基于因子图的GNSS/LiDAR外参在线估计方法

Online estimation method of GNSS/LiDAR external parameters based on factor graph

针对园区巡检、物流等固定场景下的任务,无人车的主流导航方法为预先构建场景点云地图,在运行时通过激光雷达(LiDAR)扫描的点云与地图相匹配,进而得到高精度的定位结果。为解决地图构建时位姿估计精度受外参误差影响较大的问题,提出了一种全球导航卫星系统(GNSS)/LiDAR外参在线标定算法,对传感器外参进行在线估计。通过搭建仿真平台对算法进行验证。结果表明:该算法可以有效地估计出GNSS与LiDAR之间的外参,杆臂估计精度达到厘米(cm)级,姿态角估计精度优于0.1°。

For tasks in fixed scenarios such as park inspections and logistics,the mainstream navigation method for unmanned vehicles is to pre-build a point cloud map of the scene,and match the point cloud scanned by LiDAR with the map to obtain high-precision positioning results.In order to solve the problem that the precision of pose estimation is greatly affected by the error of external parameters during map construction,a global navigation satellite system(GNSS)/LiDAR external parameter online calibration algorithm is proposed to estimate the sensor external parameters online.The algorithm is verified by building simulation platform.The results show that the algorithm can effectively estimate the external parameters between GNSS and LiDAR.The estimation precision of the lever arm reaches the centimeter(cm)level,and the estimation precision of the attitude angle is prior to 0.1°.