Here, we propose a 3D SLAM method with laser-based measurements. We extended the regionalized Gaussian process (GP) map construction algorithm, used in 2D situations, to a 3D scenario. A major improvement to the algorithm was made where multiple-direction functional relationships were assumed in a sub-region. From this improvement, a new kind of map representation is proposed, which is named as GP map. The GP map, with a point-cloud like, is a dense map representation, which can fully depict the environment structure. Based on this new map representation, we designed the state estimation method and the map update methods using simple and clear mathematics. The results show that our method can accurately estimate the trajectory and the map for various types of scenarios.
3D SLAM Method Based on Improved Regionalized Gaussian Process Map Construction
Lect. Notes Electrical Eng.
2021-10-30
13 pages
Article/Chapter (Book)
Electronic Resource
English
Regionalized Safety Performance Functions
NTIS | 2016
|Optimal identification problems of regionalized variables systems
Springer Verlag | 1986
|