碩士 / 國立虎尾科技大學 / 電機工程系碩士班 / 105 / In recent years, automatic navigation of mobile robots has been studied by many scholars. In this study, we move the robot to the target location into two behavior. One is the Compensatory Fuzzy Controller (CFC) that uses the Reinforcement Learning (RL) method after the Grouped Teaching Learning Based Optimization (GTLBO) learn, which is called the target search behavior. In this section, we will introduce two Fuzzy Logic Controllers (FLC) for obstacle avoidance behavior. The two fuzzy logic controllers along the wall are left wall-following fuzzy logic controller (LWFLC) and right wall-following fuzzy logic controller (RWFLC). In the navigation behavior of the robot, we get the output of the left and right wheels through the calculation of the CFC according to the distance between the target and the obstacle from the environment. At the same time, we use the enhanced learning method to design a fitness function, and use this fitness function to evaluate the performance of CFC in an unknown environment. In this paper, the original Teaching-Learning-Based Optimization (TLBO) algorithm is improved in the way that the learners will use the fuzzy C-means (FCM), enables algorithms to improve global search capabilities. Finally, in this paper, we use the GTLBO algorithm and other popouation-based evolutionary algorithms to verify that our proposed method can successfully complete the navigation task
| Identifer | oai:union.ndltd.org:TW/105NYPI5441027 |
| Date | January 2017 |
| Creators | Yan-Zhong Huang, 黃彥中 |
| Contributors | 陳政宏 |
| Source Sets | National Digital Library of Theses and Dissertations in Taiwan |
| Language | en_US |
| Detected Language | English |
| Type | 學位論文 ; thesis |
| Format | 72 |
Page generated in 0.0021 seconds