Intelligent Controller Design for Identification of Dynamic System and Implementation of Two-Wheeled Balancing Vehicle

• Main Authors: Kai-Chun Weng, 翁楷鈞
• Other Authors: Chih-Min Lin
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/2s9w5f

碩士 === 元智大學 === 電機工程學系甲組 === 107 === With the development of artificial intelligence,neuromorphic system is widely used to solve many engineering problems and has a profound impact. This thesis discusses the application of brain-imitated neural networks, which include the cerebellar mode articulation controller (CMAC), recurrent CMAC (RCMAC), brain emotional learning controller (BELC), integration of CMAC and BELC named as CMBE and recurrent CMBE (RCMBE) to dynamic system identification and two-wheeled balancing vehicle control. In the first example, the neural network works as the main controller, which has longer training time. In the second example, sliding mode control technique is applied and the neural network is used as an observer. The effectiveness of the observer is tested by applying the external disturbance. By comparing with other brain-imitated neural networks, RCMAC has the best learning ability and obtains enhanced performance results.