Error flash crazyflie firmware with kalman config.mk

[chuhao]

Hi

I was following the arnaud demo video https://www.youtube.com/watch?v=ZgH4bLZdq2A to flash crazyflie firmware with loco positioning system.

I customized the file " tools/make/config.mk", as

Code: Select all

SENSORS = task
ESTIMATOR = kalman

Then, compile it and it shows error at the end,

make: *** No rule to make target 'sensors_task.o', needed by 'cf2.elf'. Stop.
rm version.c

So, I tried delete "SENSORS = task" and config.mk would be

Code: Select all

ESTIMATOR = kalman

It is compiled successful, but I'm not sure how's its affect on indoor autonomous fly.

Thanks
Chuhao

[arnaud]

Hi,

Removing sendor is correct, this is a change in the firmware since the video has been made, the wiki was actually updated for this change: https://wiki.bitcraze.io/doc:lps:earlya ... e_firmware.

So there should be no problem, if you enabled the kalman filter, Crazyflie will fly in the LPS!

[chuhao]

thanks, this part should be no problem.
Do you have wiki to introduce how to edit flying trajectory? I want to try different flying path.

[arnaud]

No, there is not yet documentation on how to fly trajectories. We are currently working on finishing the documentation related to setup and higher level use case will follow later.

If you are using ROS you can write your own ros node that would move the crazyflie setpoint. A good starting point would be to start from the pulish_pose or publish_pose_teleop node.

[chuhao]

I saw there are topics to publish /cmd_vel and /goal. But I'm just worried about radio delays. If the cmd is transmitted from ground pc to drone, it requires a very robust tele channel.

[arnaud]

Well it depends of what is your goal exactly. The latency is quite good with the radio (a couple of milliseconds), so it is good enough to send position setpoint and we even had some success sending velocity and acceleration setpoint from the radio (though we then have to be very careful to have a good strong radio link).

Onboard trajectory generation has also been experimented with and it is the way to go in order to follow precise trajectory.

Using the topic /goal you can send position setpoints.