Research On Vision-Force Combined Control Strategy Of Stewart Parallel Robot

With Industry 4.0 becoming a hot topic worldwide,the robot industry has ushered in vigorous development.How to meet the needs of more accurate,more efficient and wider application scenarios is a common pursuit in the field of robot research.Robotic assembly technology plays an important role in many fields such as industry.Traditional parallel robot assembly has high precision and high carrying capacity,but lacks autonomous control capability and cannot identify targets within the work environment.This paper adds autonomy to the control system by implementing navigation control of Stewart's parallel robot through visual servo control.Meanwhile,in order to improve the efficiency and accuracy of robot vision servo control,the combination of vision servo control and force control is proposed to realize the improvement of Stewart parallel robot control accuracy,and then complete the robot assembly task.This paper is based on the project of Jilin Province Science and Technology Development Plan-' Large Workpiece Flexible Assembly and Docking Robot with Autonomous Positioning and Navigation '(Project Number : 20200401130GX),aiming at improving the control performance of robot motion,aiming at the problems of poor visibility of traditional robot control methods and low control accuracy of single control mode,the research on visual servo control and visual-force combination control of parallel robot is systematically and deeply carried out.(1)For the working conditions of large workpiece and high load,the motion control of Stewart six-DOF parallel robot is studied.The kinematic characteristics of the Stewart platform are analyzed,and the inverse kinematics solution of the Stewart platform is given.The parallel platform has the advantages of high precision and high load in the field of assembly docking,which meets the performance requirements of the robot under task conditions.(2)For the problem of autonomous positioning and navigation of robots,the image-based visual servo control method of robots is studied.Image processing adds color space conversion to improve the accuracy of target recognition.The point feature is used as the image feature information,and the image Jacobian matrix is deduced by using the camera calibration parameters and the relative relationship between the pixel coordinate system and the Stewart platform coordinate system.The change rate of the pixel of the feature point is converted to the speed of the Stewart platform struts to realize the visual servo navigation control of the Stewart platform.The Stewart platform is prone to oscillation under high load conditions,which affects the relative position relationship between the platform and the camera,and the camera calibration results are greatly affected by the illumination conditions.Therefore,when the camera calibration accuracy does not meet the task requirements,the servo control mode without camera calibration is used to drive the platform motion.Based on the nonlinear objective function minimization theory,the iterative formula of the platform struts length is obtained,and the dynamic quasi-Newton method is used to estimate the image Jacobian matrix online to realize the platform uncalibrated visual servo control.(3)In order to improve the visual servo control accuracy of Stewart platform,the visual-force control strategy of Stewart platform is studied.Based on the working conditions of Stewart platform,it is necessary to avoid the influence of large change of contact force and oscillation on the motion control of the platform.The force control adopts impedance control method.The visual servo completes the navigation control of the platform,and the force control adjusts the pose of the platform.The visual-virtual force control method of Stewart platform is proposed,and the determination method of virtual force is explained in detail.The visual-virtual force control strategy optimizes the motion performance of the platform,which not only enhances the robustness and autonomy of visual navigation control,but also realizes the flexible docking operation required for industrial intelligence.(4)Using Vrep simulation platform and Vs2019 compiler tool,the proposed Stewart platform visual servo,control algorithm combined with control strategy and system communication are written and simulated.Subsequently,the Stewart motion control experimental platform was built,and the software control interface was designed with Qt creator to realize the control visualization.Finally,the platform visual servo control experiment and the visual-virtual force combination control experiment are completed,and the experimental data are analyzed and summarized.This experiment improves the efficiency and accuracy of parallel robot navigation and assembly,and provides technical support for the application of robot in intelligent manufacturing and industrial production.