Semantic Localization Via the Matrix Permanent
Nikolay A. Atanasov, Menglong Zhu, Kostas Daniilidis, George Pappas

Citation
Nikolay A. Atanasov, Menglong Zhu, Kostas Daniilidis, George Pappas. "Semantic Localization Via the Matrix Permanent". Robotics: Science and Systems (RSS), July, 2014.

Abstract
Most approaches to robot localization rely on low-level geometric features such as points, lines, and planes. In this paper, we use object recognition to obtain semantic information from the robot's sensors. The task of the robot is to localize itself within a prior landmark-based map, in which the landmarks are annotated with semantic classes. As object recognition algorithms miss detections and produce false alarms, correct data association is central to the semantic localization problem. Instead of the traditional vector-based representations, we use random finite sets (RFS) to represent the map and the object detections. Bayesian filtering in the resulting RFS space offers a natural solution to the data association and clutter filtering problems. Our main contribution is to reduce the problem of computing the likelihood of an RFS, required in the update equation of the set-based Bayes filter, to the problem of computing a matrix permanent. It is this crucial transformation that enables us to use polynomial-time algorithms to carry out semantic Monte-Carlo localization. The performance of our approach is demonstrated in simulation and in a real environment using a deformable-part-model object detector. Comparisons are made with the traditional lidar-based geometric Monte-Carlo localization.

Electronic downloads

Citation formats  
  • HTML
    Nikolay A. Atanasov, Menglong Zhu, Kostas Daniilidis, George
    Pappas. <a
    href="http://www.terraswarm.org/pubs/305.html"
    >Semantic Localization Via the Matrix
    Permanent</a>, Robotics: Science and Systems (RSS),
    July, 2014.
  • Plain text
    Nikolay A. Atanasov, Menglong Zhu, Kostas Daniilidis, George
    Pappas. "Semantic Localization Via the Matrix
    Permanent". Robotics: Science and Systems (RSS), July,
    2014.
  • BibTeX
    @inproceedings{AtanasovZhuDaniilidisPappas14_SemanticLocalizationViaMatrixPermanent,
        author = {Nikolay A. Atanasov and Menglong Zhu and Kostas
                  Daniilidis and George Pappas},
        title = {Semantic Localization Via the Matrix Permanent},
        booktitle = {Robotics: Science and Systems (RSS)},
        month = {July},
        year = {2014},
        abstract = {Most approaches to robot localization rely on
                  low-level geometric features such as points,
                  lines, and planes. In this paper, we use object
                  recognition to obtain semantic information from
                  the robot's sensors. The task of the robot is to
                  localize itself within a prior landmark-based map,
                  in which the landmarks are annotated with semantic
                  classes. As object recognition algorithms miss
                  detections and produce false alarms, correct data
                  association is central to the semantic
                  localization problem. Instead of the traditional
                  vector-based representations, we use random finite
                  sets (RFS) to represent the map and the object
                  detections. Bayesian filtering in the resulting
                  RFS space offers a natural solution to the data
                  association and clutter filtering problems. Our
                  main contribution is to reduce the problem of
                  computing the likelihood of an RFS, required in
                  the update equation of the set-based Bayes filter,
                  to the problem of computing a matrix permanent. It
                  is this crucial transformation that enables us to
                  use polynomial-time algorithms to carry out
                  semantic Monte-Carlo localization. The performance
                  of our approach is demonstrated in simulation and
                  in a real environment using a
                  deformable-part-model object detector. Comparisons
                  are made with the traditional lidar-based
                  geometric Monte-Carlo localization.},
        URL = {http://terraswarm.org/pubs/305.html}
    }
    

Posted by Nikolay A. Atanasov on 22 Apr 2014.
Groups: services

Notice: This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright.