本文應用紅外線感測器於設計與建構漫遊式機器人之室內定位系統,使機器人在室內環境下能自我校正其位置。本文首先運用幾何圓相交法建立三角定位推論以及利用適應性類神經網路(ANFIS)訓練定位模糊推論系統,然後將此兩種定位推論系統植入機器人主控器於實際場地測試取得定位誤差,根據實驗結果探討兩種定位推論系統之優缺點並擇其佳者作為室內定位系統之位置推論法則,經實驗結果顯示以模糊推論系統作為演算法有較快速與精確之定位能力。 本文之機器人室內定位系統以地標方位辨識(AOA)為其定位方式,機器人以其上之紅外線接收模組掃瞄接收放置於已知位置地標特定之紅外線,再由機器人計算出地標之方位角度資訊,由各地標所在之方位角度資訊經由定位推論系統演算出機器人所在之位置。各地標發送之紅外線皆以訊號編碼之方式傳送以供接收端檢測與排除錯誤。 本文最後將模糊推論演算法以行為模組之方式整合於以行為式控制架構法建構之漫遊式機器人控制系統並驗證其能夠進行位置追蹤與自我位置校正。
In this paper, the behavior module of self-localization is design and constructed by using the infra-red based sensors system in the indoor environment for the wheel-based rover robot. The behavior module enables the robot to calibrate and update its position in real time base. Here, geometric circle intersection method established by using triangulation algorithm and the Fuzzy inference method by employing the ANFIS model are constructed. These two positioning and reasoning system are both implanted into the robot controller and to compare the performance through the field test. According to the results of two positioning reasoning systems, the better one was selected to be used for the design of the robot positioning module of behavior. The results show that the fuzzy inference system has a more rapid and accurate positioning capability. This indoor positioning system recognizes the baring of landmarks for positioning calculation. Placed in known locations, the landmark emits frequency modulated infrared signal which can be received by the robot with onboard infrared receiver module. The azimuth information of landmarks can then be deduced by the positioning inference system of the robot. Finally, a positioning module of behavior based on the fuzzy inference algorithm developed is integrated into a wheel-based robot rover constructed by using the Subsumption control structure for the field test. The results of experiments show that the positioning module can provide the robot with the ability of position tracking and localization.