Research On Laser SLAM And Point Cloud Correction In Location Method Of An AGV

SLAM is an important technology for robot to realize intelligent autonomous movement in unknown environment.In this paper,AGV Robot's own map construction,positioning and laser point cloud correction methods are studied.The AGV Robot system with laser radar sensor is built on the simulation platform namely Robot Operating System(ROS).RPLIDAR A2 laser radar is used as the obstacle avoidance sensor and in the same time odometer is used as the internal sensor.And meanwhile,the improved SLAM algorithm is verified through simulation and the corresponding entity machine experiments.The main work of this paper can be list as follows:First of all,according to the research background of AGV and SLAM algorithm,the advantages and disadvantages of the existing laser SLAM algorithm as well as the applicable scenarios are analyzed.On the basis of some shortcomings of the corresponding SLAM algorithm in the positioning process,an entity machine is built for experiments by combining the commonly used methods of robot positioning as well as three coordinate systems and the theory of pose transformation method.In the process,the frame platform and prototype model are analyzed.The frame platform.includes the software framework and hardware platform of AGV Robot dominated by ROS platform,while the model construction includes the analysis of robot kinematics model and track reckoning model,as well as the introduction of common map models.This paper focuses on the mathematical model of the robot in the slam process and the common filter algorithms including Kalman Filter and Particle Filter based on Bayesian method,then makes comparison and selection after the introduces of the algorithms.In view of the point cloud distortion caused by the relative motion of the robot in the slam process,this paper analyzes the solution by combining the common algorithms of robot frame registration such as ICP algorithm and PLICP algorithm.Aiming at the point cloud data distortion caused by the relative motion of robot in the slam process,considering the shortcomings of existing algorithms,an improved point cloud correction method is proposed to solve the problem of map distortion and inaccurate positioning caused by the laser point cloud distortion of robot during the SLAM process.A series of simulation and experiments are carried out to verify the algorithm through the entity machine.Combined with the improved algorithm theory proposed in this paper,the data of the experimental process and the experimental results is analyzed at last.Simulation and experimental results synchronously show that the improved method can solve the problem of point cloud distortion well with high speed and good accuracy in the same time.