Development of Autonomous Driving and Docking Control Algorithms for 3D Camera-Based Automatic Guided Vehicle (AGV) System

Abstract

This paper proposes robot driving algorithms that enable an automatic guided vehicle (AGV) used for industrial logistics to recognize the area inside the safety lines and autonomously drive within these lines. In addition, we propose a system in which a robot can distinguish between driving direction signals, docking signals, and pedestrians, to enhance the accuracy of the docking position while moving to the location for delivery or automatic charging. The average error of the arrival point was about 19.7 mm, when this robot moved to the docking position for charging, after autonomous driving. Furthermore, during autonomous driving, when a pedestrian suddenly entered the safety line in front of the robot, the average accuracies of distance recognition and near recognition measurements were approximately 96.3% and 100%, respectively.