Development And Design Of Embedded Control System For Mobile Robots Based On Micro-ROS

In recent years,under the general trend of "machine substitutions",my country’s various robotics markets have grown rapidly,and the development of the domestic robot industry has also made great progress.The mobile handling robot,which is an important branch in the field of robotics,plays an increasingly important role in production and life,which greatly improves production and life efficiency.This article uses robot-related knowledge to design a mobile robot using the micro-ROS architecture,and equipped with Cartographer SLAM to complete the construction of environmental maps in an unknown environment.The specific content is as follows:(1)The micro-ROS development environment was built on the PC side,and the design program verified that the micro-ROS node interacts with ROS2.According to the robot application scenario,the robot of the four-wheel motion structure was selected and the differential control movement was used.Perform motion analysis of mobile robot modeling.For the disadvantages of wheeled odometer,IMU is introduced for data fusion.Finally,the advantages and disadvantages of several common environmental maps were analyzed.(2)Complete the overall construction of the robot system.The control system is divided into three parts,and the implementation functions of each part are summarized.For the construction of the hardware system,the main control processor,sensor,motor drive board and other hardware are selected according to the needs of each task module.The hardware circuit design was carried out for the underlying motion control board,and the software program design of the robot system was finally completed.Based on micro-ROS architecture,it provides wired serial port and wireless network to realize the control of the robot.(3)The model of the SLAM problem and algorithms for solving the SLAMbased problem are presented.Mathematical modelling of the grid map is carried out,and experiments are conducted to build the map using a dataset providing known positioning.The mainstream Gmapping algorithm and the Cartographer algorithm were analyzed and validated using dataset simulations,and finally the two algorithms were compared on the basis of the same dataset and the Cartographer algorithm was found to be more accurate.This led to the introduction and implementation of the Cartographer algorithm in the micro-ROS-based robot system.(4)Firstly,the performance of the robot was tested,mainly including communication test,speed response test,straight ahead accuracy test,rotation accuracy test and odometer positioning test of the robot system.A robot simulation model based on XACRO files and simulation environment was built in the Gazebo environment and maps were created using the Cartographer algorithm with good results.Finally,SLAM mapping of the real environment was carried out using the robot entities and a high accuracy map was obtained by comparing the map feature point distances with the map measurements.The CPU occupancy was calculated to verify the real-time performance of the algorithm.