Study On Simultaneous Localization And Mapping Of Laser And Vision Fusion For Mobile Robot

Simultaneous Localization and Mapping(SLAM)is a core technology for mobile robots to complete autonomy and intelligence.Laser SLAM performs well in small environments,but in large environments,due to detection of loopback errors or cumulative errors If it is too large,it will cause large distortion in positioning and mapping,and its amount of information is small,so it can not recognize the environment well in the environment with high geometric similarity.On the contrary,visual SLAM is rich in information and can identify the environment very well,but the construction of the map is not accurate enough,and the calculation amount due to the rich information will be very large,and the real-time performance of the algorithm is not high.Therefore,the use of a single sensor of the robot will be greatly limited,so this paper proposes an algorithm for fusion of laser and vision.The rich semantic information of the visual sensor is used to assist the laser sensor in simultaneous positioning and mapping,thereby enhancing the robot's adaptability to change.Environmental capacity.This paper combines the advantages and disadvantages of visual SLAM and laser SLAM to study and implement the SLAM algorithm of the mobile robot that combines the two.The main work of this article is as follows:1.Research on FastSLAM algorithm for lion optimization.On the basis of expounding the principle of FastSLAM algorithm,aiming at the problems of FastSLAM algorithm particle degeneration and multiple particle loss,the idea of lion group optimization was introduced into FastSLAM algorithm for improvement,and the proposed simulation was verified by comparison with FastSLAM algorithm The feasibility of the algorithm.2.Research on SLAM algorithm of laser and vision fusion.Aiming at the defects of the traditional laser SLAM algorithm,a SLAM algorithm of laser and vision fusion is proposed.The method of using the monocular camera's pose to merge the lidar pose to improve the pose accuracy is combined with visual loopback detection and laser loopback detection.The loopback detection is more accurate to eliminate accumulated errors better,and simulation experiments are conducted using the data set.The experiment proves that the fusion algorithm has greatly improved performance compared with the traditional algorithm in terms of pose accuracy and loopback detection accuracy.3.Experimental verification of SLAM algorithm fusion of laser and vision.In the laboratory(2m×5m)and outdoor corridor(21m ×19m),the fusion algorithm and the traditional algorithm proposed in the paper were tested on-site.Through experimental comparison,it is proved that the SLAM algorithm proposed in this paper improves the problem of inaccurate loopback of the built map due to excessive cumulative error in certain environments,and improves the inaccuracy of the pose caused by the robot's single sensor due to environmental constraints problem.figure[41]table[8]reference[61]...