Topics created by Gary Holmgren

@Gary-Holmgren , we have not tried the mask with 1280x800 resolution.

Can you please try using the mask of size 160x120 and make it half white / half black for sanity testing. It is possible that the mask size has to be of these specific dimensions (and it would get resized to match the aspect ratio of the actual image, not sure).

Also, you should enable the debug prints in QVIO, as mentioned here https://forum.modalai.com/topic/3459/masking-qvio/6 and there is another potential "gotcha", the mask path is relative to working directory where qvio server starts.. but after you enable the qvio server debug prints, you will see the path from which the mask file is being loaded, as you can see from screen shots from the other post.

Unfortunately, the underlying QVIO source code is not available, so I cant check the exact behavior of the masking feature.

Alex

@Gary-Holmgren Was able to fix it with bash chmod +x /home/4.1.0.4/tools/HEXAGON_Tools/8.4.05/Tools/bin/ld.qcld (while in the docker image of course)
It appeared hexagon-link was a symbolic link and making it executable fixed the target not having permissions. Hope this helps someone with a similar issue.

@Eric-Katzfey Ahh that makes more sense, thankyou for explaining!

Hi @Gary-Holmgren,

We tested this functionality but due to insufficient demand, we did not mainline it, because it would require parameter change and could cause confusion.

https://github.com/modalai/px4-firmware/commit/abd04fafa746af512960a1de2386d5d7f6cb512a

However, you should be able to cherry pick this commit into latest px4-firmware from voxl-dev branch or whatever you are using now.

No need to modify M0065 because the interface between VOXL2 and M0065 does not specify the limits - M0065 just creates the pulse of the specific length that VOXL2 requests and M0065 does not care about the limits, so all the logic is done in VOXL2 / PX4.

After this change, the voxl2_io driver will look for the following px4 params, which have standard convention / naming. Defaults will be used if not specified.
(for x=1:8)

VOXL2_IO_FUNCx VOXL2_IO_MINx VOXL2_IO_MAXx VOXL2_IO_DISx VOXL2_IO_FAILx

Alex

@Gary-Holmgren I managed to finally get it working after flashing a few times with the web portal when the receiver and transmitter were in WiFi mode. It helped to connect instantaneously after flashing when making a binding phrase on both the receiver and transmitter.

@Gary-Holmgren Great to hear, new SDK release with that fix coming soon

@Gary-Holmgren Figured it out, I hope this helps others with a similar setup.

I didn't notice before but in the usr/bin/voxl-px4-start script there is a comment:

# qshell vl53l1x start -X -b 4

Which you can make or replace to change the DISTANCE_SENSOR condition to:

if [ "$DISTANCE_SENSOR" == "VL53L1X" ]; then # Make sure to set the parameter SENS_EN_VL53L1X to 1 qshell vl53l1x start -X -b 4 # qshell lightware_laser_serial start -d 7

Then you can change the voxl-px4-conf to:

DISTANCE_SENSOR=VL53L1X

After setting the relevant distance sensor / EKF2 params in QGC you should be able to restart voxl-px4 and the distance sensor will appear!

@Zachary-Lowell-0 Yes this did the trick, thankyou!

@Zachary-Lowell-0 Would definitely be a useful function. I am probably gonna stick with voxl2 running px4 for now since it simplifies things.

Hi @Gary-Holmgren. This is likely because you're running the build script from an x86 system (docker requires that you build on the same architecture as the base image, in this case arm64). You can solve that by cloning the repo onto the voxl and running build.sh from there.

Note: make sure your voxl has a wifi connection before running ./build.sh.

@Gary-Holmgren Thanks for the update!

@Gary-Holmgren Copying over response from another thread: https://forum.modalai.com/post/20119

@Gary-Holmgren Hi Gary, we have not thought to try that setup as HITL is typically used to develop the lower level flight controller. What are you trying to achieve?

@Gary-Holmgren Check out the AprilTag Relocalization page

Basically, every time an AprilTag is viewed, the position is updated to reflect that tag as an anchor. So, it's not precision landing as much as position refinement. You then land at the desired position which should be more accurate due to the position refinement with the fiducial marker

We haven't looked at the MAVSDK in a while, is there a specific error you're running into?

@Alex-Kushleyev Good to know, Thankyou for clarifying Alex!