The real-time implementation of the existing multi-channel Wiener filter (MWF) algorithms suffer from performance degradation due to the lack of robustness against estimation errors of the second-order statistics. The reasons are twofold: one, the estimation of the statistics relies on real voice activity detector (VAD), which often fails in adverse environments. Second, the MWF solutions involve estimation of the second order clean speech statistics, which also exaggerates the errors. This paper presents an MWF algorithm that requires neither VAD nor clean speech statistics. Performance evaluation under real scenarios shows that the proposed method outperforms the conventional MWF solution in terms of the trade-off between noise reduction and speech distortion.

In broadband beamformer applications with dynamically moving sources, it can be important to have a simple mechanism to steer the main-beam. It can also be desirable if the beampattern of the beamformer is invariant to the look direction. A number of design methods for such beamformers, based on the spherical harmonic transform, have been reported in the literature. However, these methods require the sensor positions to satisfy a certain condition which may conflict with practical considerations. This paper proposes a design method which obviates this restriction thus allowing for spherical arrays with arbitrary sensor configurations. Moreover, for a comparable level of performance and computational complexity to the existing spherical harmonic beamformers, the proposed beamformer requires fewer sensors. The trade-off is that the design of the beamformer now depends on the sensor positions. Other considerations, such as the effects of array mis-orientation and robustness, are also discussed in this paper and illustrated by design examples.

Beamformers are known to be sensitive to errors and mismatches in their array elements. This paper proposes a robust steerable broad band beamformer design using the Farrow structure and for arbitrary array geometry. The design formulation includes stochastic models describing the microphone characteristics as random variables, thus allowing flexibility for microphone errors. This method establishes a direct relationship in controlling the robustness specification in the design from given microphone characteristics. The robust design procedure optimises the mean performance of the beamformer. Design examples show significant reduction in error sensitivity in the robust design formulation.

Many popular direction-of-arrival (DOA) estimators rely on the fact that the array response vector of the array is Vandermonde, for example, that of a uniform linear array (ULA). Array interpolation is a preprocessing technique to transform the array response vector of a planar array of arbitrary geometry to that of a ULA over an angular sector. While good approximation within the target sector is attained with the various existing array interpolation approaches, the response of the interpolated array in the out-of-sector region is at best partially controlled. Accordingly, out-of-sector signals, especially those highly correlated with the in-sector signals, can degrade significantly the performance of DOA estimators that rely on the Vandermonde form to work correctly. Recently, we proposed an improved array interpolation approach that takes into account the array response over the full azimuth. In this paper, we develop the idea further and present a simple data-adaptive array interpolation scheme that can provide significantly better accuracy in the DOA estimates. We present examples to demonstrate the effectiveness of our proposal.

Many popular direction-of-arrival (DOA) estimators rely on the fact that the array response vector of the array is Vandermonde, for example, that of a uniform linear array (ULA). Array interpolation is a preprocessing technique to transform the array response vector of a planar array of arbitrary geometry to that of a ULA over an angular sector. While good approximation within the target sector is attained in the existing array interpolation approaches, the response of the interpolated array in the out-of-sector region is at best partially controlled. Accordingly, out-of-sector signals, especially those highly correlated with the in-sector signals, can degrade significantly the performance of DOA estimators (e.g., MUSIC with spatial smoothing) that rely on the Vandermonde form to work correctly. In this paper, we propose an improved array interpolation approach that takes into account the array response over the full azimuth. We present also numerical examples to demonstrate the shortcomings of the existing approaches and the effectiveness of our proposal.

We propose a sequential M-estimation algorithm as an alternative to sequential least squares. Being an approximation of the exact M-estimator, the proposed technique is robust to nonGaussian processes and outperforms sequential least squares. Simulation results demonstrate the power of the proposed sequential M-estimator.

In this paper, we generalize the well-known variable fractional delay filter structure of Farrow to include Laguerre and Kautz filters. We also show how one can incorporate amplitude response shaping into the optimization design formulation, and extend the range of variable delay to beyond the usual one sampling period. Some salient features of the new structures are demonstrated through three design examples based on the simple least squares criterion.

The Davies transformation is a method to transform the steering vector of a uniform circular array (UCA) to a vector with Vandermonde form. As this form is similar to that of the steering vector of a uniform linear array (ULA), we can apply to UCAs the many tools that have been developed for ULAs. However, the Davies transformation can be highly sensitive to perturbations of the underlying ideal array model. In this paper, we present a method for deriving a more robust transformation using novel optimization techniques. In particular, we consider its application to direction of arrival estimation in the presence of correlated signals. The effectiveness of the method is illustrated through a numerical example.

By virtue of their geometry, uniform circular arrays (UCAs) are ideally suited to provide 360 degrees of coverage in the azimuthal plane. However, in a correlated signal environment, the well-known technique of spatial smoothing to mitigate the signal cancellation effect as seen in an optimum beamformer will not work since this technique is applicable only to uniform linear arrays. We show how the transformation of Davies (1965) can be adopted to design optimum beamformers for UCAs in a correlated signal environment. We also introduce derivative constraints to improve the robustness of the optimum beamformers to mismatches between the beamformers' look direction and the actual direction-of-arrival of the desired signal. The effectiveness of our design method is illustrated by a numerical example.

