A Novel Methodology for Constructing Prediction Filters in Wavelet-Transform Domain and Its Application

  • 林 續勳

Student thesis: Master's Thesis

Abstract

A novel methodology for constructing prediction filters in wavelet-transform domain and its novel application to image interpolation is proposed in this thesis Firstly a novel optimal error compensation iterative learning observer-based tracker for the unknown nonlinear system with disturbances is proposed The novelty in this part includes the following newly presented items: 1 a new optimal analog tracker design for the continuous-time system with known disturbances 2 a new digital-redesign linear quadratic tracker (LQT) design for the sampled-data system with known disturbances 3 a new iterative learning LQT design for the repetitive continuous-time system with unknown disturbances 4 a new iterative learning LQT design for the repetitive sampled-data system with unknown disturbances 5 a new LQT design for the discrete-time system with disturbances 6 a new iterative learning LQT design for the repetitive discrete-time system with unknown disturbances 7 a new analog linear quadratic observer (LQO) design for the continuous-time system with known disturbances 8 a new digital-redesign observer design for the sampled-data system with known disturbances 9 a new iterative learning analog LQO design for the repetitive continuous-time system with unknown disturbances 10 a new iterative learning digital LQO design for the repetitive discrete-time system with unknown disturbances and 11 an extended applications of above design methodologies from the given linear systems to the unknown nonlinear system with input time delay Applications of the above mentioned design methodologies on the automatic control area are also given in this thesis to demonstrate their novelty Secondly integrated with the above newly presented methods another novel methodology for constructing prediction filters in wavelet-transform domain and its novel application to image interpolation as well as more detailed study on its characteristics are presented Corresponding illustrative examples are also given in this thesis to demonstrate their effectiveness
Date of Award2014 Sep 1
Original languageEnglish
SupervisorShu-Mei Guo (Supervisor)

Cite this

'