Open-source robot mower

A lawn mower that plans its own route.

LawnBerry Pi is an autonomous lawn mower platform for the Raspberry Pi. It steers by centimeter-level RTK GPS, follows missions you draw on a map, and reports back through a web dashboard styled like a 1980s arcade cabinet.

View the code on GitHub Project status

What it does

Mission planner

Drop waypoints on a map, reorder them, and set blade and speed for each one. Start the mission, then pause, resume or abort it at any time.

Live telemetry

GPS position, battery state and IMU readings stream to the dashboard five times a second over a WebSocket.

RTK positioning

GPS corrections arrive over NTRIP, which brings position accuracy down to the centimeter range.

Simulation mode

The whole stack runs without any hardware attached, so routes and software changes can be tried at a desk first.

Hardware

  • Raspberry Pi 5 as the main computer, with reduced-feature support for the Raspberry Pi 4B.
  • RoboHAT (RP2040) motor controller running its own firmware.
  • I2C sensors for GPS, IMU and battery monitoring.
  • Camera with a streaming service for a view from the mower.

Software

  • Python and FastAPI backend with a documented REST and WebSocket API.
  • Web dashboard served from the mower itself.
  • systemd services for running on the device.
  • Raspberry Pi OS Bookworm (64-bit) as the target system.

Project status

LawnBerry Pi works today for development, simulation and supervised testing on the mower. It is not yet rated for unattended mowing and is not a finished product. Missions that run the blade stay locked until the mower passes the final stage of its safety qualification, and blade tests are off by default.

This is a personal hobby project. It is not sold, and nothing here is an offer of a product or service.

Google account access

The project includes a small set of companion notebooks that run in Google Colab. They use an app named LawnBerry Pi to sign in with Google and work with files in the signed-in person's own Google Drive: reading files they choose, and saving files back, so that large project files can be moved between Drive and other storage that person controls.

The app runs entirely inside the user's own Colab session. It has no server of its own, and the project's author never receives your files or your Google account details.

Read the privacy policy for the full details, including how to remove the app's access.