Towards Robust Place Recognition for Robot Localization

M. Ullah, A. Pronobis, B. Caputo, Patric Jensfelt and H. Christensen

Abstract:

Localization and context interpretation are two key competences for mobile robot systems. Visual place recognition, as opposed to purely geometrical models, holds promise of higher flexibility and association of semantics to the model. Ideally, a place recognition algorithm should be robust to dynamic changes and i t should perform consistently when recognizing a room (for instance a corridor) in different geographical locations . Also, it should be able to categorize places, a crucial capability for transfer of knowledge and continuou s learning. In order to test the suitability of visual recognition algorithms for these task s, this paper presents a new database, acquired in three different labs across Europe. It contains image sequences of s everal rooms under dynamic changes, acquired at the same time with a perspective and omnidirectional camera, mounted on a socket. We assess this new database with an appearance-based algorithm that combines local features with support vec tor machines through an ad-hoc kernel. Results show the effectiveness of the approach and the value of the database.

BibTeX Entry:

@InProceedings{Ullah08a,
  author =       {M. Ullah and A. Pronobis and B. Caputo and P. Jensfelt and
H. Christensen},
  title =        {Towards Robust Place Recognition for Robot Localization},
  booktitle =    {Proc.~of the IEEE International Conference on Robotics and Automation (ICRA'08)},
  year =         2008
}

Download: pdf (1.7M)