Research On SLAM Algorithm Of Mobile Robot Based On Lidar Navigation

In recent years,the research in the field of artificial intelligence has been progressing,and the application of artificial intelligence robots has gradually entered various fields and industries,so it is an important research topic to apply to everyday indoor mobile robots.A highly intelligent mobile robot,able to move autonomously,to achieve environmental identification and path optimization,the environment can be accurately judged.The robot's autonomous navigation is based on laser scanning matching,the main idea of which is to analyze the robot's current relative position by calculating the adjacent two-frame scanning laser data,and laser scanning matching is one of the most reliable methods for mobile robot localization and mapping.In this treatise,the scan SLAM(simultaneous localization and mapping)technology of mobile robots in indoor environment are the research goal.In view of this goal,the main research content of this thesis is as follows:First of all,from 2D/3D laser SLAM,vision SLAM,deep learning and SLAM combined three aspects of the current mainstream SLAM framework.Secondly,based on the mobile robot laser simultaneous localization and mapping basic mathematical theoretical framework research.The probability model,multi-sensor data timestamp alignment,laser motion distortion correction,robot motion model,robot measurement model and probability map model in the field of laser SLAM are combed.Third,a fast and robust multi-resolution map SLAM front-end Algorithm is proposed to solve the problem that the laser matching algorithm diverges easily when the initialization noise is large.In this treatise,a high reliable laser scanning matching algorithm is designed,and the advantages of the algorithm are listed by sacrificing the cost of memory to reduce the time consumption and speed up the laser front-end matching,the main challenge is to minimize the complexity of the operation while maximizing the quality and robustness of the optimal solution for laser matching,while comparing the current mainstream SLAM front-end matching algorithms,when the initialization noise is very serious,the effect of multi-resolution probabilistic map matching algorithm is much better than ICP and ICL Algorithm.Finally,aiming at the problem of map losing one dimension information and getting the wrong matching result in the corridor environment of laser SLAM,a weighted random consistency detection corridor feature algorithm is proposed to improve the robustness of global map optimization algorithm in the backend of Laser Slam.This treatise mainly studies some problems of the current main back-end closed-loop detection algorithm of SLAM,and uses the random consistency algorithm to detect long corridors,so as to solve the problem of missing one dimension information in laser local long-line and map matching,the robustness of the Slam back-end closed-loop Detection Algorithm is greatly improved by avoiding the problem of global pose optimization error and map inconsistency.