Minimum Power Distortionless Response Beamformer Research Articles

Audio enhancement and intelligent classification of household sound events using a sparsely deployed array

A household sound event classification system consisting of an audio localization and enhancement front-end cascaded with an intelligent classification back-end is presented. The front-end is composed of a sparsely deployed microphone array and a preprocessing unit to localize the source and extract the associated signal. In the front-end, a two-stage method and a direct method are compared for localization. The two-stage method introduces a subspace algorithm to estimate the time difference of arrival, followed by a constrained least squares algorithm to determine the source location. The direct localization methods, the delay-and-sum beamformer, the minimum power distortionless response beamformer, and the multiple signal classification algorithm are compared in terms of localization performance for sparse array configuration. A modified particle swarm optimization algorithm enabled an efficient grid-search. A minimum variance distortionless response beamformer in conjunction with a minimum-mean-square-error postfilter is exploited to extract the source signals for sound event classification tasks that follow. The back-end of the system is a sound event classifier that is based on convolutional neural networks (CNNs), and convolutional long short-term memory networks Mel-spectrograms are used as the input features to the CNNs. Simulations and experiments conducted in a live room have demonstrated the strength and weakness of the direct and two-stage methods. Signal quality enhancement using the array-based front-end proves beneficial for improved classification accuracy over a single microphone.

Distortionless Beamforming Optimized With <inline-formula> <tex-math notation="LaTeX">$\ell _1$</tex-math> </inline-formula>-Norm Minimization

We propose a beamforming method that minimizes the $\ell _1$ norm of a beamformer output vector under the same distortionless constraint as that of the conventional minimum power distortionless response (MPDR) beamformer. Using the $\ell _1$ norm makes the beamformer output sparse. This leads to reducing the residual elements of the interference signal. In addition, the sensitivity of the proposed beamformer can be controlled by adding a norm constraint as in the MPDR beamformer. The proposed method improved the signal-to-interference-noise ratio by 7 dB from that of the MPDR beamformer for reverberation time $T_{60}$ = 300 ms in a simulation.

Modeling of reverberant room responses for two-dimensional spatial sound field analysis and synthesis.

A unified framework is proposed for analysis and synthesis of two-dimensional spatial sound field in reverberant environments. In the sound field analysis (SFA) phase, an unbaffled 24-element circular microphone array is utilized to encode the sound field based on the plane-wave decomposition. Depending on the sparsity of the sound sources, the SFA stage can be implemented in two manners. For sparse-source scenarios, a one-stage algorithm based on compressive sensing algorithm is utilized. Alternatively, a two-stage algorithm can be used, where the minimum power distortionless response beamformer is used to localize the sources and Tikhonov regularization algorithm is used to extract the source amplitudes. In the sound field synthesis (SFS), a 32-element rectangular loudspeaker array is employed to decode the target sound field using pressure matching technique. To establish the room response model, as required in the pressure matching step of the SFS phase, an SFA technique for nonsparse-source scenarios is utilized. Choice of regularization parameters is vital to the reproduced sound field. In the SFS phase, three SFS approaches are compared in terms of localization performance and voice reproduction quality. Experimental results obtained in a reverberant room are presented and reveal that an accurate room response model is vital to immersive rendering of the reproduced sound field.

Iterative robust adaptive beamforming

The minimum power distortionless response beamformer has a good interference rejection capability, but the desired signal will be suppressed if signal steering vector or data covariance matrix is not precise. The worst-case performance optimization-based robust adaptive beamformer (WCB) has been developed to solve this problem. However, the solution of WCB cannot be expressed in a closed form, and its performance is affected by a prior parameter, which is the steering vector error norm bound of the desired signal. In this paper, we derive an approximate diagonal loading expression of WCB. This expression reveals a feedback loop relationship between steering vector and weight vector. Then, a novel robust adaptive beamformer is developed based on the iterative implementation of this feedback loop. Theoretical analysis indicates that as the iterative step increases, the performance of the proposed beamformer gets better and the iteration converges. Furthermore, the proposed beamformer does not subject to the steering vector error norm bound constraint. Simulation examples show that the proposed beamformer has better performance than some classical and similar beamformers.

Combining DCQGMP-Based Sparse Decomposition and MPDR Beamformer for Multi-Type Interferences Mitigation for GNSS Receivers.

In the coexistence of multiple types of interfering signals, the performance of interference suppression methods based on time and frequency domains is degraded seriously, and the technique using an antenna array requires a large enough size and huge hardware costs. To combat multi-type interferences better for GNSS receivers, this paper proposes a cascaded multi-type interferences mitigation method combining improved double chain quantum genetic matching pursuit (DCQGMP)-based sparse decomposition and an MPDR beamformer. The key idea behind the proposed method is that the multiple types of interfering signals can be excised by taking advantage of their sparse features in different domains. In the first stage, the single-tone (multi-tone) and linear chirp interfering signals are canceled by sparse decomposition according to their sparsity in the over-complete dictionary. In order to improve the timeliness of matching pursuit (MP)-based sparse decomposition, a DCQGMP is introduced by combining an improved double chain quantum genetic algorithm (DCQGA) and the MP algorithm, and the DCQGMP algorithm is extended to handle the multi-channel signals according to the correlation among the signals in different channels. In the second stage, the minimum power distortionless response (MPDR) beamformer is utilized to nullify the residuary interferences (e.g., wideband Gaussian noise interferences). Several simulation results show that the proposed method can not only improve the interference mitigation degree of freedom (DoF) of the array antenna, but also effectively deal with the interference arriving from the same direction with the GNSS signal, which can be sparse represented in the over-complete dictionary. Moreover, it does not bring serious distortions into the navigation signal.

