Research On Indoor Mobile Robot Navigation Based On Lidar

Autonomous navigation of mobile robots has been a research hotspot in intelligent robots in recent years.Intelligent robots should not only be able to identify and locate the environment autonomously,but also be able to move and make decisions autonomously.In this article,a two-wheeled differential mobile robot equipped with ROS is used as the real platform,and lidar is used as the main sensor to solve the positioning,grid map creation and navigation problems of mobile robots.The main work of this article is studied from the following three aspects:1.Aiming at the problem of sensor information preprocessing in mobile robot navigation,this article pre-processes the data collected by wheel odometer and lidar,which improves the accuracy of mobile robot positioning.For the system error of the wheel odometer,this article mainly adopts the linear least square method,combined with laser radar,to linearly calibrate the pose information collected by the wheel odometer.For the problem of laser radar motion distortion,this article uses a linear interpolation method,combined with the advantages of the wheel odometer high frequency pose estimation period,to interpolate the data of each frame of the laser to remove the laser points caused by the movement of the laser radar in practice.Finally,the feasibility of the method adopted in this article is proved by experiments and data.2.Aiming at the accuracy problem of constructing mobile robot environment map,this paper studies the laser SLAM(simultaneous localization and mapping)problem based on graph optimization.Based on the principle of nonlinear least squares,this paper uses the method of Gauss-Newton optimization to optimize the pose graph of the data set,and its optimization effect proves the necessity of graph optimization.Finally,the robot's posture map is optimized in the established simulation environment,which helps the mobile robot to create a more accurate grid environment map and provides a basic environment for subsequent robots to achieve navigation.3.For the efficiency of global path planning for mobile robots,this article proposes a path planning method based on improved genetic algorithm,which effectively improve the efficiency of the global optimal path calculation of mobile robots in real-time navigation.In order to realize the global navigation of mobile robots on the grid map established based on lidar,this article designs a process to generate all the initial feasible paths by randomly selecting candidate points at the edge of obstacles and adding directivity factors,which is used to solve the global shortest path without collision.Finally,the simulation and experiment of the algorithm are implemented in this article,and compared with the effect of path planning achieved by ordinary genetic algorithms,which shows the feasibility and speed of the proposed method.