This is a known bug that was fixed in voxl-vision-px4 v0.8.8 which is currently on the development repo, scheduled to be merged into the stable repo in a week or so.

The bug is not caused by PX4, it was due the the handling of the specific mavlink packet containing console data. It should behave normally when not using the PX4 console.

I tried to recreate this with a simple little test_time program but it behaves as expected:

#include <stdio.h>
#include <time.h>

void main()
{

	time_t t = time(NULL);
	struct tm datetime = *localtime(&t);

	printf("time:\n");
	printf("%d-%d-%d_%d-%d-%d\n", datetime.tm_year+1900,datetime.tm_mon+1, datetime.tm_mday,datetime.tm_hour,datetime.tm_min,datetime.tm_sec);
	printf("DONE\n");

	return;
}

the result of starting this manually and from a systemd service is the same.

Aug 03 17:52:43 apq8096 systemd[1]: Starting test_time...
Aug 03 17:52:44 apq8096 systemd[1]: Started test_time.
Aug 03 17:52:44 apq8096 test_time[4238]: time:
Aug 03 17:52:44 apq8096 test_time[4238]: 2021-8-3_17-52-44
Aug 03 17:52:44 apq8096 test_time[4238]: DONE

Note that if you run this on boot then it will likely run before establishing an internet connection and fetching a date and time from either GPS or an NTP server. VOXL will not remember the date and time between reboots.

Also, modifying VVPX4 to log VIO data is an unnecessarily difficult way of going about this. It's also unsafe to write to a file in the same flight-critical thread that's sending PX4 high-rate commands. There is a tool called voxl-logger that likely does what you need, and would be safer to modify to do what you need than messing with VVPX4.

yocto:/etc/systemd/system$ voxl-logger -h

Tool to save data published through Modal Pipe Architecture to disk.
By default, this saves images to /data/voxl-logger/ since the /data/ parition is
the largest partition on VOXL. You can change this with the -d argument.

-c, --cam {name}            name of a camera to log, e.g. tracking
-d, --directory {dir}       name of the directory to save the log to, default
                              is /data/voxl-logger/ if argument is not provided.
-i, --imu {name}            name of an imu to log, e.g. imu1.
-k, --skip {n_to_skip}      number of samples to skip between logged samples
-h, --help                  print this help message
-n, --note {note}           optionally add a very short note to the end of the log
                              directory name, e.g. flight_with_new_motors
-o, --preset_odometry       record a preset log containing qvio, tracking cam and both
                              IMUs for testing VIO. This can be used with other pipes.
-s, --samples {samples}     maximum samples to save for one specific channel,
                              otherwise it will run untill you quit the program.
-t, --time {seconds}        time to log for in seconds.
-v, --vio {name}            name of a vio pipe to log, e.g. qvio
-z, --debug                 enable verbose debug mode

typical uses:
  To log one camera image:
    yocto:/# voxl-logger --cam tracking --samples 1 

  To log 5000 imu samples:
    yocto:/# voxl-logger -i imu1 -s 5000 --note "primary imu test"

  To log 1000 imu0 samples and 2000 imu1 samples:
    yocto:/# voxl-logger -i imu0 -s 1000 -i imu1 -s 2000

  To log every 5th imu sample (skip 4) for 5.5 seconds:
    yocto:/# voxl-logger -i imu1 --skip 4 -t 5.5

  To log tracking camera and both imus until you press ctrl-c
    yocto:/# voxl-logger --preset_odometry

  To record a typical log for testing VIO (imu + tracking) for 1 minute
    yocto:/# voxl-logger --preset_odometry --time 60 --note "log for vio replay test 1"

Here is the source code if you are curious, but it is installed by default as part of voxl-suite so you can use it out of the box. Note, we haven't used this tool in a while so if you find a bug/regression please let us know.

https://gitlab.com/voxl-public/modal-pipe-architecture/voxl-mpa-tools/-/blob/master/tools/voxl-logger.cpp

Best,
James

Hi Zach,

I just updated libmodal_pipe to 2.0.8 with a fix to cleanup the trailing characters from echoing a pipe name to the request pipe. Now you can run modal-hello-server in one window and see the output in another:

yocto:/run/mpa/hello$ echo  "test" > request  && cat test
hello160hello161hello162hello163hello164hello165hello166

Let me know if this is along the right path. I have no immediate intention for making a python library.

Best,
James

Do you have a GPS/Mag plugged in? If not please try plugging in a GPS/Mag and set the PX4 parameter SYS_HAS_MAG to 1. There is a PX4 bug that prevents VIO-only operation if a magnetometer is missing, even if the magnetometer is not in use.

We've been trying to get the bugfix approved and merged into PX4 master for over a year now but it seems to be a lost cause. For now our own branch of PX4 has the bugfix along with some custom tools for our ESCs. If your system starts working again with the magnetometer then we know somehow a standard PX4 image got onto your flight core instead of one of our own.

Regards,
James

This is a known bug that was fixed in voxl-vision-px4 v0.8.8 which is currently on the development repo, scheduled to be merged into the stable repo in a week or so.

The bug is not caused by PX4, it was due the the handling of the specific mavlink packet containing console data. It should behave normally when not using the PX4 console.

@shlee853

I'm afraid to say the VOA feature has not yet been ported forward to our new software architecture (voxl-suite v0.3.4), but it still works in the legacy software (voxl-suite 0.2.0). We are a few weeks away from finishing the new version of VOA and adding it back in, at which point we will release another voxl-suite version. Until then you can go back to voxl-suite 0.2.0, sorry for the inconvenience! We are in a bit of a transition period as we rewrite all of our legacy software in the new architecture.

Best Regards,
James

Hello Dean,

Starting with voxl-suite 0.3.4 (what you have installed) voxl_cam_ros is deprecated due to annoyances like this.

Please try its replacement node voxl_mpa_to_ros. This will publish all cameras, IMU, and some other data to ROS all with one node!

yocto:/# roslaunch voxl_mpa_to_ros voxl_mpa_to_ros.launch

in another window:

yocto:/# rostopic list
/hires/image_raw
/imu0
/imu1
/qvio/odometry
/qvio/pose
/qvio_overlay/image_raw
/rosout
/rosout_agg
/stereo/left/image_raw
/stereo/right/image_raw
/tracking/image_raw
/parameter_descriptions
/tracking/image_raw/compressedDepth/parameter_updates

Don't worry, it won't actually publish data until something subscribes to the topic so it's not wasting CPU having all those topics advertised.

Best Regards,
James

Hello, When you ran the wizard, what did you enter for the camera configuration question, it should have been option 3 (hires+stereo+tracking) for a flight deck.

Do you mind running voxl-vision-px4 manually in odometry debug mode to see if VIO is working and share the output? You can use this command:

voxl-vision-px4 --debug_odometry

Thanks!

Does this behavior only occur in PX4 position mode? Does it fly in PX4 manual mode?

Hello,

The M500 should have been sent to you all configured and working. Do you mind stopping the voxl-vision-px4 service and running voxl-test-vision-lib to see if VIO is working? You can use the following commands:

systemctl stop voxl-vision-px4
voxl-test-vision-lib -o

Hi Rapit,

You don't need the secondary qgc_ip, just one udp link from VOXL to MAVROS. In this case you would use MAVROS as another link in the chain between VOXL and QGC.

VOXL <--> MAVROS <--> QGC

Try launching mavros with the fcu_url :
fcu_url:=udp://VOXL_IP:14550@:14550

and gcs_url:
gcs_url:=udp://@127.0.0.1

You may need to change the port to QGC to something else (like 14551) to avoid a conflict.

Best,
James