Research On Mobile Robot SLAM Technology Based On ROS

Simultaneous Localization and Mapping(SLAM)could provide the means to make a mobile robot totally autonomous.SLAM is the process whereby an entity(robot,vehicle or even a central processing unit with sensor devices carried by a person)has the capacity for building a global map of the visited environment and,at the same time,utilizing this map to deduce its own location at any moment.Recently,SLAM research is gradually shifting from traditional filter framework to graph optimization framework.A large number of research literatures also showed that the performance of SLAM based on graph optimization is better and more promising.ROS is an open source distributed robot software development system.It has a wide variety of robot hardware abstraction layer interface,and also embedded some convenient functions in the robot development process,such as the debugging tools and visualization simulation tools.Therefore,ROS is gradually being widely supported and loved by robot researchers.The main content of this paper is the mobile robot SLAM technology based on ROS.The SLAM solutions based on filtering method and graph optimization method were discussed respectively.On the basis of them,the corresponding improvement measures were put forward,and the experimental results were obtained and proved to be effectiveness.The specific research work of this paper mainly includes the following aspects.Firstly,this paper reviewed the research background of SLAM and its development of domestic and foreign,introduced the application scenario and research direction of SLAM,and determined the research direction of the topic.Secondly,two methods of SLAM problems were introduced from the perspective of probabilistic state estimation and nonlinear optimization: RBPF-Based SLAM and Graph-Based SLAM.For RBPF-Based SLAM,the improvement measures based on DBSCAN clustering algorithm and APGDR resampling algorithm were proposed.According to Graph-Based SLAM,the method of extracting ORB feature points based on grid segmentation was proposed.According to the feature point distribution index of the grids,the region of the ORB feature extraction was selected and the location distribution of the feature points was homogenized by the DBSCAN algorithm.Lastly,the integrated system framework of mobile robot experiment platform was designed and built.The transplantation,coding and debugging of all functional modules were completed.A series of experiments were conducted to verify the proposed method.The experimental results show that the proposed methods are correct and effective.