DE102022204903A1 - Signal processing device for processing water sound with a directional generator - Google Patents

Abstract

A signal processing device (19) for processing water noise (22) in a watercraft is disclosed. The signal processing device comprises a signal processing processor (20) and a direction generator (21). The signal processing processor is designed to receive output signals (24) from water sound transducers (18), which convert the received water sound (22) into the output signal (24), and to determine an overall acoustic pressure P(t, xn) based on the output signals (24). , with time t and xn = (n - 1) · d, where d corresponds to the distance between two water sound transducers (18) n (with n = 1 ... N; N ≥ 2), the total pressure P increases in the frequency range transform to obtain a spectrum P(f,xn) of the total pressure, where f denotes the frequency, and determine a power density spectrum based on the spectrum P(f,xn). The direction generator is designed to carry out a direction formation of the output signals based on the power density spectrum.

Description

The invention relates to a signal processing device and a signal processing method for processing output signals (also referred to as water sound signals) from water sound transducers, which receive water sound and convert it into an associated output signal.

The signal processing of water sound signals is known in principle. For example, sonar methods can be used, for example to locate objects or to orientate a submerged underwater vehicle. Manipulation of the water sound signals, for example by identifying and eliminating noise components, is made more difficult by the complex-valued variables in the frequency range.

The object of the present invention is therefore to create an improved concept for the signal processing of water sound signals.

The task is solved by the subject matter of the independent patent claims. Further advantageous embodiments are the subject of the dependent claims.

Exemplary embodiments show a signal processing device for processing water noise in a watercraft with a signal processing processor. The signal processing processor is designed to receive output signals from water sound transducers. The water sound transducers convert the received water sound into the output signal, also known as the water sound signal. In particular, the water sound transducers are arranged equidistantly. The signal processing processor and the water transducers can also form a system.

Based on the output signals, the signal processing processor can determine a total acoustic pressure P(t, x _n ). Here t denotes the time and d the distance between two water sound transducers n (with n = 1 ... N; N ≥ 2). Furthermore, x _n = (n - 1) · d. The signal processing processor transforms the total pressure P into the frequency domain to obtain a spectrum P(f,x _n ) of the total pressure, where f denotes the frequency. Furthermore, the signal processing processor can determine a power density spectrum based on the spectrum P(f,x _n ).

bestimmt werden, wobei $F_x$

die Fouriertransformation nach x bezeichnet.To determine the power density spectrum, the signal processing processor can transform the spectrum P(f,x _n ) into the wavenumber frequency range. The spectrum P(f, k) thus generated can be converted into a power density spectrum using the magnitude square. Here k denotes the wave number. Thus, the power density spectrum can be calculated according to the formula | ${\left|F_x\left(P\left(ƒ,x_n\right)\right)\right|}^2$

be determined, whereby $F_x$

denotes the Fourier transform to x.

Furthermore, the signal processing device has a direction generator, which is designed to carry out a direction formation of the output signals based on the power density spectrum.

The idea is therefore to carry out direction formation directly based on the power density spectrum instead of the delay-and-sum beamformer. This is advantageous because the power density spectrum has become real through the absolute value formation and phase information from the output signals of the water sound transducers no longer need to be taken into account. This improves the results of signal processing and reduces the necessary computing power.

wobei c die Schallgeschwindigkeit und 0 die Richtung angibt, wählen.In embodiments, the directional generator may perform directional formation based on the power density spectrum multiplied by an angle that is independent of the output signals. This angle can be selected depending on a direction in which reception sensitivity is optimized. Ie, in the direction in which the water sound transducers should look. For example, the direction generator can define the angle as e ^{i·2·π·τ n (Θ)} , with the imaginary unit i, the frequency f and the angle information ${\tau }_n\left(\mathrm{\Theta }\right)=\frac{x_n\cdot \text{cos}\left(\mathrm{\Theta }\right)}{c},$

where c is the speed of sound and 0 is the direction.

