scispace - formally typeset
Search or ask a question
Proceedings ArticleDOI

Background noise suppression for acoustic localization by means of an adaptive energy detection approach

TL;DR: A novel approach that combines the information provided by a Gaussian energy detector (GED) with the approved localization method SRP-PHAT is presented in this paper.
Abstract: A microphone array can be employed to localize dominant acoustic sources in a given noisy environment. This capability is successfully used in good signal to noise ratio (SNR) conditions but its accuracy decreases considerably in the presence of other background noise sources. In order to counteract this effect, a novel approach that combines the information provided by a Gaussian energy detector (GED) with the approved localization method SRP-PHAT is presented in this paper. To evaluate the presented technique, several acoustic sources (speech and impulsive sounds) were considered in a variety of different scenarios to demonstrate the robustness and the accuracy of the system proposed.
Citations
More filters
Proceedings ArticleDOI
24 Aug 2009
TL;DR: A modified approach is presented as a solution that distinguishes between impulsive and non impulsive sound sources, and additionally aligns the detection window to the event.
Abstract: In this paper, we present a novel approach for detection and localization of both impulsive and non-impulsive sound sources. At first, theoretical basics of the used algorithms are presented. Subsequently, we describe a standard SRP-PHAT based localization method and discuss occurring complications, especially for impulsive sound sources. Therefore, a modified approach is presented as a solution. It distinguishes between impulsive and non impulsive sound sources, and additionally aligns the detection window to the event. The pre-classification and alignment are done with the help of an energy detector.

10 citations


Cites background or methods from "Background noise suppression for ac..."

  • ...The estimation of R ninj (τ) for each microphone pair is achieved during phases of no activity of the sound source, which are detected by using the energy detector, both described in detail in [6]....

    [...]

  • ...This is the reason why even in scenarios without any background noise, the mislocalization rate is very high [6]....

    [...]

  • ...If we assume that the noise is fully correlated and we have an ideal room with a Dirac impulse response, we can easily achieve a noise free estimation of the correlation in a given microphone pair by subtracting the correlation of the noise from the correlation of the received signal, analogously to the background noise suppression, presented in [6]:...

    [...]

  • ...While the localization of non-impulsive sound sources following the approach in [6] showed very stable results, impulsive sound sources were not localized reliably....

    [...]

  • ...For non-impulsive sound sources like speech or a mixer, on which we concentrated in the past [6], this setup delivers high correct localization rates of over 95% with a relatively small root mean square error (RMS), under both conditions, i....

    [...]

Proceedings ArticleDOI
01 Oct 2019
TL;DR: A steady-noise suppression method is presented to exclude the influence of steady noise in a sound delay estimation of the acoustic vehicle detector and remove a peak caused by the noise to minimize the influence.
Abstract: Vehicle detection is a basic component for many applications in intelligent transportation system (ITS). We are developing a low-cost vehicle detector relying on sound arrival time difference on two microphones. Our previous paper presented that our acoustic vehicle detector successfully detected vehicles with an F-measure of 83 %. However, the acoustic detector has difficulties in vehicle detection in an environment with steady noise such as rain noise.This paper presents a steady-noise suppression method for the acoustic vehicle detector. Our key idea is to exclude the influence of steady noise in a sound delay estimation. The acoustic vehicle detector estimates vehicle sound delay by finding a peak on a cross-correlation function. We theoretically analyze the influence of steady noise and remove a peak caused by the noise to minimize the influence. Experimental evaluations revealed that the steady-noise suppression method effectively reduced the noise influence and resulted in F-measures of 0.92 and 0.90 in normal and heavy rain conditions, respectively.

5 citations


Cites methods from "Background noise suppression for ac..."

  • ...A noise reduction method steered response power phase transform (SRP-PHAT) using a microphone array has also been proposed [17]....

    [...]

