本論文針對具備雙眼視覺的兩輪移動機器人系統，發展機器人物件追蹤控制行為。首先，將推導兩輪移動機器人在機體座標的動態模式，進而使用Lyapunov直接方法與逆動力學控制，設計兩種機器人的運動控制器。其次，結合雙眼視覺的距離量測與物件辨識功能，發展機器人位置式影像伺服控制。最後，利用Lyapunov控制與逆動力學控制的特性，規劃兩輪機器人的雙眼視覺輔助物件追蹤行為。本論文將規劃的物件追蹤控制行為應用在配備雙眼視覺的兩輪移動機器人系統。模擬與實測的結果顯示，所規劃的控制器可以提供兩輪機器人進行物件追蹤功能。此整合系統可應用在居家巡防機器人，使具備偵測與追蹤入侵者的能力；或是應用在家用寵物機器人，使具備偵測主人臉龐與跟隨主人的能力。 An object tracking behavior is developed in this thesis for a two-wheeled mobile robot system with a binocular vision. First stage of the research, the dynamics model of the two-wheeled robot in body-fixed coordinates is derived. Based on this dynamics model, two different motion controllers are designed using the concepts of Lyapunov’s direct method and inverse dynamics control. Second, equipped with a binocular vision system which has the capabilities of target distance measurement and object identification, a position-based visual servo control is developed for the two-wheeled mobile robot system. In final stage, an object tracking behavior is planned for the robot utilizing specific functions of Lyapunov control and inverse dynamics control. The proposed object-tracking behavior is tested and verified on a two-wheeled mobile robot system with a binocular vision. Simulation and experimental results show that the developed controllers are well-designed for a two-wheeled robot to track the face of a person. In the future, the integrated system can be applied to a household cruise robot and enable it to detect and track the invader, or to a pet robot which has the capability to detect the face of its master and follow him.