mit $\rho \left(ƒ,r,\mathrm{\Theta }\right)=\frac{\mathrm{\Gamma }\left(ƒ,x_n,\mathrm{\Theta }\right)}{\mathrm{\Gamma }\left(ƒ,x_{ref},\mathrm{\Theta }\right)}$

für r = x_n - x_ref > 0 und $\overline{\rho \left(ƒ,-r,\mathrm{\Theta }\right)}$

für r < 0 ausgedrückt werden, wobei $\mathrm{\Gamma }\left(ƒ,x_n,\mathrm{\Theta }\right)=F_x^{-1}{\left|F_x\left(P\left(ƒ,x_n\right)\right)\right|}^2\cdot e^{i\cdot 2\cdot \pi \cdot ƒ\cdot {\tau }_n\left(\mathrm{\Theta }\right)}$

mit $F_x$

als Frequenztransformation nach x und dem Schalldruck P. $\overline{\rho \left(ƒ,-r,\mathrm{\Theta }\right)}$

bezeichnet den konjugiert komplexen Wert von p(f, -r, 0). Der Array Gewinn bezeichnet den Gewinn der Messung mit dem Array unter Verwendung der Richtungsbildung (Beamforming) gegenüber der Verwendung eines einzelnen Wasserschallwandlers.In further exemplary embodiments, the direction generator can be used as an array $$aG\left(t,\mathrm{\Theta }\right)=\frac{N^2}{{\mathrm{\Sigma }}_{n=1}^N{\mathrm{\Sigma }}_{m=1}^N\rho \left(ƒ,x_n-x_m,\mathrm{\Theta }\right)}$$

with $\rho \left(ƒ,r,\mathrm{\Theta }\right)=\frac{\mathrm{\Gamma }\left(ƒ,x_n,\mathrm{\Theta }\right)}{\mathrm{\Gamma }\left(ƒ,x_{ref},\mathrm{\Theta }\right)}$

for r = x _n - x _ref > 0 and $\overline{\rho \left(ƒ,-r,\mathrm{\Theta }\right)}$

can be expressed for r < 0, where $\mathrm{\Gamma }\left(ƒ,x_n,\mathrm{\Theta }\right)=F_x^{-1}{\left|F_x\left(P\left(ƒ,x_n\right)\right)\right|}^2\cdot e^{i\cdot 2\cdot \pi \cdot ƒ\cdot {\tau }_n\left(\mathrm{\Theta }\right)}$

with $F_x$

as a frequency transformation according to x and the sound pressure P. $\overline{\rho \left(ƒ,-r,\mathrm{\Theta }\right)}$

denotes the conjugate complex value of p(f, -r, 0). The array gain refers to the gain of the measurement with the array using beamforming compared to using a single water transducer.

In further exemplary embodiments, the signal processing processor is designed to correct the determined power density spectrum by a disturbance value in order to obtain an optimized power density spectrum with reduced noise components.

Thus, any corrections to the water-sound signal, i.e. for example, if certain components are to be calculated out, the corrections can be carried out based on the power density spectrum of the water-sound signal. This is advantageous because the power density spectrum is real and phase information no longer needs to be taken into account during the modification. This avoids a whole source of errors and reduces the necessary computing power. The investigation of the method shows that avoiding this source of error significantly increases the quality of the results. Further computing power is also saved because the calculation can be carried out partially analytically and no numerical method is used.

If other approaches to audio signal processing are considered, it is noticeable that the results are only calculated numerically. It is also noticeable that these approaches to classical audio signal processing for hydroacoustics provide worse results than the results for classical audio signal processing would suggest. These approaches are based, for example, on the analysis of the water sound signal or its spectrum with a Vienna filter. In this respect, the proposed signal processing delivers better results and is more efficient in calculation.

• Γ(f,x_n) bezeichnet das Autokorrelationsspektrum und ist definiert als die inverse Ortsbereichstransformation (beispielsweise IFFT) des Leistungsdichtespektrums $$\left(\mathrm{\Gamma }\left(ƒ,x_n\right)={\mathcal{F}}_x^{-1}\left({\left|{\mathcal{F}}_x\left(P\left(ƒ,x_n\right)\right)\right|}^2\right)\right).$$
  

For a better understanding of the following exemplary embodiments, reference is made to the following definitions:

• Γ(f,x _n ) denotes the autocorrelation spectrum and is defined as the inverse spatial domain transform (e.g. IFFT) of the power density spectrum $$\left(\mathrm{\Gamma }\left(ƒ,x_n\right)={\mathcal{F}}_x^{-1}\left({\left|{\mathcal{F}}_x\left(P\left(ƒ,x_n\right)\right)\right|}^2\right)\right).$$
  

mit r = x_n - x_ref.Normalized to the autocorrelation spectrum of a reference water sound transducer, the associated autocorrelation function results $\rho \left(ƒ,r\right)=\frac{\mathrm{\Gamma }\left(ƒ,x_n\right)}{\mathrm{\Gamma }\left(ƒ,x_{ref}\right)}$

with r = x _n - x _ref .

