本論文主要目的在使用影像技術以及模糊控制器去實現自走車之即時物體辨識與追蹤。整體架構分為四個系統，分別為影像處理系統、微控制器處理系統、結構及動力驅動系統、無線傳輸系統，整合四個系統與電腦端，透過無線傳輸，做物體位置判別及控制命令的判斷。在追蹤的過程之中，自走車上的無線針孔攝影機將獲得的影像資訊，經由無線傳輸回電腦端，為了減少影像資料量，將每一張影像圖片先經過影像處理，計算之後便能得知追蹤物體與自走車之間的相對關係，再使用Matlab模擬車體的軌跡，然後比較兩種控制方法，經由影像追蹤模擬結果比較之後，本論文使用模糊控制，運用模糊控制器將得到的相對關係解出所需要的轉角及速度，再經由無線通訊模組將控制指令傳回自走車上的微控制處理晶片，進而追蹤目標物體。經由Matlab所完成之影像追蹤模擬及真實實驗的結果，證明我們所提出的概念有助於相關工程上的應用。 The main purpose of this research is utilizing image processing technology and fuzzy control to realize real-time object tracking. The proposed autonomous vehicle includes four parts, image processing system, microcontroller, mechanism and wireless communication system. When tracking the object, the CCD camera feedback the environment information to host computer. The computer recognizes the position of object by image processing and control the movement of vehicle. The simulation is provided to estimate the performance of both control method. The result shows that fuzzy control could achieve better response. The angles and speeds of the vehicle is estimated by fuzzy controller. The control command was transmitted from PC to microcontroller by wireless module. Finally, the autonomous vehicle can track the object accurately and stably. The simulation of target tracking which is performed by MATLAB also proves the feasibility of the proposed robot system.
