Research Of The Parallel Robot's Kinematic Calibration And Pose Tracking Based On Vision

The robot technology has been greatly developed with the in-depth implementation of "Made in China 2025 Plan".As one of the most important branches of robotics,the research of parallel robot technology has been taken seriously.Especially with the rapid development of machine vision technology in recent years,the fusion of parallel robot technology and machine vision technology has become a new and popular Research areas.Based on this background,this paper taked the 3-RPS parallel robot in our lab as the main object of study.Combining the machine vision technology.We realized the parallel robot's position and orientation detection,on this basis,we studied the parallel robot's kinematic calibration and pose tracking.That provides a new idea to slove the parallel robot technology problem with machine vision technology.Firstly,we studied the layout and positional relationship between the 3-RPS parallel robot and the image acquisition device.And then designed a rectangular marker for vision detection.Combinated the rectangular marker and the PNP position detection algorithm,we proposed overall scheme of the parallel robot's position detection.Based on the overall scheme of the position detection,we proposed the overall scheme of kinematic calibration and pose tracking of the parallel robot.Secondly,we proposed a method for predicting feature points' regions,which can automatically determine the regions of the rectangle marker's feature points.Combinated that with the feature points detection,we realized the recognition of rectangular markers.On this basis,we analyzed and studied a vision-based pose detection method,which realizes the detection of the parallel robot's position and orientation.Then we carried out a precision test experiment to verify the accuracy of vision detection.Thirdly,we analyzed and established the inverse kinematics model and error equation of the 3-RPS parallel robot.We also built a hardware and software experiment system based on vision for kinematic calibration of the parallel robot.Calibration experiment which including vision detection of parallel robot's pose,error parameter identification and inverse kinematic model correction was performed.Then we compared the parallel robot's control accuracy before calibration with after calibration,and the result shows that the control accuracy of the parallel robot has been improved obviously after the calibration by visual technology.Finally,we analyzed and improved the principle of the traditional pyramid-LK tracking algorithm,which makes it easier to trace the feature points of rectangular markers.By combining the improved pyramid-LK tracking algorithm with the pose detection method mentioned above,a method to realize the pose tracking of the parallel robot is obtained.We built the pose tracking experiment system of parallel robot.Then combined with the forward kinematics of the 3-RPS parallel robot,an ideal trajectory of the parallel robot's pose change can be obtained.We compared the ideal trajectory with the actual trajectory obtained by visual tracking.The comparison results show that the method proposed in this paper has a good effect on the tracking of parallel robot's pose.