An Algorithm of Factor Graph Optimization for 3D Map of Unmanned Exploration Vehicle in Tunnel Environment

Abstract

The simultaneous localization and mapping (SLAM) technology based on LiDAR has not only achieved real-time localization of unmanned exploration vehicles in unknown environments but also efficiently obtained three-dimensional geospatial information of the environment, attracting widespread attention in the field of unmanned driving. In tunnels with degraded geometric structures, poor lighting conditions, and a lack of GPS signal, relying solely on geometric information cannot register point clouds. Therefore, traditional LiDAR SLAM algorithms are difficult to apply in tunnels. To address this issue, this paper proposes a factor graph-optimized SLAM algorithm based on the fusion of iterative closest point (ICP) and normal distributions transform (NDT) algorithms. The front-end point cloud registration module and the back-end construction algorithm based on filtering and graph optimization are designed. To verify the effectiveness of the algorithm proposed in this paper, experiments were conducted on the KITTI dataset and a real tunnel. The results show that the average processing time of the proposed method in this article is about 75 ms, which can meet the real-time requirements of unmanned exploration vehicles.