Hi everyone, thanks for watching! If you have any questions for me leave them on this comment so I get notified and can try to get back to you. Cheers!
Great video! I ask myself: At what point would you "interfere" with manual control inputs now while the drone is running? Like in a drone that you fly yourself, but once you let go of the sticks, it just hovers where you left it?
sphericalsphere you could set up your hand-held remote controller to command altitude rate, yaw rate, and XY position. These would be converted in the software to the reference signals that the control loops use that I described in this video. When you’re “manually” flying your drone you’re telling it to go up or down, or to rotate, or to fly horizontally. The inner control loops are still running during this and keeping the drone stable and following your commands. As soon as you let go of the toggles the drone would immediately start to hover since you’re telling it to stay right where it is.
Great Series Brian, I myself have gone through all these while creating our own drone controller. Especially, tuning the cascaded control was the most challenging of all. It is a highly sensitive system with lot of dependencies and coupled motions. There are not much resources out there who discuss this aspect of drone control. Thanks for taking this up and eagerly waiting for your next video to see your approach to tune this controller.
Great video and explanation. I am wondering - wouldn't it be possible to use a neural network to figure it all out by itself? (of course this includes crashing the drone more or less often....)
Well, I am not sure whether I understand your question correctly, yes, NN will work for sure, however, a simple NN will not, for this kind of task I highly recommend using simple algorithm which requires less computational ability. And also, NN can be used to identify the quadcopter model more precisely than traditional models, which can make us control the drone more accurately.
If you have a model of the system, i.e., equations that predict the position and velocity of the quadcopter, given the input rotation speed of the rotors, you use reinforcement learning to train a neural network which will serve as the controller.
Great video but I mean a rate controller is normally used commanded from attitude controller. Angular rates are measured directly and quite accurately. Therefore it is possible to command a rates into inner controller. The feedback would have better rejection for disturbance as inertia change etc.
Not until Monday ... the delay is my fault. I have all of the videos completed for this series and ready to go but I want to release a companion video on my channel when drone #3 posts. My video won't be ready until Monday, hence the delay. Stay tuned though!
in this video, if the roll, pitch and yaw angles were relative to the reference frame, then we don't need the yaw feedback to the outer loop controller, right?
This series will cover the exact flight controller that ships with the Aerospace Blockset and is for the Parrot Mambo. However, the general concepts and ideas can be used for other drones as well.
Hi everyone, thanks for watching! If you have any questions for me leave them on this comment so I get notified and can try to get back to you. Cheers!
Great video! I ask myself: At what point would you "interfere" with manual control inputs now while the drone is running? Like in a drone that you fly yourself, but once you let go of the sticks, it just hovers where you left it?
sphericalsphere you could set up your hand-held remote controller to command altitude rate, yaw rate, and XY position. These would be converted in the software to the reference signals that the control loops use that I described in this video. When you’re “manually” flying your drone you’re telling it to go up or down, or to rotate, or to fly horizontally. The inner control loops are still running during this and keeping the drone stable and following your commands. As soon as you let go of the toggles the drone would immediately start to hover since you’re telling it to stay right where it is.
hi brian
i am having project to design a lidar enaled uav system for terrain mapping is it possible to use a lidar sensor in this also
Dude I follow your channel as well, I had no idea it was you in this series!
Great video Mr Brian, my question is how you can get the (x,y) coordinates without a GPS.
I was watching your channel a few weeks ago and now I've found this! Your control theory insights are the best. Cheers!
Great Series Brian, I myself have gone through all these while creating our own drone controller. Especially, tuning the cascaded control was the most challenging of all. It is a highly sensitive system with lot of dependencies and coupled motions. There are not much resources out there who discuss this aspect of drone control. Thanks for taking this up and eagerly waiting for your next video to see your approach to tune this controller.
Wow. Stunningly simple and understandable explanation. Looking forward to the next video!
Great!
waiting for the remaining lectures.
Very nicely explained and the graphics is simple but very useful in understanding.
WOW!, great video and explanation. Waiting anxiously for the other videos. Many thanks.
Very much informative ...Good work Brian.
Great video and explanation.
I am wondering - wouldn't it be possible to use a neural network to figure it all out by itself? (of course this includes crashing the drone more or less often....)
Well, I am not sure whether I understand your question correctly, yes, NN will work for sure, however, a simple NN will not, for this kind of task I highly recommend using simple algorithm which requires less computational ability. And also, NN can be used to identify the quadcopter model more precisely than traditional models, which can make us control the drone more accurately.
If you have a model of the system, i.e., equations that predict the position and velocity of the quadcopter, given the input rotation speed of the rotors, you use reinforcement learning to train a neural network which will serve as the controller.
Great video but I mean a rate controller is normally used commanded from attitude controller. Angular rates are measured directly and quite accurately. Therefore it is possible to command a rates into inner controller. The feedback would have better rejection for disturbance as inertia change etc.
this is an unbelievably great video!
What component do you use to build this system? Should it be operational amplifier ?
Great, great great video, thanks!!
So the PID control parameters that are carried out for roll, pitch and hover movements can the parameters be the position of the drone?
Awesome series, Brian. Can we expect the next video today?
Not until Monday ... the delay is my fault. I have all of the videos completed for this series and ready to go but I want to release a companion video on my channel when drone #3 posts. My video won't be ready until Monday, hence the delay. Stay tuned though!
@@BrianBDouglas Thanks for the quick response. Looking forward to it!
@@BrianBDouglas Thank you for the feedback we all were waiting for. Looking forward to it.
Great videos now how can you to fallozw the mouvement avec a cibl and which components are you going to use??
What do the transfer functions look like, and how are they used in this system?
Great videos now how can you do to fallow an object to have its position in distance and all time? and which components are you going to use??
in this video, if the roll, pitch and yaw angles were relative to the reference frame, then we don't need the yaw feedback to the outer loop controller, right?
Thanks for making these videos, keep it going please.
How to find the basic flight controller block present on 12:59
Good work!
Thanks
What should we do if roll will affect pitch and yaw and altitude at the same time?
hi Brian, is it possible to model the flight controller for other drones and Arduino self-made flight controllers?
This series will cover the exact flight controller that ships with the Aerospace Blockset and is for the Parrot Mambo. However, the general concepts and ideas can be used for other drones as well.
@@BrianBDouglas thank you very much, you explain everything clearly and it is easy to understand.
Waiting for the next video very eagerly😀
Would be nice to have a link to part 1 in the description
Thanks for the suggestion. I've added a link in the description and here is a link to the playlist page: bit.ly/2RmLZLQ
Good idea. I don't have access to make that change but I've put in the request.
where can i find that flight controller block? in 13:03?
I Brian nice video, Can you explain more the position controller
what a great teacher !!
Wow amazing
I don't understand how one toggle control both left and right speed, can any one explain this?
i picked this project as my final year project. please provide me with the starting point that from where i should start this project
I am also doing this project
Can you help me?
Great..:)