Subspace Methods for Multimicrophone Speech Dereverberation

We present a new multi-stage iterative technique for enhancing noisy speech under low signal-to-interference-ratio (SNR) environments. In the present paper, the speech is enhanced in two stages, in the first stage the noise component of the observed signal is whitened, and in the second stage a spatio-temporal power method is used to extract the desired speech component. In both the stages, the coefficient adaptation is performed using the multi-channel spatio-temporal correlation sequences of the observed data. The technique is mathematically equivalent and is computationally simpler than the existing generalized eigenvalue decomposition (GEVD) or the generalized singular value decomposition (GSVD) based techniques. Simulation results under low SNR diffuse noise scenarios indicate significant gains in SNR without introducing musical noise artifacts.

In this paper, we propose a new supervised learning algorithm, which is named the Generalized Marginal Fisher Analysis (GMFA), to utilize the advantages of the Marginal Fisher Analysis (MFA) and the Generalized Singular Value Decomposition (GSVD) techniques for face recognition. The experimental results on several standard face databases demonstrate that GMFA outperforms LDA/Fisherface, LDA/GSVD and MFA.

Bilinear equation method for unbiased identification of linear FIR systems in the presence of input and output noises

Cardiovascular time series are often affected by noise and contain components generated by different physiological sources. Least squares (LS) modelling is a powerful tool to identify these components but it provides biased estimates when the input measurements are noisy. The total least squares (TLS) model is appropriate when noise is present in all the measures. If the noise is additive, zero-mean, independent and white, TLS method results in asymptotically unbiased, consistent estimates. The authors applied the TLS method to the following applications obtaining good results: impulsive response estimation of the transfer function which relates the respiratory sinus arrhythmia with the respiratory activity; estimation of oscillatory components in RR-interval time series, cancelling of the respiratory component in the RR-interval time series. A comparison with the LS method, performed on simulated signals, showed a better impulsive response estimation and a higher resolution frequency estimate also in a highly nonstationary environment.

This paper proposes an algorithm for three-phase harmonic state estimation for power distribution networks, by using the Singular Value Decomposition (SVD) technique. As non-linear loads are distributed throughout the network, state estimation-based tools become relevant to assess the impacts of the harmonics on the grid. Moreover, the SVD technique is able to estimate the state with a reduced number of meters and, if the estimation is not possible, it indicates where monitors should be placed in order to provide full observability. The results were quite satisfactory when applied to the IEEE 13 node test system.

Modeling and forecasting human mortality are significant research topics in several disciplines because mortality rates are fundamental in planning and policy decisions. Among various techniques, Lee Carter (LC) model is one of the most popular stochastic method in human mortality modeling. The original LC model was fuzzified to eliminate the assumptions related with homoscedasticity. The existing fuzzy model makes use of ordinary least squares (OLS) technique, which prevents the model to capture the existing fluctuations in data. In this study, a revised version of fuzzy LC model utilizing singular value decomposition (SVD) technique is proposed to overcome this issue. After modeling the mortality rates, their future values are forecasted by a modified first order fuzzy time series technique. For illustration purposes, proposed method is applied to mortality data of Finland. Numerical outputs show that proposed method is statistically better in modeling mortality compared to the existing fuzzy method. In addition, the modified fuzzy time series technique generates better forecasts than the original version.

The problem of QR decomposition is considered one of the fundamental problems commonly encountered in both scientific research and engineering applications. In this paper, the QR decomposition for complex-valued time-varying matrices is analyzed and investigated. Specifically, by applying the zeroing neural dynamics (ZND) method, dimensional reduction method, equivalent transformations, Kronecker product, and vectorization techniques, a new continuous-time QR decomposition (CTQRD) model is derived and presented. Then, a novel eleven-instant Zhang et al discretization (ZeaD) formula, with fifth-order precision, is proposed and studied. Additionally, five discrete-time QR decomposition (DTQRD) models are further obtained by using the eleven-instant and other ZeaD formulas. Theoretical analysis and numerical experimental results confirmed the correctness and effectiveness of the proposed continuous and discrete ZND models.

The dynamic performance of voltage source converter-based multi-terminal high-voltage direct current (VSC-MTDC) grid for large-scale renewable energy integration is becoming a concern. This study proposes a system-level dynamic response model of wind farm (WF)-MTDC-main AC systems, and then analyses the dynamic performance based on the singular value decomposition (SVD) technique. In the modelling, an improved virtual synchronous machine control is developed, and the interaction between AC frequency and DC voltage can be readily described. Using the SVD technique, parameters of the controllers are tuned, thereby making the AC frequency and DC voltage deviations within the limitations. Some oscillation modes of the system are observed and virtual power system stabiliser is proposed to suppress the oscillations. Additionally, the oscillation can be mitigated by emulating capacitance-based inertia response. The efficiency of the proposed model and analysis is verified through the frequency-domain and time-domain results.

