F. Dellaert, D. Fox, W. Burgard, and S. Thrun

Monte Carlo Localization ForMobile Robots

Proc. of the IEEE International Conference on Robotics and Automation (ICRA'99)


 

Abstract

To navigate reliably in indoor environments, a mobile robot must know where it is. Thus, reliable position estimation is a key problem in mobile robotics. We believe that probabilisticapproaches are among the most promising candidates to providing acomprehensive and real-time solution to the robot localization problem. However, current methods still face considerable hurdles. Inparticular, the problems encountered are closely related to the typeof representation used to represent probability densities over therobot's state space. Recent work on Bayesian filtering with particle-based density representations opens up a new approach formobile robot localization, based on these principles.  In this paper we introduce the Monte Carlo Localization method, where werepresent the probability density involved by maintaining a set of samplesthat are randomly drawn from it. By using a sampling-based representation weobtain a localization method that can represent arbitrary distributions. Weshow experimentally that the resulting method isable to efficiently localizea mobile robot without knowledge of its starting location. It is faster, more accurate and less memory-intensive than earlier grid-based methods.

Download

Full paper [.ps.gz](882 kb, 7 pages)

Bibtex

@INPROCEEDINGS{Del99Mon,
  AUTHOR    = {Dellaert, F. and Fox, D. and Burgard, W. and Thrun, S.},
  TITLE     = {Monte Carlo Localization For Mobile Robots},
  BOOKTITLE = {Proc.~of the IEEE InternationalConference on Robotics \& Automation},
  YEAR      = {1998}
}