Embodiments show that the signal processing processor uses a power density spectrum of background noise as a disturbance value. Preferably, the signal processing processor can estimate the power density spectrum of the background noise as isotropic noise.

auf. c bezeichnet die Schallgeschwindigkeit im Wasser, sine bezeichnet den Sinus cardinalis, auch als Si-Funktion bekannt. Unter Verwendung der im Vorhinein eingeführten Funktionen ergibt sich das zugehörige Leistungsdichtespektrum zu $\mathrm{\Gamma }\left(ƒ,x_{ref}\right)\cdot \frac{\pi \cdot c}{2\cdot \left|ƒ\right|}\cdot rect\left(\frac{c}{4\cdot ƒ}\cdot k_r\right).$

Insoweit kann der Signalverarbeitungsprozessor das Leistungsdichtespektrum des Hintergrundrauschens mittels des Terms $\mathrm{\Gamma }\left(ƒ,x_{ref}\right)\cdot \frac{\pi \cdot c}{2\cdot \left|ƒ\right|}\cdot rect\left(\frac{c}{4\cdot ƒ}\cdot k_r\right)\text{ mit }{k}_r=\frac{2\pi }{{\lambda }_r}$

$$

schätzen und, insbesondere unter der Voraussetzung der Unkorreliertheit, das geschätzte Leistungsdichtespektrum des Hintergrundrauschens von dem Leistungsdichtespektrum des Spektrums P(f,x_n) subtrahieren. Dies ist beispielsweise vorteilhaft, um das Signal-Rausch Verhältnis in dem Wasserschallsignal zu verbessern.Isotropic noise has the autocorrelation function ${\rho }_{isO}\left(ƒ,r\right)=sinc\left(\frac{2\cdot ƒ\cdot r}{c}\right)$

on. c denotes the speed of sound in water, sine denotes the cardinal sine, also known as the Si function. Using the functions introduced in advance, the associated power density spectrum is obtained $\mathrm{\Gamma }\left(ƒ,x_{ref}\right)\cdot \frac{\pi \cdot c}{2\cdot \left|ƒ\right|}\cdot rect\left(\frac{c}{4\cdot ƒ}\cdot k_r\right).$

In this respect, the signal processing processor can calculate the power density spectrum of the background noise using the term $\mathrm{\Gamma }\left(ƒ,x_{ref}\right)\cdot \frac{\pi \cdot c}{2\cdot \left|ƒ\right|}\cdot rect\left(\frac{c}{4\cdot ƒ}\cdot k_r\right)\text{with}{k}_r=\frac{2\pi }{{\lambda }_r}$

$$

estimate and, in particular under the condition of uncorrelation, subtract the estimated power density spectrum of the background noise from the power density spectrum of the spectrum P(f,x _n ). This is advantageous, for example, to improve the signal-to-noise ratio in the water sound signal.

In further exemplary embodiments, the signal processing processor can determine a power density spectrum of a sound source that is audible in the water sound as a disturbance value. A sound source is an object that emits sound or reflects sound. To determine the power density spectrum of the sound source, the signal processing processor can determine a direction α _sig in which the sound source is located. The direction can be determined, for example, by a difference between the signal strength in the direction α _sig and the signal strengths in the direction α _sig + 3° and α _sig - 3° (preferably in each case) exceeding a threshold value.

