在行走機器人的領域裡，視覺感測器扮演著重要的角色，而一個智慧型的感測器系統必須滿足兩個需求：(1)正確地分辨出反射物的形狀(如平面，牆沿及牆角) (2)精準地對反射物定位(包括距離及方位)。因此，在本篇論文中，我們提出了一個新的動態三超音波感測器融合視覺的量測系統，其主要的目的可以正確地識別反射物並對它們做正確地定位。其中，我們提出一套新的辨識規則，使量測系統能更準確地對反射物做分類。並且，以一種新的動態估測方法進行對反射物做正確地定位。 Visual sensors play an important role in the realm of mobile robot. An intelligent sensor system must satisfy two requirements: (1) classifying the reflectors correctly (such as planes, corners and edges) (2) localizing the reflectors accurately (including distance and orientation). In the thesis, a new measuring system called the dynamic tri-ultrasonic transducer fused with vision system will be proposed. The main functions are to discriminate the reflectors correctly and localize them accurately. Furthermore, the new classification rules will be introduced to classify the reflectors more accurately and a dynamic estimation method will proposed to localize them precisely.
