Research And Implementation Of ROS-based Indoor Positioning And Path Planning System

With the continuous development of science and technology,people are no longer satisfied with the robot as an auxiliary operation tool,and expect it to be able to achieve various work independently,so the autonomous robot is proposed.The autonomous robot is designed to carry sensors and controllers to complete the work independently without external human intervention.The key technology to make the robot work autonomously is the positioning and path planning of the robot.This thesis takes autonomous robots in indoor environments as the research object,the positioning,map construction,and path planning technology of indoor mobile platform are studied,which are verified by simulation and physical objects.The main research contents are as follows:(1)Based on the ROS(Robot Operating System)robot operating system,the upper system of the platform has been modularized and split,including chassis control nodes,keyboard command release nodes,positioning mapping nodes,path planning and navigation nodes,etc.,which improves the robot software Reuse and environmental adaptability of the robot.(2)Aiming at the kinematics of the mobile platform,two-wheel differential tracked motion model is selected.Through analyze the kinematics of the model to solve the relationship between the speed input of the platform model and the pose state,and use simulation modeling to verify that the platform moves freely The feasibility.(3)For the overall design of the physical prototype system,a distributed structure of "server layer(PC)+ operating system layer(Raspberry Pi 4B)+ driver layer(STM32 driver board)" is adopted.The drive layer designs the corresponding motor drive circuit,and realizes the autonomous control of the platform through the classic speed loop PID motor control method;the server layer and the operating system layer use Ubuntu and Ubuntu mate systems respectively,with the help of the distributed network of ROS,through the server The layer realizes the reading and display of the position of the mobile platform and the map environment,and at the same time completes the release of control instructions to the drive layer;the communication network of the whole system is built by wired serial communication and wireless network communication.Thus completed the construction of the physical prototype system of the mobile platform.(4)For the indoor positioning mapping part of the mobile platform,After analyzing the two SLAM(Simultaneous Localization and Mapping)algorithms based on filtering and graph optimization,the most representative Gmapping algorithm and Cartographer algorithm in the two SLAM methods are selected for further research and test in the simulation and actual environment,respectively.The results show that the map constructed by the Cartographer algorithm,which introduces the inter-frame matching method and the loop detection method,is more accurate,with few data distortions and accumulated errors in SLAM,and the map update speed is about 4 to 5 times that of the Gmapping algorithm;In a highly similar environment,due to the loop detection technology of the Cartographer algorithm,the positioning and mapping effect will be reduced.Through the research of the Cartographer algorithm,the least square method and quadratic interpolation method are introduced in the initial sensor data input part of the algorithm for multi-sensor data fusion preprocessing to increase the reliability of pose data and avoid the data distortion of lidar.(5)For the navigation part of path planning,the Dijkstra,A~*,and JPS algorithms in the graph search method are compared and simulated,and the JPS algorithm is used as the core algorithm of global path planning,combined with platform kinematics equations to carry out the basic path Optimization and smooth motion trajectory;At the same time,the DWA local path planning algorithm is added to the global path planning.The algorithm generates local offline paths and combines online dynamic obstacle weights and path evaluation functions to achieve obstacle avoidance under local path planning.Through simulation and measured analysis,the global path planning algorithm based on JPS is more efficient,and DWA local path planning algorithm can realize dynamic obstacle avoidance.The experimental results show that,based on the ROS system,choosing a two-wheel differential crawler mobile platform,using the Cartographer graph to optimize the SLAM algorithm,and the JPS fusion DWA path planning algorithm can complete the technical requirements of indoor positioning and path planning,and realize the ROS-based Indoor positioning and path planning system.