Research On Robot Visual Guided Assembly Operation Based On Uncalibrated Visual Servoing

Visual servoing technology in recent years in the field of robot research has been rapid development and high priority.It is of great theoretical and engineering application to carry out the visual guidance and assembly technology of the robot without calibrated visual servoing.In this paper,the uncalibrated visual guidance assembly process based on support vector regression(SVR)is studied by combining the theory and experiment.Based on the characteristics of the target image and the composition of the visual system,the general scheme of the visual system is carried out.The visual guidance system of the module based on PC is designed and constructed.The visual guidance software system including the image acquisition module,the processing module and the motion control module is designed in the ROS environment.On the basis of the above hardware and software,the visual servoing control system has been studied in detail.Based on the experimental requirements of real-time control of uncalibrated visual servoing,the look-and-move closed-loop control mode is selected.The design of the control system with no fixed visual servoing is completed according to the visual mapping model and the proportional control method established under the pinhole camera model.Secondly,a uncalibrated visual servoing control strategy based on support vector regression is proposed.In order to ensure the performance of the visual guidance control system,the image feature extraction method of the uncalibrated visual guidance system based on threshold segmentation and edge extraction algorithm is studied and the image processing scheme satisfying the on-line experiment requirements is determined.Based on the support vector regression model and the visual mapping model,the Jacobian matrix model based on SVR is designed.In the C++ environment,the SVM module is used to analyze the images obtained by the collected samples.In this paper,the Jacobian matrix model based on support vector regression theory is introduced.Feature and joint angle information.The unknown parameters in the Jacobian matrix model based on SVR are obtained,and the calibration of the Jacobian matrix model based on SVR is realized.On this basis,the visualization of the uncalibrated visual vision The planning and design of the guide assembly process.Finally,the experimental platform is built and the experiment is carried out based on SVR robot without calibration visual guidance assembly.The performance of the control system is analyzed and the experimental conditions are optimized according tothe experimental results,which ensure the accuracy of the assembly target in the assembly process.Based on the visual servoing to complete another assembly target object of the crawling experiment,and constantly optimize the lighting environment and other experimental conditions to ensure that the target object of the crawl accuracy,and ultimately,the gap assembly assembly task successfully completed.The entire visual guidance assembly operation validates the effectiveness of the uncalibrated visual servoing system controller and control strategy.The study is to further develop the theoretical and experimental basis for the more efficient and intelligent application of visual servoing to service places such as homework,entertainment and leisure.