Study On Trajectory Planning Algorithm Of Industrial Robot Based On Visual Guidance

Industrial robots are now applied to all aspects of intelligent manufacturing,combining with machine vision is the inevitable trend of the development of industrial robots.The environment perception ability is enhanced of the robot and its application scope is widened by machine vision.Application of trajectory planning based on visual guidance,the adaptability of the robot is improved,the mechanical impact and vibration are reduced,and its service life is prolonged.Based on the background of assembly system,the trajectory planning algorithm for industrial robot based on visual guidance is researched.The following four aspects are researched in this paper:(1)The robot vision system is researched.The imaging principle and calibration method of monocular and binocular vision are analyzed.For monocular vision system,a template matching algorithm based on image acquisition,standard template establishment and similarity measurement is designed to locate the object.For binocular vision system,a visual algorithm based on gray threshold segmentation,Hough transform and stereo matching is designed to realize the 3D positioning of the object.(2)The kinematics modeling of industrial robot is researched.The pose description,coordinate transformation and D-H modeling method of the robot are introduced.According to the structure of the 4-DOF SCARA and the 6-DOF ABB robot,the D-H coordinate systems are built.The forward kinematics model of the robot is solved by using D-H parameters,and the inverse kinematics model is obtained by solving forward model inversely.Finally,the correctness of the models are proved through the experiment.(3)The trajectory planning algorithm based on visual guidance is researched.The existing trajectory planning algorithms of robot are studiedand compared.For the plane assembly system of four axis robot,the monocular vision is used to locate the workpieces,and the five polynomial interpolation algorithm is used for trajectory planning.The effectiveness of the algorithm is verified by the simulation of the trajectory of each joint.For the assembly system of six axis robot,binocular vision is used to complete the 3D positioning of workpieces.A trajectory planning algorithm combining linear interpolation and five polynomial interpolation is designed.The better performance of this algorithm is proved by comparing with the common interpolation algorithms.(4)The robot trajectory planning system is established.Two sets of robot assembly systems are implemented,which are SCARA robots based on monocular vision and ABB based on binocular vision.The hardware of the system is mainly composed of camera,robot and industrial computer.The software includes image processing,trajectory planning,Socket communication,target assembly program and human-machine interface.The camera calibration,workpiece location and robot assembly experiments are designed to verify the function of the system.The experimental results show that the calibration errors of monocular and binocular vision are 0.0337 and 0.001 pixels;the monocular vision recognition rate is 98%,the maximum location error is 5 pixels,and the matching time is 20ms;the maximum error of the binocular vision location is6 pixels,the height measurement error is 0.5mm,and the positioning time is100 ms.The workpieces can be accurately assembled by the robot in the experiment,which proves that online trajectory planning of industrial robot based on machine vision can be realized.