Show off any cool projects you’re working on!
Guidelines:
- Please keep it to one thread per project.
- Remember that image uploads are not allowed, please link to external sites instead (e.g. imgur).
Show off any cool projects you’re working on!
Guidelines:
My robot chassis for articubot using laser cut acrylic and 3d printed parts
My differential drive robot using ROS 2 Jazzy, with Raspberry Pi 5.
repo: https://github.com/YJ0528/minibot
Hope this helps those who are trying ROS 2 Jazzy as well.
I have a repository based on Articulated Robotics code, ported to Ubuntu 24.04, ROS Jazzy, Gazebo Harmonic.
The SIM part can be easily run on a desktop Ubuntu 24.04 computer, here are instructions:
It is work in progress and is frequently updated.
ike a lot of you, I learned ROS 2 partly through Josh’s videos. Along the way I built up a notepad full of commands: which workspace to source, which launch file, which args, which terminal goes first. And whenever something didn’t work, I lost an evening finding out it was a stale overlay, a QoS mismatch or a missing /clock.
So I built ROSHub, a small local web dashboard (localhost:8090) that does two things:
1. Manages your ROS 2 projects. Point it at a GitHub URL or a local folder and it finds the packages and launch files statically, without running any of your code. It also detects multiple workspaces (e.g. humble_ws + jazzy_ws) and works out which simulator you’re using. It then gives you a step-by-step bring-up: a clean colcon build, dependency checks (it shows you the exact rosdep/apt command and never installs silently), and readiness gates that wait for /clock, topics and ports before moving on.
2. Helps you debug. The Project Doctor runs a quick (~2 s) or deep (~5 s) check covering build state, environment overlays, RMW/Zenoh, orphaned or duplicate simulators, QoS mismatches, /clock, /odom and the TF tree. There’s also a “Can it move?” probe: it finds who’s listening for velocity commands, works out Twist vs TwistStamped, drives gently (0.1 m/s for 1.5 s) and checks that /odom actually changed. If something’s wrong, one click exports a Markdown report you can paste into an issue, a forum post like this one, or an LLM.
It’s pure Python (stdlib + PyYAML), so there’s nothing heavy to install. There’s a verified example profile for TurtleBot3 + Gazebo Harmonic and an attach profile for Isaac Sim.
It’s an early alpha (0.1.0-alpha), so I’d really value:
Thanks for reading, and happy to answer questions!