rocketry and remote controled aircraft when our airfield was closed because somebody felt a need to build another hospital. I suspect others have this trouble so I wondered if you have any tips to share to help your viewers find a place to indulge in this hobby or what restrictive laws we might encounter in our own neighborhoods.
My best advice is to always have a kill switch for safety, and consider looking into smaller aircraft that you can fly below the tree line-if ever confronted about where you’re flying, your best defense is “I’m sorry, I didn’t know” Happy flying!
Your knowledge about code and programming is so far beyond me, it's laughable. It would take me a year to comprehend what you so casually understand so easily.
![Inverted Pendulum on a Cart [Control Bootcamp]](/img/n.gif)
6:55 that sound is so cute, it melts my heart.
5:40 I like how the drone immediately went for peter. Terminator pendulum mode engaged!
Might’ve been intentional ;)
Your results are incredible! I was fully expecting something a lot less stable.
I was surprised too haha
Its just as stable on top as it is on bottom. Its called the rocket fallacy when you think other wise.
@@NicholasRehm - I'd love to see you do something with plasma actuators and ducted fans. There are plenty of RC "jet" models out there that use ducted fans in place of turbines. But I'd always heard that ducts create increased drag problems. Meanwhile, I'd also read that plasma actuators which ionize the air can improve the flow characteristics and reduce drag. I'd like to know if it's possible to put the plasma actuators around the duct inlets to reduce the drag and improve the performance of the aircraft.
Now that is spectacular engineering 👏👏
I do fishing drones, for years we have known an encouraged people to use longer pendulums, (suspending the loads further down) basically just to slow the pendulum down enough for the standard off-the-shelf GPS position guided drone to cope better with the swinging mass.
The two masses in motion can cause some serious problems especially when the suspended mass is on a short/fast pendulum and it’s mass big in relation to the quad’s own mass.
Seriously well done on your projects 👏👏
Thanks for the comment! Drone fishing sounds like a fun retirement plan down the road...
@@NicholasRehm give it a try, it has been gaining mass popularity since inception in 2013/4, I currently have around 30k guys around the globe fishing with Gannets. Applications will keep growing, smart guys like you have endless opportunities to contribute to the advancement in this new world in flight.
Drones are now where mobile phones were around the 2000’s, look what we can now do with a cellphone, then think what drones will be able to do in 2040?
Wow, how could anyone not like this video!?! So interesting, and I'm jealous of your engineering skills. Next up, a dual pendulum drone...?(lol) Seriously, great work, thanks!
Really appreciate your compliments. Dual pendulum would be a heck of a challenge!
@@NicholasRehm did you fix the problem?
I know where the problem is...
Great video 👍👍👍
Being a drone pilot myself (camera) I'm not talented enough for fpv, but you Sir take it to the NEXT LEVEL....well done.
This channel will go places!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Nice work!
I decide to take a quick break from my control systems lab assignment and end up watching a video on an amazing control system. I can't get away from this stuff man.
Keep those poles in the left hand plane my dude
Great job! I really admire your working/research style and skill. Thank you for sharing this project with us!
Well edited and what a great video keep it up Nick
That was great. Thanks for putting it up
This is a challenge I would like to try. I've seen examples of using a vision system to control it, but this allows for much more freedom of movement. I enjoyed the video and I have a desire to try it someday.
No idea what you’re saying but it’s awesome to see!
This is ingenious
I loved how you used the joystick to get the angle of the stick 👏
Gosh the little *zip* the props make off the grass sound like a tiny adorable ouch from the drone
The control system built are amazing, great flight, I am student learning control systems this was cool demonstration 🔥🔥🔥
Best of luck to you and your studies. Controls is tough but you got this
@@NicholasRehm Thanks
I never saw theese type of experiment...so good... interesting to watch ✨
Great video. Awesome topic 👌🏼
Amazing. Reminds me of Grad School doing controls engineering on Matlab
Very cool. Thanks!
Brilliant!
