Research On Tracking And Aiming Control Method Of Eye-in-Hand Robot

Eye-in-hand robot detection and visual tracking of specific targets is an important research content in the field of robotics.The traditional robot vision tracking control method based on calibration technology is limited by the calibration results of the vision system and robot system.To overcome the impact of calibration accuracy on the system and avoid the cumbersome calibration process and realize the rapid aiming and tracking function of the eye-in-hand robot to a specific target,this thesis proposes an eye-in-hand robot for a multi-degree-of-freedom robotic arm with a vision system installed at the end.Tracking and aiming control system,which can quickly and accurately detect specific targets,and control the robot to achieve aiming and tracking of specific targets.The main research content of this paper is as follows:(1)Aiming at the problem of eye-in-hand robot for target detection of small target drones,this thesis studies and proposes an improved YOLOv7 algorithm for target detection of small target drones.For the small target of the UAV,it is proposed to use the convolutional attention mechanism module to improve the YOLOv7 network structure,optimize the network model,and strengthen the ability of the network model to extract features.Finally,the effectiveness of the proposed improved YOLOv7 network model is verified by comparative experiments,and its performance is better than that of YOLOv7.(2)Aiming at the problem of uncalibrated visual servo control of eye-in-hand robot,starting from uncalibrated visual servo,this thesis studies the uncalibrated visual servo control based on quasi-Newton method and the uncalibrated visual servo control method of eye-in-hand robot based on Gaussian quasi-Newton method.And an improved eye-inhand robot calibration-free visual servo control method based on the least square method.The effectiveness of the two methods is verified by experiments,and the better performance of the eye-in-hand uncalibrated visual servo control method improved by the least squares method is verified by comparative analysis.Finally,the uncalibrated visual servo of the eye-in-hand robot improved based on the least square method is tested on the UR5 robot,and the effectiveness and feasibility of the algorithm are verified.(3)Aiming at the trajectory tracking control problem of the eye-in-hand robot,the inverse kinematics modeling and dynamics modeling of the manipulator of the eye-inhand robot are first carried out,and then the PID control of the robot is studied,and the problems existing in the sliding mode variable structure control of the robot are studied.Convergence speed and chattering problems,a method of robot sliding mode control using improved exponential reaching law and nonlinear sliding mode surface is proposed,and the stability analysis is carried out by using Lyapunov function.According to the dynamic model of the established UR5 robot,the corresponding control scheme is designed,and the control of the proposed control scheme is verified by tracking simulation experiments and comparative analysis of the end trajectory of the eye-in-hand robot on the MATLAB and Coppelia Sim robot simulation platform performance.(4)Aiming at the problem of eye-in-hand robot vision detection and tracking,the overall scheme and control process of eye-in-hand robot tracking and aiming control system are designed.Then,the experimental platform required for the experiment is introduced.Finally,the simulation experiment analysis of the proposed eye-in-hand robot tracking and aiming control method is carried out to verify the effectiveness of the proposed method.