Tuning crazyflie for smooth, stable and responsive flight

[Call_me_a_Cab]

Erm, really basic question here....

Where is "config.mk"?

[chad]

Erm, really basic question here....

Where is "config.mk"?

Hi @Call_me_a_Cab, there is a sample one (config.mk.example) in crazyflie-firmware/tools/make. You can either use that one and rename it config.mk or just create a new file called config.mk and add the parameter. The main Makefile in the crazyflie-firmware directory will pick it up from crazyflie-firmware/tools/make.

[Call_me_a_Cab]

Thanks Chad,

I read that and thought... Yeah that makes some sense..

However, within the VM 0.8 in eclipse in crazyflie-firmware I do not have a tools folder!?

God this is coming difficultly to me haha

[nanobot9000]

I noticed the same thing. I copied a tools folder from one of the other projects and put it in the firmware project. Not sure why there is not a tools folder or if what I did was OK, but I think it worked.

[chad]

No sweat. I think the 0.8 VM has aged out a bit and the crazyflie-firmware is on the wrong branch. Try this:

In the crazyflie-firmware directory:

Code: Select all

git branch

It might say "crazyflie2". If so, do this:

Code: Select all

git checkout master
git pull
git branch

It should say "master". Then, check and see if you have the directory.

[Call_me_a_Cab]

Sorted!

Thanks a lot

[theseankelly]

I'm so glad this has been useful! PID tuning is more of a dark art than a science. Lots of guess and test. I encourage everyone to play with the params and see how they affect flight. I'm certain there's a better set of coefficients than what I came up with!!

[nanobot9000]

Thanks for the tip on pulling down the updated master - that fixed the missing tools folder for sure.

I am new to this git business and am enjoying learning. I do have what is probably a silly question, but after I create the config.mk with the appropriate statement how do I know for sure that the build was correct aside from the behavior once I load it? Is there some number in this example that tells me it used my config.mk as part of enabling the experimental PID?

Code: Select all

Crazyflie 2.0 build!
Build 97:c2a88215d5dd (2016.02-97) CLEAN
Version extracted from git
Crazyloader build!
   text	   data	    bss	    dec	    hex	filename
  99784	   1668	  35956	 137408	  218c0	cf2.elf
rm version.c
bitcraze@bitcraze-vm:~/projects/crazyflie-firmware$ 

Thanks,
Mike

[liwanggt]

I have noticed that the VM is on the "crazyflie2 branch" by default, rather than the "master branch". I also noticed that the crazyflie 2.0 I recently bought came with the "crazyflie2 branch" firmware. My question is: what is the difference between those two branches in terms of attitude estimation and control algorithms.

From the hovering experiments I performed using Crazyflie ROS driver and Optitrack, I found the "crazyflie2 branch" firmware performing slightly better than the "master branch" firmware. The difference is observed when I tried to make it hover at a fixed point. I haven't tested their difference in following given trajectories yet.

Li

No sweat. I think the 0.8 VM has aged out a bit and the crazyflie-firmware is on the wrong branch. Try this:

In the crazyflie-firmware directory:

Code: Select all

git branch

It might say "crazyflie2". If so, do this:

Code: Select all

git checkout master
git pull
git branch

It should say "master". Then, check and see if you have the directory.

[chad]

@liwanggt

The VM is a little old now. The crazyflie2 branch was for initial work on the Crazyflie 2. Once the Crazyflie 2 and Crazyflie Nano (1.0) firmwares were merged, all work continued on master. The last commit on the crazyflie2 branch was back on Jul 20, 2015 - nearly a year ago!

There are significant differences between the two branches - crazyflie2 is 286 commits behind the master branch! You can look through the comparison on GitHub.