WOW, fantastic video, and you’ve got a fantastic brain to match👍
Really appreciate your compliments. Cheers
You have to put a 360 cam on top of the Stick.😉👍
Well good. It'd make a great camera mount..
Awesome
This is fabulous.
I'd love to see you do something with plasma actuators and ducted fans. There are plenty of RC "jet" models out there that use ducted fans in place of turbines. But I'd always heard that ducts create increased drag problems. Meanwhile, I'd also read that plasma actuators which ionize the air can improve the flow characteristics and reduce drag. I'd like to know if it's possible to put the plasma actuators around the duct inlets to reduce the drag and improve the performance of the aircraft.
That sounds absolutely insane. I’m in.
@@NicholasRehm - awesome! FYI, there's a real light sport aircraft that's recently come out called the *PJ-2 Dreamer.* It's styled to look like a real jet fighter, but it uses a piston engine to power twin ducted fans, in order to make it more affordable to operate, and also to comply with LSA restrictions on engines. But its speed tops out at 200mph due to the drag on the flow inside the ducts: ua-cam.com/video/OkyUr-3ZtQ8/v-deo.html
Rafaello d'andrea (cofounder of Amazon Robotics, Professor at ETH Zurich) showed similar acrobatic experiments with drone in a ted talk around 10 years ago
They used motion capture cameras
@Nicholas Rehm, I had a challenge but I am now realizing the double pendulum pid stabilized videos I've been seeing are not three dimensional. I was going to challenge you to get double pendy with it... but that's a lot of maths when the three dimensional double pendulum cannot freely swing through itself in reality (i cannot immediately think of a way); if we did have access to such a device that would be my challenge. Double Pendulum in all the axis's flying around scaring folk and crashing into flowerpots after a totally working for a good while. Hey, thanks for sharing; this was quality content.
As long as it stays relatively vertical and doesn't swing down like you mentioned, the double pendulum on a drone is absolutely do-able. My ability to get a controller working well enough is a different story haha
@@NicholasRehm are signal IO'ing with pwm, serial, other? There are some protocols that could provide far faster than pwm/serial bus signaling. Thank you for replying to my double pendy suggestion.
@@slowvoltage Its not so much an issue of computation/communications speeds, but precision of the drone controller in general
@@NicholasRehm as in the "tune" of the quad's flight controller? I... dare I say, I specialize in betaflight tuning from 80 to 800 grams without blackbox and obviously benefit from the non-heuristic approach, using black box logs... A good pattern for pure logists, have someone you trust throw your rig like a frisbee and you pretend your in Space Camp the movie and do your best (it's easier done than articulated) to regain control and then do normal flight patterns and the grocery list of isolated movements by way of your partner signal audibly for each change to the next movement based on timing; these would be reviewed as a time series set to evaluate for the various (scripted) movement event times. I prefer the heuristic approach personally.
As an engineer, I will argue that floating point maths are best codified and abstracted for quick changes with unknown variations on the over all theme, the tune, so to speak. Am I way off?
@@slowvoltage more so a problem of compounding error and reduced responsiveness as you go down the line of controlled variables from drone rate -> drone angle -> lower pendulum angle -> upper pendulum angle
Pretty cool.
Do a comparison of lift with and without ducts around the prop blades.
Smart
Top ..💥👏👏👏👏👏
What about trying to fix the tip of the pendulum and flying drone to mimic real pendulum motion in air with invisible pivot?? How hard is that?
Another question: How do you tune your control algorithm for the inverted pendulum? Are you using Simulink, after modeling the drone? Do you maybe have some hints in this direction, and maybe some code?
Trial and error & making sure each term in the PID controller is correcting in the right direction and at an approximately right magnitude that lets me go for a test flight. Then more refining PID gains to try to tease out more controller tracking performance
Aww the drone gave a little squeak when it hit the ground 😅
He deserves better than the abuse I put him through
Have seen a drone balance a pole on a pole and it could even flip the pole on the pole , on tedx
I would have expected the joystick potentiometers to give better measurements than an MPU 6050....interesting
If they were fully linear and calibrated properly, then maybe. Too much work to do that :)
Great channel, great projects. Did you consider using the Arduino Nano RP 2040? It has a fast processor and the IMU included on board ... looks like a perfect platform for these projects. And a GPS module too, to top it up...
