Aristotle University Of Thessaloniki (AUTH)
Έλεγχος συστημάτων υψηλής αβεβαιότητας με ανάδραση εξόδου
Abstract
dc:descriptionOutput feedback control of unknown, nonlinear systems comprises a scientific challenge nowadays, since most control systems appearing in nature are highly uncertain and nonlinear, while in real-world applications, the full-state measurement assumption is rarely satisfied. Therefore the need to design new, efficient control strategies, using the minimum possible information regarding the system structure and the signals available for control, naturally arises. In the case of full state measurement, a plethora of efficient methodologies that use powerful mathematical tools such as Lyapunov theory, has led to constructive solutions surpassing the complexity of the structure of the controlled systems. However, during the last decades, although state feedback control of nonlinear systems evolves rapidly, respective progress in output feedback control has been much slower. The main reason is that in nonlinear systems, unlike the linear case, the well known separation principle is not directly applicable. Therefore, designing nonlinear observers that produce accurate estimates of the unknown states and successfully incorporating them in the closed loop is a difficult task and refers to very restrictive classes of nonlinear systems. Furthermore, in the case of completely unknown systems, de\-si\-gning an asymptotically convergent observer is extremely demanding since no information exists for the system structure. It is therefore clear that the field of output feedback control of unknown nonlinear systems includes many open research issues. The most important is that as the class of the nonlinear systems becomes more general, the complexity of the respective control schemes is increased, while their effectiveness is decreased. The design and implementation of asymptotically convergent observers becomes harder and their integration in the closed loop system, while ensuring the stability of the overall design, very pretentious. In this thesis we focus on the solution of the output feedback control of unknown nonlinear systems problem. Our aim is on one hand to generalize the class of the considered systems met in the literature, on the other to enforce desired behaviour on the control system and ensure closed-loop robustness against uncertainties originating from the same or the environment. This is achieved using two basic techniques: a) output feedback passivation, b) designing and including nonlinear filters on the closed loop. On the first part of the thesis, solution of the output feedback control problem of unknown nonlinear systems is obtained using the well known Passivity Based Control (PBC) theory. Subsequently, the control scheme is robustified with respect to unmodelled dynamics and is then redesigned in order to guarantee convergence of the output tracking error to a predefined arbitrarily small residual set, with convergence rate no less than a prespecified value, while exhibiting maximum overshoot less than a sufficiently small preassigned constant (Prescribed Performance Control - PPC), in the presence of additive disturbances and dynamic uncertainties. On the second part, aiming at the expansion of the class of the considered systems, an alternative approach is presented for the considered problem by neglecting output feedback strict passivity assumption and including specially designed dynamical systems (filters) on the closed loop. Specifically, a PPC output feedback control design is presented for nonlinear, partially unknown strict feedback systems. Finally, the problem is generalized for the class of completely unknown, affine in the control multi-input multi-output nonlinear systems. Extensive simulation studies, on both mathematical and physical systems that are widely used in the relative literature, clarify and verify the theoretical findings. Throughout the thesis, in order to cope with increased level of uncertainties, linear-in-the-weights neural networks are used as universal approximators.
Degree
thesis:*- Grantor dc:publisher
- Aristotle University Of Thessaloniki (AUTH)
- Year dc:date
- 2012
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Kostarigka, Artemis
- Κωσταρίγκα, Άρτεμις
Subjects
dc:subject × 14- Προσαρμοστικός έλεγχος
- Ανάδραση εξόδου
- Νευρωνικά δίκτυα
- Adaptive control
- Output feedback
- Neural networks
- Φυσικές Επιστήμες
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Επιστήμες Μηχανικού και Τεχνολογία
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Natural Sciences
- Computer and Information Sciences
- Engineering and Technology
- Electrical Engineering, Electronic Engineering, Information Engineering
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/31020