Research On Manipulators Handing A Flexible Payload Based On Visual Servo System

In recent years,with the rapid development of science and technology,the robot has been application and it plays a more and more important role in industrial production,in order to further improve the degree of the robot's environmental adaptation ability and intelligent,the research of the robotic visual servo system is more and more important.Especially in the area of aerospace,printed circuit board,the automotive industry,the large objects or thin metal and deformation objects have been application,so the traditional theory and technology of manipulators are no longer applicable.And for this kind of flexible object operation and precise control has always been a big problem in the study of industrial robots,also restricts the degree of automation of electronic,automobile and other related industries.In this thesis,it is used the theory of visual servo to research the elastic object,and this thesis has the following several aspects:Firstly,for the visual servo manipulator system,this system includes many models,like a dynamic model of robot,kinematics model of robot and visual model,we should fully understand the control theories of the visual servo manipulator system based on the above models of the system.In addition,we use the feature points of image as the output feedback and the error between the feature point and the desired points as the input information of the system,then the joint,the joint speed and the joint angular acceleration are as the state feedback to design the robot controller based on the robot dynamic model.Secondly,to research the visual servo system of the camera uncalibrated,the camera intrinsic and extrinsic parameters and the robot dynamics model are included uncertain part,by selecting the robot dynamics parameters that can be linearized,the uncertainties of the robot dynamics model can be translated into uncertainties of the parameters by selected.The adaptive method is used to design adaptive parameters updating laws and to estimate the camera and robot dynamics parameters online real-time,and use the estimations,the feature information,the visual signal and the robot joint for the joint controller design that can drive robot rapidly and directly to complete the task.Finally,to research the visual servo system of the camera uncalibrated control the elastic objects,but the camera intrinsic and extrinsic parameters and the elastic model are uncertain.The structures of deformation object are used connection between the physical properties of object and the end-effector of manipulator to expression by the mathematical model.Then,the adaptive method is used method to design adaptive parameters updating laws and to estimate the deformation object parameters online real-time,and use the estimations,the feature information,the visual signal and the updating laws to design the control input that can drive manipulator rapidly and directly to complete the point to point task.Last,we use the MATLAB and the Lyapunov function to prove robust.