Research On Localization And Navigation Of Mobile Robot Based On SLAM

Autonomous mobile robot is one of the hot spots of current research in China.Aiming at the problem of "autonomous" ability of robots,the current research is mainly divided into two parts: 1)Simultaneous Localization and Map-Building(SLAM,Simultaneous Localization and Map);2)Navigation on the created map.SLAM provides environment map data for autonomous navigation of robots and makes it possible for robots to sense the environment and complete autonomous movement.The navigation is based on the known map for path planning,positioning and control to make the robot autonomously move to the designated target point.Based on the advantages of the open source ROS(Robot Operating System)in code reuse and expansion,this paper makes an in-depth study of the above two aspects.According to the needs of this study,Firstly,the Kinetic version of ROS system and Gazebo virtual physical simulation environment are built on the computer,construct the URDF Robot(Unified Robot Description Format)model,write a laser,IMU,UWB(ultra wideband technology positioning),encoder sensor software plug-ins(plugins)to build a set of complete test scenarios and Robot site hardware and software.Secondly,Map-construction in unknown indoor environment(SLAM)is realized by Gmapping an open source project.Finally,the positioning algorithm is optimized,and the controller design is completed to realize the autonomous navigation of the robot under the SLAM technology.This paper mainly focuses on the positioning and relocation problems in the navigation process,as well as navigation control issues.First of all,using the principle of extended Kalman filter,data fusion of sensors UWB,IMU and odometer is carried out to achieve relocation.In order to ensure the positioning accuracy,the fused positioning data and laser data are fused by Monte Carlo particle filter,and a more accurate position of the robot is estimated by predicting and evaluating a large number of particles.Aiming at the problem that the existing navigation methods of mobile robots are difficult to guarantee the accuracy and timeliness of navigation at the same time,this paper designs two kinds of path following control algorithms,namely PID controller and linear quadratic optimal controller,which can improve the efficiency and accuracy of mobile robots in large-scale environment,and implements them in software.