DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | Lee, Soo-Young | - |
dc.contributor.advisor | 이수영 | - |
dc.contributor.author | Lee, Min-Ho | - |
dc.contributor.author | 이민호 | - |
dc.date.accessioned | 2011-12-14 | - |
dc.date.available | 2011-12-14 | - |
dc.date.issued | 1995 | - |
dc.identifier.uri | http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=101751&flag=dissertation | - |
dc.identifier.uri | http://hdl.handle.net/10203/36306 | - |
dc.description | 학위논문(박사) - 한국과학기술원 : 전기및전자공학과, 1995.8, [ xii, 138 p. ] | - |
dc.description.abstract | Based on the approximating and the generalizing capabilities of the feedforward multilayer neural networks, many methodologies of the neuro-control have been developed for identification and control of the nonlinear dynamic systems with unknown charactreristics. Although these approaches give a possibility for construction of an efficient controller of a system which is difficult to be controlled by the conventional methods, it is still necessary for the neuro-control methods to contain the control concepts such as stability, robustness, and optimality for real applications. On the other hand, a structured information of a system may be available in real situations, and it is more benefit to use the information for construction of the system controller. In order to handle not only the unstructured numerical data but also the structured informations, a new neural network model must be developed by incorporating the fuzzy logic. In this dissertation, new identification and control methods using neural and neuro-fuzzy networks are developed for real applications, which depend on a priori informations of the systems. The developed control methods satisfy the stability with neural system identifier and controller training algorithm. Moreover, a robustness of the control scheme is introduced to the neuro-control by combining the feedback error learning method with the sliding mode control theory. It also has the advantages that the steady state error of conventional sliding mode controller with boundary layer technique is reduced by the on-line training process of the neural controller, and the control performances can be more or less maintained by the actions of the neural controller even though the initial assumptions of uncertainty bound are violated. A new neuro-fuzzy network is developed to deal with the structured informations of a system. The proposed neuro-fuzzy network can modify and creat the fuzzy rules of a system, and also be used for finding the invers... | eng |
dc.language | eng | - |
dc.publisher | 한국과학기술원 | - |
dc.subject | Control | - |
dc.subject | Identification | - |
dc.subject | Neuro-fuzzy Network | - |
dc.subject | Neural Network | - |
dc.subject | Artificial Heart | - |
dc.subject | 인공 심장 | - |
dc.subject | 제어 | - |
dc.subject | 동정화 | - |
dc.subject | 뉴로-퍼지 | - |
dc.subject | 신경회로망 | - |
dc.title | Identification and control of uncertain systems using neural and neuro-fuzzy networks | - |
dc.title.alternative | 신경회로망과 뉴로-퍼지를 이용한 불확실 시스템의 동정화와 제어 | - |
dc.type | Thesis(Ph.D) | - |
dc.identifier.CNRN | 101751/325007 | - |
dc.description.department | 한국과학기술원 : 전기및전자공학과, | - |
dc.identifier.uid | 000925253 | - |
dc.contributor.localauthor | Lee, Soo-Young | - |
dc.contributor.localauthor | 이수영 | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.