Development of Compact Microphone Array for Direction-of-Arrival Estimation
Direction-of-arrival estimates are required in many applications such as automatic video camera steering and multiparty teleconferencing for beam forming and steering to suppress noise and reverberation and improve speech intelligibility. Ambient noise an
- PDF / 1,715,696 Bytes
- 7 Pages / 439.37 x 666.142 pts Page_size
- 7 Downloads / 224 Views
bstract Direction-of-arrival estimates are required in many applications such as automatic video camera steering and multiparty teleconferencing for beam forming and steering to suppress noise and reverberation and improve speech intelligibility. Ambient noise and multiple reflections of the acoustic source signal significantly degrade the performance of time-difference-of-arrival (TDOA) methods to localize the sound source using only two microphones. In this work, we investigate the performance of a multichannel cross-correlation coefficient (MCCC) algorithm for the estimation of the direction-of-arrival (DOA) of an acoustic source in the presence of significant levels of both noise and reverberation. Simulations and initial experimental results confirm that the DOA estimation robustness and complexity is suitable for a practical micro-phone array using miniature MEMS microphones and an FPGA implementation of the MCCC algorithm.
Keywords Direction of arrival estimation Microphone arrays cessing algorithms Time of arrival estimation
Signal pro-
1 Introduction The basic idea to solve the DOA problem is to determine the time difference of a sound source signal arriving at two microphone locations. If the sound source is located in the far-field and the distance b between the two microphones is known, T. Quoˆ´cV~o U. Klein (&) School of Electrical Engineering, International University, Vietnam National University, HCMC, Ho Chi Minh City, Vietnam e-mail: [email protected]
J. J. (Jong Hyuk) Park et al. (eds.), Multimedia and Ubiquitous Engineering, Lecture Notes in Electrical Engineering 240, DOI: 10.1007/978-94-007-6738-6_146, Springer Science+Business Media Dordrecht(Outside the USA) 2013
1189
1190
T. Quoˆ´c V~ o and U. Klein
the TDOA s12 ¼ ðb cos hÞ=va is directly related to the DOA angle h. Here, va is the sound velocity in air. In order to improve the estimate of the DOA in noisy and reverberant environments many algorithms have been proposed using multiple microphones. The linear spatial prediction method [1], the multichannel cross-correlation coefficient algorithm [1], and the broadband multiple signal classification (MUSIC) method [2] all employ the correlation of the aligned microphone signals to estimate the TDOA. The minimum entropy (ME) method [3] uses higher order statistics that could be more suitable for non-Gaussian source signals such as speech. The most reliable TDOA estimation performance in reverberant environments is achieved by adaptive blind multichannel identification (ABMCI) [4], which relies on the blind identification of the real rever-berant impulse response functions of the SIMO system, consisting of the single signal source and multiple microphones. Both ME and ABMCI are considered to give better results than cross-correlation based techniques, although at the cost of higher computational requirements, thereby increasing the hardware complexity and cost. Because of its relatively low computational complexity the MCCC method has been selected for a low-cost hardware imple
Data Loading...
