LIO-Vehicle: A Tightly-Coupled Vehicle Dynamics Extension of LiDAR Inertial Odometry
- 9 November 2021
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 7 (1) , 446-453
Abstract
We propose LIO-Vehicle, a new tightly-coupled vehicle dynamics extension of LiDAR inertial odometry (LIO) method that provides highly accurate, robust, and real-time vehicle trajectory estimation. Since most existing LiDAR-based localization methods are not specifically proposed for vehicles, they do not take the motion constraints of ground robots into account. And they may not work well in structure-less areas, such as tunnels and narrow corridors. For LIO in these LiDAR-degraded circumstances, inertial sensors alone are unable to sustain reliable long-term accuracy due to the accumulation of errors without external periodic corrections. Therefore, it is necessary to introduce other low-cost sensors and vehicle motion constraints to build a more accurate and robust odometry algorithm. In this letter, we use wheel speedometer and steering angle sensor measurements to establish a two-degree-of-freedom vehicle dynamics model and then construct a preintegration factor based on the model’s output. At the backend, we add the vehicle dynamics preintegration results, IMU preintegration measurements, and LiDAR odometry results to a factor graph and get the optimized result with the help of sliding window optimization. The experiments show that the proposed method can achieve higher positioning accuracy compared with the existing LiDAR inertial odometry methods and it can significantly mitigate navigation error in harsh areas where environmental features are insufficient for LiDAR odometry.Keywords
Funding Information
- National Key Research and Development Program of China (2020YFA0711402)
- National Natural Science Foundation of China (U1764261, 41801335, 41871370)
This publication has 26 references indexed in Scilit:
- Real-time loop closure in 2D LIDAR SLAM2016
- A robust and modular multi-sensor fusion approach applied to MAV navigation2013
- A Quadratic-Complexity Observability-Constrained Unscented Kalman Filter for SLAMIEEE Transactions on Robotics, 2013
- Real-time motion tracking on a cellphone using inertial sensing and a rolling-shutter camera2013
- iSAM2: Incremental smoothing and mapping using the Bayes treeThe International Journal of Robotics Research, 2011
- Dead reckoning of mobile robot in complex terrain based on proprioceptive sensors2008
- A New Online-Identification Algorithm for Odometer's Scale Factor2006
- Sensor fusion for precise autonomous vehicle navigation in outdoor semi-structured environments2005
- Localization via ultra-wideband radios: a look at positioning aspects for future sensor networksIEEE Signal Processing Magazine, 2005
- Effects of Model Complexity on the Performance of Automated Vehicle Steering Controllers: Model Development, Validation and ComparisonVehicle System Dynamics, 1995