Research On SLAM Technology Of Mobile Robot Constrained By Artificial Landmarks

With the rapid development of intelligent robot technology,the process of intelligent and real-time reform in the surveying and mapping industry has been accelerated.Among them,the Simultaneous Localiz Action and Mapping(SLAM)technology of robots in the unknown environment has become a hot research direction in the entire industry,which has promoted the development of traditional surveying and mapping in the direction of real-time dynamics.However,the classical SLAM algorithm only relies on visual odometry.The location state estimation is easily affected by environmental factors such as lighting,environmental textures and dynamic scenes.Therefore,this paper proposes a method based on artificial landmark constraints to conduct synchronous positioning and mapping technology research on robots in unknown environments.The main research results are as follows:(1)For the conversion problem between multiple coordinate systems in the motion constraint system,this paper discusses the coordinate expression and mutual conversion relationship in different coordinate systems,and realizes the unification of the coordinate relationship;for the self-constraint problem of the rotation matrix,The mathematical form of pose expression based on Lie group Lie algebra is discussed,and the local nonlinear optimization of the robot pose is carried out to obtain the local optimal solution.(2)In order to solve the problems of misrecognition and missed recognition of the vision sensor in the process of AprilTag code detection and recognition,based on image preprocessing,this paper improves the filtering of the image,improves the retention of image edge information,and combines quadrilateral optimization.The algorithm and Canny operator improve the accuracy of image edge extraction,and realize the acquisition of artificial landmark position information through recognition and decoding.According to different forms of AprilTag codes in practical application scenarios,verification experiments of recognition robustness and accuracy are carried out.The experimental results show that compared with the traditional recognition algorithm,the algorithm in this paper has higher recognition robustness and accuracy,and the recognition success rate is increased by 5.40%,which provides a guarantee for the subsequent constrained robot positioning.(3)Aiming at the error accumulation problem of classical SLAM technology,this paper proposes a SLAM localization method for mobile robots constrained by artificial landmarks.The AprilTag code laid out in the scene is accurately identified by the camera,combined with the motion state of the robot estimated by the visual odometry,the motion pose constraint model is established,the global optimization of the motion state of the robot is realized,the optimal pose is obtained,and the positioning method and Accuracy comparison experiments of classic SLAM localization methods.The experimental results show that the constrained localization method proposed in this paper reduces the root mean square error by 5.51%and 38.23%,respectively,compared with the classical SLAM localization method under two different motion forms,namely linear and circular,which verifies the effectiveness of the proposed method.