基于图像的视觉伺服机器人系统-控制科学与工程专业论文.docxVIP

哈尔滨工业大学工学硕士学位论文 哈尔滨工业大学工学硕士学位论文 - - II - Abstract Robot vision system has increasingly become a hotspot nowadays, it connect vision control with robot space movement to conduct servo task. Visual servo of robot is a domain that is concerned with image processing, computer vision, kinematics, dynamics, control science and real-time calculation. Robot vision system has tremendous significance and wide application future in industry and education. The paper discussed the working principle of visual servo robot based on figure and the vision servo system are designed and tested. The key problem in the system is the obtain of Image Jacobian matrix desertion are made after careful analysis that conventional method can hardly deal with wide range image servo problems so an improved Image Jacobian matrix method are used. Another key point for the prosperity of the system is the effect of image processing. The method of edge detection is used to get the value of image and after processing in the up PC to get an control value, in order to get visual feedback. For the practical part of the system, body of mechanical hand motivated by DC motor is built and PWM system is designed to drive the DC motor. The result from experiment proved the fine character of the drive circuit and protection circuit are used in PWM system to protect the mechanical hand in case that it overturned. At last, debugging of the whole system is done, I/O board PCI1711 is used as the connection between sensors and joint controller of robot and the control computer. A real-time windows target emulator is build in Simulink interface so that it become possible to connect the control system to the mechanical arm, and at last accomplish the two closed loop control which contain velocity closed loop and position closed loop. Keywords Visual servo; Image Jacobian matrix; edge detection; PWM - - III - 目录 HYPERLINK \l _bookmark0 摘要 Ⅰ HYPERLINK \l _bookmark0 Abstract II HYPERLINK \l _bookmark1 第 1 章 绪论 1 HYPER