Research And Implementation On Mobile Manipulator Motion Control Based On Uncalibrated Visual Serving

In the combination of manipulator and visual perception system,manipulator visual servo system build visual closed loop control circuit through visual perception information.That makes the robot system can adapt to more complex working environment,be a more intelligent.Which,the calibration visual servo control method based on the unknown camera parameters is widely used in manipulator visual servo system,such as the image Jacobin matrix method,the intelligent fitting method and on some special visual layout,the direct control method based on image error.With the application of the robot arm servo system,the wider the working environment is,and the working environment is becoming more and more complicated,then the fixed robot arm is no longer meet the demand.The mobile robot arm servo system on the carrier platform is paid more and more attention.There is still an unresolved difficulty in the study of the visual servo system of the mobile manipulator relative to the fixed manipulator where the working environment of the mobile manipulator is dynamic and system parameters are subject to dynamic environment interference during the course of the movement,where the mobile arm control system is required to have strong antiinterference and robustness.Based on position positioning of the non-calibration visual servo task of the mobile robot arm system in the Cartesian space,in this paper,the following main work are conducted as following:(1)Due to the eye-to-hand binocular vision sensing system,a mobile servo controller with no calibration of the mobile arm is proposed.The controller adapts the specific image features of the visual system,and completes the intelligent control of the mobile platform and the robot directly owing to the image.The stability of the controller is proved,and the input and output characteristics of the controller are given to the binocular vision system.(2)For the eye-to-hand binocular vision system,the input and output characteristics of the binocular vision system is analyzed qualitatively and quantitatively based on the established visual image model from the camera image space to the Cartesian space.Combined with the visual system requirements of proposed mobile arm without calibration of the visual servo intelligent controller,the two cameras gets the ideal installation point and margin at this time,meanwhile conducting the experiment.Theoretical analysis and experimental verification show that the eye-to-hand binocular vision system has a large installation margin at the ideal mounting point,which can meet the requirements of the proposed controller,and has strong robustness and antijamming.(3)Use the controller designed in this paper to complete the physical experiment based on the two-wheel differential wheel mobile platform,Million learn through six degrees of freedom arm and mobile arm servo servo physical platform built by Ordinary camera in order to verify guiding significance and practical feasibility in the non-calibration of the visual servo system of the analysis and conclusion of he visual mapping of binocular vision system in this study.