Onboard Robust Nonlinear Control for Multiple Multirotor UAVs

DSpace Repositorium (Manakin basiert)


Dateien:

Zitierfähiger Link (URI): http://hdl.handle.net/10900/83028
http://nbn-resolving.de/urn:nbn:de:bsz:21-dspace-830289
http://dx.doi.org/10.15496/publikation-24419
Dokumentart: Dissertation
Erscheinungsdatum: 2018-07-12
Sprache: Englisch
Fakultät: 7 Mathematisch-Naturwissenschaftliche Fakultät
Fachbereich: Informatik
Gutachter: Zell, Andreas (Prof. Dr.)
Tag der mündl. Prüfung: 2018-06-26
DDC-Klassifikation: 004 - Informatik
Schlagworte: Robotik , Flugkörper , Robuste Regelung , Planung , Nichtlineare Regelung
Freie Schlagwörter: Nichtlineare Regelung
Flugroboter
Formationskontrolle
Trajektorienplanung
Robuste Regelung
UAV
Aerial Robotics
Nonlinear Control
Robust Control
Trajectory Planning
Formation Control
Lizenz: http://tobias-lib.uni-tuebingen.de/doku/lic_mit_pod.php?la=de http://tobias-lib.uni-tuebingen.de/doku/lic_mit_pod.php?la=en
Gedruckte Kopie bestellen: Print-on-Demand
Zur Langanzeige

Inhaltszusammenfassung:

In dieser Arbeit konzentrieren wir uns auf die Entwicklung und Validierung von robusten nichtlinearen On-Bord Steuerungsansatzen für mehrere unbemannte Multirotor-Luftfahrzeuge (UAVs), mit dem Ziel, nicht triviale Aufgaben zu erledigen wie z.B. Wegfolge mit aggressiven Manovern, Navigation in komplexen Umgebungen mit Hindernissen und Formationsflug in einer Gruppe. Um diese anspruchsvollen Missionen zu erfullen liegt unser Hauptaugenmerk bei der Stabilität der Flugsteuerung für aggressive UAV Manöver mit steilen Lagewinkeln. Des weiteren ist Kontroll-robustheit sehr wünschenswert, um die Multirotor-UAVs unter Beeinflussung sicher und genau zu steuern. Daruber hinaus ist die Effizienz des Kontrollalgorithmus ein wichtiges Element für die Onboard-Implementierung mit eingeschrankter Rechenfähigkeit. Abschliessend ist das Potenzial, gleichzeitig eine Gruppe von UAVs in stabiler Weise zu kontrollieren, erforderlich. All dies motiviert uns zur Arbeit an den folgenden Aspekten: Zuerst behandeln wir das Problem der robusten Steuerung nichttrivialer Manöver eines Multirotor UAV unter Störeinfluss. Ein komplettes Framework wird entwickelt, welches dem UAV ermöglicht diese anspruchsvollen Aufgaben zu bewältigen. Es beinhaltet einem nichtlinearen Lageregler, basierend auf der Lösung von globalen Ausgangsrege lungsproblemen für Starrkörperrotationen SO(3), einem backstepping basierten Positionsregler, einen sechsdimensionalen (6D) wrench observer um die unbekannten Kraftund Drehmomenteinflusse zu schätzen, sowie einem Online-Trajektorienplaner basierend auf Model Predictive Control (MPC). Wir weisen die starken Konvergenzcharakteristiken der vorgeschlagenen Methode nach, sowohl in der Theorie als auchmittels intensiver Real-roboter-Experimente, mit aggressiver Wegpunktnavigation und Wegfindungsaufgaben in extremer Fluglage in Gegenwart externer Einflüsse, z.B. Windböen. Als nächstes bearbeiten wir das Problem der autonomen Navigation eines Multirotor UAV in komplexen Szenarien. Wir stellen einen effektiven und robusten Steuerungsansatz dar, nämlich eine schnelle MPC-Methode mit der Einbeziehung von nichtlinearer Einschränkungen zur Hindernisvermeidung, und implmenetieren diese an Bord des UAV mit 50Hz. Der entwickelte Ansatz ermöglicht die Navigation eines Multirotor UAVs in 3D-Umgebungen mit mehreren Hindernissen, wobei autonom entschieden wir, über oder um die zufällig gelegenen Hindernisse zu fliegen. Das dritte Problem, das in unserer Arbeit angesprochen wird, ist die Bildungssteuerung für eine Gruppe von Multirotor UAVs. Wir lösen dieses Problem, indem wir einen verteilten Formationskontrollalgorithmus für mehrere UAVs auf der Grundlage der Lösung des Retraction Balancing Problems vorschlagen. Der Algorithmus bringt die ganze Gruppe von UAVs gleichzeitig auf eine vorgeschriebene Untermanigfaltigkeit, welche die Formation in asymtotisch stabiler Weise in 2D- und 3D-Umgebungen bestimmt. Wir validieren unseren vorgeschlagenen Algorithmus uber eine Reihe von Hardware-in-the- ¨ Loop-Simulationen und Real-Roboter-Experimente mit verschiedenen Formationsvarianten in beliebigen zeitveränderlichen (z. B. expandierenden, schrumpfenden oder bewegten) Formen. In den eigentlichen Experimenten wurden bis zu 4 Multirotoren eingesetzt, um beliebige dreieckige, rechteckige und kreisförmige Formen zu bilden, die vom Bediener über eine Mensch-Roboter-Interaktionsvorrichtung vorgezeichnet wurden. Wir haben auch virtuelle Tests mit bis zu 6 Onboard-Computern durchgeführt, um eine sphärische Formation und eine Formation zu erreichen, die sich durch Hindernisse. bewegt.

