Digital Redesign of the Observer-Based Decentralized Adaptive Tracker for Sampled-Data Nonlinear Large-Scale System with MIMO Subsystems: Evolutionary Programming Approach

• Main Authors: You-Yao Chiu, 裘友堯
• Other Authors: Jason Sheng-Hong Tsai
• Format: Others
• Language: en_US
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/28441590188399772838

碩士 === 國立成功大學 === 電機工程學系碩博士班 === 97 === In this thesis, a novel digital redesign of the observer-based decentralized adaptive tracker for sampled-data nonlinear large-scale system consisting of nonlinear multi-input multi-output subsystems, using evolutionary programming to further improve the tracking performance for ill-condition systems, is proposed. Based on the given sampled-data large scale nonlinear system consisting of nonlinear multi-input multi-output interconnected subsystems, the decentralized two-stage design is proposed to construct a decoupled well-design reference model, so that the output response of the well-design reference model will well track any trajectory specified at sampling instant, which may not be presented by the analytic reference model initially, and it may not be bounded in a quite large range. Then, the other digital-redesign decentralized adaptive tracker is proposed, so that states of the digitally controlled sampled-data large-scale system closely match the ones of the well-design reference model with the closed-loop decoupling property. As a result, it yields the output of the digitally controlled sampled-data large scale system tracks well the trajectory, which may not be presented by the analytic reference model initially. When the state of the system is not measurable, an observer-based decentralized adaptive tracker is proposed. Besides, the evolutionary programming (EP) is applied to tune the observer gain to further improve the state estimation and tracking performance for the ill-conditional system. Finally, illustrative examples are given to demonstrate the effectiveness of the proposed methodology.