Deep Farmer Robotics

In a world that that is quickly changing from the evolving weather to new requirements or demands for truly organic products with no chemicals, there is a need to innovate quickly in the field of weeding - one of the most important and expensive activities farmers face today.

In this project you will develop a robotic AI agent in AWS Robomaker that kills individual weeds with laser. The agent must be developed in AWS Robomaker for the ROS2 Foxy distribution.

Weed perception system

The perception system of the agent includes a weed detection pipeline powered by this dataset You will use a pipeline similar to this. Your task is to replicate the results presented in terms of performance (eg. F1 score) and author a report explaining the implementation and the challenges you met.

Navigation perception system

TBD

ROS2 Robot Model

You will need to construct the robot model such as this one.

Gazebo environment model

You need to construct the model of the field that the robot will be used on. Generative modeling can be used to create weeds and plants.

Planner

TBD