An End-to-End Multi-Robot Framework for Weed Control in Agricultural Fields

Weed management is one of the major concerns for crop production in agricultural fields. In this work, we developed a novel multi-robot framework to control weeds in an agricultural field using two groups of mobile robots and a base station. Our framework consists of three modules: (i) the coverage module will take the field boundary as input and generate an optimal coverage path for a group of survey robots; (ii) the weed detection module will collect images of field locations along the coverage path, analyze the collected images, and return the weed locations; and (iii) the planning module will compute an optimal path and send commands to a group of sprayer robots to selectively spray herbicides at the weed locations. All three modules will run on the base station in a step-by-step manner.