Florent Le Courtois

Compressed sensing for wideband wavenumber tracking in dispersive shallow water

In shallow water zones and at low frequency, seabed and water column properties can be estimated from the acoustic wavenumbers using inversion algorithms. When considering horizontal line arrays (HLA) and narrowband sources, the wavenumbers can be evaluated with classic spectral analysis methods. In this paper, a compressed sensing (CS) method for sparse recovery of the wavenumbers is proposed. This takes advantage of the few propagating modes and allows for spectral estimation when short HLA are used. The CS representation improves the wavenumber estimation, compared to the Fourier transform. However, for small arrays and several propagating modes, the CS generates interferences and does not allow proper wavenumber estimation. When considering broadband sources, it is possible to combine the wavenumbers estimated at several frequencies in order to build a frequency-wavenumber (f - k) representation. In this case, a post-processing tracking operation which improves the f - k resolution is presented. This relies on a general approach of waveguide physics and uses a particle filtering (PF) algorithm to track the wavenumbers. The consecutive use of CS and PF leads to a better wavenumber estimation. This methodology can be used for sources that are not at an end-fire position. It is illustrated by simulations and successfully applied on the Shallow Water 2006 data using the 32 sensor SHARK array.

Autoregressive model for high-resolution wavenumber estimation in a shallow water environment using a broadband source

In a shallow water environment, wavenumbers can be estimated by computing time and spatial Fourier transforms of horizontal array measurements. The frequency-wavenumber representation allows wide band estimation but a sufficient number of hydrophones are required for accurate wavenumber resolution. This paper presents the application of an autoregressive (AR) model to compute the high resolution wavenumber spectrum. The smallest number of required sensors for the AR model is found using a stabilization diagram. The method is validated on simulated and experimental data. The wavenumbers are accurately estimated over a wide frequency band using fewer sensors than are needed for the spatial Fourier Transform.

Reconstruction of Dispersion Curves in the Frequency-Wavenumber Domain Using Compressed Sensing on a Random Array

In underwater acoustics, shallow-water environments act as modal dispersive waveguides when considering low-frequency sources, and propagation can be described by modal theory. In this context, propagated signals are composed of few modal components, each of them propagating according to its own wavenumber. Frequency-wavenumber

Wavenumber tracking in a low resolution frequency-wavenumber representation using particle filtering

(f-k)

Ambient noise dynamics in a heavy shipping area

representations are classical methods allowing modal separation. However, they require large horizontal line sensor arrays aligned with the source. In this paper, to reduce the number of sensors, a sparse model is proposed and combined with prior knowledge on the wavenumber physics. The method resorts to a state-of-the-art Bayesian algorithm exploiting a Bernoulli–Gaussian model. The latter, well suited to the sparse representations, makes possible a natural integration of prior information through a wise choice of the Bernoulli parameters. The performance of the method is quantified on simulated data and finally assessed through a successful application on real data.

Modeling the shipping noise in uncertain environment for marine space planning

In underwater acoustics, shallow water environments (d <; 200 m) act as dispersive waveguide when considering low-frequency sources (f <; 250 Hz), and propagation is described by modal theory. Propagated signals are usually multicomponent, and the group delay of each mode (each component) is dispersive and varies with mode number. The waveguide dispersion is characterized by modal wavenumbers, which are widely used as inputs of inversion algorithms to estimate environmental properties. Considering a horizontal array and a source along the axis of the array, wavenumber estimation is equivalent to spectral analysis in the range dimension. A large number of hydrophones (i.e. range samples) is thus required to perform an accurate wavenumber (i.e. spectral) estimation. This paper proposes an original approach for estimating the wavenumbers using a short array and a broadband low-frequency source. The wavenumbers are tracked in the frequency-wavenumber (f - k) domain using particle filtering. The waveguide physics provides generic system and state equations to model the f - k diagram. In particular, it is possible to define an iterative relationship for wavenumber at two consecutive frequencies using the dispersion relation, which holds true in every waveguide. The proposed method provides interesting results on simulated data using 10 hydrophones. It is validated on experimental data recorded in the North Sea.

Taking into account uncertainties in environmental impact assessment of underwater anthropogenic noise

The management of underwater noise within the European Unions waters is a significant component (Descriptor 11) of the Marine Strategy Framework Directive (MSFD). The indicator related to continuous noise, is the noise levels in two one-third octave bands centered at 63Hz and 125Hz. This paper presents an analysis of underwater noise in the Celtic Sea, a heavy shipping area which also hosts the seasonal Ushant thermal front. In addition to the MSFD recommended frequency bands, the analysis was extended to lower and upper frequency bands. Temporal and spatial variations as well as the influence of the properties of the water column on the noise levels were assessed. The noise levels in the area had a high dynamic range and generally exceeded 100dB re 1μPa. Finally, the results highlighted that oceanic mooring must be designed to minimize the pseudo-noise and consider the water column physical properties.

Modal wavenumber estimation using a Bayesian compressed sensing algorithm

The shipping noise is a major component of the underwater soundscape at low frequency. The increase of the fleet size at world scale for several decades have arisen the potential impacts of noise pollution on marine ecosystems. Modeling the shipping noise levels at large scale became an important task of marine environmental policies. However, because of fluctuating and unknown environmental parameters, the numerical model may present a bias in the estimated levels up to several tens of dB. To tackle this problem, this paper relies on a semi-empirical formulation of the noise level standard deviation, related to environmental mismatch. The model provides monthly noise atlas at world scale, using statistical distribution of shipping. Results are presented using worldwide ship traffic data for the years 2003, 2012, and 2016. It provides a relevant tool to monitor the ambient noise evolution and have been applied for the evaluation of the Marine Strategy Directive Framework.

Optimisation par algorithme génétique de la géométrie d'antenne pour la localisation de sources

The management of underwater anthropogenic noise is becoming a significant component of marine policies. Rules to limit noise exposure in a way that marine animals are not adversely affected are expected to be set up in the future. However, modeling or measuring underwater noise levels (orother suitable metrics) can still be subject to a lot of uncertainties due to difficulty in estimatation or measuring key parameters as source pressure levels and waveguide features. On the other hand, the audition threshold values inferred from bioacoustics studies may also lack of statistical robustness and are difficult to generalize. The combination of these two major sources of uncertainties may lead to misestimate the noise exposure risk and may hinder marine space planning efficiency. This work aims at developing a framework for impact studies while considering uncertainties on the acoustic metrics and confidence in the threshold values. It relies on probabilistic description of the acoustics pressure, as proposed...

Assessing risk of underwater noise impact on marine mammals throughout a new methodology

In shallow water, low-frequency acoustic propagation is described by modal theory. In this context, the environment acts as a dispersive waveguide, and the geoacoustic properties of the seabed can be inferred from the modal wavenumbers. When considering horizontal aperture (using a horizontal array or a towed source), wavenumber estimation is a well-known problem, equivalent to spectral analysis. In this paper, wavenumber estimation is revisited using Compressed Sensing (CS). Our method benefits from two strong physical hypotheses. Only a few modes are propagating so that the wavenumber spectrum is sparse. Moreover, if the source is broadband, the wavenumbers can be related from one frequency to the next using a general dispersion relationship. Our method resorts to a state-of-the-art Bayesian algorithm exploiting a Bernoulli-Gaussian model. The latter, well-suited to the sparse representations, makes possible a natural integration of the prior dispersion information through a wise choice of the Bernoulli...