Specs look good minus i/o capabilities...not so god for lots of different sensors and RC equipment
Hello, I'd like to build a drone and make the same experiments you did in this video. Could you please post the drone specs? I'd like to buy the same motors and controller. Are these the same you post on your site in the "store" section?
They are basically the same!
@@NicholasRehm The drone aperture is 250mm? Do you have a name/reference? The motors are iFlight XING-E Pro 2207 2450KV 4S ? Would a 6ch radio like the fs-i6 be enough for this experiment (the price differential is quite significant). For the remaining, I hope I can figure it out.
Have you tried porting the flight controller to less powerful HW, such as the f405? Is it too limited (almost 4 times lower clock frequency than the Teensy v4, and less powerful architecture - Cortex M4 vs M7). I'm pretty interested in having a hackable FC where I would implement my algorithms. In your presentation "How the Flight Controller Code Works" you say the control loop runs at 2kHz. Do you know how much free time you have per loop iteration? That is, how much function loopRate takes of the 500usec loop iteration? Finally: I assume the Arduino interface brings a lot of comfort. Do you know how much it penalizes performance? I see that Betaflight runs its PID at 4k on such HW (8k may be too demanding), but I assume input integration is run at a lower frequency, whereas here everything runs at the same frequecy.
There’s about 150 out of 500us overhead free in the 2khz loop. The whole thing about dRehmFlight with the teensy is that it’s just a simple arduino project, so yea there’s some performance & feature sacrifices compared to betaflight, but it’s still “good enough” and offers the benefits of easy integration of new hacky stuff. Hardcore drone racing guys would hate how ‘sloppy’ dRehmFlight feels. But for fun little VTOL projects where performance doesn’t matter as much, it’s more than enough
More than 35 minute? I'm curious about your batteries Man.
nice work, can I buy a copy of your knowledge I have to know what you do :)
Just hit that subscribe button and I’ll keep trying to share everything I know
A może insta360 oneX zamontuj na szczycie tej tyczki. Powinny wyjść świetne ujęcia
I took a shortcut, got lost, back tracked and then ended up walking down a few different paths.
Hi Nicholas, I was just getting involved in model rocketry and remote controled aircraft when our airfield was closed because somebody felt a need to build another hospital. I suspect others have this trouble so I wondered if you have any tips to share to help your viewers find a place to indulge in this hobby or what restrictive laws we might encounter in our own neighborhoods.
My best advice is to always have a kill switch for safety, and consider looking into smaller aircraft that you can fly below the tree line-if ever confronted about where you’re flying, your best defense is “I’m sorry, I didn’t know”
Happy flying!
Terry Davis would have done this easily in HolyC
Oh more helicopter fun. I wonder if the Dufus Level Earth Observer pop in to make silly claims.
Change out the rod for a cup holder and add autonomous control and attempt to deliver a cup of coffee to someone .
Hahah love that idea
Now do it with a double spherical inverted pendulum
would fpga's be a good next step for flight controllers?
the tinyFPGA BX seems pretty cool
5:58 why u gotta do me like that
why is this comment not pinned
but the shortcut learned your a lesson or two
PLEAASSEE stick a 360 camera on this 👅🤤 that would be siiickkk
That would be pretty cool wouldn't it? Need to find a cheap 360 cam now...
Poor little drone, let out a little whimper at impact.
Pfff make a stabilized double pendulum
Alright
Make drone landing on a QR code
Oooh. Sounds scientific. 🤩
Sensei, how do i become like you? haha
no, seriously.
Your knowledge about code and programming is so far beyond me, it's laughable.
It would take me a year to comprehend what you so casually understand so easily.
Amazing!