Democritus University of Thrace (DUTH)
Development of adaptive fuzzy multiple models based control techniques for nonlinear systems
Abstract
dc:descriptionIn this thesis, some new architectures for the adaptive fuzzy control of uncertain dynamic nonlinear systems based on Takagi-Sugeno (T-S) multiple models are presented. The difficulties that arise from the nonlinear nature of the plants to be controlled are treated by using fuzzy logic techniques. More specifically, by using the T-S fuzzy modeling techniques the nonlinear system can be described by many linear subsystems which are then fuzzy blended in order to express the original nonlinear plant. The main advantage of using T-S models is that they offer to the designer the possibility to utilize linear control techniques in order to control the global nonlinear system. Another advantage of T-S models is that they have the ‘universal function approximation’ property, in the sense that they are able to approximate every smooth nonlinear function to any degree of accuracy in a convex compact region provided that sufficient fuzzy rules are available for the description of the dynamical system. Apart from the nonlinear nature of the systems, there is another factor that imposes difficulties to the control problem solution; that is the partial knowledge of the system. This means that the system must be controlled and remain stable although some of its parameters are unknown. Adaptive control techniques are used to provide the answer to this problem. This thesis focuses on the indirect adaptive control techniques, which means that the T-S fuzzy model of a system is estimated on-line. Although fuzzy indirect adaptive control approaches are very efficient, their main drawbacks are that the use of only a single adaptive identification model, may lead to instability or poor performance problems, especially when the parameter uncertainty is very high, the parameters are changing rapidly and discontinuously and the initial estimations for the original parameters of the plant are highly inaccurate. Motivated by the weakness of some methods which use a single fuzzy identification model, we overcome the aforementioned drawbacks by introducing some T-S multiple model-based adaptive switching control architectures that are suitable for control problems which incorporate the above characteristics. More specifically in the proposed architectures of this thesis, instead of using a single identification model, we use more than one T-S identification adaptive models. One fuzzy adaptive or fixed feedback linearization controller corresponds to each identification model and a switch defines the controller that will provide the input to the system. By using a switching rule, the control signal to be applied is determined at every time instant by the model which best approximates the plant. At every step the parameters of the adaptive identification models and controllers are adapted simultaneously so that the asymptotic stability and a satisfactory performance for the system are ensured. In addition, the parameters adaptation is restricted in a way that there are no singularity problems in the control input. Lyapunov stability theory is used in order to obtain the adaptive law for the multiple models parameters, ensuring the asymptotic stability and the robustness of the system while a modification in this law keeps the control input away from singularities. All the control architectures that are presented in this thesis provide very good control performance and their efficiency is indicated by some simulation results.
Degree
thesis:*- Grantor dc:publisher
- Democritus University of Thrace (DUTH)
- Year dc:date
- 2018
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Sofianos, Nikolaos
- Σοφιανός, Νικόλαος
Subjects
dc:subject × 10- Ασαφής προσαρμοστικός έλεγχος
- Έλεγχος πολλαπλών μοντέλων
- Εύρωστος έλεγχος
- Fuzzy adaptive control
- Multiple models control
- Robust control
- Επιστήμες Μηχανικού και Τεχνολογία
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Engineering and Technology
- Electrical Engineering, Electronic Engineering, Information Engineering
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/43729