Research On Map Construction Method Of Indoor Mobile Robot Based On RGB-D Camera

With the continuous development of science and technology,the related research in the field of robotics has been widely concerned.The technology of simultaneous localization and map building(SLAM)for indoor mobile robots is the premise of autonomous navigation,and also the basis for many functional robots to perform other tasks.Therefore,it is of great significance to study the map construction method of indoor mobile robot.The main contents of this paper are as follows:(1)Firstly,the mobile robot platform and all kinds of sensors selected in this paper are studied.In this paper,the cost is reduced by using Kinect 2.0,which is an RGB-D camera,and the calculation efficiency is improved by using TX2 computer module.The principle of Kinect 2.0 camera using depth map to calculate point cloud is analyzed,and the transformation relationship between image pixel coordinate system and world coordinate system is obtained.Then according to the mobile robot selected in this paper,the robot coordinate system,robot motion model and robot observation model are established.Then,the paper studies the ROS operating system used by the mobile robot,completes the design of the software framework,analyzes the relationship among the four coordinate systems involved in the robot system,and obtains the coordinate system conversion relationship of the robot in this paper.(2)The basic theory of SLAM algorithm is analyzed,then RBPF-SLAM algorithm and Gmapping SLAM algorithm are studied,and the related theoretical derivation is carried out.After that,the improved Gmapping SLAM algorithm based on niche genetic optimization is proposed as the mapping method in this paper.The method uses crossover and mutation operations to update particles,and sets the upper and lower threshold of the effective particle number to improve the diversity of particles,so as to achieve the effect of improving the mapping accuracy.(3)In this paper,the real scene mapping experiment is carried out,the RGB image and depth image are obtained by Kinect 2.0 camera,and the image data is converted into lidar data.Then,the feasibility of Kinect 2.0 camera used to build two-dimensional planar map is verified by comparing two different real scenes,and the experimental data shows the effectiveness of the improved algorithm.At the same time,the comparative experiments of different particle numbers are carried out,and the optimal particle number suitable for the experimental equipment and experimental environment is 80.Through the specific research in this paper,it can effectively improve the drawing accuracy of the mobile robot,and reduce the cost,which lays the foundation for the mobile robot to complete more advanced tasks,and has important practical application value.