UWU-Blobfish

Code for a submersible.

Installation

If using the Nvidia Jetson or a Nvidia GPU, using the ISAAC ROS Docker Environment method is recommended. Otherwise, refer to the documentation on Google Drive.

ISAAC ROS Docker Environment

Ensure Docker for your OS is installed. For recent versions of Windows and Docker, Nvidia GPU support should be built-in. For Linux, follow the Nvidia Docker installation instructions.

In a terminal (current directory doesn't matter), run:

./activate.sh

This will open a shell inside the Docker container with a pre-defined environment, and the entire repository mounted as /workspaces/isaac_ros-dev. Subsequent runs of activate.sh re-use the same container instead of creating a new one. To reset the environment, remove the container by running the following command on the host machine:

docker rm -f isaac_ros_dev-container

Note: Resetting the environment is necessary if any Dockerfile is modified, or if CONFIG_IMAGE_KEY inside .isaac_ros_common-config is changed, to rebuild the container so that changes take effect.

Note: For an alternative workflow using VSCode Dev Containers (which support Python Intellisense), you can use "> Dev Containers: Attach to Running Container...", see: https://code.visualstudio.com/docs/devcontainers/attach-container.

Native ROS2

Refer to the documentation on Google Drive.

Developer Guide

New ROS2 packages should be added to src/.
Third-party packages should be added via git submodule to thirdparty/.
- Fork third-party packages to our org if you need to make modifications.
Arduino code should be added to arduino/.
Create a file named COLCON_IGNORE (i.e., touch COLCON_IGNORE) in folders containing ROS2 packages you don't want colcon to build.

Folder Structure

See respective README.md files in each folder for more information.

arduino/: Arduino code.
docker/: Additional image layers for the ISAAC ROS Docker Environment.
src/: First-party ROS2 packages.
thirdparty/: Third-party ROS2 packages added as git submodules.

Camera Stuff

Cause the realsense viewer doesn't work on my (J-H) container for some reason, plus we now have that totally crap oversaturation issue with the D455 (see last page of https://www.intelrealsense.com/download/13629/), we are using a regular usb cam. So here's the ros2 usb cam node:

sudo apt install ros-humble-usb-cam
ros2 run usb_cam usb_cam_node_exe --ros-args -p video_device:=/dev/video4

# Video recording node
ros2 run debug_cv record_vid --ros-args -p img_topic:=/image_raw

# CV testing mode (using your webcam)
ros2 launch debug_cv test.launch.py

See literally the source code for the parameters: https://github.com/ros-drivers/usb_cam/blob/70ed391a979287bad056c9e75bad8c2001a98f2b/src/ros2/usb_cam_node.cpp#L65-L85

Surprisingly, most of the node's parameters can be changed live via rqt, so that's a good way to figure out what does what and calibrate.

SOAR-UWU / UWU-Blobfish