Font Size: a A A

Design Of Image-based Visual Control System Of Robotic Arms

Posted on:2012-09-13Degree:MasterType:Thesis
Country:ChinaCandidate:S N LiFull Text:PDF
GTID:2218330362450502Subject:Control Science and Engineering
Abstract/Summary:PDF Full Text Request
Robot is a machine which can imitate human to complete various tasks. Along with the rapid development of science and technology, robot has been applied to different areas of industrial production. In the meantime, the technical level of robot is improving. Robot is required to have strong adaptability and high intelligence in sense, thinking, decision making and acting. Vision robot can fulfill these requirements and thus becomes a research hot spot in robot domain.The original control system of PUMA 560 has been aged and is not open-ended. This paper re-designs its control system and adds in visual control system in order to obtain an open-ended teaching-purpose equipment. This paper contains four parts, which are the drive circuit design, joints control system design, visual subsystem design and system integration and test.PUMA560 is a 6 DOF manipulator. It has 6 rotary joints which can imitate waist, shoulder, elbow and wrist action of human. Before designing the control system, this paper isolates and locates the input port of the control signal, employs the original protection circuit and obtains the position feedback signal. PWM method is used in the drive circuit. The debugging of joint control system is done under Matlab. Real-time simulation diagram is built in simulink, and the connection between manipulator and control computer is done though PCI1711 data acquisition board. The model of every joint is built and based on the models, this paper designs the joint control system which includes rate and position double closed-loop. Davinci experimental platform DM6446 test box is used to complete image acquisition, image processing and image feature points extraction. The research of the visual servo system in this paper is based on image, and the visual controller design is based on the Jacobian matrix. The derivation process of Jacobian matrix is given and the Jacobian visual positioning controller based on image is designed. The derivation results are used in the control simulation under Matlab/simulink.The results of the simulation show that the proposed method has ideal control effect and feasibility in a small range. However, the visual controller based on Jacobian matrix is effective only in part of the task space. The response speed and control precision still need further research.
Keywords/Search Tags:Visual Servo, PUMA560, Robotic arm, Image processing
PDF Full Text Request
Related items