danilo-pejovic
commented
1 year ago Hey,
glad you are learning ROS and hopefully RAE can help out with that. I would suggest you start with tutorials from ros2 documentation - it shouldn't take you too long and you will get most of the stuff you need from them (understanding topics, services, nodes,...). Other than that, knowledge of Python and C++ are also necessary to really start learning ROS - Python is decent for prototyping but to get real performance out of nodes C++ is very important. I would also suggest playing around with gazebo - it is relatively simple simulation environment that works relatively well with ROS, at least in my case working in simulation environment helped me a lot at the start.
As far as RAE is concerned I would just start with very basic examples shown on README. Assuming you disabled the agent (so it is not fighting for hardware resources with ros stack) following command:
ros2 launch rae_bringup robot.launch.py
Should run whole stack inside prebuilt docker images you can download. When cameras are run you should be able to see image streams both on the robot and on your host computer. Editing this config file inside that image will let you change parameters you want to use for cameras (for example by default we will just publish streams of 2 out of 4 cameras). Most of the information about camera config can be found on depthai-ros repo as RAE pipeline are another version of those pipelines - that repo can also be very useful when learning ROS (especially bit more advanced concepts) things such as composable nodes are very important part of getting best performance out of ROS2.
I would suggest connecting to a robot via USB while you are learning, to avoid messing to much with network but obviously as you would want to drive it around you can connect to it via AP.
Where do I begin? I'm a senior engineer (learning python) but I'm unsure how to begin.