by Udo Frese and Lutz Schröder
Abstract:
We present an improved version of the treemap SLAM algorithm which uses Cholesky factors for representing Gaussians and a Hierarchical Tree Partitioning algorithm derived from the established Kernighan-Lin heuristic for graph bisection. We demonstrate the algorithm's efficiency by mapping a simulated building with 1032271 landmarks. In the end, we close a million-landmarks loop in 21, providing an estimate for 10000 selected landmarks close to the robot, or in 442 for computing a full estimate.
Reference:
Udo Frese and Lutz Schröder: Closing a Million-Landmarks Loop, In Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems, Beijing, 2006. [preprint]
Bibtex Entry:
@InProceedings{fre06c,
author = {Udo Frese and Lutz Schr{\"o}der},
title = {Closing a Million-Landmarks Loop},
year = {2006},
booktitle = {Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems, Beijing},
url = {http://www.informatik.uni-bremen.de/agebv/en/ClosingAMillionLandmarksLoop},
comment = { <a href = "http://www.informatik.uni-bremen.de/agebv/downloads/published/freseiros06.pdf"> [preprint] </a>},
abstract = {We present an improved version of the treemap SLAM algorithm which uses Cholesky factors for representing Gaussians and a Hierarchical Tree Partitioning algorithm derived from the established Kernighan-Lin heuristic for graph bisection. We demonstrate the algorithm's efficiency by mapping a simulated building with 1032271 landmarks. In the end, we close a million-landmarks loop in 21, providing an estimate for 10000 selected landmarks close to the robot, or in 442 for computing a full estimate.},
}