In this paper, the optimum (in the minimum MSE (MMSE) sense) transfer functions of the receive filter pair in a carrierless AM/PM (CAP) transceiver system are derived under the conditions that the input noise has an arbitrary spectrum shape and the baud-rate sampler has a fixed but unknown sampling phase. An expression for the corresponding MMSE at the slicer is also given. Although similar studies have been reported in the open literature, the analysis results were not consistent with the corresponding simulation results. In this work, we show that our theoretical results are in agreement with the simulation results.

Uniform circular arrays (UCAs) are naturally suited to provide 360 degrees of coverage in the azimuthal plane. In this paper, we describe a new approach for synthesizing array patterns with guaranteed maximum sidelobe levels for any look direction. The advantage of this approach is that it is computationally efficient which makes it eminently suitable for real-time beamforming and beamsteering applications. The approach is based on the Dolph-Chebyshev method for uniform linear arrays except here, the Dolph-Chebyshev method is applied to the transformed array response vector of the UCAs.

Carrierless amplitude and phase (CAP) is one of the modulation schemes that has been proposed to accomplish the task of transporting high bit-rate data over cheap copper wire pairs. In this paper, we reexamined some CAP receiver structures that have been proposed in the literature and showed that the modified quadrature amplitude modulation receiver can offer a significant advantage over the other receiver structures through its greater immunity to timing phase errors. Our recommendation is based on an analysis of the eye diagram of a received CAP signal.

By using the acoustic scattering form function, a method for classifying underwater spherical shell targets in an open ocean environment is proposed. The resulting backscatter signal from an incident wideband signal is used to illustrate some of the salient scattering features such as mid-frequency enhancement (MFE). A shell classification technique is then developed. An affine transformation of a template form function is used in the classification scheme. It is shown that the proposed scheme is robust against uncertainties in the material properties and noise.

Shepherd and McWhirter (1993) proposed a QRD-RLS algorithm for adaptive filtering with linear constraints. In this paper, the numerical properties of this algorithm are considered. In particular, it is shown that the computed weight vector satisfies a set of constraints which are perturbed from the original ones, the amount of perturbation being dependent on the wordlength. The linearly constrained FLS algorithm of Resende et al. (1996) is also studied. Simulation results show that this algorithm is numerically unstable, even in the absence of explosive divergence.

The use of wideband active sonar array processing to estimate the range, velocity and bearing of a target has received much interest in the literature recently. Although increased attention has been focused on wideband correlation processing for estimating range and velocity, array directivity patterns are almost always computed and interpreted under the narrowband signal assumption. This paper considers the target bearing estimation problem using the wideband correlation approach. Via this approach, it is shown how an optimum set of array weights can be selected for a known transmitted signal. The optimization procedure also provides robustness against errors in the array structure.

This paper investigates two techniques for extending the narrowband (NB) eigenvector method for towed array shape estimation to process broadband sources. These techniques are the weighted incoherent broadband processor (WIB) and the weighted quasi-coherent broadband processor (WQB). It is demonstrated, both analytically and via computer simulations, that the two broadband processors can dramatically improve the bias and variance of the receiver position estimates, compared to the NB technique. The results of computer simulations also demonstrate that the WQB method has a lower SNR threshold than the NB and incoherent methods.

The eigenvector method for estimating the positions of the receivers of an ocean-towed array is based on the eigendecomposition of the array signal correlation matrix to find the phase delays between the array receivers. Previous work has shown that for reasonable SNR, the bias and variance of the phase estimates is relatively independent of the number of receivers in the array. This suggests that the computational cost of the eigenvector method could be substantially improved by partitioning the array receivers into groups of smaller sub-arrays and then applying the eigenvector method to each sub-array. This paper introduces the interleaved partitioned eigenvector method, and shows that it significantly reduces the computational cost without adversely affecting the quality of the position estimates. Numerical examples substantiate the theoretical work of this paper and also demonstrate the improvement in beamforming when employing the shape estimation algorithm.

In many adaptive signal processing applications, it is often desired to impose linear and quadratic constraints on the adaptive filter weights in order to meet certain performance criteria. The paper presents a modification of the well-known adaptive RLS or FLS algorithm to achieve this. By way of illustration, the paper considers an adaptive narrowband beamformer with first and second order spatial derivative constraints. The performance of the algorithm is studied via computer simulations.

The eigenvector method for estimating the positions of the receivers of an ocean-towed array is based on the eigendecomposition of the array signal correlation matrix to find the phase delays between the array receivers. This paper derives the mean, variance and covariance of the phase delays when the correlation matrix is estimated from receiver data. It is shown that the variance of the relative phase delays is inversely proportional to the number of snapshots and SNR but relatively independent of the size of the correlation matrix. The theoretical derivations are supported by numerical simulations.>

A method frequently employed to optimise the performance of a broadband antenna array processor is to find the set of processor weights which minimises its mean output power subject to linear constraints on the processor weights. These linear constraints determine the frequency response of the array processor in the look direction. To obtain robustness against mismatch between the array look direction and the actual signal direction, derivative constraints can be further imposed. However, derivative constraints corresponding to necessary and sufficient (NS) conditions for a maximally flat spatial power response in the look direction are in general quadratic. This limits their value for real-time processing. This paper presents several simpler derivative constraints. These constraints correspond only to sufficient conditions for maximal flatness but perform almost as well as the NS constraints and are better suited to real-time implementation.>