$\mathrm{\Gamma }\left(ƒ,x_{ref}\right)\cdot \frac{\epsilon }{{\left({\tilde{k}}_r-\frac{ƒ\cdot \text{cos}\left({\mathrm{\alpha }}_{sig}\cdot \frac{\pi }{180}\right)}{c}\right)}^2+{\epsilon }^2}$

ergeben. Der Signalverarbeitungsprozessor kann das Leistungsdichtespektrum der Schallquelle demnach mittels des Terms $$\mathrm{\Gamma }\left(ƒ,x_{ref}\right)\cdot \frac{\epsilon }{{\left({\tilde{k}}_r-\frac{ƒ\cdot \text{cos}\left({\mathrm{\alpha }}_{sig}\cdot \frac{\pi }{180}\right)}{c}\right)}^2+{\epsilon }^2}$$

mit ${\tilde{k}}_r=\frac{1}{{\lambda }_r}=\frac{k_r}{2\pi }$

schätzen und das geschätzte Leistungsdichtespektrum der Schallquelle von dem Leistungsdichtespektrum des Spektrums P(f,x_n) subtrahieren. ε bezeichnet einen Parameter, der die Breite der Ziellinie der Schallquelle abbildet. ε kann ≤ $\le \frac{1}{75},$

bevorzugt ≤ $\le \frac{1}{100},$

besonders bevorzugt ≤ $\le \frac{1}{120}$

gewählt werden.In embodiments, the signal processing processor may estimate the power density spectrum of the sound source as the target line of a wavenumber frequency diagram. The power density spectrum of this direction-dependent target line can be estimated using a delta distribution. For example, a power density spectrum can arise $$

$\mathrm{\Gamma }\left(ƒ,x_{ref}\right)\cdot \frac{\epsilon }{{\left({\tilde{k}}_r-\frac{ƒ\cdot \text{cos}\left({\mathrm{\alpha }}_{siG}\cdot \frac{\pi }{180}\right)}{c}\right)}^2+{\epsilon }^2}$

result. The signal processing processor can therefore determine the power density spectrum of the sound source using the term $$\mathrm{\Gamma }\left(ƒ,x_{ref}\right)\cdot \frac{\epsilon }{{\left({\tilde{k}}_r-\frac{ƒ\cdot \text{cos}\left({\mathrm{\alpha }}_{siG}\cdot \frac{\pi }{180}\right)}{c}\right)}^2+{\epsilon }^2}$$

with ${\tilde{k}}_r=\frac{1}{{\lambda }_r}=\frac{k_r}{2\pi }$

estimate and subtract the estimated power density spectrum of the sound source from the power density spectrum of the spectrum P(f,x _n ). ε denotes a parameter that represents the width of the target line of the sound source. ε can ≤ $\le \frac{1}{75},$

preferably ≤ $\le \frac{1}{100},$

particularly preferably ≤ $\le \frac{1}{120}$

to get voted.

The total sound pressure P(t,x _n ) now consists of the intrinsic noise, the background noise and the noise generated by one or more sound sources. For customer acceptance of the watercraft, it is now important to determine the watercraft's own noise, as this must not exceed a predetermined threshold value. The intrinsic noise can be determined by subtracting the power density spectrum of the sound source(s) and the power density spectrum of the background noise from the power density spectrum of the total sound pressure to obtain the optimized power density spectrum.

In exemplary embodiments, it is possible to carry out direction formation directly on the optimized power density spectrum instead of the delay-and-sum beamformer. The signal processing device can thus comprise a direction generator which is designed to carry out a direction formation of the output signals based on the optimized power density spectrum. This is advantageous because the phase information of the output signals necessary for using the delay-and-sum beamformer is not required. This reduces the computing power required to carry out the direction formation. Furthermore, it is thus possible for the directional generator to obtain an improved signal-to-noise ratio by using the optimized power density spectrum, for example by previously eliminating background noise, and thus also enable improved target detection.

Furthermore, a corresponding method for processing water noise in a watercraft is disclosed. The method includes the steps: -receiving output signals from water sound transducers, which convert the received water sound into the output signal; -Determine a total acoustic pressure P(t, x _n ) based on the output signals, with time t and x _n = (n - 1) - d, where d is the distance between two water sound transducers n (with n = 1 ... N ; N ≥ 2) corresponds; -Transforming the total pressure P into the frequency domain to obtain a spectrum P(f,x _n ) of the total pressure, where f denotes the frequency; -Determining a power density spectrum based on the spectrum P(f,x _n ); - Performing directional formation of the output signals based on the power density spectrum.

• 1: ein schematisches Blockdiagramm eines Systems mit einem Signalprozessor zur Verarbeitung von Wasserschall;
• 2: ein schematisches Frequenz-Wellenzahl Diagramm.

