Research On Mapping Technology Of Mobile Robots For Unknown Environment

Map construction in an unknown environment is the primary prerequisite for mobile robots to achieve autonomous navigation,and is also the focus of research in the fields of computer vision and intelligent robots.Synchronous positioning and map construction,that is,SLAM technology,is the key to solving the problem of map construction,but it cannot be ignored that SLAM technology using a single sensor still has many limitations.For example,visual sensors often lose features in complex situations.The detection range of single-line lidar is limited.Therefore,this paper proposes an improved method of binocular vision fused with inertial measurement unit(IMU)data,and on this basis,an in-depth study of the fusion scheme of laser and vision is conducted to further enhance the mobile robot's ability to perceive the unknown environment.The main research contents and innovative achievements of this paper are summarized as follows:1.Establish the kinematics model and odometer model of the two-wheel differential mobile robot,analyze the sensor models such as the binocular camera and lidar used in this article,and conduct an in-depth study on the calibration method between the sensors,and complete Calibration of the sensor.2.On the basis of completing ORB feature point extraction and matching,the EPn P method is used to achieve camera motion estimation.At the same time,the back-end optimization method based on the graph optimization framework is used for optimization,and a complete visual SLAM system is constructed.Conduct experiments on public data sets.Finally,several classic laser SLAM algorithms are studied,and comparative experiments are performed based on the Gazebo simulation platform.3.Aiming at the defect that the binocular vision system is easy to lose feature information,an improved method for fusing vision and IMU data is proposed,that is,preintegrating the IMU data first,and then using non-linear optimization methods to fuse the vision and IMU data,and finally Based on the Euroc dataset,experiments were conducted to test the robustness and accuracy of the method.4.In view of the limitation of the detection range of single-line lidar,a laser-vision fusion method based on grid map is proposed,that is,the three-dimensional stereoscopic information obtained by improved vision is used to supplement certain data for lidar.Not only can the map be directly used for robot navigation,but its map information is more complete,which effectively enhances the mobile robot's ability to perceive the unknown environment.5.Use mobile platforms,binocular cameras and other equipment to build an experimental platform,and verify the map construction method proposed in this article.The results show that the multi-sensor fusion scheme proposed in this paper can effectively use the feature information in the unknown environment and construct a more accurate and complete environment map.