攝影機在日常生活中隨處可見,各種使用攝影機作為感測器的系統在工業上已經越來越多。利用雙CCD系統模擬人類雙眼,嘗試讓機器人能看到目標外,也能判斷目標距離,在行動上得到更多元化的表現。 本文專注於機器手臂研究,藉由一組仿人視覺系統與機器手臂,達到對特定目標進行指定任務。攝影機系統在經過影像處理、立體視覺、機器人運動學等演算法,讓從攝影機擷取的影像成為命令傳給機器人,令其發現一空間中目標並對其發出追蹤運動。運動控制部分則有四顆伺服馬達提供手臂運動,並交由ADLINK的PCI-8134運動控制卡負責馬達方面的回授信號,特別針對手臂對目標姿態問題,使用一向量分析找出手臂合適的位姿,來讓手臂以適當的位向追蹤目標。 MATLAB因強大的矩陣能力常用於工程運算,利用MEX檔案格式作為程式銜接以整合C語言環境,簡化軟硬體傳輸介面、增加資訊交換彈性來達成即時控制。
We use camera everywhere in our life. There are many kinds of systems which regard the camera as a sensor in the industry. Using a pair of CCDs system to simulate human’s eyes search for the target. It’s not only let robot to see what it is, but also could measure the distance. It makes robot diversification. This paper studies the robotic arm to achieve the assignment from a pair of CCDs and a motor-drived arm. The camera system can pick up the image transmits command control robotic arm from image processing, stereo vision, robotic kinematics, and shoots the tracing exercise which discovers the target in workspace. In motion control, four servo motors support the robotic arm use. PCI-8134 motion controlling card of ADLINK controls the feedback signals of servos. We direct the posture matrix of the robotic arm using the vector analysis. It gives the motor arm tracing target uses as appropriate direct. MATLAB makes huge powerful be used of engineering calculation. The MEX-File combines MATLAB with C language how can simplify the hardware I/O device interface, increase the information exchange to accomplishing the robotic arm control.