Preferred embodiments of the present invention are explained below with reference to the accompanying drawings. Show it:

• 1 : a schematic block diagram of a system with a signal processor for processing water sound;
• 2 : a schematic frequency-wavenumber diagram.

Before exemplary embodiments of the present invention are explained in more detail below with reference to the drawings, it should be noted that identical, functionally identical or equivalent elements, objects and/or structures are provided with the same reference numerals in the different figures, so that those shown in different exemplary embodiments Description of these elements is interchangeable or can be applied to one another.

1 shows a schematic block diagram of a system 16 with three water sound transducers 18 and a signal processing device that has a signal processing processor 20 and a directional generator 21. The water sound transducers receive water sound 22 and convert the water sound into a (usually electrical) output signal 24, the water sound signal. The signal processing processor 20 receives the output signals 24 and processes them. A corresponding processed water sound signal 26, or one water sound signal per output signal, can be provided at the output of the signal processing processor. This can be, for example, the optimized power density spectrum or a signal derived from it.

The directional images 21 can determine the direction of the processed water sound signals 26. This makes it possible to increase the sensitivity of the water sound transducers in one direction and thus allow the water sound transducers to look virtually in one direction or to pivot virtually. At the exit of the direction former 21 the corresponding direction signals 28 can be tapped.

2 shows an exemplary diagram in which an intensity distribution of water sound can be represented. The intensity is plotted per frequency and wave number. This means that a grid of frequency versus wave number with intensities of water sound is preferably displayed. The number of points in the grid is determined by the sampling rate (resolution of the frequency axis) and the number and spacing of the water sound transducers in the line array (resolution of the wavenumber axis).

The diagram is divided by a V-shaped dividing line 30. Within the V 30, ie in an area 32, lie the sound waves that hit the water sound transducers at a relative speed that is greater than the speed of sound in the water. On the V, ie on the dividing line 30, lie the sound waves that travel over the water sound transducers at a speed that corresponds exactly to the speed of sound in the water. Targets or target tracks result in a straight line within the V. An example is a target 34 in 2 shown.

The disclosed (water) sound transducers are designed for use under water, especially in the sea. The sound transducers can convert water sound into an electrical signal (e.g. voltage or current) corresponding to the sound pressure, the water sound signal. It is also possible for the sound transducers to convert an applied electrical voltage into water sound. The sound transducers can therefore be used as water sound receivers and/or as water sound transmitters. The sound transducers can have a piezoelectric material, for example a piezoceramic, as the sensory material. The sound transducers can be used for (active and/or passive) sonar (sound navigation and ranging). The sound transducers are preferably not suitable for medical applications or are not used for medical applications.

Although some aspects have been described in connection with a device, it is understood that these aspects also represent a description of the corresponding method, so that a block or a component of a device is also to be understood as a corresponding method step or as a feature of a method step. Analogously, aspects that have been described in connection with or as a method step also represent a description of a corresponding block or detail or feature of a corresponding device.

The embodiments described above are merely illustrative of the principles of the present invention. It will be understood that modifications and variations of the arrangements and details described herein will occur to others skilled in the art. Therefore, it is intended that the invention be limited only by the scope of the following claims and not by the specific details presented from the description and explanation of the exemplary embodiments herein.

List of reference symbols:

16

systems

18

Water sound transducer

19

Signal processing device

20

Signal processing processor

21

Direction Former

22

Water noise

24

Output signal

26

processed water sound signal

28

Directional signals

Claims (14)

Signal processing device (19) for processing water sound (22) in a watercraft, the signal processing device (19) having the following features: a signal processing processor (20) which is designed to receive output signals (24) from water sound transducers (18) which convert received water sound (22) into the output signal (24); - based on the output signals (24) to determine a total acoustic pressure P(t, x _n ), with time t and x _n = (n - 1) - d, where d is the distance between two water sound transducers (18) n (with n = 1 ... N; N ≥ 2) corresponds; - transform the total pressure P into the frequency domain to obtain a spectrum P(f,x _n ) of the total pressure, where f denotes the frequency; - determine a power density spectrum based on the spectrum P(f, x _n ); and a directional generator configured to perform directional formation of the output signals based on the power density spectrum. Signal processing device (19) according to Claim 1 : - wherein the signal processing processor is designed to correct the power density spectrum by a disturbance value in order to obtain an optimized power density spectrum with reduced noise components. Signal processing device (19) according to one of the preceding claims: - wherein the direction generator is designed to carry out the direction formation based on the power density spectrum multiplied by an angle that is independent of the output signals. Signal processing device (19) according to Claim 3 ; - wherein the directional generator is designed to select the angle depending on a direction in which a reception sensitivity is optimized. Signal processing device (19) according to Claim 4 ; - where the direction generator is designed, the angle as e ^{i · 2 · π · f · τ n (Θ)} , with the imaginary unit i, the frequency f and the angle information $$

