在移動機器人自主運動控制的研究課題上,定位是相當重要的,尤其在未知的環境下,必須要先建立環境地圖才能對機器人做定位。當機器人在未知的環境下從任一的位置出發,在建立地圖的同時,必需利用已建立的地圖來更新本身的位置,因此建圖與定位這兩者是相輔相成的,究竟是先有完善的地圖來進行定位,還是先有精確位置估測再來進行建圖,這也就是所謂的〝雞生蛋〞與〝蛋生雞〞的問題。本論文使用基於拓展卡爾曼濾波器的SLAM演算法(Simultaneous Localization and Mapping)來解決這問題,根據馬可夫假設把定位與建圖分開來討論,讓這兩者有時間關係性,使輪型機器人Segway能夠使用雷射測距儀(Laser Range Finder)在未知環境下建立特徵地圖並且設定移動座標讓機器人在朝目標點自主移動的過程中同步定位。
The localization is the very important problem for a mobile robot in the unknown environment. It is necessary to modify the exact position of the robot by the new map if the robot moves from any start position. So, localization and mapping of a mobile robot are complement each other in the practical application. Sometimes, it is also called chicken/egg problem. The objective of this thesis is to make a SLAM by the EKF theory. The localization and mapping are discussed separately under the assumption of Markov. The data of a laser range finder are used to develop the feature map. The feature map and the corresponding pre-set way point are applied to calculate the SLAM by the moving robot in the unknown environmen