Using the Condensation Algorithm for Robust, Vision-Based Mobile Robot Localization
Abstract
To navigate reliably in indoor environments, a mobile \nrobot must know where it is. This includes both the ability \nof globally localizing the robot from scratch, as well \nas tracking the robot’s position once its location is known. \nVision has long been advertised as providing a solution to \nthese problems, but we still lack efficient solutions in unmodified \nenvironments. Many existing approaches require \nmodification of the environment to function properly, and \nthose that work within unmodified environments seldomly \naddress the problem of global localization. \nIn this paper we present a novel, vision-based localization \nmethod based on the CONDENSATION algorithm \n[17, 18], a Bayesian filtering method that uses a sampling-based \ndensity representation. We show how the CONDENSATION \nalgorithm can be used in a novel way to track the \nposition of the camera platform rather than tracking an object \nin the scene. In addition, it can also be used to globally \nlocalize the camera platform, given a visual map of the environment. \nBased on these two observations, we present a vision-based \nrobot localization method that provides a solution to \na difficult and open problem in the mobile robotics community. \nAs evidence for the viability of our approach, we show \nboth global localization and tracking results in the context \nof a state of the art robotics application.
Cite
Text
Dellaert et al. "Using the Condensation Algorithm for Robust, Vision-Based Mobile Robot Localization." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 1999. doi:10.1109/CVPR.1999.784976Markdown
[Dellaert et al. "Using the Condensation Algorithm for Robust, Vision-Based Mobile Robot Localization." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 1999.](https://mlanthology.org/cvpr/1999/dellaert1999cvpr-using/) doi:10.1109/CVPR.1999.784976BibTeX
@inproceedings{dellaert1999cvpr-using,
title = {{Using the Condensation Algorithm for Robust, Vision-Based Mobile Robot Localization}},
author = {Dellaert, Frank and Burgard, Wolfram and Fox, Dieter and Thrun, Sebastian},
booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition},
year = {1999},
pages = {2588-},
doi = {10.1109/CVPR.1999.784976},
url = {https://mlanthology.org/cvpr/1999/dellaert1999cvpr-using/}
}