Research On Autonomous Navigation Mobile Robot Basedon ROS And Laser SLAM

The e-commerce “front-storage” mode is an emerging warehousing model proposed in the context of “new retail”,which puts new demands on warehousing and handling methods.With the development of mobile robots and artificial intelligence,the “goods to people” handling method based on mobile robots is replacing the traditional “people arrival” handling method.However,current mobile robots in the "goods to people" handling mode often adopt road sign navigation methods such as two-dimensional code and magnetic wires.Due to the poor path flexibility and the time-consuming and laborious construction,such methods are difficult to adapt to the flexible business needs of the “front-storage” mode.In addition,the current development of mobile robots is mostly based on dedicated software systems,and the openness is poor,and the portability and scalability of the system are greatly limited.In this regard,this paper studies the more autonomous and open warehouse mobile robot based on ROS and laser SLAM under the “goods to people” handling mode.The main research contents and results are as follows:(1)According to the characteristics and requirements of the "goods to people" handling method,a mobile robot platform was designed and developed,in which the ROS architecture was adopted for the design of motion control,mapping positioning,navigation system structure and software function system.The basic test results of the platform chassis show that the position control error of the robot developed in this paper is less than 5 cm within 5 meters and can meet the needs of construction and navigation in the storage environment.(2)A method of constructing indoor environment map based on laser SLAM algorithm is studied and completed,in which Gmapping and Cartographer algorithms are used for mapping and comparison.The experimental results show that the Cartographer algorithm has higher mapping accuracy and is more suitable for the storage environment.(3)A real-time location method for warehousing mobile robot based on LSTM is proposed.The method realizes the precise positioning of mobile robots in the warehouse environment by preprocessing,model training and application of lidar sensors and map data.The experimental results show that the positioning method is superior to the commonly used adaptive Monte Carlo method and the CNN-based method,and the average positioning error is less than 10 cm,meet the positioning requirements of mobile robots in the storage environment.(4)Finally,the mobile robot navigation system of “goods to people” was designed and implemented.The system introduces the cost map to convert the static map into a navigation map,and performs path planning through the A* or Dijkstra algorithm in the cost map.Then the path is completed according to the TEB algorithm.The experimental results show that the navigation system can successfully complete the storage environment.The trajectory error is less than 15 cm and end point error is less than 10 cm,and meet navigation requirements in a warehouse environment.