Research On Image Based Vision Servo Manipulator Control

Visual servo is a hot subject in the field of robotics that has been gradually developed in recent years,attracting more and more scholars' research.Visual Servo is an interdisciplinary mix of disciplines,involving a large number of other disciplines.Image-based visual servoing control method is one of the key components of visual servoing control,and because it does not need complex two-dimensional to three-dimensional reconstruction,but by using the camera to obtain without any processing of pixel information design corresponding algorithms make robots work,which is the focus of research in related fields.This thesis focuses on the IB VS algorithms,models,etc.made a more detailed study.First of all,the first chapter introduces the development of visual servo technology in recent years as well as some of the research results obtained so far.Chapter 2 introduces the basics of visual servoing.Chapter 3 uses a calibration-free visual servoing method in order to avoid the complicated camera calibration process,thus simplifying the workload and improving the control effect.The Kalman filter method is used to estimate the image Jacobian matrix.In order to avoid the mutual coupling between the camera's rotational motion and the camera's translational motion,the camera speed is decoupled.In general,the singularity of the matrix can sometimes be introduced in the case of using a point as a target.And,if the camera calibration may introduce calibration error.Chapter 4 designs an Extreme Learning Machine estimator to estimate the velocity of the robotic end effector,and in order to make the system converge faster without the instability of the system due to excessive speed,Using fuzzy inference theory,a kind of controller whose proportion coefficient can change with the system is designed.To verify the viability and effectiveness of the Extreme Learning Machine designed,it was applied to the vision consisting of a Logitech C310 webcam,a DENSO VP-6242M six-degree-of-freedom robot,a Windows 7 64-bit computer,and a Quanser control box Servo-based control system,the fact that the design of the algorithm in kind can also be very good results.