本論文主要利用雙眼視覺技術,用以研製具機器視覺之機械手臂系統。文中描述硬體架構與七個自由度(DOF)的機械手臂之機構設計,利用直流馬達為制動器,並在各個關節安裝電位計作為感測器,利用電壓訊號變化計算出直流馬達的轉動角度,再經由核心處理器C8051F120整合回授訊號與系統資訊,並傳回電腦運算機械手臂軌跡,達成機械手臂軌跡追蹤的目標。爲了實現影像追蹤,首先利用攝影機校正方法(camera calibration),求出雙攝影機內外參數並予以校正,並結合雙眼立體視覺(binocular stereo vision)求出目標物在世界座標系統中的質心位置。接著利用逆向運動學(inverse kinematics)推導機械手臂的運動模型,藉由已知的姿態與位置,求得各個關節應轉動的角度,對回授訊號做控制補償,再由模糊邏輯控制器(fuzzy logic controller)產生命令,以達成機械手臂軌跡追蹤。最後藉由實驗結果驗證,本文所設計的系統能有效地且正確地達成追蹤並抓取目標物之要求。
This study describes the design and implementation of a stereo vision-based robot arm system. First, the hardware architecture and the mechanism of the seven degree of freedom (DOF) robot arm are described. The actuators of robot arm are the seven DC motors with potentiometer. The angle of rotation is calculated by voltage signal change. Second, the stereo vision system can estimate the position of the world coordinate system. A camera calibration algorithm is used for estimating two wed cameras’ intrinsic and extrinsic parameters. Given the desired position and orientations, the desired angle of joints can be solved by using inverse kinematics (IK). In order to achieve tracking control, the fuzzy logic controller (FLC) is implemented. Finally, the experiment results demonstrate tracking performance.