Neuro-Fuzzy-Based Soft Computing System and Its Application to a Submersible Vehicle Control System
碩士 === 長庚大學 === 電機工程研究所 === 91 === In this thesis, soft-computing approaches are studied for the control application of an Unmanned Free-Swimming Submersible Vehicle (UFSSV). To alleviate the multifarious design procedure in traditional control approach, a fuzzy logic approach is used for...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | en_US |
Published: |
2003
|
Online Access: | http://ndltd.ncl.edu.tw/handle/71943952319993393826 |
id |
ndltd-TW-091CGU00442018 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-091CGU004420182016-06-24T04:15:56Z http://ndltd.ncl.edu.tw/handle/71943952319993393826 Neuro-Fuzzy-Based Soft Computing System and Its Application to a Submersible Vehicle Control System 模糊類神經軟式計算系統及其潛水艇控制應用 Yu-Chieh Chen 陳玉潔 碩士 長庚大學 電機工程研究所 91 In this thesis, soft-computing approaches are studied for the control application of an Unmanned Free-Swimming Submersible Vehicle (UFSSV). To alleviate the multifarious design procedure in traditional control approach, a fuzzy logic approach is used for the design of controller. For the design of a fuzzy logic controller (FLC), expertise, engineering experience and judgment can be used into the design of knowledge base in the FLC. The supreme merits of an FLC are in its simplicity, understandability of fuzzy rules, and model-free approach by which a plant is viewed as a black box and only the input/output of the plant is needed to the FLC for observation. However, during the design procedure of the FLC, position and width of fuzzy sets are designed manually to obtain good performance. For more advanced study, a learning algorithm is applied to modify the inner parameters of self-learning neuro-fuzzy controller to get better system response. And finally, a neuro-fuzzy-based soft computing system can learn its rule base structure and parameters from input/output training data. There is no fuzzy IF-THEN rule in the system initially. The fuzzy control policy is set up automatically during learning process. There are two phases in the self-organization learning process, which are structure learning and parameter learning. The initial control policy of the neuro-fuzzy system for control application is generated using a clustering algorithm. With the well-known random optimization method, the NFS can learn its parameters for application. The proposed approach for self-organization of system structure and self-adjustment of system parameter can be applied on control problems of both linear and nonlinear plants. Chunshien Li 李俊賢 2003 學位論文 ; thesis 121 en_US |
collection |
NDLTD |
language |
en_US |
format |
Others
|
sources |
NDLTD |
description |
碩士 === 長庚大學 === 電機工程研究所 === 91 === In this thesis, soft-computing approaches are studied for the control application of an Unmanned Free-Swimming Submersible Vehicle (UFSSV). To alleviate the multifarious design procedure in traditional control approach, a fuzzy logic approach is used for the design of controller. For the design of a fuzzy logic controller (FLC), expertise, engineering experience and judgment can be used into the design of knowledge base in the FLC. The supreme merits of an FLC are in its simplicity, understandability of fuzzy rules, and model-free approach by which a plant is viewed as a black box and only the input/output of the plant is needed to the FLC for observation. However, during the design procedure of the FLC, position and width of fuzzy sets are designed manually to obtain good performance. For more advanced study, a learning algorithm is applied to modify the inner parameters of self-learning neuro-fuzzy controller to get better system response. And finally, a neuro-fuzzy-based soft computing system can learn its rule base structure and parameters from input/output training data. There is no fuzzy IF-THEN rule in the system initially. The fuzzy control policy is set up automatically during learning process. There are two phases in the self-organization learning process, which are structure learning and parameter learning. The initial control policy of the neuro-fuzzy system for control application is generated using a clustering algorithm. With the well-known random optimization method, the NFS can learn its parameters for application. The proposed approach for self-organization of system structure and self-adjustment of system parameter can be applied on control problems of both linear and nonlinear plants.
|
author2 |
Chunshien Li |
author_facet |
Chunshien Li Yu-Chieh Chen 陳玉潔 |
author |
Yu-Chieh Chen 陳玉潔 |
spellingShingle |
Yu-Chieh Chen 陳玉潔 Neuro-Fuzzy-Based Soft Computing System and Its Application to a Submersible Vehicle Control System |
author_sort |
Yu-Chieh Chen |
title |
Neuro-Fuzzy-Based Soft Computing System and Its Application to a Submersible Vehicle Control System |
title_short |
Neuro-Fuzzy-Based Soft Computing System and Its Application to a Submersible Vehicle Control System |
title_full |
Neuro-Fuzzy-Based Soft Computing System and Its Application to a Submersible Vehicle Control System |
title_fullStr |
Neuro-Fuzzy-Based Soft Computing System and Its Application to a Submersible Vehicle Control System |
title_full_unstemmed |
Neuro-Fuzzy-Based Soft Computing System and Its Application to a Submersible Vehicle Control System |
title_sort |
neuro-fuzzy-based soft computing system and its application to a submersible vehicle control system |
publishDate |
2003 |
url |
http://ndltd.ncl.edu.tw/handle/71943952319993393826 |
work_keys_str_mv |
AT yuchiehchen neurofuzzybasedsoftcomputingsystemanditsapplicationtoasubmersiblevehiclecontrolsystem AT chényùjié neurofuzzybasedsoftcomputingsystemanditsapplicationtoasubmersiblevehiclecontrolsystem AT yuchiehchen móhúlèishénjīngruǎnshìjìsuànxìtǒngjíqíqiánshuǐtǐngkòngzhìyīngyòng AT chényùjié móhúlèishénjīngruǎnshìjìsuànxìtǒngjíqíqiánshuǐtǐngkòngzhìyīngyòng |
_version_ |
1718323015198965760 |