Research On Laser SLAM Algorithm For Autopilot

In the context of the current vigorous promotion of artificial intelligence and intelligent transportation,the automatic driving technology is well known before everyone.Aiming at automatic driving,the thesis chooses laser radar as the leading factor and takes intelligent car as an example to simulate automatic driving for relevant experiments.The details are as follows:(1)Analyze and compare the application and current situation of the research on the algorithm of laser radar localization and mapping(SLAM)at home and abroad.The movement process of the smart car and sensors such as laser radar are modeled,and the classification,flow and framework of SLAM algorithm are summarized.(2)Secondly,the thesis studies several laser SLAM algorithms based on feature estimation methods,and then uses the RP LIDAR A1 laser radar carried on the smart car for positioning and map construction,and analyzes its advantages and disadvantages.The laser SLAM algorithm is improved,ICP algorithm is used for front-end matching process,and Matlab is used for simulation and error analysis.(3)The performance and error of multisensor fusion methods based on extended Kalman filter and lossless Kalman filter are compared,and each part of sensor information is fused.Build a ROS-based smart car experimental platform,and use the multi-sensor fusion laser SLAM algorithm to build a map.By comparing and analyzing the map built by a single laser SLAM algorithm and multi-sensor fusion SLAM algorithm,the effectiveness and robustness of the multi-sensor fusion method in the thesis are proved.