Nonlinear Control of Unmanned Aerial Vehicles : Systems With an Attitude
Author
Summary, in English
This thesis deals with the general problem of controlling rigid-body systems through space, with a special focus on unmanned aerial vehicles (UAVs). Several promising UAV control algorithms have been developed over the past decades, enabling truly astounding feats of agility when combined with modern sensing technologies. However, these control algorithms typically come without global stability guarantees when implemented with estimation algorithms. Such control systems work well most of the time, but when introducing the UAVs more widely in society, it becomes paramount to prove that stability is ensured regardless of how the control system is initialized.
The main motivation of the research lies in providing such (almost) global stability guarantees for an entire UAV control system. We develop algorithms that are implementable in practice and for which (almost) all initial errors result in perfect tracking of a reference trajectory. In doing so, both the tracking and the estimation errors are shown to be bounded in time along (almost) all solutions of the closed-loop system. In other words, if the initialization is sound and the initial errors are small, they will remain small and decrease in time, and even if the initial errors are large, they will not increase with time.
As the field of UAV control is mature, this thesis starts by reviewing some of the most promising approaches to date in Part I. The ambition is to clarify how various controllers are related, provide intuition, and demonstrate how they work in practice. These ideas subsequently form the foundation on which a new result is derived, referred to as a nonlinear filtered output feedback. This represents a diametrically different approach to the control system synthesis. Instead of a disjoint controller/estimator design, the proposed method is comprised of two controller/estimator pairs, which when combined through a special interconnection term yields a system with favorable stability properties.
While the first part of the thesis deals with theoretical controller design,
Part II concerns application examples, demonstrating how the theory can solve challenging problems in modern society. In particular, we consider the problem of circumnavigation for search and rescue missions and show how UAVs can gather data from radioactive sites to estimate radiation intensity.
The main motivation of the research lies in providing such (almost) global stability guarantees for an entire UAV control system. We develop algorithms that are implementable in practice and for which (almost) all initial errors result in perfect tracking of a reference trajectory. In doing so, both the tracking and the estimation errors are shown to be bounded in time along (almost) all solutions of the closed-loop system. In other words, if the initialization is sound and the initial errors are small, they will remain small and decrease in time, and even if the initial errors are large, they will not increase with time.
As the field of UAV control is mature, this thesis starts by reviewing some of the most promising approaches to date in Part I. The ambition is to clarify how various controllers are related, provide intuition, and demonstrate how they work in practice. These ideas subsequently form the foundation on which a new result is derived, referred to as a nonlinear filtered output feedback. This represents a diametrically different approach to the control system synthesis. Instead of a disjoint controller/estimator design, the proposed method is comprised of two controller/estimator pairs, which when combined through a special interconnection term yields a system with favorable stability properties.
While the first part of the thesis deals with theoretical controller design,
Part II concerns application examples, demonstrating how the theory can solve challenging problems in modern society. In particular, we consider the problem of circumnavigation for search and rescue missions and show how UAVs can gather data from radioactive sites to estimate radiation intensity.
Department/s
Publishing year
2021-10-18
Language
English
Full text
- Available as PDF - 49 MB
- Download statistics
Document type
Dissertation
Publisher
Department of Automatic Control, Lund University
Topic
- Control Engineering
- Robotics
Keywords
- Nonlinear Control
- Lyapunov Methods
- Aerospace
- Aerial Vehicles
- Output Feedback
- Control Theory
Status
Published
Project
- Semantic Mapping and Visual Navigation for Smart Robots
Supervisor
- Anders Robertsson
- Bo Bernhardsson
- Zhiyong Sun
ISBN/ISSN/Other
- ISBN: 978-91-8039-047-7
- ISBN: 978-91-8039-048-4
Defence date
12 November 2021
Defence time
13:15
Defence place
Lecture hall KC:A, Kemicentrum, Naturvetarvägen 14, Faculty of Engineering LTH, Lund University, Lund. Zoom: https://lu-se.zoom.us/j/66304404918?pwd=OGg2enQ4cGZvZ1lJRkorM3RLZCtydz09
Opponent
- Taeyoung Lee (Prof.)