Interval Type-2 Fuzzy Logic Controller Design with Artificial Neural Network Reduction for Stewart Platform

• Main Authors: Lin,Wei-Zhang, 林維樟
• Other Authors: Huang,Chin-I
• Format: Others
• Language: zh-TW
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/u4rvhu

碩士 === 國立高雄第一科技大學 === 電機工程研究所碩士班 === 106 === Abstract In this study, the controllers are based on type-II fuzzy logic that are employed of Gaussian membership functions to design the fuzzy sets. And using the neural network reduction algorithm is used to replace Karnik-Mendel (KM) and Enhanced Karnik-Mendel (EKM) algorithm in reduction process. Due to the KM algorithm and EKM algorithm will repeatedly perform iterative calculations and consume a lot of computational cost, the neural network-based reduction algorithm(NNR) which can eliminate the iterative process of the reduction algorithm and improve the overall system. In addition to performance, and reduce the overall system hardware needs. For comparison of performance in the Stewart platform, the efficiency of reduction algorithms (NNR, KM and EKM) algorithms of the interval type-II fuzzy sets are employed for reduction process. Keywords: Stewart Platform, Interval Type-2 Fuzzy Logic, Karnik-Mendel(KM) algorithms, Enhanced Karnik-Mendel(EKM) algorithms, Neural Network(NNR) algorithms, Back-Propagation Neural Network(BPN) algorithms.