本論文探討並實踐一個應用於二維天車系統的視覺回授控制器。這個控制器使二維天車系統具備定位和抑制搖擺。滑動模式控制在本論文中作為控制器,並與模糊邏輯控制器結合,用以提昇滑動模式控制器的性能。藉由模糊邏輯控制器,使滑動模式函數的斜率可適應於系統特徵。模糊邏輯控制器不但決定了滑動模式控制器的輸出,也消除了雜音的問題。此外,在本論文中也應用了快速且平順運輸的動作計畫方法,能讓天車速度的改變緩慢,並降低加速度。 基於色彩直方圖的視覺追蹤,在本論文中作為即時的視覺追蹤的方式。其追蹤了繩索的角度與負載的位置。色彩直方圖方法對於影像識別的快速辨識與定位,與移動追蹤區域的策略,都被用來增加系統的運算速度。最後,結合了視覺追蹤與天車控制器,成為一個具有定位與抑制搖擺的強健視覺回授二維天車控制器。
This thesis developed a visual feedback controller which used a CCD to control a two dimensional crane system. A control scheme was developed for the position and anti-swing of a two dimensional overhead crane system. The sliding mode control was proposed as the controller, combined with the fuzzy logic controller to enhance the performance of the sliding mode controller. The slope of the sliding mode function is adaptable to the system characteristic by fuzzy logic controller. The output of the sliding mode controller is decided through the fuzzy logic controller either to cancel the chattering problem. Moreover, a motion planning method for the rapid and smooth transportation, which make the slowly change of the crane’s speed and reduce the acceleration is also employed. The visual tracking based on color histogram is proposed to handle the real-time visual tracking. It tracks down the angle of the rope and the position of the load. The fast recognition and localization of the color histogram method for image pattern recognition and the strategy using moving tracking area is used to increase the calculation speed. Finally, it combines the visual tracking and the crane controller, to be a robust visual feedback controller for the position and anti-swing of a two dimensional crane system.