The field of LiDAR-based localization and mapping is experiencing significant advancements, with a focus on improving accuracy, efficiency, and robustness. Recent developments have led to the creation of novel methods that can handle challenging environments, such as those with sparse features, repetitive geometric structures, and high-frequency motion. These innovations have resulted in more accurate and reliable mapping and localization systems. Notable contributions include the use of non-line-of-sight perception, Gaussian semantic fields, and degeneracy-aware multi-metric approaches.

Some particularly noteworthy papers include:

• FORM, which proposes a fixed-lag odometry method with reparative mapping, allowing for real-time performance and active correction of the local map.
• SuperEx, which introduces a framework that integrates non-line-of-sight sensing directly into the mapping-exploration loop, enabling more efficient exploration and mapping in unknown indoor environments.
• Gaussian Semantic Field for One-shot LiDAR Global Localization, which presents a lightweight tri-layered scene graph for semantic disambiguation and one-shot localization.
• DAMM-LOAM, which proposes a degeneracy-aware multi-metric LiDAR odometry and mapping module, improving mapping accuracy through point cloud classification and degeneracy-based weighted least squares-based ICP algorithm.
• PlanarMesh, which presents a novel incremental, mesh-based LiDAR reconstruction system that adaptively adjusts mesh resolution to achieve compact, detailed reconstructions in real-time.