Design And Implementation Of Autonomous Grabbing Mobile Robot System Based On ROS

With the advancement of human living standards,mobile grasping robots are playing an increasingly important role.The task of a mobile grasping robot refers to the combination of an autonomous mobile platform and a robotic arm,and uses the information provided by related sensors to complete the perception of the unknown environment,the recognition of the target pose,and the grasping task.Mobile grasping robots are widely used in various fields due to their small size,flexible action,and convenient operation.At present,the research on mobile grasping robots has problems such as poor scalability,weak algorithm integration,slow processing speed,and insufficient system stability.This paper mainly studies the key issues of mobile grasping robot SLAM,path planning,target object pose estimation and manipulator motion grasping,and builds an autonomous mobile grasping robot platform based on ROS to verify related algorithms and functions.This paper divides the design of mobile grasping robot system into four modules according to functions: image acquisition,visual navigation,pose estimation and grasping control.The image acquisition module is composed of a camera sensor and an intelligent development board to provide a stable input source.The visual navigation module of the system is realized by using mobile robot,lidar and image acquisition module to further combine the navigation function package set of ROS and occupancy grid mapping algorithm.Among them,the path planning of the robot is realized by designing ROS motion control nodes through Dijkstra and A* path planning algorithms.With the development of acceleration chips and convolutional neural networks,image processing based on deep learning has higher speed and accuracy than traditional methods.This paper designs and uses the RGB-CNN-based pose estimation algorithm to accurately estimate the 3D pose of the target object,and transplants the algorithm to the intelligent high-speed development board to further complete the pose detection module of the mobile grasping robot.Further,based on the ROS-based robot motion planning framework Move It,a grab control function package is designed to complete the grab control module of the system.This paper designs a multi-machine communication mobile grasping robot system based on ROS.Multiple hosts distribute and calculate different contents to improve the stability and scalability of the system.In order to better detect and control the system,this paper designs an online monitoring software based on Qt and librviz library to improve the observability and controllability of the system.Finally,with the help of monitoring software,the system's mapping,navigation,pose estimation and grabbing modules were tested and tested,which verified the effectiveness and robustness of the system.