Design And Research Of Indoor Autonomous Mobile Grabbing Robot Platform Based On ROS

In recent years,with the rapid development of computer technology,sensor technology and artificial intelligence,indoor mobile robot technology has also continuously moved toward independent intelligence,such as warehouse logistics robots,welcome robots,meal delivery robots and sweeping robots.However,most of these mobile robots use infrared guidance,magnetic strip navigation and two-dimensional code navigation technology,and most of them have the disadvantages of fixed track,non-autonomous movement,dependence on special environment,time-consuming installation and high cost.ROS(Robot Operating System),an open source robotic operating system,provides a complete and open robot research and development platform,providing an easy-to-use framework and a library of many open source algorithms,allowing users to quickly develop advanced robotic applications.Therefore,based on the ROS system,this paper designs autonomous mobile crawling robot platform based on Navigation and MoveIt architecture,and tests and experiments on the function.Autonomous mobile grasping robot is a kind of mobile platform with manipulator.It combines the advantages of mobile platform with large range of motion and high degree of freedom of manipulator.Its autonomous function comes from the perception of lidar and depth camera,ROS system and upper level planning of open source algorithm.Robot host computer controller uses raspberry pie 3B and PC,and uses master-slave computer and remote control mode,which alleviates the computing pressure of a single platform and facilitates robot debugging and human-computer interaction.The lower controller of the robot uses Arbotix_M and KFBC4S1 to control the motion of the manipulator and chassis and to receive data feedback.HLS_LFCD lidar and Kinect_V1 depth camera are used for mobile positioning,navigation and target recognition respectively.For mobile navigation platform,the ROS Navigation architecture is used to build the robot navigation system.Specifically,two-wheel differential motion model is used to calculate Odom,lidar and Hector are used to realize SLAM mapping,AMCL is used to locate the robot,and Move_base node is used to realize path planning.The results are sent to the lower computer controller to realize the navigation of the mobile platform.For the navigation platform of the manipulator,MoveIt architecture and depth camera are used to build the navigation system of the manipulator.The platform uses camera to obtain RGB and Depth diagrams,uses template matching to realize target recognition and positioning,uses PCL point cloud to reconstruct Depth images,calculates the three-dimensional coordinates of the center of shape of the identified object,and sends the results to the manipulator control system,then writes the Cartesian motion planning program,and calls on OMPL's RTConnect algorithm to carry out the manipulator motion planning.The manipulator combines visual positioning and grabbing operation.Finally,according to the functional requirements of the robot design,a series of tests and experiments are carried out using the robot platform to verify the feasibility and stability of the robot platform,which lays a foundation for the subsequent design and development of more accurate,efficient and stable industrial products.