Stefania Repetto

A stochastic approach to the synthesis of a robust frequency-invariant filter-and-sum beamformer

Frequency-invariant beam patterns (FIBPs) are often required by systems using an array of sensors to process broadband signals. Although several methods have been proposed to design a broadband beamformer (typically characterized by a finite impulse response (FIR) filter for each sensor) with an FIBP, until now, the case in which the spatial aperture is shorter than the involved wavelengths has very rarely been considered. In such a case, the use of a superdirective beam pattern is essential to attaining an efficient system. In this context, robustness to array imperfections and random errors is a very crucial feature. In this paper, a method to design a robust broadband beamformer that produces an FIBP for a data-independent superdirective array is proposed and compared with other potential approaches. The method generates a far-field beam pattern that reproduces the desired profile over a very wide frequency band (also if the array is shorter than the wavelength) and is based on a stochastic approach to the direct synthesis of the FIR filters. The very simple implementation and the resulting robustness of the attained filter-and-sum beamformer to array imperfections increase the applicability of the system. This fact is particularly important in the context of a audio signal processing carried out by microphone arrays, which is the main application considered in this paper.

Devising an Affordable Sonar System for Underwater 3-D Vision

In this paper, the preliminary design and assessment of a high-resolution 3-D acoustic imaging system based on a sparse planar array of sensors, which is particularly intended for underwater applications, are presented. Critical issues in the development of high-resolution 3-D sonar systems are 1) the cost of hardware, which is associated with the huge number of sensors that compose the planar array, and 2) the computational burden of processing the signals that were gathered. Here, such problems are overcome by the optimized synthesis of an aperiodic sparse array that allows the device to operate at different frequencies, yielding an acceptable sidelobe level and a good tradeoff between the field of view and the resolution. The array optimization is performed using an efficient stochastic method, in which the number of sensors is minimized, whereas their positions and weights are simultaneously optimized. To test the validity of the designed system, the signals that the sparse array received in response to the insonification of a scene with a wideband pulse are simulated, and voxel-based beamforming is applied to generate the 3-D image. The obtained images show high fidelity to the geometrical characteristics of the scene, in accordance with the expected performance of the 3-D sonar system.

Designing superdirective microphone arrays with a frequency-invariant beam pattern

Frequency-invariant beam patterns are often required in systems using an array of sensors to process broadband signals. Although several methods have been proposed to design a broadband beamformer [typically realized with a finite-impulse-response (FIR) filter for each sensor] with a frequency-invariant beam pattern (FIBP), until now the case in which the spatial aperture is shorter than the involved wavelengths has very rarely been addressed. In such a case, the use of a superdirective beam pattern is essential for attaining an efficient system. In this paper, a novel method to design a broadband beamformer that produces an FIBP for a data-independent superdirective array is proposed and compared with other potential approaches. The method generates a far-field beam pattern that reproduces the desired profile over a very wide frequency band, also if the array is shorter than the wavelength. Two steps are necessary: 1) the generation of many apodizing windows at different frequency values by a stochastic method and 2) the synthesis of the FIR filters with the Parks-McClellan technique. At the end of the design chain, the very simple implementation and the robustness of the attained broadband beamformer to array imperfections increases the applicability of the system, for instance, in audio signal processing using microphone arrays

Design and Assessment of a Low-Cost 3-D Sonar Imaging System Based on a Sparse Array

In this paper, the design and the assessment of a high-resolution three-dimensional acoustic imaging system based on a sparse planar array of sensors are presented. The aim is to generate useful acoustic 3D images in underwater context. Towards this end, a planar array is mandatory, as a linear aperture does not allow one to discriminate signals coming from a 3D space. One critical issue in the development of high-resolution 3D sonar systems is the hardware cost associated to the necessary huge number of sensors. In this paper, an innovative 3D imaging system able to operate at different resolution levels is proposed that is based on a single sparse planar array consisting of only 584 elements. Such a limited number of sensors represents an important stage in designing 3D acoustic imaging systems, making feasible the achieving of a drastic reduction in both costs and successive processing associated to the system. The array optimization is performed by an efficient stochastic method based on the simulated annealing algorithm, in which the positions and the weights of the array elements are optimized simultaneously. To test the validity of the proposed system, the signals received by the sparse array as the response of a given scene insonifled by a pulse are simulated. To move from the simulated signals to the 3D image of the scene, a voxel-based beamforming in the time domain is designed. To assess the proposed acoustic imaging system, several complex scenes are taken into account and images are obtained that exhibit a high fidelity to the geometrical and physical characteristics of the assumed underwater environments

Processing and Analysis of Underwater Acoustic Images Generated by Mechanically Scanned Sonar Systems

The processing and analysis of images generated by mechanically scanned sonar systems have received poor attention despite their widespread application. In this paper, some efficient methods for acoustic image enhancement and automatic object detection are presented and assessed using a large set of experimental data collected at sea with commercial sonar systems. Specifically, a set of methods for increasing the quality of the gray-level images produced by a fan-shaped-beam sonar is introduced. Such a set includes a dynamic brightness assignment, a precise interpolation, a speckle-reduction filter, and a contrast-enhancement block. Two versions of a template-matching-based method that allows the automatic detection of a simple object contained in a region scanned with a pencil-beam sonar are also proposed and assessed. The main difficulty to be coped with in this field is related to the sparseness of the binary maps generated by this sonar system. The performance and robustness of the proposed techniques have been evaluated using real data that provided satisfactory results for both the image-enhancement and the object-detection tasks. Moreover, the computational burden of most of the proposed techniques turned out to be quite limited, and their real-time implementation with a standard computer architecture could be estimated.

