Ears of the robot: Direction of arrival estimation based on pattern recognition using robot-mounted microphones

    Research output: Contribution to journalArticle

    2 Citations (Scopus)

    Abstract

    We propose a new type of direction-of-arrival estimation method for robot audition that is free from strict head related transfer function estimation. The proposed method is based on statistical pattern recognition that employs a ratio of power spectrum amplitudes occurring for a microphone pair as a feature vector. It does not require any phase information explicitly, which is frequently used in conventional techniques, because the phase information is unreliable for the case in which strong reflections and diffractions occur around the microphones. The feature vectors we adopted can treat these influences naturally. The effectiveness of the proposed method was shown from direction-of-arrival estimation tests for 19 kinds of directions: 92.4% of errors were reduced compared with the conventional phase-based method.

    Original languageEnglish
    Pages (from-to)1522-1530
    Number of pages9
    JournalIEICE Transactions on Information and Systems
    VolumeE91-D
    Issue number5
    DOIs
    Publication statusPublished - 2008 May

    Fingerprint

    Direction of arrival
    Microphones
    Pattern recognition
    Robots
    Audition
    Power spectrum
    Transfer functions
    Diffraction

    Keywords

    • Crosspower-spectrum phase
    • Direction-of-arrival estimation
    • Multiple signal classification
    • Pattern recognition
    • Robot audition

    ASJC Scopus subject areas

    • Electrical and Electronic Engineering
    • Software
    • Artificial Intelligence
    • Hardware and Architecture
    • Computer Vision and Pattern Recognition

    Cite this

    @article{1380180722f1450ba02f0cbd565713c5,
    title = "Ears of the robot: Direction of arrival estimation based on pattern recognition using robot-mounted microphones",
    abstract = "We propose a new type of direction-of-arrival estimation method for robot audition that is free from strict head related transfer function estimation. The proposed method is based on statistical pattern recognition that employs a ratio of power spectrum amplitudes occurring for a microphone pair as a feature vector. It does not require any phase information explicitly, which is frequently used in conventional techniques, because the phase information is unreliable for the case in which strong reflections and diffractions occur around the microphones. The feature vectors we adopted can treat these influences naturally. The effectiveness of the proposed method was shown from direction-of-arrival estimation tests for 19 kinds of directions: 92.4{\%} of errors were reduced compared with the conventional phase-based method.",
    keywords = "Crosspower-spectrum phase, Direction-of-arrival estimation, Multiple signal classification, Pattern recognition, Robot audition",
    author = "Naoya Mochiki and Tetsuji Ogawa and Tetsunori Kobayashi",
    year = "2008",
    month = "5",
    doi = "10.1093/ietisy/e91-d.5.1522",
    language = "English",
    volume = "E91-D",
    pages = "1522--1530",
    journal = "IEICE Transactions on Information and Systems",
    issn = "0916-8532",
    publisher = "Maruzen Co., Ltd/Maruzen Kabushikikaisha",
    number = "5",

    }

    TY - JOUR

    T1 - Ears of the robot

    T2 - Direction of arrival estimation based on pattern recognition using robot-mounted microphones

    AU - Mochiki, Naoya

    AU - Ogawa, Tetsuji

    AU - Kobayashi, Tetsunori

    PY - 2008/5

    Y1 - 2008/5

    N2 - We propose a new type of direction-of-arrival estimation method for robot audition that is free from strict head related transfer function estimation. The proposed method is based on statistical pattern recognition that employs a ratio of power spectrum amplitudes occurring for a microphone pair as a feature vector. It does not require any phase information explicitly, which is frequently used in conventional techniques, because the phase information is unreliable for the case in which strong reflections and diffractions occur around the microphones. The feature vectors we adopted can treat these influences naturally. The effectiveness of the proposed method was shown from direction-of-arrival estimation tests for 19 kinds of directions: 92.4% of errors were reduced compared with the conventional phase-based method.

    AB - We propose a new type of direction-of-arrival estimation method for robot audition that is free from strict head related transfer function estimation. The proposed method is based on statistical pattern recognition that employs a ratio of power spectrum amplitudes occurring for a microphone pair as a feature vector. It does not require any phase information explicitly, which is frequently used in conventional techniques, because the phase information is unreliable for the case in which strong reflections and diffractions occur around the microphones. The feature vectors we adopted can treat these influences naturally. The effectiveness of the proposed method was shown from direction-of-arrival estimation tests for 19 kinds of directions: 92.4% of errors were reduced compared with the conventional phase-based method.

    KW - Crosspower-spectrum phase

    KW - Direction-of-arrival estimation

    KW - Multiple signal classification

    KW - Pattern recognition

    KW - Robot audition

    UR - http://www.scopus.com/inward/record.url?scp=68149125169&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=68149125169&partnerID=8YFLogxK

    U2 - 10.1093/ietisy/e91-d.5.1522

    DO - 10.1093/ietisy/e91-d.5.1522

    M3 - Article

    AN - SCOPUS:68149125169

    VL - E91-D

    SP - 1522

    EP - 1530

    JO - IEICE Transactions on Information and Systems

    JF - IEICE Transactions on Information and Systems

    SN - 0916-8532

    IS - 5

    ER -