Summary

• add a new tutorial under ros/ros2_jazzy_turtlebot3_diffdrive_pixi for an end-to-end ROS 2 Jazzy + pixi workflow
• include a headless TurtleBot3 Gazebo launch file and a live diff-drive calibration node focused on wheel modifier parameters
• provide pixi tasks for simulation, calibration, teleop/commanding, MCAP bag recording, Foxglove bridge, and MCAP replay using foxglove_mcap_player
• add helper scripts to clone/build foxglove_mcap_player and replay a user-selected MCAP file
• register the tutorial in the repository root README.md (generated via .utils/generate_readme.py)

Details

This tutorial is designed around a practical loop:

1. Start headless simulation (pixi run sim)
2. Run live calibration (pixi run calibrate-sim)
3. Record MCAP bag (pixi run bag-record)
4. Replay with foxglove_mcap_player (pixi run mcap-play with MCAP_FILE=...)
5. Calibrate replayed command stream live in Foxglove (pixi run calibrate-replay)

The calibration node exposes ROS parameters:

• wheel_radius_multiplier
• wheel_separation_multiplier
• cmd_timeout

and pre-compensates Twist commands so these wheel geometry modifiers can be tuned live in Foxglove.

Notes

• tutorial keeps generated artifacts out of git via local .gitignore (bags/, external/, __pycache__/)
• foxglove_mcap_player is built from source because it is not available as a RoboStack Jazzy package
• TurtleBot3 Jazzy Gazebo uses static SDF diff-drive wheel values; calibration is performed at runtime by parameterized command pre-compensation