Design and Development of a Visual - SLAM based Automated Guided Vehicle

Abstract

The automobile industry is currently undergoing fast change as big manufacturers devote their attention to the development of future self- driving vehicles or automated guided vehicles. As a result, mobile robots and current automotive technologies are integrating to create the area of autonomous navigation. The camera sensor is the one garnering the most attention in terms of autonomy sensor research due to the large amount of information it provides for object detection and classification algorithms in addition to depth, coloring, size, and texture recognition. In the last ten years, there have been significant advancements in computing capabilities and image analysis algorithms. The requirement for developing low-level motion controllers grew as a consequence of the increase in machine vision performance and capabilities. This article discusses mecanum wheel mobile robotic applications of vision and SLAM-based motion controller. When using sensors like cameras or Lidar, autonomous mobile robots usually use the vision- based & simultaneous localization and mapping (SLAM) technology. The four-wheeled autonomous robot model that was utilized in the studywas created using a Jetson board with high performance processor, a depth camera, and LIDAR sensors along with ROS's RVIZ software.