Quadcopter PID Debugging

feiyinglala

Quadcopter PID Debugging [Copy link]

 

I made a drone myself. After debugging all the modules, I need to debug the PID and attitude parameters. Is there a simple solution to build a debugging environment? I saw a barbecue-style rack online (as shown below), but I don't have the conditions to build one here. Does anyone have a better method?

Here is a video of the first version of the real machine

第1版无人机2016

okhxyyo

I'll call two experts for you. One of them shared his DIY quadcopter work @lb8820265, and the other shared a wonderful explanation of PID control @DDZZ669. They may be able to help you.

okhxyyo

I'll call two experts for you. One of them shared his DIY quadcopter work @lb8820265, and the other shared a wonderful explanation of PID control @DDZZ669 . They may be able to help you.

feiyinglala

okhxyyo posted on 2022-11-9 16:33 I will call two masters for you. One of them shared the quadcopter DIY work @lb8820265, and the other shared a very exciting PID control explanation...

Thanks to the administrator, hehehe

吾妻思萌

Is this a Taobao kit? It looks similar to the one I had before.

lb8820265

Is the drone made by the poster a flying drone or one with GPS?

If it is a racing drone debugging, the biggest difficulty is to place the sensor flat. If it is not placed flat, it will fly to one side and accelerate in one direction. The second most difficult part is the calibration of the electric speed controller and the motor. If the calibration is not performed at the beginning, the motor will fly to one side at the beginning due to different forces at the beginning of the rotation stage. Although it is still flat, it has an initial velocity flying to one side.

After completing the above two difficulties, you can start adjusting the parameters. These parameters are generally available in the open source flight control system, which can be used as a reference. The only parameter of the cross-country drone is the attitude PID. In fact, you don’t have to have that stand to adjust the parameters. I put a cushion under the drone, or make a tripod, so there is no need to worry about the drone falling and damaging it. Use the remote control to quickly add fuel, and release the fuel immediately when the drone reaches a certain height. As long as it can take off nearly vertically, the parameter debugging is completed.

If it is a drone with GPS, in addition to the attitude PID loop, there is also a position PID loop. After completing the debugging of the attitude PID loop, adjust the position PID loop.

The above are my suggestions, I hope it helps

DDZZ669

I haven't played with a quadcopter yet. For the PID debugging of the balancing car, basically no additional equipment is needed, just adjust it by hand. Since the quadcopter can run in any direction, it is necessary to use a stand to fix it and then adjust the PID in different directions when debugging it. In addition, the propellers are more dangerous when they rotate, and they may break after falling over, so they should be fixed during the initial debugging.

The device in the picture has the quadcopter fixed on a stick, leaving only one degree of freedom for rotation, which should be used to adjust the pid in a certain direction of rotation. If you don't have such a professional device, you can also build a simple device yourself, such as finding a stick and fixing it on an object, such as between two tables, etc. You can use local materials according to the situation. The small yellow platform in the middle of the shelf in the picture may be a universal joint underneath, which can rotate in any direction. After the pid in the three directions is adjusted, it can be fixed on it to perform overall pid debugging in all directions to test whether the quadcopter can maintain balance. For this, I didn't think of a simple alternative device. If you just don't want the quadcopter to lose control and run far, tie a rope underneath to limit its flying distance, and put some soft things on the ground to prevent the quadcopter from breaking.

feiyinglala

Azuma Simeng posted on 2022-11-9 19:13 Is this a Taobao kit? It looks similar to the one I had before

The pictures are just for explaining the debugging scene, not the actual pictures of my machine. I only bought the rack ESC, and the circuit board and code were all made by myself.

feiyinglala

lb8820265 posted on 2022-11-9 20:42 Is the drone made by the poster a cross-country drone or one with GPS? If it is a cross-country drone, the biggest difficulty in debugging is to place the sensor flat. If it is not flat, ...

It's not a flying drone, it should be the one with GPS you mentioned. Your debugging idea is very good. I only considered the calibration of the sensor before, not the motor. When I actually debugged it, I hung a barbell on each corner.

feiyinglala

DDZZ669 posted on 2022-11-9 22:27 I haven't played with a quadcopter yet. For the PID debugging of the balancing car, basically no additional device is required. Just adjust it by hand. Is this because of the quadcopter...

Oh, thank you. I once wrote a program wrong, and the plane took off diagonally and hit the wall, and all four wings were broken. Later, I didn't dare to let it go free, so I hung four barbells on it.

吾妻思萌

Azuma Simeng posted on 2022-11-9 19:13 Is this a Taobao kit? It looks like the one I had before

The picture is just used to illustrate the debugging scene, not the actual picture of my machine. I only bought the rack and ESC, and the circuit board and code were all made by myself.

6. I made it myself. You must be a master. If you want to debug it, you can spread a quilt in an open space outdoors. Mine was broken by me.

feiyinglala

Azuma Simeng posted on 2022-11-10 18:25 The picture is just used to illustrate the debugging scene, not the actual picture of my machine. I only bought the rack ESC, and the circuit board and code were all made by myself. ...

I started working on it in 2015. I didn't choose the right CPU for the first version, so I used STM32F051. I finally completed one in 2017. It was only last year that I made up my mind to make the second version. I don't have many friends to communicate with, and it's purely a hobby. I do it in my spare time. I'm still adjusting the camera.

lugl4313820

This is awesome. I look forward to sharing the final precise algorithm!

feiyinglala

lugl4313820 posted on 2022-11-12 08:17 This is amazing, I look forward to sharing the final precise algorithm!

I am a novice, and I have no idea about attitude and flight control algorithms.

通途科技

Learned a lot, was educated, the content is rich, it is worth watching!