Large-Model and Generative-Intelligence Agricultural Robot Systems*

Abstract

In order to reduce the problem of excessive reliance on labor force in agriculture and achieve automated intelligent operations, a large-model and generative-intelligence agricultural robot systems (LGARS) is proposed. The structure of LGARS is composed of a perception module, a robot cooperation module, a human-computer interaction module, and a robot system center. By collecting, analyzing, and forecasting data from diverse sensors, the systems generate task and control signals for the generative intelligent agricultural robots. Subsequently, the robots coefficiently complete the assigned tasks. The unique contribution of the proposed LGARS is that the user just needs to send a superior task instruction, and all the other task planning, task generation, autonomous decisions, task decomposition and the specific operation process are conducted by the robot. It can greatly reduce labor costs, improve production efficiency and work accuracy.