Font Size: a A A

Research On Robot Control System Based On Visual Servo

Posted on:2017-03-21Degree:MasterType:Thesis
Country:ChinaCandidate:W B LiFull Text:PDF
GTID:2278330488461207Subject:Mechanical Manufacturing and Automation
Abstract/Summary:PDF Full Text Request
With the rapid development of robotic, the more complex the tasks of robot are, the higher the requirement of robot intelligence become. Robot visual serving control means the visual information is applied to implement the closed-loop control of robot; it is a hot topic in the research of intelligent robot in recent years. In this article, the study focused on servo system overall scheme, mechanical arm kinematics, the camera calibration technique, Visual servo subsystem and the design visual controller.The mechanical arm, Laboratory developed independently, as the research object, the kinematics of robotic arm is analyzed, the mechanical arm is modeled and simulated by the Robotics, the simulation results verify the correctness of mechanical arm kinematics equation.Aiming at the shortcomings of the calibration toolbox, this paper proposes a camera calibration method combining Harris corner detection algorithm with Zhang two-step camera calibration. The industrial camera is calibrated by using calibration toolbox and the proposed approach, the results show that the error of the new method comparing with calibration toolbox is small, and can well realize the calibration of camera.According to the characteristics of the target, the paper puts forward an image processing scheme apply to this project. The image is preprocessed to extract target edge and the target characteristics are solved by fitting the ellipse edge, the ellipse fitting error is analyzed and laid the foundation for the subsequent visual controller designAiming at the shortcomings of visual controller designed by conventional methods, the visual controller is designed by BP neural network; the neural network model is simulated by using different network parameters and training algorithm, the simulation results show that using the BP neural network fitting visual controller has high simulation precision. The mechanical arm visual servo system is debugged; the experimental results show that the servo system has high precision of servo that meets the design requirements of the subject.
Keywords/Search Tags:visual servo, robotic arm kinematics, image processing, neural network
PDF Full Text Request
Related items