Lalan Kumar

Robust Multi-Source Localization Over Planar Arrays Using MUSIC-Group Delay Spectrum

Subspace-based source localization methods utilize the spectral magnitude of the MUltiple SIgnal Classification (MUSIC) method. However, in all these methods, a large number of sensors are required to resolve closely spaced sources. A novel method for high resolution source localization based on the group delay of MUSIC is described in this work. The method can resolve both the azimuth and elevation angles of closely spaced sources using a minimal number of sensors over a planar array. At the direction of arrival (DOA) of the desired source, a transition is observed in the phase spectrum of MUSIC. The negative differential of the phase spectrum also called group delay, results in a peak at the DOA. The proposed MUSIC-Group delay spectrum defined as product of MUSIC-Magnitude (MM) and group delay spectra, resolves spatially close sources even under reverberation owing to its spatial additive property. This is illustrated by performing spectral analysis of the MUSIC-Group delay function under reverberant environments. A mathematical proof for the spatial additive property of group delay spectrum is also provided. Source localization error analysis, sensor perturbation analysis, and Cramér-Rao bound (CRB) analysis are then performed to verify the robustness of the MUSIC-Group delay method. Experiments on speech enhancement and distant speech recognition are also conducted on spatialized TIMIT and MONC databases. Experimental results obtained using objective performance measures and word error rates (WER) indicate reasonable robustness when compared to conventional source localization methods in literature.

Stochastic Cramér-Rao Bound Analysis for DOA Estimation in Spherical Harmonics Domain

Cramér-Rao bound (CRB) has been formulated in earlier work for linear, planar and 3-D array configurations. The formulations developed in prior work, make use of the standard spatial data model. In this paper, the existence of CRB for the spherical harmonics data model is first verified. Subsequently, an expression for stochastic CRB is derived for direction of arrival (DOA) estimation in spherical harmonics domain. The stochastic CRBs for azimuth and elevation are plotted at various signal to noise ratios (SNRs) and snapshots. It is noted that a lower bound on the CRB is attained at high SNR. A similar observation is made when larger number of snapshots are used.

Near-field source localization using spherical microphone array

Source localization using spherical microphone arrays has received attention due to the ease of array processing in the spherical harmonics (SH) domain with no spatial ambiguity. In this paper, we address the issue of near-field source localization using a spherical microphone array. In particular, three methods that jointly estimate the range and bearing of multiple sources in the spherical array framework, are proposed. Two subspace-based methods called the Spherical Harmonic MUltiple SIgnal Classification (SH-MUSIC) and the Spherical Harmonics MUSIC-Group Delay (SH-MGD) for near field source localization, are first presented. Additionally, a method for near-field source localization using the Spherical Harmonic MVDR (SH-MVDR) is also formulated. Experiments on near-field source localization are conducted using a spherical microphone array at various SNR. The SH-MGD is able to resolve closely spaced sources when compared to other methods.

Near-Field Acoustic Source Localization and Beamforming in Spherical Harmonics Domain

In this paper, we address the issue of near-field source localization using spherical microphone array. The spherical array has been widely used for far-field source localization due to ease of array processing in spherical harmonics domain. Various methods for far-field source localization has been reformulated in spherical harmonics domain. However, near-field source localization that involves joint estimation of range and bearing of the sources has hitherto not been investigated. In this paper, the near-field data model is developed in spherical harmonics domain. In particular, three methods that jointly estimate the range and bearing of multiple sources in the spherical array framework are proposed. Two subspace-based methods called the Spherical Harmonic MUltiple SIgnal Classification (SH-MUSIC) and the Spherical Harmonics MUSIC-Group Delay (SH-MGD) for near-field source localization are first presented. In addition, a method for near-field source localization and beamforming using Spherical Harmonic MVDR (SH-MVDR) is also formulated. Formulation and analysis of Cramér-Rao bound for near-field sources is presented in spherical harmonics domain. Various source localization experiments were conducted on simulated and signal acquired over spherical microphone array in an anechoic chamber. Root-mean-square error and probability of resolution are utilized as measures to evaluate the proposed methods. The significance and practical application of the proposed methods is discussed using experiment on interference suppression. The near-field SH-MVDR beamforming is utilized in this context.

Robust source localization and tracking using MUSIC-Group delay spectrum over spherical arrays

In this paper, a novel method of robust source localization using MUSIC-Group delay (MUSIC-GD) spectrum computed over spherical harmonic components is described. Our earlier work on the MUSIC-GD spectrum has focused on uniform linear array (ULA) and uniform circular array (UCA) for resolving closely spaced speech sources using minimal number of sensors under reverberant environments. However, this work tries to utilize the advantages of the MUSIC-GD spectrum in a spherical harmonics framework that is computationally simple and more accurate. The MUSIC-GD spectrum for spherical harmonic components is first defined. Its advantages in high resolution DOA estimation are also discussed. Several experiments are conducted for 3-D source localization in reverberant environments and the performance of the MUSIC-GD is compared to other conventional methods. Additional experiments on source tracking are also conducted. The results obtained from the MUSIC-GD computed over spherical arrays are motivating enough to further investigate the method for multiple source tracking in reverberant environments.

