Research On Uncalibrated Visual Servoing For Manipulator Tracking Control

The visual servoing control of the manipulator is a closed-loop control method that uses the visual sensor to detect the position and orientation of the target and the manipulator,so as to realize the positioning,tracking and other operations of the manipulator to the target.The visual servoing technology of manipulator has a broad development prospect,which is widely used in complex and changeable intelligent manufacturing tasks and heavy and dangerous space tasks.It is a key technology that needs to be applied and developed in the frontier intelligent fields such as intelligent manufacturing,aerospace and so on.However,the visual servoing has the disadvantages of tedious and time-consuming calibration which is also easy to lose efficacy.In order to overcome the above shortcomings,people put forward uncalibrated visual servoing.Uncalibrated visual servoing which is unnecessary to calibrate can control the movement of the manipulator through the research of control rate.It has great advantages in control cost,convenience and performance and so on.It also meets the future needs of convenient,efficient and flexible intelligent operation.However,the application research of uncalibrated visual servoing lags far behind the theoretical research,and there are still many obstacles to apply uncalibrated visual servo in actual production.For example,when tracking objects,how to shorten tracking time to meet production efficiency needs?Compared with laser radar and other sensors,vision sensor covers more information and requires more computation.How to ensure real-time tracking in practical applications to meet the real-time requirements of production applications?Around the above questions,this thesis focuses on the research of uncalibrated visual servoing.The main contents of this thesis include:(1)The author investigates the background and significance of the research on uncalibrated visual servoing of robot,and expounds the development status of industrial robot and the research status of uncalibrated visual servoing domestically and abroad.(2)The basic theory of visual servoing is discussed,including coordinate system space transformation,robot kinematics modeling method,camera model and jacobian matrix derivation method.(3)The author designs the calibration-based visual servoing control system for ABB manipulator and studies the calibration method of visual servoing for manipulator.The calibration-based visual servoing control system is designed firstly.Then the theory and experiment process of camera calibration and hand eye calibration are discussed.Besides,the paw control module and image recognition module of the system are designed.Finally,the performance of the calibration-based visual servoing control system is verified by physical experiment.(4)The theory of uncalibrated visual servoing are researched and the uncalibrated visual servoing control system for ABB manipulator is designed.The Guass Newton algorithm,dynamic Broyden algorithm and recursive least square algorithm of uncalibrated visual servoing are derived;the theory and implementation of ORB and CamShift image recognition algorithm are studied;the realization of communication between different operating systems of computer and manipulator controller is described;In simulation experiment and physical experiment,the manipulator which is under uncalibrated visual servoing system's control completes the real-time tracking of dynamic target.(5)Aiming at the problem of low tracking efficiency and poor real-time performance of uncalibrated visual servoing in the application of dynamic object tracking,a fast tracking strategy algorithm of uncalibrated visual servoing control system is proposed.This thesis introduces the background and content of the research and describes the principle of the algorithm.In simulation experiment and physical experiment,the advantages of the improved algorithm in the aspects of rapidity,real-time and anti-interference are verified.In this thesis,the feasibility and effectiveness of the whole design scheme of the calibration-based visual servoing control system and the uncalibrated visual servoing control system for ABB manipulator are verified through the physical experiment.Through simulation experiment and physical experiment,the fast tracking algorithm is proved to be fast,real-time and anti-interference.