Plane-assisted indoor Lidar SLAM

Indoor LiDAR SLAM is essential for autonomous navigation, robotic mapping, and virtual environment reconstruction. In this work, we propose a plane-assisted LiDAR SLAM framework that leverages planar features as auxiliary pose constraints rather than primary registration primitives. A robust planar feature extraction algorithm is introduced, specifically designed for multi-line LiDAR sensors, enabling reliable detection of dominant planar structures in indoor environments. Planar features are integrated into both odometry and global pose optimization by constructing residuals between plane observations, which serve as geometric constraints to suppress drift and improve mapping accuracy. In addition, we propose a registration failure detection and correction mechanism based on plane occlusion analysis, which effectively identifies and mitigates errors arising from misaligned point cloud registration. Extensive experiments conducted in challenging indoor environments (including stairwells, narrow corridors, and multi-room areas) demonstrate that the proposed approach significantly outperforms state-of-the-art approaches in terms of SLAM accuracy and reliability.