@nickk14 Awesome, glad to hear it! Yeah, we've actually mounted fans to our boards that run multiple models simultaneously because the thermals definitely get up there. Might be worth thinking about depending on your application. All the best, let me know if I can help out in any other way. Thomas

@Eric-Katzfey as far as I know docker image in the voxl2 ecosystem is used for installing stuff like a different version of Ros or opencv or some other dependency. In my case I modified a lot of back end services and the .conf files, that along with some of the Ros dependencies make the process of cloning a drone atleast 1.5 hr long. Making a docker image for ros dependencies and installing it would shave off only half off that time the other half I would still need to reconfigure voxl2 sdk as my other drone. For 4-5 drones it's fine, but for more 5 it's not

@Moderator Great, I noticed in SDK 1.2 there is now support for dual tracking cameras in the camera server, can both feeds be fed into the QVIO service to allow for the use of both cameras in VIO?

@Manu-Verma , please see the details in the link below. Please note that Modalai ESC calibration is not the same thing as the typical ESC calibration. QGC does not support calibrating ModalAI ESCs and it needs to be done using voxl-esc tools running either directly on VOXL2 or a Linux PC. https://gitlab.com/voxl-public/voxl-sdk/utilities/voxl-esc/-/blob/dev/voxl-esc-tools/calibration.md

@Moderator thank you I appreciate it.

Hello, I am trying to get pretty much the same thing working on the VOXL 2 with a Yolov5 model that only has one class. I am seeing similar outputs when I deploy it to the drone. @Siyu-Chen, what steps did you go through to get this working on your device? The model inferences just fine on my computer. I think it has something to do with the pre/post-processing. Thanks in advance for your help.

@boga If SDLOG_PROFILE doesn't have what you are looking for then you can set up a custom set of log topics / rates from a file. You don't need to recompile PX4 to use it. https://docs.px4.io/main/en/dev_log/logging.html#sd-card-configuration

@Wei-Cui How did you wipe the /data?

@david-devassy I'd recommend sending in an RMA: https://www.modalai.com/pages/rma and we'll get things sorted

Hello LaborzaubererTHI, You can see the video now as it has been set as unlisted. https://www.youtube.com/watch?v=jjCKZmrqrVQ Sorry for that. Regards

@Ajay-Bidyarthy Only the inference time should be affected by your choice of hardware. The other two should be the same regardless of your choice. Thomas

@Eric-Katzfey So you probably just need to do the magnetometer calibration procedure to be ready for flight.

@Eric-Katzfey said in Take-Off and Land command each fly to xy=[0,0] location instead of on the spot: @LaborzaubererTHI Sounds like it may be similar to: https://forum.modalai.com/topic/2533/failsafe-landing-bug-in-px4-1-14/19?_=1713800679314 Indeed it does.

@Dawid-Mościcki Perhaps you can ask the QGC development team for some help in narrowing down what is causing QGC to crash. If you think there is a missing parameter then in your /usr/bin/voxl-px4-start script add a new line to touch the parameter so that QGC will see it. (e.g. param touch SDLOG_MODE)

I reflashed firmware and reconfigured drone manually (not using automatic scripts which platform detection) and problem vanished. Probably there is something wrong with detecting actual drone version in those automated scripts.

If any1 will have similar problems in future don't follow guide : https://docs.modalai.com/sentinel-user-guide-pre-flight-setup/ See those docs : https://docs.modalai.com/voxl-elrs/ All You need to do is voxl-elrs -b, then on aperture go to ExpressLRS and choose bind Can be closed.

@Eric-Katzfey thanks mate for your answer!

@Dawid-Mościcki Wow amazing I will give this a try very soon. Thank you so much!

@Alex-Kushleyev Thank you! This worked perfectly, got all of the motors spinning