Group delay based methods for speech source localization over circular arrays

Conventional sub space based approaches for source localization use the spectral magnitude of MUSIC. In this paper, a group delay based method for source localization of spatially close speech sources over circular arrays, with minimal number of sensors is proposed. This approach is based on the MUSIC-Group delay spectrum and can be used to accurately estimate both azimuth and elevation angles of spatially close sources. Both simulated and real speech signal measurements are acquired over a circular array and the DOA estimation is carried out for several trials. The accuracy of the proposed approach is illustrated by using two dimensional scatter plots for a single source, and average error distribution plots for multiple sources. The high resolution property of this method is explained using the additive property of the MUSIC-Group delay spectrum. The proposed method is also evaluated under sensor perturbation errors. Experiments on distant speech recognition are conducted using the proposed approach on sentences from the TIMIT database acquired over circular arrays. The MUSIC-Group delay method indicates reasonable reduction in word error rates when compared to the standard MUSIC-Magnitude method as noted from these experiments.

The spherical harmonics root-music

Spherical harmonics root-MUSIC (MUltiple SIgnal Classification) technique for source localization using spherical microphone array is presented in this paper. Earlier work on root-MUSIC is limited to linear and planar arrays. Root-MUSIC for planar array utilizes the concept of manifold separation and beamspace transformation. In this paper, the Vandermonde structure of array manifold for a particular order is proved. Hence, the validity of root-MUSIC in the spherical harmonics domain is confirmed. The proposed method is evaluated by using simulated experiments on source localization. Root mean square error analysis and statistical analysis are presented. The experimental measures at various signal to noise ratios (SNRs) show the robustness of the proposed method. The method is also verified by using experiment on real signal acquired over spherical microphone array.

Representation and modeling of spherical harmonics manifold for source localization

Source localization has been studied in the spatial domain using differential geometry in earlier work. However, parameters of the sensor array manifold have hitherto not been investigated for source localization in spherical harmonics domain. The objective of this work is to represent and model the manifold surface using differential geometry. The system model for source localization over a spherical harmonic manifold is first formulated. Subsequently, the manifold parameters are modeled in the spherical harmonics domain. Source localization methods using MUSIC and MVDR over the spherical harmonics manifold are developed. Experiments on source localization using a spherical microphone array indicate high resolution in noise.

Robust two dimensional source localization using the MUSIC-Group delay spectrum

Subspace-based methods require a large number of sensors for localization of closely spaced sources since the spectral magnitude of Multiple Signal Classification (MUSIC) is used. However, the MUSIC-Group delay (MUSIC-GD) method has been used earlier to resolve closely spaced sources with a limited number of sensors. In this work, the MUSIC-GD method is used in high resolution azimuth and elevation estimation of spatially close sources under reverberant environments over a planar array. The efficiency of the MUSIC-GD method in effectively resolving closely spaced sources, even when the noise eigenvalues change considerably under reverberation, is described and illustrated. Localization error analysis is performed on the proposed method and its performance is illustrated using two dimensional scatter plots. Cramer-Rao lower bound (CRB) analysis is also performed and the CRB is compared with the Root Mean Square Error (RMSE) of the proposed method. Large vocabulary speaker dependent speech recognition experiments are conducted on sentences from the TIMIT database acquired over a planar microphone array. The proposed MUSIC-GD method indicates reasonable improvements in terms of localization and the Cramer-Rao lower bound error analysis. A reasonable reduction is also observed in terms of word error rate (WER) from the experiments conducted on distant speech recognition.

On the Rapid Prototyping of a Portable Multi Media Acquisition System for Intelligent Meeting Capture

Emerging multi-modal signal processing applications require a sustained effort on the part of the developer to realize and deploy an application. A rapid prototyping platform will reduce the effort, cost, and time required to develop and deploy an application. In this paper, a rapid prototyping platform is developed for realizing a multi-modal signal processing application that involve real time interfacing of multi-modal signals both at the input and the output. The platform allows the designer to simulate various applications and produce the required product only after entire testing has been done. A portable intelligent meeting capture system that can be rapidly deployed in smart meeting rooms is implemented on this platform. The setup consists of a microphone array which computes the two-dimensional direction of arrival (DOA). The azimuth and the elevation angles are computed using advanced signal processing algorithms like GCC-PHAT, MUSIC which are implemented on the Real Time Operating System (RT-OS). The DOAs are communicated to a wireless networked camera which steers in real time towards the active speaker. Performance evaluation of the rapidly prototyped system is tested in real time meetings in terms of average error deviations in the DOA. The accuracy of the results indicate further miniaturization of the system. The possibilities of using this platform for developing multi-modal signal processing in general is also described.