Proceedings ArticleDOI
01 Mar 2019
TL;DR: A wind noise suppressor is presented for the acoustic vehicle detector to remove frequency components corresponding to wind noise that can be derived from a part of frequency components of vehicle sound signals.
Abstract: Vehicle detection is one of the fundamental tasks in the ITS (intelligent transport system). We are developing an acoustic vehicle detector that relies on two microphones at a sidewalk [1]. The vehicle detector successfully detected vehicles as well as their traveling directions. However, the detector has difficulties in vehicle detection in a high wind condition due to wind noise. This paper presents a wind noise suppressor for the acoustic vehicle detector. Our simple idea is to remove frequency components corresponding to wind noise. Our acoustic vehicle detector relies on TDOA (time difference of arrival) of sound signals on two microphones to detect vehicles, which can be derived from a part of frequency components of vehicle sound signals. We experimentally analyze frequency components of wind noise and design a filter to reduce wind noise. Initial experimental evaluations revel that our vehicle detector with a wind noise suppressor successfully detected vehicles with an F-measure of 0.77 in a normal wind condition.

3 citations


Cites methods from "Background noise suppression for ac..."

  • ...proposed a noise reduction method using a microphone array [6]....

    [...]

Proceedings Article
01 Aug 2009
TL;DR: A multiple energy detector structure (MED) is utilized and features called temporal MED (TMED) features are introduced to enable a pre-classification for the differentiation between impulsive and non-impulsive acoustic events.
Abstract: In this work, a classification method using a novel approach for acoustic feature extraction is proposed. Therefore, a multiple energy detector structure (MED) is utilized and features called temporal MED (TMED) features are introduced. The usage of an energy detector enables a pre-classification for the differentiation between impulsive and non-impulsive acoustic events. The actual classification task can be performed by using a MED. Furthermore, investigations regarding the classification accuracy using more than one microphone are presented.

2 citations


Cites methods from "Background noise suppression for ac..."

  • ...In this work, a simple energy detector (ED) test given in (3) is used to distinguish between the background noise and the novel events that are to be classified [6]: yT y σ2w H1 > H0 λ, (3) where σ2w is the noise variance and y is the observation vector supposing uncorrelated samples....

    [...]

DissertationDOI
12 Sep 2011
TL;DR: A multiple energy detector, based on successive subdivisions of the original observation interval, is presented, revealing the advantages derived from utilizing this novel structure, making it a worthwhile alternative to the single detector when a signi cant mismatch is present between the Original observation length and the actual duration of the signal.
Abstract: This thesis is dedicated to the development of new energy detectors employed in the detection of unknown signals in the presence of non-Gaussian and non-independent noise samples. To this end, an extensive study has been conducted on di erent energy detection structures, and novel techniques have been proposed which are capable of dealing with these problematic situations. The energy detector is proposed as an optimum solution to detect uncorrelated Gaussian signals, or as a generalized likelihood ratio test to detect entirely unknown signals. In both cases, the background noise must be uncorrelated Gaussian. However, energy detectors degrade when the noise does not ful ll these characteristics. Therefore, two extensions are proposed. The rst is the extended energy detector, which deals with the problem of non-Gaussian noise; and the second is the preprocessed extended energy detector, used when the noise also possesses non-independent samples. A generalization of the matched subspace lter is likewise proposed based on a modi cation of the Rao test. In order to evaluate the expected improvement of these extensions with respect to the classical energy detector, a signalto- noise ratio enhancement factor is de ned and employed to illustrate the improvement achieved in detection. Furthermore, we demonstrate how the uncertainty introduced by the unknown signal duration can decrease the performance of the energy detector. In order to improve this behavior, a multiple energy detector, based on successive subdivisions of the original observation interval, is presented. This novel detection technique leads to a layered structure of energy detectors whose observation vectors are matched to di erent intervals of signal duration. The corresponding probabilities of false alarm and detection are derived for a particular subdivision strategy, and the required procedures for their general application to other possible cases are indicated. The experiments reveal the advantages derived from utilizing this novel structure, making it a worthwhile alternative to the single detector when a signi cant mismatch is present between the original observation length and the actual duration of the signal.

1 citations


Cites methods from "Background noise suppression for ac..."

