Design And Research Of Sorting Robot Based On Machine Vision

Robotic sorting technology is widely used in fields such as farms,industrial production,logistics and warehousing,and industrial robots are the main performence for sorting.Traditional industrial robots are bulky,low in flexibility,and work in a structured environment.Once the work task changes or the environment changes,it need to be recompiled.Therefore,it is of great significance to use lightweight sorting robots combined with machine vision to replace traditional industrial robots to perform sorting tasks.Aiming at the problems existing in the traditional industrial robot sorting system,this paper designs a lightweight sorting robot,builds a sorting robot experimental platform based on machine vision,and completes the sorting of small castings.The main research contents are as follows:(1)In order to solve the problems of low flexibility,poor portability and low intelligence in the traditional industrial robot sorting system,a design scheme of sorting robot based on machine vision is proposed.The innovative tools in TRIZ theory is used to analyze the problems existing in the traditional industrial robot sorting system,the design scheme of the sorting robot body based on the integrated drive and control joints is obtained,which provides a basis for the design scheme of the sorting robot based on machine vision.(2)In order to improve the overall scheme of the sorting robot based on machine vision,the ontology design of the sorting robot is completed.Firstly,the topological analysis of the robot mechanism is carried out to complete the design of the robot structure.Secondly,the maximum torque of the joint load under the actual load is obtained by using the ADAMS simulation software,and the selection of the servo motor and harmonic reducer is completed.The strength of the parts such as the plate is checked to verify the rationality of the structural design.Finally,the production of the sorting robot is completed.(3)In order to study the kinematic characteristics of the sorting robot body,kinematics analysis and dynamic simulation are carried out.Firstly,the kinematic modeling analysis of the sorting robot body is carried out by using the standard and improved D-H method,and the forward and inverse solutions of the robot in the link coordinate system are completed.Then the trajectory planning of the sorting robot body is carried out in joint space and Cartesian space respectively.Finally,the dynamic simulation of the sorting robot body is carried out,which verifies the rationality of the design and selection.(4)In order to complete the construction of the sorting experiment platform,the overall scheme of the sorting robot is designed.Firstly,the system composition of the sorting robot is perfected,and the design concept is proposed.Secondly,the end effector,conveying system,pneumatic system and electrical system of the sorting robot are designed,and a flexible system based on the sorting robot is designed.Then the vision system and robot parameters are calibrated.Finally,a sorting robot experimental platform based on machine vision is built.(5)In order to verify the effectiveness of the sorting experimental platform,simulation and recognition grasping experiments are carried out.Firstly,a simulation platform is built based on the Robot Operating System(ROS)robot operating system.Secondly,the software design of the sorting robot control system and the overall control system is completed.Linemod’s template matching algorithm and deep learning algorithm based Darknet are used to identifie and pose the target object respectively,and then the appropriate algorithm is select to complete the grasping and sorting of small castings through comparative analysis.Finally,the experimental results are analyzed and summarized.In the end,the research content and innovation points of the full text are summarized,and the shortcomings of the research are pointed out,and then some suggestions are given for future research work.Figure [107] Table [17] Reference [147]...