- 學位論文
自主行動機器人無線感測網路定位技術之發展
Development of Wireless Sensor Network Localization Techniques for an Autonomous Mobile Robot
摘要
一般而言，服務型機器人的作業環境大都以室內為主，其中室內定位技術對服務型機器人而言是最關鍵的技術之一，唯有具備精確定位功能的服務型機器人，才能即時知道自身所處的精確位置，才有辦法確切執行指派的服務工作。近年來，由於無線感測網路具有架構簡單、小尺寸、低成本、低消耗功率、相互連結性高等特點，已廣泛應用於自動化設備或環境安全的監控系統上，此外，利用無線感測網路技術建置室內定位系統亦明顯增加。本文的研究目的即在發展結合無線感測網路技術及立體視覺演算法之自主行動服務型機器人的精確室內定位技術。礙於室內環境的限制，根據文獻，無線感測網路定位的精確度約在2 m範圍內，因此，本文提出以機器人本身的雙眼視覺系統，輔以立體視覺演算法進行定位補償以縮小誤差。根據實驗結果顯示，只要無線感測網路的定位誤差能控制在2 m範圍內，利用立體視覺演算法輔助定位，則將可縮小誤差到5%以內，即10 cm內。
並列摘要
In general, service robots used mainly in indoor environments. Therefore, one of the most important techonology for service robots is the indoor localization. The service robot is able to know precisely its position and preform exactly assigned tasks if the robot has the capability of precise positioning. Recently, due to its simple structure, small size, low cost, low power consumption, and easy to connect with each other, the wireless sensor network has been widely used in monitoring and control system for factory automation equipment and environmental security. Moreover, building an indoor localization system based on wireless sensor network technology has increased noticeably. The purpose of this study is to develop a precise positioning technology for an autonomous mobile service robot by integrating wireless sensor network with stereovision algorithm. According to the literature, the positioning accuracy of wireless sensor network is about 2 m under indoor environmental constraints. Hence, we propose a compensation method in this study by using the robot’s binocular vision system with stereovision algorithm to reduce positioning error. Experimental results show that if the positioning accuracy of wireless sensor network can be controlled within 2 m, we can reduce the positioning error to no more than 5%, that is, within 10 cm.