本論文研究目的是為了實現多關節機械手臂的控制設計。本機械手的系統硬體架構包含一隻多關節機械手和兩顆CCD。該機械手包含四顆伺服馬達和一隻由兩顆伺服馬達所組成的機械手掌。雙CCD裝置在機器人的頭部上。機械手的控制目的為以雙CCD觀測目標，利用電腦去計算目標在三維空間的位置，然後控制機械手各個關節旋轉角度，以正確到達目標並完成抓取任務。控制流程上，以逆向運動學的概念建立機械手的運動控制，而機械手的實際位置則可由雙CCD裝置觀測而得。當機械手臂抵達目標物附近，加入模糊控制器來調整手掌位置並縮小其與目標物之間的位置誤差，直到兩者之間的誤差能滿足精確抓取目標物的需求為止。除此之外，加入手掌力回饋抓取機制和手掌平行桌面機制來偵測手掌是否確實抓取目標物並且避免機械手撞到桌面。最後，機械手能夠完成以下三大目標。目標一，抓取雞蛋並放入碗中；目標二，抓取裝水的瓶子並準確將水倒入馬克杯中；目標三，精確按壓電梯面板上不同位置的按鈕。最終本研究以多次的實際實驗來驗證本論文所提出的多關節機械手臂控制設計的有效性。 This thesis studies and implements the control designs for a multi-joint robot arm. The robot arm system hardware contains a multi-joint robot arm with a gripper and two CCDs. The robot arm contains four servo motors and the palm is composed of two motors as a gripper. Two CCD are set on the head of the robot. The control process of the robot arm is using two CCD to detect the target and using PC to calculate the target’s position in 3D space. Then the robot arm is control to reach and capture the target. In the process, the inverse kinematics (IK) concept is utilized to manipulate the robot arm, in which the practical gripper position can be messured by two CCD. When the robot arm reaches around of the target, fuzzy control is proposed to adjust the gripper position and shorten the position error such that the position error can meet the needs of capturing the target precisely. Furthermore, in order to detect whether the gripper captures the target and avoid the robot arm hitting the table, the force-sensory-feedback object grasping control, and the obstacle avoided are added. After all, the robot arm can achieve the following three objectives. The first is to hold an egg and put it onto a bowl; the second is to grasp a bottle filled with water and pour the water into a cup, and the last one is to press the exact number of elevator buttons. Several practical experimental results will demonstrate the effectiveness of the proposed control design.
