本論文是以網路攝影機監視機器人所有移動的範圍,將影像資訊傳送回電腦控制端,並應用Borland C++ Builder (BCB)為平台設計人機介面做影像處理與模糊邏輯控制運算,將所得的機器人驅動速度透過無線傳輸傳達給機器人,使機器人依照控制系統執行運作,再經由網路攝影機監控,使得整體系統形成閉迴路架構,進而對目標物做追蹤。 影像處理主要目的在於辨別機器人行進方向以及尋找目標物的所在位置。在機器人及目標物上貼上有顏色的判斷標誌,辨別方式運用了色彩圖像處理,讓系統能較準確的辨認標誌,並減輕光線亮度變化所造成的影響,減少外來雜訊的干擾,加強系統辨識的強健性。 控制方面使用模糊邏輯控制,系統藉由影像處理所得的角度與距離作為模糊邏輯控制的輸入,並建立適當的規則庫推論出機器人所需速度,再以實驗調整控制器參數,達到快速以及準確的目標追蹤目的。最後透過目標追蹤和閃避障礙物等實驗來驗證本論文中對目標追蹤的效果。
This paper uses a webcam to monitor the movements of a wheeled-robot within a prescribed region. Robot and target images are sent to a computer, where image processing and fuzzy logic controller are implemented. Control signal from fuzzy logic controller is transmitted wirelessly to the robot for controlling the robot motions. New robot and target positions are picked up again by the webcam, thus forming a closed loop tracking control of the robot. Image processing in the computer is conducted mainly for computing the direction of the robot and the distance between robot and the target. Different color labels are attached to robot and target for easy identification of the robot and target locations, in which HSI color model is used for diminishing the effect of light intensity variations, hence ensuring a robust system identification. Fuzzy logic controller for computing the robot commands is employed, where angle and distance between the robot and target are inputs and the drive speed for the robot is the output variable. Fuzzy control rules and control parameters are designed and tuned carefully for accurately tracking a target. Experimental results show the effectiveness of the controller for tracking a target with and without obstacle in the path.