Research On Mobile Robot Navigation And Positioning Technology Based On LiDAR/IMU Fusion

With the continuous improvement of social living standards,indoor facilities are continuously improved,and the demand for indoor positioning is also increasing.At present,outdoor positioning can already be achieved through GPS or Beidou navigation positioning technology,but indoors,the GNSS signal was relatively weak,resulting in inaccurate positioning.Therefore,a low-cost and high-precision indoor positioning technology was required for accurate positioning.Due to the rapid development of sensor technology,there was a better way to solve the indoor positioning problem of mobile robots.However,through single sensor positioning,there will be some interference from external environmental factors and its own reasons,which makes the positioning accuracy not high.Recently,researchers have focused on solving indoor positioning problems and putting more research indoors using multiple sensors for fusion navigation simultaneous location and mapping(SLAM)technology.Due to the use of a single LiDAR sensor for positioning,there will be the problem of insufficient accuracy.This paper designs a mobile robot navigation and positioning platform using LiDAR/IMU(inertial measurement unit)sensor fusion to study indoor positioning.The main contributions of this article are as follows:1)Based on the principle of point cloud feature point detection algorithm,under the same environment,three algorithms of SIFT,Harris and Voxel-SIFT are used for experiments,and the feature point detection results are obtained.Then,the number of feature points and time-consuming situation of these three algorithms are compared and analyzed.The experimental results show that in terms of point cloud feature point extraction,the Voxel-SIFT feature point extraction algorithm used in this study is more effective than the other two algorithms,with more detections and less time-consuming.2)The improved point cloud registration algorithm using voxel grid,which improves the two shortcomings of the ICP registration algorithm.One is to reduce the time used by the algorithm.The second is to improve the final error convergence value.3)After summarizing the related theories and methods of indoor SLAM of LiDAR/IMU fusion,this paper comparatively analyzes the positioning trajectory of LiDAR/IMU combination and the result of LiDAR positioning trajectory through experiments.Then,they compare the experimental results obtained in the real environment and find that the positioning trajectory using the LiDAR/IMU fusion navigation algorithm is closer to the actual environment trajectory,and the positioning accuracy is higher.