Adaptive Robust Cerebellar Model Articulation Controller Design for Uncertain Nonlinear Systems

• Main Authors: Chiu-Hsiung Chen, 陳 邱 雄
• Other Authors: 林志民
• Format: Others
• Language: en_US
• Online Access: http://ndltd.ncl.edu.tw/handle/59429583464596542448

博士 === 元智大學 === 電機工程學系 === 95 === The purpose of this dissertation is to develop the adaptive robust cerebellar model articulation controller (CMAC) based on the dynamic characteristics of recurrent CMAC (RCMAC), and to integrate it with adaptive control, sliding mode control and robust control technologies for the control application to uncertain nonlinear systems. According to Lyapunov synthesis approach, the adaptive tuning laws of RCMAC can be derived and the system stability can be guaranteed. This dissertation introduces the structures of CMAC and RCMAC first. Then, the adaptive RCMACs are developed for the single-input single-output (SISO) nonlinear control systems; and they are applied to a ship heading control, a car-following control, a linear ultrasonic motor (LUSM) position control and a chaotic circuit control. Moreover, in multi-input multi-output (MIMO) control system design; this dissertation also proposes the adaptive robust control systems for the uncertain nonlinear MIMO systems. In this designs, RCMAC can be used as the main controller or the uncertainty estimator. The developed MIMO RCMAC adaptive robust control systems are then applied to a nonlinear chaotic circuit and a mass-spring-damper system. Furthermore, an RCMAC fault tolerant robust control of a biped robot is also presented. From the simulation and experimental results, the control schemes proposed in this dissertation have been shown to achieve satisfactory control performance for the considered nonlinear systems.