Inverse Perspective Mapping Implemented for Humanoid Robots to Simultaneous Localization And Mapping

• Main Authors: Tz-fan Chang, 張子凡
• Other Authors: Kuo-yang Tu
• Format: Others
• Language: zh-TW
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/55311882087510072598

碩士 === 國立高雄第一科技大學 === 系統資訊與控制研究所 === 101 === To realize artificial intelligence, localization is the fundamental work of robot development. However, localization must relies a map. Therefore, in this thesis the humanoid robot focuses on Simultaneous Localization And Mapping (SLAM). However, it’s hard to have a stable basic coordinate system to fix a vision system in a humanoid robot. Inverse Perspective mapping is thus applied to build the map for humanoid robot localization. Based on probability theory, Monte Carlo Localization (MCL) is used for robot localization. Finally, the experiments for a humanoid robot to build a map for localization are also included.