Autonomous flight of an unmanned aerial vehicle (UAV) based on vision SLAM

• Main Author: Li, Jingning
• Other Authors: Low Kin Huat
• Format: Final Year Project
• Language: English
• Published: 2017
• Subjects: DRNTU::Engineering::Aeronautical engineering
• Online Access: http://hdl.handle.net/10356/71039
• Institution: Nanyang Technological University

Much active research has been done to look for new localization methods for robotics like Unmanned Arial Vehicles (UAVs) and increasing research efforts have been spent on computer vision as well as Simultaneous Localization and Mapping (SLAM) over the past 10 years. This paper presents the development process of an Unmanned Aerial System, which uses a camera as the position sensor together with an open-source VSLAM program to obtain position estimation data and an on-board computer for high-level flight control. The first part of the report presents a literature review on computer vision and the selection of open source VSLAM programs. After a combination of an on-board computer, a camera, and the VSLAM program was confirmed, a quadcopter platform was developed to carry all necessary hardware onboard, which includes the vision system, a flight controller unit for lower level controls, remote communication modules, and power supplies for all electrical devices. A higher-level control framework was also developed on the computer which gathers information from various sources and controls the behavior of the UAS. Flight tests were also conducted with the UAS and the test results were discussed in this report. At last, some precautions and advice for future development were presented at the end of the report.