許陳鑑洪欽銘林士勛2019-09-042015-02-102019-09-042012http://etds.lib.ntnu.edu.tw/cgi-bin/gs32/gsweb.cgi?o=dstdcdr&s=id=%22GN0698700339%22.&%22.id.&http://rportal.lib.ntnu.edu.tw:80/handle/20.500.12235/99197本論文係以裝載在Pioneer3-DX雙輪自走車之超音波感測器進行環境偵測，利用所蒐集到的感測器資料建立室內未知環境地圖，並以格點地圖方式呈現。由於超音波感測器所蒐集到之環境量測數據具有不確定性，因此本論文提出一模糊資料融合的方法，利用實驗結果所歸納出的超音波感測器模型，以解決超音波所具有的角度不確定性及多次反射易造成距離誤判的缺陷問題，再以模糊邏輯運算將感測器資料加以融合。隨著自走車的移動，地圖內的格點資訊將不斷地被計算與更新，最後可獲得一完整之障礙物環境格點地圖，可提供移動式機器人做為定位、導航或路徑規劃之依據，增加其自主運行的能力。文末以本校科技學院及演化控制實驗室外走廊等局部環境，利用自走車建置環境地圖，實驗結果證實本論文所提出的模糊資料融合方法所建構之未知環境地圖之可行性。This thesis investigates the use of ranging data collected from the ultrasonic sensors mounted on a two-wheel mobile robot, Pioneer3-DX, to build occupancy grid maps of an unknown indoor environment based on fuzzy sensor fusion. Because of uncertainties inevitably encountered by using ultrasonic sensors, a more reliable sensor model is designed to solve the problems of angle uncertainties and multiple reflections. To address the problems due to measurement uncertainties of the ultrasonic sensors, a fuzzy logic approach is proposed to construct the grid map, where the information of the grids are continually computed and undated through fuzzy logic operations. As long as the environmental map is obtained, where every grid in the map is represented as the possibility of occupancy, it can be used for localization, navigation, or path planning to strengthen the autonomy of mobile robots. To validate the feasibility of the proposed approach, we also conduct experiments to build maps in the Technology Building of the University.地圖建構模糊邏輯超音波感測器模型感測器融合不確定性機器人導航Map buildingfuzzy logicultrasonicsensor modelsensor fusionuncertaintyrobotnavigation