Abstract:

In this thesis, we focus on developing and validating onboard robust nonlinear control approaches for multiple multirotor unmanned aerial vehicles (UAVs), for the promise of achieving nontrivial tasks, such as path following with aggressive maneuvers, navigation in complex environments with obstacles, and formation in group. To fulfill these challenging missions, the first concern comes with the stability of flight control for the aggressive UAV maneuvers with large tilted angles. In addition, robustness of control is highly desired in order to lead the multirotor UAVs to safe and accurate performance under disturbances. Furthermore, efficiency of control algorithm is a crucial element for the onboard implementation with limited computational capability. Finally, the potential to simultaneously control a group of UAVs in a stable fashion is required. All of these concerns motivate our work in this thesis in the following aspects. We first propose the problem of robust control for the nontrivial maneuvers of a multirotor UAV under disturbances. A complete framework is developed to enable the UAV to achieve the challenging tasks, which consists of a nonlinear attitude controller based on the solution of global output regulation problems for the rigid body rotations SO(3), a backstepping-like position controller, a six-dimensional (6D) wrench observer to estimate the unknown force and torque disturbances, and an online trajectory planner based on a model predictive control (MPC) method. We prove the strong convergence properties of the proposed method both in theory and via intensive real-robot experiments of aggressive waypoint navigation and large-tilted path following tasks in the presence of external disturbances, e.g. wind gusts. Secondly, we propose the problem of autonomous navigation of a multirotor UAV in complex scenarios. We present an effective and robust control approach, namely a fast MPC method with the inclusion of nonlinear obstacle avoiding constraints, and implement it onboard the UAV at 50Hz. The developed approach enables the navigation for a multirotor UAV in 3D environments with multiple obstacles, by autonomously deciding to fly over or around the randomly located obstacles. The third problem that is addressed in our work is formation control for a group of multirotor UAVs. We solve this problem by proposing a distributed formation control algorithm for multiple UAVs based on the solution of retraction balancing problem. The algorithm brings the whole group of UAVs simultaneously to a prescribed submanifold that determines the formation shape in an asymptotically stable fashion in 2D and 3D environments. We validate our proposed algorithm via a series of hardware-in-the-loop simulations and real-robot experiments in various formation cases of arbitrary time-varying (e.g. expanding, shrinking or moving) shapes. In the actual experiments, up to 4 multirotors have been implemented to form arbitrary triangular, rectangular and circular shapes drawn by the operator via a human-robot-interaction device. We have also carried out virtual tests using up to 6 onboard computers to achieve a spherical formation and a formation moving through obstacles.

Das Dokument erscheint in: