Robotic Warehouse Management System

Abstract

This study presents two approaches to the prevention of inter-robot collisions. The first approach is to develop trajectory planning and motion control algorithms. The second approach is to reduce the number of auxiliary robots as much as possible. The rigidly programmed systems are less flexible and adaptive than systems capable of independent data analysis and pattern identification. Therefore, this study uses the neural network for robot training and an analytical module (AN) to make decisions regarding the quantity of robots. The AN assisted and non-assisted management systems were examined under the two scenarios, namely the steady and random increment of applications. In both scenarios, using the AN reduced the number of auxiliary robots and, consequently, robot collisions in the operating area. This can help to reduce the warehouse maintenance costs and improve manufacturing scalability. Therefore, the proposed robotic management system has the potential to enhance warehouse efficiency.