@Alex-Kushleyev This worked! Thank you

Hello @itzikwds, Sorry for the delay. We can extend pympa to support the ai_detection class. Here is as example of an application that receives the detection data : https://gitlab.com/voxl-public/voxl-sdk/utilities/voxl-mpa-tools/-/blob/dev/tools/voxl-inspect-detections.c However, since you need this in python, pympa would need to be extended to support the detection data. Can you please confirm that you still need this functionality or did you find another solution? Alex

@tom Thank you so much. It is working now with your suggestion. I think the README file of voxl-tflite-server need to be updated.

@Ege-Yüceel Wow, this is a very ambitious plan! Keep in mind that VIO needs not only camera images but also IMU input. Camera images are normally supplied by the voxl-camera-server on MPA pipes. So, you would need to stop voxl-camera-server and start your own application that takes the images from Gazebo and places them into correctly named MPA pipes. The images would have to be formatted properly, etc. Likewise for IMU samples that normally come from voxl-imu-server. I think this will be quite difficult to get working properly but theoretically possible.

Solved using SDK 1.4

Hi @taiwohazeez, The " /fmu/in" prefix is used for sending data to the autopilot - you're never going to receive data from the autopilot listening to that topic. Look at the "/fmu/out" topics to see messages coming from the autopilot. Best, Ted

Hi @teddy-zaremba, The fusing part was a bit tricky, but it works now. Thank you!

Hi @Judoor-0, Cool! Feel free to make a MR or post the link to your code if you want someone to review it.

@Jeffrey-Mao Hello Jeffery, We are currently working on implementing SLAM for the Starling 2 platform. Our initial approach involved using the available ROS 2 topics provided by voxl_mpa_to_ros2, installing standard SLAM packages, and using RViz for visualization. However, we've encountered a major issue: most of the topics do not have a frame_id defined, and there is no existing transform (TF) tree that can be used with RViz. So far, we've only been able to visualize the tof_pc (Time-of-Flight point cloud) data. To do this, we had to manually publish a static transform between the TOF frame (which is named world) and map. However, in this setup, the drone's location does not update dynamically. To address this, we're considering writing a transformation node to define the relationships between coordinate frames (e.g., map, odom, base_link, tof, etc.). However, developing this from scratch is proving to be time-consuming and error-prone. I also noticed that you were able to generate a map using voxl-mapper and et the map in ROS2 with the mpa_ros2 interface. This could be a much more efficient path for us to pursue. Could you please explain in detail: What exact steps you followed to export the map using voxl-mapper? Is it possible to regenerate or update the map in real-time in RViz, or export it to another visualization tool? Thanks

After some additional investigation I was able to resolve this by replacing the include: #include <c_library_v2/common/mavlink.h> with: #include <c_library_v2/development/mavlink.h> in actuator_test.cpp

@JackHammer no problem. This should be the esc param file for the Starling 2 Max motor : https://gitlab.com/voxl-public/voxl-sdk/utilities/voxl-esc/-/blob/dev/voxl-esc-params/D0012_Starling_MAX/esc_params_m0129_f2203_5_1500kv_7x3.7x3.xml The param file should also be present in /usr/share/modalai/voxl-esc-params Do you know how to update the esc parameters? systemctl stop voxl-px4 cd /usr/share/modalai/voxl-esc-tools ./voxl-esc-updload-params.py -p <full path to params file> After updating the params, you can run a "power ramp test", as shown here : https://gitlab.com/voxl-public/voxl-sdk/utilities/voxl-esc/-/blob/dev/voxl-esc-tools/doc/low_kv_motor_tuning.md?ref_type=heads#power-ramp-test . Use id=255 to spin all motors (and you can limit power to 80%) ./voxl-esc-spin.py --id 255 --power 80 --ramp-time 3.0 --timeout 3.5 --enable-plot 1 --cmd-rate 250 Since you can't display the plot on voxl2, a plot will be saved as html, which you get download from voxl2 and view in a browser. You can do it without propellers (for safety). If all the motors are working properly, the four rpms will be very close to the same as the power ramps (within 100-200 rpm). Otherwise, you may have a bad motor that was damaged during the crash. https://www.modalai.com/products/starling-2-max-replacement-parts?variant=49704826896688 Alex

@Riley-Dove would you have any insight on this? Thanks!

@GlennTee Could you please describe how did you connect to VSCode with using the SSH Client extension?

@Judoor-0 You can use QGroundControl v5 and later. Then once your drone coots and connected to the QGC, mouse click on the desirable position and select Set Estimator Origin option. This will be your [0,0,0] in local frame and [lat, lon, alt] in global [image: 1752642917216-24ef863e-534d-4894-b99d-d1cacd609954-untitled.jpg]

Hi @Ege-Yüceel, We usually create a volume to those headers when doing something like this but you can install the libmodal pipe package inside the docker container like this if you want. wget http://voxl-packages.modalai.com/dists/qrb5165/sdk-1.5/binary-arm64/libmodal-pipe_2.13.1_arm64.deb dpkg -i libmodal-pipe_2.13.1_arm64.deb **Note change sdk-1.5 to whatever sdk version you're on

@teddy-zaremba Nevermind, I was able to separate the images by modifying the voxl-camera-server.conf file to treat each camera as separate and it seems to be working now