  • ...To evaluate both the performance of this combined technique and the improvement introduced in the localization phase (in comparison to the case without background noise suppression), two recording sets were tested in a kitchen scenario where people and a robot interact [59]....

    [...]

  • ...Several experiments were subsequently conducted; the resulting modi ed localization algorithm was evaluated and revealed considerable improvements [59]....

    [...]

  • ...In order to counteract this e ect, a novel approach is used: it implements a background noise suppression algorithm based on the ED to improve the localization method SRP-PHAT, as described in [59]....

    [...]

References
More filters
Journal ArticleDOI
TL;DR: The Fundamentals of Statistical Signal Processing: Estimation Theory as mentioned in this paper is a seminal work in the field of statistical signal processing, and it has been used extensively in many applications.
Abstract: (1995). Fundamentals of Statistical Signal Processing: Estimation Theory. Technometrics: Vol. 37, No. 4, pp. 465-466.

14,342 citations


"Background noise suppression for ac..." refers background in this paper

  • ...Index Terms— energy detector, background noise suppression, SRP-PHAT, acoustic localization....

    [...]

  • ...In this case several active sound sources can exist in the robots proximity, for example in a kitchen, which contains many different appliances that can be acoustically observed....

    [...]

Journal ArticleDOI
TL;DR: In this paper, a maximum likelihood estimator is developed for determining time delay between signals received at two spatially separated sensors in the presence of uncorrelated noise, where the role of the prefilters is to accentuate the signal passed to the correlator at frequencies for which the signal-to-noise (S/N) ratio is highest and suppress the noise power.
Abstract: A maximum likelihood (ML) estimator is developed for determining time delay between signals received at two spatially separated sensors in the presence of uncorrelated noise. This ML estimator can be realized as a pair of receiver prefilters followed by a cross correlator. The time argument at which the correlator achieves a maximum is the delay estimate. The ML estimator is compared with several other proposed processors of similar form. Under certain conditions the ML estimator is shown to be identical to one proposed by Hannan and Thomson [10] and MacDonald and Schultheiss [21]. Qualitatively, the role of the prefilters is to accentuate the signal passed to the correlator at frequencies for which the signal-to-noise (S/N) ratio is highest and, simultaneously, to suppress the noise power. The same type of prefiltering is provided by the generalized Eckart filter, which maximizes the S/N ratio of the correlator output. For low S/N ratio, the ML estimator is shown to be equivalent to Eckart prefiltering.

4,317 citations


"Background noise suppression for ac..." refers background in this paper

  • ...By using the energy detector it is possible to distinguish between a stationary background noise source and another active source....

    [...]

Journal ArticleDOI
TL;DR: There are a multitude of applications where novelty detection is extremely important including signal processing, computer vision, pattern recognition, data mining, and robotics.

1,457 citations


"Background noise suppression for ac..." refers background in this paper

  • ...Index Terms— energy detector, background noise suppression, SRP-PHAT, acoustic localization....

    [...]

Book ChapterDOI
01 Jan 2001
TL;DR: This chapter summarizes the current field and comments on the general merits and shortcomings of each genre, and presents a new localization method that is significantly more robust to acoustical conditions, particularly reverberation effects, than the traditional localization techniques in use today.
Abstract: Talker localization with microphone arrays has received significant attention lately as a means for the automated tracking of individuals in an enclosure and as a necessary component of any general purpose speech capture system. Several algorithmic approaches are available for speech source localization with multi-channel data. This chapter summarizes the current field and comments on the general merits and shortcomings of each genre. A new localization method is then presented in detail. By utilizing key features of existing methods, this new algorithm is shown to be significantly more robust to acoustical conditions, particularly reverberation effects, than the traditional localization techniques in use today.

649 citations


"Background noise suppression for ac..." refers methods in this paper

  • ...When a desired sound source is detected the mean of the last H SRP-PHAT computations of the noise, stored in the buffer, is computed and subtracted from the current SRP-PHAT estimation in the fol- lowing way: P (BNS)(t, s) = P (t, s) − 1 H H∑ i=1 B(i, s), (10) where the resulting Power-Field P…...

    [...]