LiDAR-IMU Tightly-Coupled SLAM Method Based on IEKF and Loop Closure Detection
Simultaneous Localization and Mapping (SLAM) technology based on LiDAR can achieve real-time robot positioning and establish environmental maps in unknown environments. LiDAR odometry can achieve accurate pose estimation in short distances or small-scale environments, but the accuracy will decrease...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2024-01-01
|
| Series: | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10430368/ |
| _version_ | 1797238864151052288 |
|---|---|
| author | Huimin Pan Dongfeng Liu Jingzheng Ren Tianxiong Huang Huijun Yang |
| author_facet | Huimin Pan Dongfeng Liu Jingzheng Ren Tianxiong Huang Huijun Yang |
| author_sort | Huimin Pan |
| collection | DOAJ |
| description | Simultaneous Localization and Mapping (SLAM) technology based on LiDAR can achieve real-time robot positioning and establish environmental maps in unknown environments. LiDAR odometry can achieve accurate pose estimation in short distances or small-scale environments, but the accuracy will decrease with the accumulation of errors. At the same time, under scenes with insufffcient structural features, point cloud-based LiDAR SLAM will show degradation phenomena, leading to failures in localization and mapping. We propose a loop closure detection method based on fusion of Inertial Measurement Unit (IMU) and LiDAR information to reduce system cumulative errors and solve environment degradation problems. Firstly, by fusing LiDAR key feature and IMU information, an odometry calculation method is proposed based on Iterative Extended Kalman Filtering (IEKF) to improve the accuracy of FAST-LIO initial pose. Then in loop closure detection, by considering geometric and intensity information simultaneously, a new scan context global descriptor is constructed from dedistortion feature points of front-end IMU to enhance the accuracy of loop closure detection for original point cloud descriptors. Finally, GTSAM is used for global optimization and GPS constraint is introduced to reduce trajectory drifts, obtaining globally consistent trajectories and mapping. Compared with existing LiDAR SLAM on KITTI dataset and self-collected dataset, proposed method has smaller trajectory errors on KITTI sequences, which reduced by 15% than baseline FAST-LIO2, and average time consumption of loop closure detection is reduced by about 50% than SC-ALOAM, and mapping drifts is reduced, which enhanced mapping accuracy and robustness while ensuring global consistency of constructed maps. |
| first_indexed | 2024-04-24T17:42:25Z |
| format | Article |
| id | doaj.art-ff0d4abcc700478a965f8620b0a154cc |
| institution | Directory Open Access Journal |
| issn | 2151-1535 |
| language | English |
| last_indexed | 2024-04-24T17:42:25Z |
| publishDate | 2024-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
| spelling | doaj.art-ff0d4abcc700478a965f8620b0a154cc2024-03-27T23:00:07ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing2151-15352024-01-01176986700110.1109/JSTARS.2024.335753610430368LiDAR-IMU Tightly-Coupled SLAM Method Based on IEKF and Loop Closure DetectionHuimin Pan0https://orcid.org/0009-0002-7988-9318Dongfeng Liu1https://orcid.org/0009-0009-1000-4789Jingzheng Ren2https://orcid.org/0009-0008-7042-8826Tianxiong Huang3https://orcid.org/0009-0007-1297-2834Huijun Yang4https://orcid.org/0000-0002-4226-2493College of Information Engineering, Northwest A&F University, Yangling, ChinaShenzhen Agricultural Science and Technology Promotion Center, Nanshan, ChinaCollege of Information Engineering, Northwest A&F University, Yangling, ChinaShenzhen Agricultural Science and Technology Promotion Center, Nanshan, ChinaCollege of Information Engineering, Northwest A&F University, Yangling, ChinaSimultaneous Localization and Mapping (SLAM) technology based on LiDAR can achieve real-time robot positioning and establish environmental maps in unknown environments. LiDAR odometry can achieve accurate pose estimation in short distances or small-scale environments, but the accuracy will decrease with the accumulation of errors. At the same time, under scenes with insufffcient structural features, point cloud-based LiDAR SLAM will show degradation phenomena, leading to failures in localization and mapping. We propose a loop closure detection method based on fusion of Inertial Measurement Unit (IMU) and LiDAR information to reduce system cumulative errors and solve environment degradation problems. Firstly, by fusing LiDAR key feature and IMU information, an odometry calculation method is proposed based on Iterative Extended Kalman Filtering (IEKF) to improve the accuracy of FAST-LIO initial pose. Then in loop closure detection, by considering geometric and intensity information simultaneously, a new scan context global descriptor is constructed from dedistortion feature points of front-end IMU to enhance the accuracy of loop closure detection for original point cloud descriptors. Finally, GTSAM is used for global optimization and GPS constraint is introduced to reduce trajectory drifts, obtaining globally consistent trajectories and mapping. Compared with existing LiDAR SLAM on KITTI dataset and self-collected dataset, proposed method has smaller trajectory errors on KITTI sequences, which reduced by 15% than baseline FAST-LIO2, and average time consumption of loop closure detection is reduced by about 50% than SC-ALOAM, and mapping drifts is reduced, which enhanced mapping accuracy and robustness while ensuring global consistency of constructed maps.https://ieeexplore.ieee.org/document/10430368/Inertial measurement unit (IMU)iterative extended Kalman filtering (IEKF)LiDAR simultaneous localization and mapping (SLAM)loop closure detection |
| spellingShingle | Huimin Pan Dongfeng Liu Jingzheng Ren Tianxiong Huang Huijun Yang LiDAR-IMU Tightly-Coupled SLAM Method Based on IEKF and Loop Closure Detection IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing Inertial measurement unit (IMU) iterative extended Kalman filtering (IEKF) LiDAR simultaneous localization and mapping (SLAM) loop closure detection |
| title | LiDAR-IMU Tightly-Coupled SLAM Method Based on IEKF and Loop Closure Detection |
| title_full | LiDAR-IMU Tightly-Coupled SLAM Method Based on IEKF and Loop Closure Detection |
| title_fullStr | LiDAR-IMU Tightly-Coupled SLAM Method Based on IEKF and Loop Closure Detection |
| title_full_unstemmed | LiDAR-IMU Tightly-Coupled SLAM Method Based on IEKF and Loop Closure Detection |
| title_short | LiDAR-IMU Tightly-Coupled SLAM Method Based on IEKF and Loop Closure Detection |
| title_sort | lidar imu tightly coupled slam method based on iekf and loop closure detection |
| topic | Inertial measurement unit (IMU) iterative extended Kalman filtering (IEKF) LiDAR simultaneous localization and mapping (SLAM) loop closure detection |
| url | https://ieeexplore.ieee.org/document/10430368/ |
| work_keys_str_mv | AT huiminpan lidarimutightlycoupledslammethodbasedoniekfandloopclosuredetection AT dongfengliu lidarimutightlycoupledslammethodbasedoniekfandloopclosuredetection AT jingzhengren lidarimutightlycoupledslammethodbasedoniekfandloopclosuredetection AT tianxionghuang lidarimutightlycoupledslammethodbasedoniekfandloopclosuredetection AT huijunyang lidarimutightlycoupledslammethodbasedoniekfandloopclosuredetection |