Research On Modeling And Tracking Control Technology Of Visual Servo System Of Robot

Due to the cost of labor increase and the technology of robot improved,robots are increasingly used in daily production and life.However,currently,most of the robots fulfill task through traditional teaching method.The environment should be contrived to suit the robot.The integration process is so complex that the application of the robot is limited greatly.This limitation is largely due to the inherent lack of sensory capability of environment and self decision in robot system.The camera is one of the main ways of robots to perceive the external environment.For this reason,this paper studies the visual servo control technology of robot.In order to solve the problem of positioning of static target and tracking of moving target,this parper mainly studies the system of target tracking and positioning which based on hybrid visual servo and switch Kalman filter.The system can improve the robot's working accuracy and reduce the tracking error effectively.This project comes from the national science and technology major projects “high grade CNC machine tool and basic manufacturing equipment”,Guangdong province science and technology major special projects “Intelligent robot integration key technology and application demonstration”.Guangzhou science and technology major special projects “robot control technology and application technology research”,the research goal of the project is to locate and track the target object through the vision sensor,so as to improve the positioning accuracy and work flexibility of the robot and expand the application of the robot in actual production and life.At the topic of static positioning,the friction of robot joints will deterior the accuracy of robot positioning,in order to solve this problem,this paper studies the modeling and compensation of robot joint friction,establish the relationship between joint friction force and joint velocity,analyze the characteristics of friction of robot joint.This paper uses the Lugre friction model to describe the friction of robot joint and identify the parameters of Lugre model.Because it is difficult to obtain the accurate friction parameters,the friction parameters and dynamic parameters are updated and corrected according to the actual trajectory tracking error by an adaptive method,and the global stability is proved by Lyapunov method.The robot hand-eye calibration algorithm is studied to calibrate the SCARA robot hand-eye system.The hybrid visual servo method is studied.The translation motion is controlled in the image coordinate system and the rotation motion is controlled in the Cartesian coordinate system to obtain the decoupling control of the tanslational and rotational movements.The accuracy of the camera is used to improve the robot's accuracy.At the topic of tracking of moving object,this paper studys the principle of tracking of moving target,and analyze the tracking error.In order to solve the problem of delay of the visual system,the Kalman filter is used to estimate the moving state of target,the velocity of target is obtained and input to the visual servo controller to reduce the tracking error.The reason for the degraded performance of the Kalman filter is analyzed when the target motion state is abruptly changed.A switched Kalman filter based on a motion monitor is proposed to improve the estimation performance of the Kalman filter.This paper designs and develops a SCARA robot vision servo control platform based on Kithara real-time system.In order to build a real-time visual servo controller,the four CPUs of the industrial controller are divided so that different control tasks occupy different CPUs and do not interfere with each other.The control system includes three realtime modules which running on kithara real-time kernl and one non-realtime module which running on windows system.Realtime module incluede SCARA robot motion control module,image acquisition module and image processing module which based on Halcon software.Non-realtime module is human-computer interaction.This paper use EtherCAT protocol to control the robot,input and output modules of Beckhoff,linear modules.The robot forward inverse algorithm,joint space interpolation algorithm,template matching algorithm based halcon,switch Kalman filter algorithm and visual servo control algorithm are written in C++.A friction identification and compensation experiment of SCARA robot was performed on this platform to verify that the compensation of friction of robot joint can improve robot positioning accuracy effectively.A static target positioning experiment based on hybrid visual servoing was performed to verify that the visual servo control improve the robot's job accuracy by camera.A tracking experiment of motion target based on the classical Kalman filter was performed to reduce the tracking error.A moving target tracking experiment based on the switch Kalman filter was performed to suppress the tracking error caused by the sudden change of the target motion.The research results of this paper are as follows: A method of joint friction identification that can be used for workspace-limited robots is proposed,and the joint friction of SCARA robot is identified.An adaptive control strategy is used to adjust parameters of dynamics model and friction model of robot to achieve satisfactory control performance.A robot vision servo control system was developed.Using the composite visual servo method to improve the robot's working accuracy,the decoupling control of translational motion and rotational motion was realized.Kalman filter is used to estimate the motion state of the target,which reduces the tracking error in tracking process.The switch kalman filter baesd motion monitor is uesd to improve the estimation performance of Kalman filter.The research results of this paper can improve the flexibility of the robot,improve the robot's working accuracy,and expand the application of the robot.