Full metadata record
DC FieldValueLanguage
dc.contributor.author夏紹基en_US
dc.contributor.authorShaw-Ji Shiahen_US
dc.contributor.author楊谷洋en_US
dc.contributor.authorKuu-Young Youngen_US
dc.date.accessioned2014-12-12T02:11:48Z-
dc.date.available2014-12-12T02:11:48Z-
dc.date.issued1993en_US
dc.identifier.urihttp://140.113.39.130/cdrfb3/record/nctu/#NT820327051en_US
dc.identifier.urihttp://hdl.handle.net/11536/57769-
dc.description.abstract在本篇論文中,我們提出一個具有學習機構的控制策略來做機器人的運動 控制。一般來說,使用學習控制器來控制機器人運動時,所遭遇的一個最 主要的問題就是學習空間過大。在我們提出的方法中,首先我們將學習空 間加以分區,再使用模糊系統在各別的區域中進行學習;最後,我們使用 CMAC網路將各別區域中學習完成的模糊系統的參數加以整合來處理整個學 習空間中的運動軌跡。對於必須重覆執行的運動軌跡及負載要求,所提出 的策略及控制器僅需進行一次的學習,而不會受到學習其它運動軌跡及負 載要求的影響。因此,所提出的控制策略可以大大降低龐大的學習空間所 造成學習及應用上的困難。 In this thesis, we propose a novel scheme for governing general robot motions by using learning mechanisms. One of the main problems in using learning controllers for robot motion control is that the learning space for excuting general motions is too large. In the proposed scheme, the learning space will first be divided. A fuzzy system is used for learning in the divided region. A CMAC-type neural network is then used to generalize the parameters of the fuzzy system, which are appropriate for the control in each local region, to deal with the whole learning space. The learning process is performed only once to deal with various trajectories under different load conditions. Therefore, the learning effort is dramatically reduced for general robot motion control.zh_TW
dc.language.isozh_TWen_US
dc.subject學習機構; 學習空間; 模糊系統; 整合;zh_TW
dc.subjectLearning Mechanisms; Learning Space; Fuzzy System; Generalization;en_US
dc.title運用學習方式之機器人運動控制策略zh_TW
dc.titleRobot Motion Control By Using Learning Mechanismsen_US
dc.typeThesisen_US
dc.contributor.department電控工程研究所zh_TW
Appears in Collections:Thesis