Autonomous technologies of global vision-based golf ball collection robot

摘要

This paper presents the autonomous technologies of global vision-based golf ball collection robot. Since the autonomous golf ball collection robot is the very complex system which is composed of mechanical systems, power systems, and electronics systems, it requires many times of trial and errors for the development, which requires very high cost. Our strategy is the derivation of the optimal system architecture through the development of diverse functions on the emulator with a simple robot. For this, we use a global camera, which takes images of the whole ground and transfers to the server in real time. A special mark is installed on the robot to indicate the robot's position and orientation. The optimal paths to scan the dense ball regions are calculated using Traveling Salesman (TS) algorithm. Also, the robot collects balls in each dense ball region along with a spiral scanning trace. Development details and their experimental results are included in this paper.