• DocumentCode
    3119767
  • Title

    Circular microphone array for robot´s audition

  • Author

    Tamai, Yuki ; Kagami, Satoshi ; Amemiya, Yutaka ; Sasaki, Yoko ; Mizoguchi, Hiroshi ; Takano, Tachio

  • Author_Institution
    Tokyo Univ. of Sci., Chiba, Japan
  • fYear
    2004
  • fDate
    24-27 Oct. 2004
  • Firstpage
    565
  • Abstract
    The paper describes a "32-channel circular microphone array" which can localize individual sounds from many sound sources in 360 degrees. This microphone array is used in robot audition. Because the microphone array has many microphones, it has tolerance for not only environmental noise but also noise caused by the robot. Moreover, it can localize and separate sound sources in all directions. Sound localization by the microphone array is achieved by a delay-and-sum beamforming algorithm. A dedicated IEEE 1394 32-channel analog-to-digital (A/D) board with microphone amplifier was developed for the microphone array in order to keep a fixed cycle such as a sampling rate of 11 kHz. Sound localization is evaluated by a sound pressure distribution map. A sound localization experiment was conducted with the microphone array mounted on a Nomad mobile robot. One discontinuous sound source, such as music, can be localized accurately. The error margin of sound localization is within about 3°. In the future, we will place our microphone array on humanoid robots to confirm system performance in human environment.
  • Keywords
    acoustic noise; array signal processing; audio signal processing; microphone arrays; mobile robots; random noise; source separation; 11 kHz; IEEE 1394 analog-to-digital board; Nomad mobile robot; circular microphone array; delay-and-sum beamforming algorithm; discontinuous sound source; environmental noise tolerance; microphone amplifier; robot audition; robot noise tolerance; sound localization; sound pressure distribution map; sound source separation; Acoustic noise; Array signal processing; Delay; Humanoid robots; Microphone arrays; Mobile robots; Music; Sampling methods; System performance; Working environment noise;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Sensors, 2004. Proceedings of IEEE
  • Print_ISBN
    0-7803-8692-2
  • Type

    conf

  • DOI
    10.1109/ICSENS.2004.1426228
  • Filename
    1426228