The WEEDINATOR Project Brings Autonomous Agriculture to the Masses

Modern commercial farming is almost indistinguishable from what it was even just a century or two ago. Not only does it rely on heavy equipment, but it also relies on automation. Autonomous tractors, for example, can work efficiently around the clock, which is becoming ever more critical as the world population increases. How can small family farms keep up? The WEEDINATOR project is an open source effort to bridge the gap.

As the name suggests, the primary focus of the WEEDINATOR project is weeding. The goal is to make an accessible robotic tractor system for weeding operations, which is more affordable and DIY-friendly than the commercial solutions on the market. But though weeding (really cultivation) is the main goal, a WEEDINATOR robot could perform other tasks.

This project has been ongoing for several years now — at least since 2017. But the GOAT INDUSTRIES team just released a new video teasing new developments. That video is a bit on the edgy side, in the Anonymous kind of way. But if you can get past that, there is a lot of promising potential on display.

Most of this focuses on the electronics, rather the vehicle itself—that can be a conventional tractor modified for this purpose. Those electronics are divided into three modules: two separate control boxes and a safety camera system.

The first control box is for low-level control via an STM microcontroller, relays, and motor controllers. The second control box is for high-level control and communication, via a Raspberry Pi or similar SBC with a 4G modem. That second control box also looks at crops through cameras and analyzes them with machine learning models. The safety cameras are for detecting people, which can override everything else and stop the robot immediately.

If you’d like to learn more, you can check out the WEEDINATOR website and the GitHub page.