Earticle

This paper presents a fast Time Difference of Arrival (TDOA) estimation for sound source localization. TDOA is the time difference between the arrival times of a signal at two sensors. We propose a partial cross correlation method to increase the speed of TDOA estimation for sound source localization. We do this by predicting which part of the cross correlation function contains the required TDOA value with the help of the signal energies, and then we compute the cross correlation function in that direction only. Experiments show approximately 50% reduction in the cross correlation computation time thereby increasing the speed of TDOA computation. This makes it very relevant for real world surveillance.

This paper proposes a simple sound source localization (SSL) method based on signal energies comparison and partial cross correlation for TDOA computation. Many sound source localization methods include multiple TDOA computations in order to eliminate front-back confusion. Multiple TDOA computations however increase the methods’ computation times which need to be as minimal as possible for real-time applications. Our aim in this paper is to achieve the same results of localization using fewer computations. Using three microphones, we first compare signal energies to predict which quadrant the sound source is in, and then we use partial cross correlation to estimate the TDOA value before computing the azimuth value. Also, we apply a threshold value to reinforce our prediction method. Our experimental results show that the proposed method has less computation time; spending approximately 30% less time than previous three microphone methods.

This paper proposes a fast sound source localization method using a TDOA sign–based algorithm. We present an L-shaped microphone set-up which creates four major regions in the range of 0°~360° by the intersection of the positive and negative regions of the individual microphone pairs. Then, we make an initial source region prediction based on the signs of two TDOA estimates before computing the azimuth value. Also, we apply a threshold and angle comparison to tackle the existing front-back confusion problem. Our experimental results show that the proposed method is comparable in accuracy to previous three microphone array methods; however, it takes a shorter computation time because we compute only two TDOA values.

A sound source localization method is presented herein around the utilization with the sound signal for device fault diagnosis. Firstly, Butterworth wavelet and its filter banks were designed aimed at the background noise separation in the industrial environment, it is proved that the Butterworth wavelet filter can extract the weak fault features in source signals without frequency aliasing phenomenon. Then the algorithms of sound source localization was studied in depth, the generalized cross-correlation(GCC) based time delay estimation(TDE) was presented. Finally the coordinates of sound source location was obtained in the quaternion microphone array model. The experiment results show that in the industrial environment, the error value of positioning accuracy is less than 4cm within an appropriate distance, it can meet the needs of practical application, meanwhile the reliability and the real-time performance is satisfactory.

Acoustic Emission (AE) source localization is a useful implement to diagnose the incipient faults in rotating machine. This paper proposes an improved Near-field Multiple Signal Classification method using four-order cumulants (NFC-MUSIC) to accurately locate the faults position. In order to overcome dispersion effect and revised velocity, the feature sub narrow band is extracted by Wavelet Packet reconstitution (WPR) and Modal Plate Wave Theory (MPWT). For multi-source decorrelation and increasing localized resolution, the four-order cumulants of observed signal can be selected for localization. The experiment results indicate that the improved method can accurately locate multi-rubbing faults. It is an efficient way to assist incipient rubbing faults diagnosis.

Sound Source Localization (SSL) has a wide application in speech separation, recognition and enhancement. Binaural sound source localization based on human spatial hearing mechanism is an important research field of SSL. The recent binaural SSL research is focused on the system robust against noise and reverberation. In order to improve the localization performance in degraded environment, this paper proposes an algorithm to adaptively select the ‘good’ sub-bands to compute the binaural localization cues. Firstly, sub-band Signal-Noise Ratio (SNR) is estimated based on the auto-correlation matrix of binaural sound signals. Then, Inter-aural Time Difference (ITD) is computed by adaptively selecting the sub-bands which have the high SNR. Since the ITD is calculated through the sub-bands which are less affected by the noise, the sound source azimuth is estimated more accurate. The simulation results show that compared to the conventional binaural SSL algorithm, the localization accuracy of the proposed algorithm has been improved significantly.