Fuzzy PID Controller with Neural Network for Controlling the magnetic levitation system

カテゴリ: 部門大会

論文No: GS8-1

グループ名: 【C】平成18年電気学会電子・情報・システム部門大会講演論文集

発行日: 2006/09/05

タイトル(英語): Fuzzy PID Controller with Neural Network for Controlling the magnetic levitation system

著者名: Agus Trisanto(千葉大学),Yasser Muhammad(千葉大学),呂建明 (千葉大学),谷萩 隆(千葉大学)

著者名(英語): Agus Trisanto(Chiba University),Yasser Muhammad(Chiba University),Jianming Lu(Chiba University),Takashi Yahagi(Chiba University)

キーワード: fuzzy|pid|neural network

要約(日本語): This paper describes the fuzzy PID (FPID) Controller with the Neural Network (NN) for controlling the nonlinear magnetic levitation system. The magnetic levitation systems are extensively used in various fields, such as high-speed maglev passenger trains, frictionless bearings, levitation of wind tunnel models, vibration isolation tables, etc. Duo to the open-loop instability and inherent non-linearlities, therefore the control problem is quite challenging to control engineers. In the other hand, FPID controllers has been successfully implemented in past to many linear and nonlinear process. But it is difficult to setting the gains of the FPID controller. Here, we propose the NN to assist the FPID controller. NN is used to compensate for incorrectly selecting gains of the FPID controller. The control output is the sum of the FPID controller output and the NN output. When the gains of the FPID controller are incorrect, then the NN takes over the controller, otherwise the NN does not operate. A simple structure for the NN is designed to fulfill the need of for fast computing and on-line training. To confirm the effectiveness of our proposed method, we compare the FPID controller and the proposed method through experimentation.

PDFファイルサイズ: 2,697 Kバイト