Research On Grasping And Assembling Of 7-DOF Robot Based On Vision

The seven degrees of freedom robot has the flexibility of avoiding obstacles and can complete the tasks of grasping and assembly in complex environments.The integration of machine vision technology and robot technology can improve the automation and intelligence level of industrial robots.In this paper,the grasping and assembly process of a vision-based seven-degree-of-freedom robot is studied by combining machine vision technology with the Sawyer seven-degree-of-freedom robot.The main work includes:Firstly,the kinematics of the Sawyer robot is analyzed,and the forward kinematics model is established by D-H parameter method.In the problem of solving inverse kinematics,the Newton Raphson iterative method is optimized.The pseudo inverse method and damped least square method are used to avoid the singular configuration.By adding the damping coefficient,the problem of increasing the end error of the robot when the pseudo inverse method is close to the singular configuration is solved.The correctness of the kinematics solution is verified through simulation in the Robot Toolbox of MATLAB.Secondly,the identification and positioning of the shaft hole are studied.According to the imaging principle of the camera,the Kinect V2 camera is calibrated by Zhang Zhengyou calibration method,and the color image and depth image information of the axis are obtained.The 3D template of the axis is created,and the image information obtained by the camera is matched with the image information of the template by using the 3D template matching algorithm to realize the identification and positioning of the axis.Aiming at the image information of the hole obtained by the Sawyer wrist camera,a method of creating an image mask using HSV image is proposed.The contour of the hole is extracted by mask operation and image processing,and the identification and positioning of the hole are realized by shape recognition and minimum circumscribed moment.Then,the path planning of the Sawyer robot is studied so that the robot can avoid obstacles during the grasping process.Aiming at the problem that the expansion direction of the new nodes in bidirectional rapid random expanding tree(RRT-Connect)algorithm is very random,an artificial potential field is introduced to improve the RRT-Connect algorithm.A gravitational point is added between the starting point and the target point so that the new node expands in the direction of the gravitational point,reducing the search for useless space.Simulation experiments are carried out in the two-dimensional plane space and the Sawyer robot workspace respectively.The improved RRT-Connect algorithm can effectively shorten the path search time and improve the quality of path planning while avoiding obstacles.Finally,a vision-based seven-degree-of-freedom robot grasping and assembly system is designed,and a grasping and assembly platform is built.The pose of the axis in the robot base coordinate is obtained through hand-eye calibration,and the grasping path is planned by the improved RRT-Connect algorithm.The robot obstacle avoidance and grasping experiments are completed.The relative position of the shaft hole is judged by the force fed back by the torque sensor at the end of the Sawyer robot so that the robot can adjust the position of the shaft,and the shaft hole assembly experiment is completed,which verifies the effectiveness and feasibility of the designed system.