- 學位論文
擴展型卡爾曼濾波器於輪型機器人之即時定位與地圖建構實現
Simultaneous Localization and Mapping of a Wheel Mobile Robot via Extended Kalman Filter
摘要
本論文主要目的為機器人在未知環境中估測出自身位置並且即時繪製出環境地圖。首先透過雷射測距得到機器人與環境的距離，使用直角判斷演算法抓取環境特徵，並由特徵關聯辨識資料相關程度建立地標資料庫。接著使用擴展型卡爾曼濾波，資訊融合電子羅盤、輪型編碼器、雷射距離，估測出機器人自身位置與特徵地標位置，透過補償偏移量，加快擴展型卡爾曼濾波收斂速度跟精準度，最後利用地圖繪製完成SLAM及驗證其理論。在實作方面，本文脫離Pioneer MobileRobots機器人限制的開發平台，將單晶片與電腦端結合，來獲取其各優點；最後再經由實驗測試與分析數據來驗證所提出方法的可行性。
並列摘要
The objective of this thesis is to realize simultaneous localization and mapping of mobile robots. This means that the robot can estimate its position in an unknown environment and build the environment map. After using a laser rangefinder to measure the environment of distance, the environment of right-angle landmark is detected by this paper algorithm. The detected landmarks are associated in a database. Then Extended Kalman filter estimates robotic position by fusing encoders, electronic compass and laser rangefinder. Improvement of precision and convergence speed is achieved by adding sensor estimation. To be different with traditional Pioneer MobileRobots, we make a new control board to remove MobileRobots development environmental limit and to improve SLAM efficiency by combining computer and micro-controller.