two channel for pitch

[curiousmind]

hi guys, i want to do project on rigidly connected drone, means two drones(two receiver,one transmitter, two flight controller,this stuff are fixed please don't argue on that ) rigidly connected via 1 metre long PVC pipe. for that pitch ,throttle and roll will be same as normal drone. but to perform yaw operation i need two opposite pitch, so how i do that?
thanks in advance....

 

[JasonK]

your going to have a hard time even getting that off the ground, with the rigid connection between the 2 drones, the Flight controllers are going to fight and struggle against each other.

to read the main sticks differently on the 2 FC, your likely going to need to modify the firmware and make something custom. From what I see, they always read the first 4 channels for AETR and expect positive to have a specific meaning, so you would need to reverse that meaning on one of the FCs, so you could get the expected behavior.

The thing with the rigid connection is that your effectively making 1 quad and having each 'side' behave as a completely separate object is going to cause the FC to 'freak out' because of the additional forces from the boom.

 

[CappyAmeric]

If you want an octocopter, set up ONE FC for 8 motors and mount it in the center.

 

[curiousmind]

your going to have a hard time even getting that off the ground, with the rigid connection between the 2 drones, the Flight controllers are going to fight and struggle against each other.

to read the main sticks differently on the 2 FC, your likely going to need to modify the firmware and make something custom. From what I see, they always read the first 4 channels for AETR and expect positive to have a specific meaning, so you would need to reverse that meaning on one of the FCs, so you could get the expected behavior.

The thing with the rigid connection is that your effectively making 1 quad and having each 'side' behave as a completely separate object is going to cause the FC to 'freak out' because of the additional forces from the boom.

thanks i really appreciate you for help ,i almost developed algorithm which prevent flight controllers to not fight and struggle against each other

you are right, i need to change firmware. can you suggest which FC i use for this project because i don't know much about flight controllers

i don't understand FC 'break out'. if you put some light on it , so that i can understand what do you mean by that.

 

[curiousmind]

If you want an octocopter, set up ONE FC for 8 motors and mount it in the center.

yea, you are right but that is not what i want for my project.

 

[JasonK]

thanks i really appreciate you for help ,i almost developed algorithm which prevent flight controllers to not fight and struggle against each other

you are right, i need to change firmware. can you suggest which FC i use for this project because i don't know much about flight controllers

i don't understand FC 'break out'. if you put some light on it , so that i can understand what do you mean by that.

"freak out" not "break out"

is typically a PID loop runaway. For example my whoop class drone will pin itself to walls if it bumps them the right way [at least in angle mode] because it tries to level itself so strong that it just gets pinned to the wall, I have to disarm to get it to fall off. Another example is if a quad takes off to the moon [goes up quickly even if the throttle is zero or low because it isn't tunned right and is trying to level itself... over doing it, and oscillating itself up up and away]

 

[CappyAmeric]

Think of it this way: Two kids think it would make perfect sense to take two bicycles, weld them together side by side and then they could ride them together like a four wheeled bike. The problem is quickly seen why it is not practical: the kids can‘t pedal equally, and they can’t possibly coordinate the steering of two sets of handlebars.

If you want an octocopter, you need to use only one FC. If you don’t know much about FCs, I am wondering how you are planning to write the algorithm needed to control 2 FCs and keep them in sync.

 

[curiousmind]

"freak out" not "break out"

is typically a PID loop runaway. For example my whoop class drone will pin itself to walls if it bumps them the right way [at least in angle mode] because it tries to level itself so strong that it just gets pinned to the wall, I have to disarm to get it to fall off. Another example is if a quad takes off to the moon [goes up quickly even if the throttle is zero or low because it isn't tunned right and is trying to level itself... over doing it, and oscillating itself up up and away]

thanks i got your point

 

[curiousmind]

Think of it this way: Two kids think it would make perfect sense to take two bicycles, weld them together side by side and then they could ride them together like a four wheeled bike. The problem is quickly seen why it is not practical: the kids can‘t pedal equally, and they can’t possibly coordinate the steering of two sets of handlebars.

If you want an octocopter, you need to use only one FC. If you don’t know much about FCs, I am wondering how you are planning to write the algorithm needed to control 2 FCs and keep them in sync.

i don't want octocoptor, i will explain you

 

[CappyAmeric]

i don't want octocoptor, i will explain you

Two quadcopters joined together with rigid PVC is an octocopter. 4 motors x 2 = 8 motors = octocopter. Maybe you have something else in mind from what you’ve said.

 

[curiousmind]

Two quadcopters joined together with rigid PVC is an octocopter. 4 motors x 2 = 8 motors = octocopter. Maybe you have something else in mind from what you’ve said.

now i think you got my point

 

[JasonK]

so you mean this:
https://spectrum.ieee.org/automaton/robotics/drones/microdrones-future-of-warehouse-automation

you realize that isn't "2 drones connected by a rigid PVC pipe," correct?

