Research On Robot Visual Servo And Tracking System

Intelligent robot visual servo is defined as this: using image features from the image sensors for feedback information as a feedback loop control of the robot's position. namely using the robot vision sensor to indirectly detect the current position and orientation relative to the target body or its relative position and attitude, in this based on the positioning of the robot control or tracking.. Robot visual servo system's distinguishing feature compared to other robot control system is the introduction of visual information as a feedback sensor signal.In the image processing stage, real-time images are captured processed to extract the target image. Firstly analyzing the moment of the images,the image features is acquired such as position and velocity in the image planer. Secondly we designed a linear and nonlinear integrated trajectory predictor to estimate the motion state in the next period in video real-time tracking.For Eye-in-hand structure of the visual system configuration, the image-based visual servo control theory is researched, using image features for feedback in servo control so as to tracking the moving target by robot. The error between the expected image features and the current features is the input to the Jacobian matrix ,which shows the relationship between the bias in pixel planer and the robot joint space. To design the visual servo controller for tracking ,the Model and the coordinate translations of the robot and camera is analyzed, upon that the Jacobian matrix is deduced for our experimental system. then using the inverse Jacobian matrix of IBVS servo designed a visual servo controller, and verify the simulation and experimental study of the feasibility of the algorithm, such as real-time performance and accuracy.During the visual servo job using Image features, assume that the target was moving in a plane, through analyzing the system model of the Eye-in-Hand robot configuration which consist of camera model and robot model, image Jacobian matrix based on image moments features and target predictive centric feature is derived. Than use the inverse Jacobian matrix and the IBVS servo method for the visual servo controller designing. On this basis, taking into account the delay of the image processing system and servo control algorithm, with predictive compensation Jacobian image PI feedback controller, without reducing the tracking accuracy, and improve the response speed of the system. Finally, the toolbox of MAT LAB visual servo simulation algorithm with two degrees of freedom tracking, positioning and tracking experiment in the build of MOTOMAN-UP6 robot experimental system could quickly and efficiently tracking on moving targets in the plane, the system response and tracking accuracy speed meet the requirements .