Research On Improved Visual SLAM Algorithm For Mobile Robot Based On ROS

The research of mobile robots began in the late 1960s.In recent years,with the continuous development of high and new technology,mobile robot research is developing more and more rapidly,more and more aspects of research are involved,and its application is more and more extensive.Mobile robot in the unknown environment to achieve simultaneous localization and environmental map building(SLAM)process is currently a major field of robotics research,and visual SLAM research is an important research direction in this field.This dissertation presents an improved visual SLAM algorithm for mobile robots based on RGB-D vision sensor.Firstly,we introduce the development history and research status of mobile robot,and put forward the key problem in the process of intelligent mobile robot,that is,location and surrounding environment.Then we introduce the SLAM algorithm which is mainly analyzed in this dissertation.The SLAM algorithm can be divided into two categories:laser and vision.Compared with the two categories,although the laser algorithm is early,theoretical research and engineering reference are relatively mature,but there are still a lot of shortcomings.Therefore,this paper focuses on the research of visual SLAM algorithm.It includes the basic knowledge of the algorithm,the process and so on.Secondly,the RGB-D visual SLAM algorithm based on the Kinect is introduced in detail.The front-end and the back-end of the algorithm are described in the steps.For example,three feature points detection algorithms,SIFT,SURF and ORB are introduced.By comparing and analyzing the advantages and disadvantages of each algorithm,we select ORB algorithm to extract and match feature points.A semi random loop closures detection method is proposed in this dissertation.This method improves the accuracy of matching and reduces the amount of computation by changing the extraction of key frames.Finally,using the improved SLAM algorithm in the laboratory environment to test the actual environment,using the wheeled vehicle as the carrier,the robot operating system(ROS)as the experimental platform,to verify the feasibility and practicability of the improved SLAM algorithm which is proposed in this dissertation.