Research And Application Of The Uncalibrated Visual Servo System In Monocular Manipulator

In recent years,with the rapid development of machine vision technology,computer control theory and other technologies,visual features information can be applied to the guidance of robot manipulator.The manipulator visual servo system integrates real-time image processing technology,robot dynamics and kinematics,computer control theory,system identification theory,real-time OS and etc,.Through using visual features information as a feedback control signal,traditional manipulator system become more intelligent.Traditional manipulator visual servo system needs precise calibration of visual sensor model and manipulator model which is a complicated process and is not robust towards complex environment.Therefore,uncalibrated visual servo system has become a new research direction in recent years,which means it does not need any other calibration process in advance.The system only requires image feature information,manipulator control law and online estimation of Jacobian matrix method to guide the manipulator to reach the desired position.The main work and the innovation in this paper can be summarized as follows:(1)Discussed the research background and significance of the manipulator visual servo system.Then we classify the visual servo system into different parts based on their structure.Thirdly,Summarizing the research development of the visual servo system and listing the main research content of this paper.(2)Described the basic theory of monocular manipulator vision servo system,.As for the monocular manipulator uncalibrated visual servo system,we presents some important contents,including task function,image feature extraction algorithm,online estimation of image Jacobian matrix method and visual servo control law design.(3)The following chapters will be discussed from four aspects.As for the task function,we list some drawbacks of the image coordination method and projective homography method.Then we tackle the problems by introducing a novel method of 6elements in image plane.Listing 3 different geometric image features extraction algorithm,including point feature,line feature and ellipse feature.We adopt an effective ellipse detector in the development of visual servo system.As for the control law and the onlineestimation of image Jacobian matrix,we list some basic common-used methods.Finally,simulation and experiments based on the robotics toolbox and vision toolbox were set up on the Matlab platform.And we built up a manipulator visual servo system in the ubuntu OS which integrates Linux-CNC,OpenCV and ViSP open source platform.