autonomous-ai/personal-ai-robot

Personal AI Robot is a build log for a small mobile robot put together from off-the-shelf parts: a Kobuki driving base, a Raspberry Pi 5, a Slamtec RPLidar 2D scanner, and an Orbbec Astra depth camera. The repository holds the hardware notes, wiring diagrams, and setup steps the author followed to take the platform from a box of parts to something that can drive itself around a room. The robot runs ROS 2 Jazzy on Ubuntu 24.04 on the Pi, with the sensors and the base all plugged in over USB. A workstation on the same WiFi network runs the rviz2 visualization tool so the operator can watch what the robot sees. One wiring note repeated in the README: do not draw Pi power from the Kobuki USB-B port, because it cannot supply the 5 amps the Pi needs. Documentation is split into eight numbered sections covering hardware assembly, the Pi 5 install with ROS 2, building the Kobuki workspace, getting the lidar and depth camera running, a first drive test, deep-learning options on the Pi, troubleshooting, and optional remote access through Tailscale and NoMachine. Several of these sections are marked as still to be updated. A separate bill of materials page lists every part with photos and quantities. The headline items are the Kobuki base, an 8 GB Raspberry Pi 5, the RPLidar, the Astra camera, and a 12 volt to 5 volt step-down converter that powers the Pi over USB-C. The author also clones the project with git submodules so that the Kobuki and sensor driver source code lands under the external folder. The README states the setup has been tested on specific firmware and hardware versions, including the Pi 5 8 GB, Ubuntu 24.04.3 LTS arm64, Kobuki firmware 1.2, RPLidar C1, and a specific Astra serial number. The credits thank the kobuki-base maintainers along with Slamtec and Orbbec for their driver SDKs.