Performance analysis of GNSS multipath mitigation using antenna arrays

Multipath affects the shape of the correlation function and results in biased pseudorange measurements and erroneous navigation solutions. Antenna array processing, which uses signal spatial characteristics, is an effective method to mitigate various types of interference signals. However, the performance of most of the distortionless beamforming techniques degrades in multipath conditions due to the correlation between the desired Line of Sight (LOS) signal and multipath signals. This paper characterizes the performance of different beamforming techniques to mitigate multipath signals through the processing and analysis of simulated and actual data. The main novelty is the investigation of multipath mitigation performance of practically realizable antenna array-based GNSS receivers when the beamforming process is completely integrated into the tracking module after de-spreading. Beamforming techniques such as Delay And Sum (DAS) beamforming, Minimum Power Distortionless Response (MPDR) with and without spatial smoothing are considered. A novel multi-antenna simulator test-bed is developed to generate multipath signals for a multi-antenna platform. A software multi-antenna GPS receiver incorporating different beamforming techniques is then developed to generate pseudorange measurements and position solutions. Carrier-to-Noise ratio (C/N0), pseudorange errors and position solutions before and after beamforming are compared to show the effectiveness of different beamforming techniques to mitigate multipath. Results with simulated and actual GPS signals show improved performance using the MPDR beamformer with spatial smoothing. The utilization of spatial processing results in a pseudorange error reduction of up to 60 % and a position error reduction of up to 30 %.

Interference and multipath mitigation utilising a two‐stage beamformer for global navigation satellite systems applications

The performance of location-based services provided by global navigation satellite systems is compromised by interference and multipath propagations. Although time/frequency interference suppression methods have been widely studied in the literature, they fail to cope with wideband interference signals. Instead, techniques utilising several antenna elements can be employed to mitigate both narrowband and broadband interference signals. However, the performance of beamforming techniques utilising antenna arrays severely degrades in dealing with correlated and coherent multipath components which cause signal cancellation phenomenon and temporal correlation matrix rank deficiency. This study proposes a two-stage beamformer to jointly deal with interference and multipath signals. In the first stage, before the despreading process, by applying the subspace method, the interference subspace is estimated and used as a constraint for the optimisation problem in the next stage. In the second stage, a modified version of the minimum power distortionless response beamformer employing several overlapping sub-arrays called the minimum difference output power method is utilised to mitigate the correlated multipath components. The proposed beamformer can deal with the signal cancellation phenomenon and temporal correlation matrix rank deficiency. Several simulation examples and a real data test are provided to illustrate the effectiveness of the proposed beamformer. Results show that the proposed method is able to put deep nulls in the direction of the narrowband and wideband interference signals, and significantly reduces the multipath-induced time of the arrival error.

Beampattern Synthesis With Blind Interference Suppression

We suggest an adaptive blind interference suppression (BIS) beamformer that synthesizes a desired beampattern. For a strong interference, the principal eigenvector of the data sample covariance matrix approaches the array response at the interference direction. The beamformer weight vector minimizes the distance from a desired beampattern subject to a zero response to this principal eigenvector. For a strong interference, uncorrelated with the desired signal, the asymptotic array gain (AG) of the BIS beamformer achieves the AG of the minimum power distortionless response (MPDR) beamformer. Simulations demonstrate that, for a finite number of samples, the AG of the BIS beamformer outperforms that of the MPDR beamformer and that it is not sensitive to sensor gain and phase errors.

Robust speech recognition based on independent vector analysis using harmonic frequency dependency

This paper describes an algorithm that enhances speech by independent vector analysis (IVA) using harmonic frequency dependency for robust speech recognition. While the conventional IVA exploits the full-band uniform dependencies of each source signal, a harmonic clique model is introduced to improve the enhancement performance by modeling strong dependencies among multiples of fundamental frequencies. An IVA-based learning algorithm is derived to consider the non-holonomic constraint and the minimal distortion principle to reduce the unavoidable distortion of IVA, and the minimum power distortionless response beamformer is used as a pre-processing step. In addition, the algorithm compares the log-spectral features of the enhanced speech and observed noisy speech to identify time–frequency segments corrupted by noise and restores those with the cluster-based missing feature reconstruction technique. Experimental results demonstrate that the proposed method enhances recognition performance significantly in noisy environments, especially with competing interference.

Maximum likelihood estimation of direction of arrival using an acoustic vector-sensor

A vector-sensor consisting of a monopole sensor collocated with orthogonally oriented dipole sensors is used for direction of arrival (DOA) estimation in the presence of an isotropic noise-field or internal device noise. A maximum likelihood (ML) DOA estimator is derived and subsequently shown to be a special case of DOA estimation by means of a search for the direction of maximum steered response power (SRP). The problem of SRP maximization with respect to a vector-sensor can be solved with a computationally inexpensive algorithm. The ML estimator achieves asymptotic efficiency and thus outperforms existing estimators with respect to the mean square angular error (MSAE) measure. The beampattern associated with the ML estimator is shown to be identical to that used by the minimum power distortionless response beamformer for the purpose of signal enhancement.