Bitcraze Forums

Hi,recently I'm trying crazyflie_ros.
I use the command
roslaunch crazyflie_demo hover_single.launch
and an error occurs:|
[crazyflie_server-1] process has died [pid 20730, exit code -6, cmd /home/liruocheng/catkin_ws/devel/lib/crazyflie_driver/crazyflie_server __name:=crazyflie_server __log:=/home/liruocheng/.ros/log/3e8cb1e8-53d2-11eb-96f0-f4939ff49e28/crazyflie_server-1.log].
log file: /home/liruocheng/.ros/log/3e8cb1e8-53d2-11eb-96f0-f4939ff49e28/crazyflie_server-1*.log

I restart the cf and the error still occurs...however, the problem was solved when I connected the cfclient once. But it will happen again when I restart the cf, so I have to connect the cfclient first before running the cf on ros...

I wonder why this kind of problems occurs. Right now I guess it may be the problem of the firmware. I have changed the firmware's makeflie and force the estimator as "kalman"
## These are set by the platform (see tools/make/platforms/*.mk), can be overwritten here
#ESTIMATOR ?= any
ESTIMATOR = kalman
CONTROLLER ?= Any # one of Any, PID, Mellinger, INDI
#CONTROLLER = Mellinger
POWER_DISTRIBUTION ?= stock
Then I do "make clean ---- make ------ make cload "

This problem will disappear when I use the official release 2020.09 -firmware-cf2-2020.09.zip.

Looking for your help,thanks a lot!!!!

I was able to reproduce this issue and it seems to be caused by an incorrect log TOC. As workaround, you can run to remove the log TOC cache files and re-run again.

In the future, it would be better to open an issue on github at https://github.com/whoenig/crazyflie_ros/issues, including the full error message.

I hope that helps!

Hi professor hoenig:
I have tried this method just now, but it is no use....I can only solve this problem by using the official release of the cf firmware, but this can not support cf flying without flow deck. I notice one thing that when using the official release the ESTIMATOR is "complementary".
And I still wonder the localization solution in your work "trajectory planning for quadrotor swarms"(my research interest is also multi-robot planning and I have read this paper thousands of times XD)
In this work you employ the "led ring"deck, which should be put at the bottom of the cf, so I think there is no room to flow deck.
I wonder how to achieve this.
Thanks a lot!!

Estimator: In all recent firmwares the estimator is actually picked dynamically. If you have the flow deck installed, ill will automatically switch to the EKF. The compile time flag you use, simply forces the estimator to be always EKF (instead of dynamically chosen).

Paper: That paper uses the Crazyswarm, not crazyflie_ros. It was flown in a space with a VICON motion capture system and the EKF state estimator.

Error: Do you have any other outstanding changes or the DEBUG mode enabled in your config.mk file of your firmware? Some of the printouts shown in the screenshot (enumeration of all deck drivers) seem non-standard to me.

Hi professor hoenig:
I have solved this problem just now!! As you say, I have changed the config.mk
PLATFORM=CF2
DEBUG=1
I have added this two lines in this flie and when I execute "make" there are many "warnings" in the cmd window...

Then I change the DEBUG=0 and remake the firmware and there are no warnings any more and the error disappear

Now I'm wondering why...

There is one extra warning flag enabled in the compiler by the Makefile when compiling in debug mode, this is why you are seeing more warnings in debug mode.

Thanks a lot! I have solved this problem.
I'm trying flying trajectory now!