A quick skim on that article shows that they are doing a form of SLAM _and_ are talking to each other constantly to coordinate their motions meaning they are _not_ 2 fully independent drones while carrying the object, they are acting as one.

Also watching the video, it looks like the mount point between the drone and bar is flexable and at/near the CG of the drone, which minimizes the forces applied to it from the other drone.

 

[CappyAmeric]

A quick skim on that article shows that they are doing a form of SLAM _and_ are talking to each other constantly to coordinate their motions meaning they are _not_ 2 fully independent drones while carrying the object, they are acting as one.

As we’ve said, custom FCs and custom firmware. The constant synchronization would cause havoc with a mini-quad FC. This is not a simple joining of 2 drones.

From the article, the purpose for multiple drones was to validate position using independent non-GPS sensors - not for lifting capabilities. I think DJI‘s consumer-level sensor technology has likely leap-frogged far over whatever this academic exercise from 2018 was trying to accomplish. Knowing the state of technology in 2021, and after reading the article, I had to ask, “What is the point?”

 

[curiousmind]

so you mean this:
https://spectrum.ieee.org/automaton/robotics/drones/microdrones-future-of-warehouse-automation

you realize that isn't "2 drones connected by a rigid PVC pipe," correct?

A quick skim on that article shows that they are doing a form of SLAM _and_ are talking to each other constantly to coordinate their motions meaning they are _not_ 2 fully independent drones while carrying the object, they are acting as one.

Also watching the video, it looks like the mount point between the drone and bar is flexable and at/near the CG of the drone, which minimizes the forces applied to it from the other drone.

we know that how drone can controlled by pitch,roll ,yaw and throttle and we also
going to use same. Flying two individual rigidly connected drone is very difficult
because of control and stability reasons.

if we consider case of rigidly connected two drones which have same angle as showed in image-1,they both can not decide whether to go up or down. This confusion make them unstable. So in order to solve this problem we have to make them cooperative by adding algorithm.

As we can see in image-2 to solve this problem we have to develop algorithm which is based on horizontal angle of drone. In order to understand how this algorithm works, we have to consider one assumption that all the angles in positive side of y axis taken positive and all the angles in negative side of y-axis taken negative. So algorithm is following, as show in image-2 if left drone get feedback from sensor that angle is positive then that drone reduce their thrust and if right drone get feedback from sensor that angle is negative then that drone increase their thrust. But this solution also have overshooting and undershooting problem but we can solve this problem by PID control and tuning. we can use same algorithm in pitch if one drone go little bit fast compare to other.

 

[JasonK]

@curiousmind - FWIW, I have an Electrical & Computer Engineering Degree and did control system theory stuff. You don't need to explain the _theory_ to me or the high level algorithm to me on how something like that _could_ work. If I wanted to, I could replicate the experiment shown in that link, I just don't have a desire (nor large amounts of spare time to spend on something that I don't have a strong desire to work on) to do so.

you _can't_ just solve that problem purely in a PID loop running separately on each drone as they have to move in a coordinated maner (and more coordinated then them both receiving the same input signal from your TX).

 

[curiousmind]

As we’ve said, custom FCs and custom firmware. The constant synchronization would cause havoc with a mini-quad FC. This is not a simple joining of 2 drones.

From the article, the purpose for multiple drones was to validate position using independent non-GPS sensors - not for lifting capabilities. I think DJI‘s consumer-level sensor technology has likely leap-frogged far over whatever this academic exercise from 2018 was trying to accomplish. Knowing the state of technology in 2021, and after reading the article, I had to ask, “What is the point?”

right, i have to change firmware but do i really need custom FC because i don't have electronic background. i know that this project is challenging and takes a lot time but i really want to do this project.

 

[JasonK]

right, i have to change firmware but do i really need custom FC because i don't have electronic background. i know that this project is challenging and takes a lot time but i really want to do this project.

Do you have a programming/control theory background... if not, your going to have a hard time writing the needed custom FC software.

 

[curiousmind]

@curiousmind - FWIW, I have an Electrical & Computer Engineering Degree and did control system theory stuff. You don't need to explain the _theory_ to me or the high level algorithm to me on how something like that _could_ work. If I wanted to, I could replicate the experiment shown in that link, I just don't have a desire (nor large amounts of spare time to spend on something that I don't have a strong desire to work on) to do so.

you _can't_ just solve that problem purely in a PID loop running separately on each drone as they have to move in a coordinated maner (and more coordinated then them both receiving the same input signal from your TX).

yea , you are right.
if purely PID loop not enough then what else needed.

 

[JasonK]

yea , you are right.
if purely PID loop not enough then what else needed.

look up SLAM ["SLAM robotics" should get you started] and you would need to get a way for the 2 drones to share information between each other.

 

[curiousmind]

Do you have a programming/control theory background... if not, your going to have a hard time writing the needed custom FC software.

i know c++ and little bit python and i am learning control theory