Borshchova, Iryna (2017) Vision-based automatic landing of a rotary UAV. Doctoral (PhD) thesis, Memorial University of Newfoundland.
[English]
PDF
- Accepted Version
Available under License - The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission. Download (3MB) |
Abstract
A hybrid-like (continuous and discrete-event) approach to controlling a small multi-rotor unmanned aerial system (UAS) while landing on a moving platform is described. The landing scheme is based on positioning visual markers on a landing platform in a detectable pattern. After the onboard camera detects the object pattern, the inner control algorithm sends visual-based servo-commands to align the multi-rotor with the targets. This method is less computationally complex as it uses color-based object detection applied to a geometric pattern instead of feature tracking algorithms, and has the advantage of not requiring the distance to the objects to be calculated. The continuous approach accounts for the UAV and the platform rolling/pitching/yawing, which is essential for a real-time landing on a moving target such as a ship. A discrete-event supervisor working in parallel with the inner controller is designed to assist the automatic landing of a multi-rotor UAV on a moving target. This supervisory control strategy allows the pilot and crew to make time-critical decisions when exceptions, such as losing targets from the field of view, occur. The developed supervisor improves the low-level vision-based auto-landing system and high-level human-machine interface. The proposed hybrid-like approach was tested in simulation using a quadcopter model in Virtual Robotics Experimentation Platform (V-REP) working in parallel with Robot Operating System (ROS). Finally, this method was validated in a series of real-time experiments with indoor and outdoor quadcopters landing on both static and moving platforms. The developed prototype system has demonstrated the capability of landing within 25 cm of the desired point of touchdown. This auto-landing system is small (100 x 100 mm), light-weight (100 g), and consumes little power (under 2 W).
Item Type: | Thesis (Doctoral (PhD)) |
---|---|
URI: | http://research.library.mun.ca/id/eprint/12868 |
Item ID: | 12868 |
Additional Information: | Includes bibliographical references (pages 116-128). |
Keywords: | UAVs, automatic landing, image-based visual servoing, discrete-event supervisory control |
Department(s): | Engineering and Applied Science, Faculty of |
Date: | October 2017 |
Date Type: | Submission |
Library of Congress Subject Heading: | Drone aircraft--Automatic control; Drone aircraft--Design and construction; Landing aids (Aeronautics) |
Actions (login required)
View Item |