Debugging with Foxglove
Last updated
Was this helpful?
Last updated
Was this helpful?
Was this helpful?
This tutorial describes how to perform debugging using Foxglove.
After you have completed the initial setup for CogniPilot on a native ubuntu machine, open a new terminal and run "build_foxglove" to install foxglove on your machines.
Note: If this command is not found, restart your system and try again.
Select "y" when asked for clone repositories using git with already setup github ssh keys.
Select "1" when asked for which release to be used - airy (1) or main (2).
NOTE: When working with hardware buggy, the debug topics must be added to the default_value of topic_whitelist in b3rb_robot/b3rb_bringup/launch/robot.launch.py as well, as shown in the steps below. Then build, source and run. The changes made on cranium must be done on NavQPlus and not the UBUNTU HOST MACHINE.
Camera Images: We can publish debug images from "b3rb_ros_line_follower/b3rb_ros_line_follower/b3rb_ros_edge_vectors.py". For publishing debug images from other ROS2 nodes, follow similar process.
Create a new publisher in the init function as shown in the attached image. We need to create a new topic and in our case we have chosen "/debug_images/vector_image".
Now, pass this publisher (self.publisher_vector_image) and the image as arguments to the "publish_debug_image" function whenever you want to output your debug images.
Open a new terminal and run the following commands.
cd ~/cognipilot/cranium
colcon build
source install/setup.bash
Append the same topic name "/debug_images/vector_image" to the default_value of topic_whitelist in electrode/src/electrode/launch/electrode.launch.py as shown in the attached image.
When working with hardware buggy append the same topic name "/debug_images/vector_image" to the default_value of topic_whitelist in b3rb_robot/b3rb_bringup/launch/robot.launch.py. The snippet of this argument is attached in the image below.
Open a new terminal and run the following commands.
Open a new terminal and run the following commands.
cd ~/cognipilot/electrode
colcon build
source install/setup.bash
LIDAR Data:
Append "/scan" to the default_value of topic_whitelist in electrode/src/electrode/launch/electrode.launch.py as shown in the attached image.
When working with hardware buggy append "/scan" to the default_value of topic_whitelist in b3rb_robot/b3rb_bringup/launch/robot.launch.py. The snippet of this argument is attached in the image below.
Open a new terminal and run the following command after starting the simulation (which is done using the command "ros2 launch b3rb_gz_bringup sil.launch.py world:=<track_name>")
ros2 launch electrode electrode.launch.py sim:=True
Then open connection to the url "ws://localhost:8765"
Camera Images.
Create a new Image panel as shown in the attached image.
Then click on panel settings and select the topic for which you created a publisher in the "prerequisites" section.
LIDAR Data.
Create a new 3D panel by clicking on the add panel button from the top-left corner as shown in the attached image.
colcon build
source install/setup.bash
Open a new terminal and run the following commands.
cd ~/cognipilot/cranium
colcon build
source install/setup.bash
Open a new terminal and run the following commands.
cd ~/cognipilot/electrode
colcon build
source install/setup.bash
Now this panel will display the image whenever you publish a debug image from your code to the said topic.
You may switch to 2D camera since the LIDAR is 2D.
Then click on panel settings.
Toggle visibility of "/scan" under Topics.
Then select "lidar_link" in Display frame under Frame.
If you don't immediately see the "lidar_link" option available, then close the settings panel, play the simulation for a few seconds, then try again.
You may close the panel settings after this.
Please see the attached images for sample output.
