Robust Transport and Assembly Automation System Using Vision-Based Control Using UAV

Abstract

We propose an automated system that moves blocks to a specific location, such as a block map, using Unmanned Aerial Vehicle (UAV). The system was developed in a simulation environment using Gazebo simulator, and the test environment was set with different GPS noise and illuminance in order to resemble the real environment. To increase the accuracy of the system, we used an algorithm that picks up and places blocks using vision-based controls using deep learning-based object detection and recognition, not just GPS based control. In addition, when picking up and placing blocks, it is easy to execute by using a magnet module rather than a typical manipulator. Our proposed system in simulation experiments in various environments consistently showed excellent performance. The system seems to have robust performance in a more difficult environment than existing robots and has simplicity and efficiency in execution. © 2021, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.