Collaborative Harvest Robot

Abstract

Agriculture is a field of activity characterized by physical labor, which is why many efforts are being made to introduce robotics on farms. In the specific case of vineyards, and harvesting, the difficulty lies in approximating the harvesting times of the grape harvester without damaging the product and consequently offering a cost-effective solution. This paper presents an operator tracking strategy for a prototype mobile, collaborative robotic platform to address the harvesting task with a human-robot collaboration that combines the manual dexterity of the harvester with the carrying capacity of a robot. The system designed and developed is applicable in the field of manual fruit picking. A commercial mobile robot platform has been used and a structure adapted to the transport and weighing of the contents of the box has been built. In addition, an RGB-D camera and UWB sensors have been integrated. A system for tracking a specific operator while maintaining a safety distance has been integrated into the robot. The system runs on ROS. The results show that the robot correctly performs the operation of tracking the grape harvester and is robust in the event of major changes in lighting.