本研究建立以視覺輔助的同時定位與建圖(Simultaneous Localization and Mapping, SLAM)系統,提供機器人建立環境地圖與自我定位之用。本研究的內容分為兩部分:包括視覺式EKF SLAM架構的建立,以及人形機器人步態的規劃與實現。發展的同時定位與建圖系統應用在具備前視型攝影機的小型人形機器人系統上。 In this thesis, the vision-based Simultaneous Localization and Mapping (SLAM) be established, and provided for the robot to do self-localization and build up the environmental map. This research is divided into two parts, including the design of the vision-based EKF-SLAM and the gait planning of the humanoid robot. The developed SLAM system be applied to a small-size humanoid robot system with frontal camera.