Robotic arm Object Detection System

Abstract

Most of the automation technology in today's factories produces fixed products over fixed production lines; this is called fixed automation. However, many companies today have short product cycles, many product varieties, and small batch sizes, and so fixed automation is not a practical solution. Therefore, our research team aims to develop a smart robotic arm system that can work in flexible automation. Towards that goal, this article describes an innovative gripping system for the screwdriver. The screwdriver gripping system uses faster region-based convolutional network to detect the gripping target, and it uses image processing methods to find the best gripping point of the object; the next command then gives the arm its control decision.