Design Of Semi Virtual Manipulator And Its Application Scenario Based On GAZEBO

Nowadays,the research on robot is in full swing,and has made great progress,and the research results are also widely used in the public domain.Robot arm is a kind of mature robot,which integrates the knowledge of computer,electronic information,control engineering and other fields.It is the best carrier of artificial intelligence learning,and has been widely used in the daily experimental teaching of colleges and universities.However,due to the high cost of mechanical arm equipment,most colleges and universities will be limited by the number of equipment,and the high frequency of equipment use leads to high damage rate,and the construction cost of university laboratory is high.At the same time,the application scenarios and development environment of fully virtual experiment are not real enough.In order to make the robot arm better used in university teaching experiment,not affected by the equipment and other factors,this paper adopts the semi virtual simulation method,changes the operation of the solid manipulator to the control of the virtual model,and designs the virtual interface version as the Control Middleware of the mechanical arm.The control instructions are transmitted to the simulation system by the interface board,and the machine vision technology is used to obtain Through the design of simulation model,the real scene is transformed into the simulation background,and the target object is detected by machine vision,and the position coordinates of the target object are transmitted to the simulation control program to control the motion of the virtual manipulator.After the experimental verification,the virtual manipulator in this design can successfully move to the target object in the simulation environment after receiving the position information of the target object.It can be concluded that the research of this subject has great value for the application of mechanical arm to teaching experiment in Colleges and universities.The main contents of this paper are as follows:(1)This paper analyzes the kinematics modeling of the manipulator,studies its forward kinematics model,inverse kinematics model and Cartesian space trajectory planning.It can not only control the robot arm to move to the target position by transferring the rotation angle of each joint,but also can reverse plan the rotation trajectory points of all joints in time by transferring the expected position information of the top of the manipulator,and finally achieves the expected goal Location.(2)Through the URDF file,the virtual manipulator model is created,and the physical properties such as inertial characteristics of the manipulator are described.The current international mainstream robot control system ROS system is used to control the manipulator,and the moveit software package in ROS is used to configure all the settings required for the manipulator motion planning.In order to better control the simulation manipulator,moveit controller is designed to manage the customized joint motion files of the manipulator,which can control and start all the joints of the manipulator.In order to realize the interaction between virtual manipulator rotation control and simulation display,a control architecture based on ROS is designed to realize the real-time communication between ROS and gazebo.After motion planning by moveit software package,the motion of virtual manipulator can be displayed in gazebo simulation software.(3)In order to realize the control of semi virtual manipulator,a virtual interface card is designed.The virtual interface card has two-way function.It can not only transmit control instructions to virtual manipulator,but also transfer the running state of manipulator to manipulator rotation control program.In the practical application process,the joint motion command sent by the program running in the embedded device can be converted into GPIO pin pulse signal through the virtual interface card,and then transmitted to the simulation platform through the interface to control the virtual manipulator movement.(4)In order to realize the combination of real scene and virtual simulation,machine vision function is added to the simulation design.The simulation model file is created,and the real scene images captured by the camera are imported into the simulation platform as the simulation background of the virtual manipulator motion.Fast r-cnn algorithm is used to identify the target object.Zhang's calibration method is used to obtain the camera internal parameters and distortion parameters.Hand eye calibration is used to obtain the transformation correlation between the position information of the target object in the camera coordinates and the robot coordinates.Finally,the virtual manipulator can obtain the correct target pose and plan the motion.Through the experiment to detect the position of various fruits in reality,and transfer the position information of the target to the rotation control program.After path planning,the manipulator in the simulation software successfully turns to the position of the target fruit in the simulation background.