Research And Implementation Of Automatic Navigation Of Warehouse Robot Based On ROS

The application scenarios of most robots are relatively simple,it is difficult to adapt to changes in the external environment,and they do not have the adaptability to multiple comprehensive scenarios.Therefore,this paper will propose to build a mobile robot based on ROS for autonomous navigation and distribution in the warehouse environment to improve the efficiency of goods circulation in the logistics industry.This paper builds mobile robots and models based on ROS,and conducts research on autonomous navigation inside the warehouse.Firstly,the characteristics of ROS operating system are briefly described,and then the whole system is systematically analyzed.According to the two-wheel differential robot model,mathematical modeling and analysis of the overall software and hardware and some important hardware modules are carried out,and the model principle is deduced to prepare for the subsequent autonomous navigation and map construction of the robot.Aiming at the problem of mapping in the indoor environment of mobile robots,two typical SLAM methods in the field of lidar are deeply studied: the Gmapping SLAM method based on RBPF and the Cartographer SLAM method based on graph optimization,and the principles of the two methods are discussed in detail.Two comprehensive scenarios were built in the ROS simulation tool Gazebo,and the Gmapping SLAM and Cartographer SLAM methods were used for robot mapping and positioning for these two comprehensive scenarios,and the results of the mapping were evaluated.Using the existing robot platform,the robot was used in the laboratory to scan and map the overall layout of the laboratory,and the indoor two-dimensional plane map was successfully drawn,which verifies the feasibility of the mobile robot for SLAM mapping in the warehouse scene.sex.Aiming at the requirement that the robot can avoid obstacles in real time in the process of autonomous navigation,the global trajectory planning methods Dijkstra and A* are firstly analyzed in depth,and the performance of each is briefly analyzed,and the two-way A* is proposed to further improve the search efficiency.In order to solve the problem of low efficiency of A* route modification,dynamic A* is introduced to solve the problem of low efficiency of A*when modifying the route.In this paper,the dynamic window method is mainly studied,the principle of the dynamic window method can be analyzed and deduced,and different trajectory evaluation parameters and the radius of the obstacle are modified in matlab,which verifies the feasibility of the dynamic window method with different parameters for local trajectory planning.and select the optimal parameters as the local trajectory planning method of this project.The autonomous navigation framework of the robot is introduced,and then the preparation of autonomous navigation experiments is started.Firstly,the warehouse environment is built in Gazebo,and the map construction and autonomous navigation experiments are carried out respectively,and the feasibility of autonomous navigation in the simulation environment is verified.Two sets of autonomous navigation experiments were also done in the laboratory using the robot,and all of them could be completed.In order to verify the feasibility of real-time obstacle avoidance,a single dynamic obstacle avoidance experiment was conducted to verify the feasibility of the parameters of the dynamic window method.