We present algorithms to determine the number of nonzeros in each row and column of the factors of a sparse matrix, for both the QR factorization and the LU factorization with partial pivoting. The algorithms use only the nonzero structure of the input matrix, and run in time nearly linear in the number of nonzeros in that matrix. They may be used to set up data structures or schedule parallel operations in advance of the numerical factorization. The row and column counts we compute are upper bounds on the actual counts. If the input matrix is strong Hall and there is no coincidental numerical cancellation, the counts are exact for QR factorization and are the tightest bounds possible for LU factorization. These algorithms are based on our earlier work on computing row and column counts for sparse Cholesky factorization, plus an efficient method to compute the column elimination tree of a sparse matrix without explicitly forming the product of the matrix and its transpose.

Sphere Decoder (SD) is widely being used in Multiple Input Multiple Output (MIMO) systems to reduce the complexity of the system while obtaining near Maximum Likelihood (ML) performance. The complexity of the system increases with the increase in antenna configuration or the constellation size. Some pre-processing is a fundamental prerequisite in iterative detectors to reduce the system complexity by focusing the received signal energy to reduce the effect of inter-symbol interference. The QR Decomposition (QRD) of communication channel matrices in the pre-processor stage is an important issue to ensure good performance of the subsequent steps of decoding thus a QRD) is commonly used in many MIMO detection algorithms. A sorted QR decomposition (SQRD) is an advanced algorithm that improves the performance of MIMO detection. In this paper the efficiency of QRD and SQRD methods in terms of computational complexity, error rate performance and the FPGA resources utilized is presented. The main contribution of this work is a comparison of hardware implementations of the QRD and SQRD system. QRD for 4x4 MIMO system is implemented on various target FPGA platforms to compare their area utilization.

A subspace decomposition based algorithm for joint frequency of arrival (FOA) and direction of arrival (DOA) estimation is presented in this paper. The proposed method termed as QR-TLS ESPRIT, estimates the signal subspace by decomposing the received signal in a specially formed Toeplitz matrix. The structure of the matrix favors the application of QR factorization, which yields the signal subspace at the expense of less number of computations. A Total Least Squares (TLS) method is applied to obtain the frequency and angle estimates from the signal space. The proposed QR-TLS ESPRIT method provides several benefits: 1) It avoids the computations of applying computationally complex Eigen Value Decomposition (EVD) or Singular Value Decomposition (SVD) methods to obtain signal subspace alike conventional methods, 2) QR decomposition method, which is intrinsically applicable to numerical rank deficient problems, is used to extract signal subspace from the observation data matrix, 3) the TLS method is integrated with the QR which enhances the estimation accuracy in the presence of noisy measurements, 4) the proposed method can estimate the FOA and DOA from a small snapshot length of data which further reduces the computational time, and 5) the method works for both coherent as well as noncoherent sources without employing pre-whitening or spatial smoothing techniques. The performance of the proposed method is evaluated through simulations and the results are illustrated with respect to various received signal strength.

TLS Method for Pulse Transfer Function Estimation in the Presence of Input and Output Noise

Fricative consonants: acoustic and X-ray measurements

It has been shown that the principal subspace-based multichannel Wiener filter (MWF) provides better performance than the conventional MWF for suppressing interference in the case of a single target source. It can efficiently estimate the target speech component in the principal subspace which estimates the acoustic transfer function up to a scaling factor. However, as the input signal-to-interference ratio (SIR) becomes lower, larger errors are incurred in the estimation of the acoustic transfer function by the principal subspace method, degrading the performance in interference suppression. In order to alleviate this problem, a principal subspace modification method was proposed in previous work. The principal subspace modification reduces the estimation error of the acoustic transfer function vector at low SIRs. In this work, a frequency-band dependent interpolation technique is further employed for the principal subspace modification. The speech recognition test is also conducted using the Sphinx-4 system and demonstrates the practical usefulness of the proposed method as a front processing for the speech recognizer in a distant-talking and interferer-present environment.

The QR Decomposition (QRD) of communication channel matrices is a fundamental prerequisite to several detection schemes in Multiple-Input Multiple-Output (MIMO) communication systems. Herein, the main feature of the QRD is to transform the non-causal system into a causal system, where consequently efficient detection algorithms based on the Successive Interference Cancellation (SIC) or Sphere Decoder (SD) become possible. Also, QRD can be used as a light but efficient antenna selection scheme. In this paper, we address the study of the QRD methods and compare their efficiency in terms of computational complexity and error rate performance. Moreover, a particular attention is paid to the parallelism of the QRD algorithms since it reduces the latency of the matrix factorization.