| Titre : |
The application of digital adaptive modelling to the identification of system parameters by using Lyapunov functions |
| Type de document : |
texte imprimé |
| Auteurs : |
M. Rima, Auteur ; J. K. M. MacCormac, Directeur de thèse |
| Editeur : |
Bath [Royaume-Uni] : University of Bath |
| Année de publication : |
1984 |
| Importance : |
121 f. |
| Présentation : |
ill. |
| Format : |
30 cm |
| Note générale : |
Mémoire de Master : Electronique : Angleterre, University of Bath : 1984
Bibliogr. f. 134 - 136 . Annexes |
| Langues : |
Anglais (eng) |
| Index. décimale : |
Ms01184 |
| Résumé : |
The purpose of this project is to design a digital adaptive model for the identification of system parameters by using Lyapunov functions.
The structure of the unknown system (plant) is already assumed known or identified.
A learning model approach is used to identify the plant parameters.
It consists of the following elements; the unknown plant, the learning model and the identification mechanism.
The techniques used in this thesis are applicable to higher order time-variant system identification.
However to investigate and develop these methods a second order single input/single output linear plant was simulated for identification purposes.
The learning model is time-varying second order linear differential equation, it is used as a part of the control system itself.
The identification mechanism is derived from a Lyapunov function which ensures stability and at the same time convergence of the learning model parameters to those of the unknown plant.
Applying Lyapunov function the plant is assumed to have an accessible state vector.
In order to achieve digital identification, the explicit time-varying learning model and the identification mechanism are converted from the time-domain to the discrete-domain applying the bilinear transformation and frequency prewarping and then the cascade method.
The micromaster refers to the microcomputer system used.
It is based on a ZILOG Z-80, eight bit, microprocessor, and it is used as a digital controller.
It executes the time-varying digital models which are used for the identification scheme.
The communication between the analog unknown plant and the micro-computer system is achieved using a programmable interface board constructed during the project.
Due to computing speed limitations it was agreed to apply the technique to a plant with a low natural frequency to demonstrate the feasibility of such a technique.
Recent microprocessor development made the identification method suitable for higher bandwidth systems.
The results indicate that Lyapunov technique using adaptive methods are capable of producing sufficiently rapid and accurate identification.
Further research is required on a method of selection of the weighting matrix and convergence constant to ensure optimum identification.
These factors could possibly be reoptimised as a function of the identification plant parameters for subsequent identification of a time-varying plant. |
The application of digital adaptive modelling to the identification of system parameters by using Lyapunov functions [texte imprimé] / M. Rima, Auteur ; J. K. M. MacCormac, Directeur de thèse . - Bath [Royaume-Uni] : University of Bath, 1984 . - 121 f. : ill. ; 30 cm. Mémoire de Master : Electronique : Angleterre, University of Bath : 1984
Bibliogr. f. 134 - 136 . Annexes Langues : Anglais ( eng)
| Index. décimale : |
Ms01184 |
| Résumé : |
The purpose of this project is to design a digital adaptive model for the identification of system parameters by using Lyapunov functions.
The structure of the unknown system (plant) is already assumed known or identified.
A learning model approach is used to identify the plant parameters.
It consists of the following elements; the unknown plant, the learning model and the identification mechanism.
The techniques used in this thesis are applicable to higher order time-variant system identification.
However to investigate and develop these methods a second order single input/single output linear plant was simulated for identification purposes.
The learning model is time-varying second order linear differential equation, it is used as a part of the control system itself.
The identification mechanism is derived from a Lyapunov function which ensures stability and at the same time convergence of the learning model parameters to those of the unknown plant.
Applying Lyapunov function the plant is assumed to have an accessible state vector.
In order to achieve digital identification, the explicit time-varying learning model and the identification mechanism are converted from the time-domain to the discrete-domain applying the bilinear transformation and frequency prewarping and then the cascade method.
The micromaster refers to the microcomputer system used.
It is based on a ZILOG Z-80, eight bit, microprocessor, and it is used as a digital controller.
It executes the time-varying digital models which are used for the identification scheme.
The communication between the analog unknown plant and the micro-computer system is achieved using a programmable interface board constructed during the project.
Due to computing speed limitations it was agreed to apply the technique to a plant with a low natural frequency to demonstrate the feasibility of such a technique.
Recent microprocessor development made the identification method suitable for higher bandwidth systems.
The results indicate that Lyapunov technique using adaptive methods are capable of producing sufficiently rapid and accurate identification.
Further research is required on a method of selection of the weighting matrix and convergence constant to ensure optimum identification.
These factors could possibly be reoptimised as a function of the identification plant parameters for subsequent identification of a time-varying plant. |
|
Ms00112 
Ajouter le résultat dans votre panier Faire une suggestion Affiner la rechercheThe application of digital adaptive modelling to the identification of system parameters by using Lyapunov functions (1984)
