Abstract: Simultaneous localization and mapping is one of the key technologies to realize unmanned driving in unmanned driving technology. The accuracy and applicability of mapping still need to be further improved. In this paper, the loop closure detection part of the LeGO-LOAM (Lightweight and Groud-Optimized Lidar Odometry and Mapping) algorithm is improved, the KD (K-Dimensional) tree is combined with the context scanning algorithm, and the external parameters of the radar-inertial measurement unit (IMU) are re-calibrated. By establishing an actual scene and comparing the adjusted algorithm with the previously unadjusted algorithm, it can be seen that the mapping accuracy and applicability of the adjusted algorithm have been significantly improved.

Key words: Simultaneous localization and mapping; LeGO-LOAM; Unmanned vehicle; Actual scenarios

摘要： 同步定位与建图是无人驾驶技术中实现无人化的关键技术之一。建图的精度以及适用 性仍需进一步提高，论文通过对轻量化和地面优化的激光雷里程计和地图（LeGO-LOAM） 算法的回环检测部分进行改进，将 KD（K-Dimensional）树与上下文扫描算法进行结合，并 且对雷达-惯性测量单元（IMU）外参进行重新标定。通过建立实际场景将调整后的算法与之 前未调整的算法进行比较，可以看出，调整后的算法的建图精度与适用性有了很明显的提高。

关键词: 同步定位与建图；LeGO-LOAM；无人驾驶汽车；实际场景