P‐2.3: Simultaneous localization and Mapping Method Based on Mechanical Lidar

Abstract

With the progress of society and the development of science and technology, automatic driving, as the advanced stage of assisted driving, has become one of the most promising research directions in the future. The premise of realizing automatic driving is to accurately and efficiently perceive the surrounding environment, and SLAM technology can fuse multiple sensors to estimate the position and position of the vehicle, which is widely used in the construction of environmental maps. In real life, the environment in the process of car driving is often composed of dynamic people and vehicles. However, the SLAM algorithm based on static scene is easy to form wrong matching relationship when it performs point cloud matching in dynamic scene. Therefore, based on the dynamic object environment, this paper adopts the multi‐sensor fusion method, improves the frontend odometer on the basis of LIO‐SAM algorithm, adds semantic information to the lidar odometer, removes the dynamic object and improves the accuracy of point cloud registration. The experimental results show that the proposed method can filter the dynamic objects in the dropped point cloud data in the dynamic scene, and the mapping effect is better than that of the pure geometric method.