Mobile robot navigation based on visual landmark recognition

M Mata, JM Armingol, A de la Escatera, MA Salichs

    Research output: Chapter in Book/Report/Conference proceedingChapter


    This paper presents a vision-based landmark recognition system for use in mobile robot navigation tasks. It is able to learn new landmarks with very little human intervention. The recognition system can read text inside landmarks, when present. The system uses genetic algorithms (GA) for both learning and recognition processes. Some of the landmarks learnt has been tested for indoor-outdoor topological navigation. The presented experimental illustrate the success of the proposed algorithm. The resulting localization system has been integrated in a complex navigation system onboard on a B21 mobile robot.
    Original languageEnglish
    Title of host publicationIntelligent Autonomous Vehicles 2001: a Proceedings Volume from the 4th IFAC Symposium, Sapporo, Japan, 5-7 September 2001 (IFAC Proceedings Volumes)
    EditorsH. Asama, H. Inoue
    PublisherElsevier B.V.
    Number of pages6
    ISBN (Print)9780080438993
    Publication statusPublished - 2002

    Publication series

    NameIFAC Proceedings Volumes
    ISSN (Electronic)1474-6670


    • mobile robots
    • genetic algorithms
    • computer vision
    • landmark-based navigation


    Dive into the research topics of 'Mobile robot navigation based on visual landmark recognition'. Together they form a unique fingerprint.

    Cite this