${\tau }_n\left(\mathrm{\Theta }\right)=\frac{x_n\cdot \text{cos}\left(\mathrm{\Theta }\right)}{c},$

where c is the speed of sound and 0 is the direction. Signal processing device (19) according to one of the preceding claims, - wherein the signal processing processor (20) is designed to transform the spectrum P(f,x _n ) into the spatial frequency range to determine the power density spectrum in order to obtain a spatial spectrum P(f, k). and to form the absolute square of the spatial spectrum to obtain the power density spectrum, where k denotes the wave number. Signal processing device (19) according to one of Claims 2 until 6 , - wherein the signal processing processor (20) is designed to use a power density spectrum of background noise as the disturbance value. Signal processing device (19) according to Claim 7 , - wherein the signal processing processor (20) is designed to estimate the power density spectrum of the background noise as isotropic noise. Signal processing device (19) according to Claim 6 or 7 , - wherein the signal processing processor (20) is designed to determine the power density spectrum of the background noise using the term $$\mathrm{\Gamma }\left(ƒ,x_{ref}\right)\cdot \frac{\pi \cdot c}{2\cdot \left|ƒ\right|}\cdot rect\left(\frac{c}{4\cdot ƒ}\cdot k_r\right)$$

with $k_r=\frac{2\pi }{{\lambda }_r}$

to estimate and subtract the estimated power density spectrum of the background noise from the power density spectrum of the spectrum P(f, x _n ), where Γ(f, x _ref ) denotes the autocorrelation spectrum of a reference water sound transducer and c denotes the speed of water sound. Signal processing device (19) according to one of Claims 2 until 9 , - wherein the signal processing processor (20) is designed to determine a power density spectrum of a sound source as a disturbance value, which is audible in the water sound (22). Signal processing device (19) according to Claim 10 , - wherein the signal processing processor (20) is designed to determine a direction α _sig in which the sound source is located in order to determine the power density spectrum of the sound source. Signal processing device (19) according to Claim 10 or 11 , - wherein the signal processing processor (20) is designed to estimate the power density spectrum of the sound source as the target line of a wavenumber frequency diagram. Signal processing device (19) according to one of Claims 10 until 12 , - wherein the signal processing processor (20) is designed to determine the power density spectrum of the sound source using the term $$\mathrm{\Gamma }\left(ƒ,x_{ref}\right)\cdot \frac{\epsilon }{{\left({\tilde{k}}_r-\frac{ƒ\cdot \text{cos}\left({\mathrm{\alpha }}_{siG}\cdot \frac{\pi }{180}\right)}{c}\right)}^2+{\epsilon }^2}$$

with ${\tilde{k}}_r=\frac{1}{{\lambda }_r}=\frac{k_r}{2\pi }$

to estimate and subtract the estimated power density spectrum of the sound source from the power density spectrum of the spectrum P(f,x _n ), where Γ(f,x _ref ) is the autocorrelation spectrum of a reference water sound transducer, c is the speed of water sound and ε is a parameter that represents the width of a target line of the Depicts sound source. Method for processing water sound (22) in a watercraft with the following steps, - Receiving output signals from water sound transducers, which convert the received water sound (22) into the output signal; - Determining a total acoustic pressure P(t,x _n ) based on the output signals, with time t and x _n = (n - 1) - d, where d is the distance between two water sound transducers (18) n (with n = 1. .. N; N ≥ 2) corresponds; - transforming the total pressure P into the frequency domain to obtain a spectrum P(f,x _n ) of the total pressure, where f denotes the frequency; - Determining a power density spectrum based on the spectrum P(f, x _n ); - Performing directional formation of the output signals based on the power density spectrum.

Images