High-resolution 3-D imaging by a sparse array: array optimization and image simulation

In this paper, the design of a high-resolution three-dimensional acoustic imaging system based on a sparse planar array of sensors is presented. The aim is to generate useful acoustic 3D images in underwater context. Towards this end, a planar array is mandatory, as a linear aperture does not allow one to discriminate signals coming from a 3D space. One critical issue in the development of high-resolution 3D sonar systems is the hardware cost associated to the necessary huge number of sensors. In this paper, an innovative 3D imaging system able to operate at different resolution levels is proposed that is based on a single sparse planar array consisting of only 584 elements. Such a limited number of sensors represents an important stage in designing 3D acoustic imaging systems, making feasible the achieving of a drastic reduction in both costs and successive processing associated to the system. The array optimization is performed by an efficient stochastic method based on the simulated annealing algorithm, in which the positions and the weights of the array elements are optimized simultaneously. To test the validity of the proposed system, the signals received by the sparse array as the response of a given scene insonified by a pulse are simulated. To move from the simulated signals to the 3D image of the scene, a voxel-based beamforming in the time domain is designed. Images are obtained that exhibit a high fidelity to the geometrical and physical characteristics of the assumed underwater environment.

Spatially selective binaural hearing aids

Traditional hearing aids are limited by the absence of spatial selectivity. Superdirective microphone array can recover such limit, performing a spatial filtering to achieve an augmented SNR. We present Glassense, a platform hosting a double microphone array connected to a processing unit and mounted on the frame of common glasses. The platform has the potential of delivering binaural spatially selective audio inputs, allowing ecological pointing of acoustic sources through head motion. The designed microphone arrays exhibit a gain suitable to improve the speech reception threshold of hearing-impaired subjects, obtained on a lightweight and scalable hardware setup.

Global synthesis of superdirective frequency‐invariant beam patterns

Frequency-invariant beam patterns are often required by systems using an array of sensors to process broadband signals. If the spatial aperture is shorter than the involved wavelengths, the use of a superdirective beam pattern is essential to attain an efficient system. In this context, robustness to array imperfections is a crucial feature. In the literature, only a few approaches have been proposed to design a robust, superdirective, frequency-invariant beamformer, based on a filter and sum architecture: in all of them, the frequency invariance is achieved imposing an a priori desired beam pattern. However the choice of a suitable desired beam pattern is not trivial and depends on the specific design case: an improper selection of the desired beam pattern can produce unsatisfactory performances. We propose a new method of global synthesis, computationally inexpensive, allowing to design a robust broadband beam pattern with an optimal trade-off between the frequency invariance and the directivity, without the need of imposing a priori a desired beam pattern. The results show that the synthesized beam patterns have a directivity, a frequency-invariance, and a robustness that are very similar to or better than those of the beam patterns obtained by the literature methods.

Frequency-Invariant Beamforming in Very Short Arrays for AUVs Instrumentation

Frequency-invariant beam patterns are often required in systems using an array of sensors to acquire and measure broadband signals. Although several methods have been proposed to design a broadband beamformer (typically realized by an FIR filter for each sensor) with a frequency-invariant beam pattern, till now the case in which the spatial aperture is shorter than the involved wavelengths has not been addressed. In such a case, the use of a super directive beam pattern is essential to attain an efficient system. In this paper, a novel method to design a broadband beamforming that produces a frequency-invariant beam pattern for a data-independent super directive array is proposed and compared with other potential approaches. The method generates a far-field beam pattern that reproduces the desired profile over a very wide frequency band, also if the array is shorter than the wavelength. Two steps are necessary: (1) the generation of many apodizing windows at different frequency values by a stochastic method; (2) the synthesis of the FIR filters with the Parks-McClellan technique. At the end of the design chain, the very simple implementation and the robustness to array imperfections of the attained broadband beamformer increases the applicability of the system, for instance, in underwater acoustic systems devoted to communication, measurement and telemetry deployed on small vehicles

Frequency invariant beamforming in very short arrays

Frequency invariant beam patterns are often required in systems using an array of sensors to process broadband signals. Although several methods have been proposed to design a broadband beamformer (typically realized by an FIR filter for each sensor) with a frequency invariant beam pattern, till now the case in which the spatial aperture is shorter than the involved wavelengths has not been addressed. In such a case, the use of a superdirective beam pattern is essential to attain an efficient system. In this paper, a novel method to design a broadband beamforming that produces a frequency invariant beam pattern for a data-independent superdirective array is proposed and compared with other potential approaches. The method generates a far-field beam pattern that reproduces the desired profile over a very wide frequency band, also if the array is shorter than the wavelength. Two steps are necessary: (1) the generation of many apodizing windows at different frequency values by a stochastic method; (2) the synthesis of the FIR filters with the Parks-McClellan technique. At the end of the design chain, the very simple implementation and the robustness to array imperfections of the attained broadband beamformer increases the applicability of the system, for instance, in underwater communications for small vehicles.