Research on SLAM navigation of wheeled mobile robot based on ROS

Abstract

In order to get a better mapping effect and use it in robot navigation, this paper first established the kinematic model and dynamic model of the wheeled mobile robot used in the experimental research, and determined the relationship between the relevant parameters of the mobile robot experimental equipment. After comparing and analyzing the advantages and disadvantages of different SLAM (Simultaneous Localization and Mapping) algorithms such as HectorSLAM, LagoSLAM, Cartographer, and GMapping, Monte Carlo adaptive positioning and laser mapping (GMapping) algorithms were selected for better results. And aiming at the problem that the traditional GMapping algorithm has no loopback and low accuracy, a solution of increase the closed-loop detection link to improvement is proposed. This GMapping algorithm is implemented through the ROS operating system. This algorithm improves the accuracy of mapping by changing the internal scanning registration algorithm. Finally, the improved algorithm and robot autonomous navigation technology are combined in experiments, and the accuracy of the improved algorithm's mapping is verified through many experiments.