@Dronodev
ov_overlay is now available under the VIO tab in voxl-portal. Please make sure to update your system to the latest released SDK.

Hi @Pelle-Bredegg-Hermansen
I'll be happy to review your design.

I'll email you now to start the exchange.
Thanks!

Hi @cbay
Sorry, we do not share our PCB design data.

The best we did share is our Breakout Board Design noted here but that is SCH and an Altium Starter Kit w/o libs. Some symbols are in there you can try to export, but not many are in there.
https://docs.modalai.com/expansion-design-guide/#voxl-2-custom-expansion-pcb

And we have this design which we "kinda-sorta" opened a little bit for some schematic guidance.
https://docs.modalai.com/voxl2-dev-test-board/

Hope this helped.

Of note, if you do a board that is meant to mate with VOXL 2 J3/J5 or the camera group connectors J6-J8, I will review your schematic as a courtesy. Just reach out again and tag me, and I can email you privately so you are not posting your design on a public forum.

@sja123 No, unfortunately, we do not currently support boot from SD card.

@magnetocircuit Yes, this is supported. In the config file for px4 /etc/modalai/voxl-px4.conf set ESC=VOXL2_IO_PWM_ESC.

@Aaron-Porter Hello there,

voxl-evaluate-vins is currently not used by the Autonomy team @ ModalAI, we opted for our new benchmark solution, a bit more consistent in terms of statistical rigor and better integration with voxl-portal. Still, when replaying logs for VINS, it is imperative that one replays them using the ION pipes, as our VINS solution is optimized for cl_mem ptrs. Note that when replaying ION pipes, it is recommend to stop camera-server before replaying -- as clean killing ION pipes via the voxl-replay args is currently being optimized. This should resolve the common segfaults experienced during the replay/benchmark operation itself.

https://gitlab.com/voxl-public/voxl-sdk/utilities/voxl-benchmark-vio

Finally, I would recommend updating your VINS to the latest available tag on Gitlab

@Sarika-Sharma ,

Can you please take a look at output of dmesg -w, and paste here only the part that prints after you start the camera server. Use the flex cable that is able to detect all 4 cameras, but streaming does not work. You should try to view the stream in voxl-portal so that the camera is actually activated in this test. Also please post the output of voxl-camera-server.

Alex

@Piyush-Singh ,

The voxl-portal does not handle well streaming large images (including in multi-view, where the images are actually appearing small). The voxl-portal back-end uses software JPEG encoding, so one of your streams (hires_large) is 4208x3120, which takes LOT of cpu processing to encode to jpeg.

You can see that the voxl-portal process uses 260% cpu (2.6 full cores), which is a lot.

There is no practical need for streaming small and large uncompressed images to voxl-portal, so you can just enable small video stream, which is much smaller.

Also, if you select an individual camera stream, you can enable advanced camera controls (click on the small check box in lower left) and then you can adjust the JPEG encode quality from default 70 to something much lower in order to use less cpu. This setting will be applied to all cameras and you can go back to multi view after that. Note that the jpeg encode quality only affects voxl-portal view, not h264/h265 encoding.

Please try.

Alex

@colombe-ai I'm not sure how much we're going to be able to help to get this going but I'd start with that magnetometer. Put an oscilloscope in the lines and see if you see any activity.

@jmltt , ok, thanks for the details. I will check if we have any of those assemblies in stock, or what the best course of action in your case.

Meanwhile, if you already have the shorter cable, are you able to use it for checking if the USB issues are resolved by switching the USB cable?

Alex