User CL Simulation Testing Setup

This setup is very similar to the tutorial on setting up a local environment for development documented in our documentation.

As such you can lookup some of the needed requirements there.

0. Requirements

  • have an LDAP mafiasi account for access to the CLs

  • have ros2 aliases setup (see linked docs)

  • have github ssh access setup for bitbots_meta (see linked docs)

1. Setup and download our software

  • SSH into the cl0* with your mafiasi user

  • setup bitbots_meta on nvme nas share

cd /srv/ssd_nvm/
mkdir -p "$USER/colcon_ws/src"
cd "$USER/colcon_ws/src"
git clone git@github.com:bit-bots/bitbots_meta.git && cd bitbots_meta
make pull-init

  • setup a python virtualenv on the nvme share for faster pip package installation

cd "/srv/ssd_nvm/$USER"
python -m virtualenv env
source env/bin/activate
pip install -r colcon_ws/src/bitbots_meta/requirements.txt

  • set PATH and COLCON_WS (see section 5), but in this case COLCON_WS="/srv/ssd_nvm/$USER/colcon_ws" as we are not working in our $HOME

2. Compile the packages

If while testing you are changing code or updating bitbots_meta via make pull-all, this step needs to be done again. For compilation of the whole meta repository run cba, which is an alias for: cd $COLCON_WS; colcon build --symlink-install --continue-on-error After a successful run, before we are able to use any ros commands we now need to source colcon built sources with sa, which is an alias for: source "/opt/ros/rolling/setup.$SHELL" && source "$COLCON_WS/install/setup.$SHELL"

3. Run Webots Simulation

We can start the Webots simulator with the following command: rl bitbots_bringup simulator_teamplayer.launch game_controller:=false This should start the simulation environment in the Webots simulator, while also starting all necessary nodes of the robot software (walking, vision, etc.). In the simulator we should see a field with a single robot.

With game_controller:=false we ensure, that the game_controller_listener is not started as well, but instead we will simulate the current gamestate by our own script (in another terminal): rr humanoid_league_game_controller sim_gamestate.py

Which allows us to simulate the current gamestate and different phases of the game. Now everything is ready for some simulation testing.

4. Testing and Debugging

By changing the simulated gamestate and seeing how the robot reacts, we can test our behavior. If there are issues with the robots behavior, they most likely have to do with DSD configuration or different parallelism handling in ROS 2. To be able to visualize the current DSD execution, we can start rqt and in the Plugins menu select RoboCup -> DSD-Visualization. This will show us the current DSD execution and can help in finding deadlocks or behavior execution logic issues.

TODO: extend this section