Patent Translate Powered by EPO and Google Notice This translation is machine-generated. It cannot be guaranteed that it is intelligible, accurate, complete, reliable or fit for specific purposes. Critical decisions, such as commercially relevant or financial decisions, should not be based on machine-translation output. DESCRIPTION JPH11146494 [0001] BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an acoustic antenna comprising a plurality of discrete acoustic transducers, and in particular to an acoustic receiving antenna, ie an acoustic receiving antenna comprising a plurality of acoustic sensors or microphones. It is about If the reciprocity theorem applies, the invention also applies to acoustic transmitting antennas. [0002] The main purpose of an acoustic receiving antenna is to reduce all reception impairments while retaining the required information, ie the information transmitted by the loudspeaker or the required sound source. . [0003] In order to better understand the difficulties which the present invention aims to overcome, an acoustic antenna taking into account the case of an antenna with an indefinite geometry constituted by an acoustic sensor having an indefinite directivity diagram. Conventional theoretical studies of arrays will be developed in the future. [0004] The acoustic signal received by the antenna sensor is (1) another transmitter; (2) multi-path propagation; (3) echo in many cases; (4) electronic noise of the sensor and amplifier; (5) probably digital Given by the quantitative noise of the process. 05-05-2019 1 [0005] Linear additive models are conceivable, ie non-linear degradation is not taken into account. Subsequently, perturbations (1) to (3) will be referred to as "spatially coherent" or simply "coherent", while perturbations (4) and (5) will be "incoherent" It is called. [0006] The performance of the antenna in terms of coherent perturbations is given by its directivity diagram. It is assumed that the loudspeaker is located in the near field, which means that the points in space are related as an alternative regardless of the relevant direction. The coherent perturbation source is assumed to be in the far field. [0007] An equation is applied that represents the improvement of the signal to the interferometric perturbation ratio under the assumption of a diffuse field as compared to an omnidirectional sensor located at the nearest antenna sensor. The reflection is treated as an image source. Therefore, it is sufficient to know the free sound field propagation method and the directivity diagram of each sensor. [0008] A typical model for propagation is as follows. Signal from sensor m also applies for Xm observation T time Up, m Directionality Sp of sensor m in the direction of source p Signal Dp, m 05-05-2019 2 transmitted at source p Distance between source p-sensor m Propagation velocity In order to simplify the incoherent noise, (electrical and quantitative noise) calculations on the bm (t) sensor m, a frequency field is input. Observation in the X, S, B frequency field, transmitted signal and noise f frequency [0009] Antenna processing can be understood as a scalar product in the frequency field. The signal at the output of the process is given by: [0010] It is assumed that the required source required is the source p = 1. Conventional antenna processing rephases the signal if it is necessary to weight the sensor for the purpose of establishing a compromise between the opening of the main lobe and the level of the secondary lobe and the sum of them It is comprised by computing. This can be expressed as a set of coefficients as follows. In this case gm (f) is real and positive. Thus, at the output the following equation is obtained: [0011] The three terms of the sum correspond to the incoming signal, to the interfering perturbations and to the incoherent noise, respectively. This equation can be used for any linear processing purpose where complex values are allowed for gm (f). For the purpose of obtaining a directivity factor, the position of the perturbation source must be able to change, for example, p = 2, and the mean value of the remaining part of the perturbation signal has to be calculated. The amplitude factor is introduced first, and its last term acts to obtain a factor that is independent of the value of the distance if it is large enough. The following equation is then obtained with λ = c / f. The complex gain of the incoming signal is The complex gain of the coherent perturbation signal is as follows. The directivity factors are as follows. The following vector number system is applied. The following equation is then obtained: This gives the following equation: [0012] 05-05-2019 3 As already indicated, these equations are based on a propagation model which is very well adapted in free fields without obstacles. The propagation model can be replaced by measurements in order to fit this operation to situations where the model does not prove to be accurate enough. In the present case, the vector d2 (f) represents the measured propagation vector. [0013] This result can be generalized by introducing the weighting U (f, ψ, θ) of the integral second error according to the direction. [0014] Incoherent noise is irrelevant to one sensor to the other, and its power is equal to all sensors. Incoherent noise reduction is described as follows in this example. [0015] From this consideration, conventional delay / weighting / sum value processing, far-field focusing can be inferred. For a linear antenna with uniform spacing d of the sensor, the complex gain of the coherent perturbation signal G2 is as follows, and the directivity diagram for a given frequency can be plotted by changing ψ. [0016] This conventional treatment method has been the subject of much research since 1946. In the technical journal entitled "I. R. E. Measures against Waves and Electrons", Vol. 34, No. 6, June, 1946, pp. 335-348. The Ndorf method is known. In this method, the sensors are equidistantly spaced and their sensitivity is such as to obtain a response with a predetermined level of main lobes and a substantially equal number of lower levels of multiple secondary lobes. Set according to the coefficients of the polynomial. Because only a portion of the sensor sensitivity is used, the array produces a response with a signal / noise ratio that is less than the ratio when the full 05-05-2019 4 sensitivity of each sensor is used. Moreover, the performance of the antenna is degraded if the distance between the sensors is too large or too small compared to the length of the wavelength. [0017] Most recently, the document FR-A-2472326 describes how to optimize the geometry of a linear acoustic antenna with a conventional summation of sensor signals. It can be considered to be for a delay / sum linear antenna with variable spacing. This antenna only works well near frequencies in narrow bands, and the antenna is relatively large in wavelength. [0018] Furthermore, very recently, the document FR-A-2722637 describes an antenna geometry in which the sensors are distributed in the horizontal plane of the concave line towards the loudspeaker. The signals from the sensors are summed in phase fashion. The antennas are each characterized by a specific spacing between the sensors and are each divided into sub-antennas assigned to a portion of the frequency band. At low frequencies we still encounter difficulties. [0019] This type of conventional processing has been studied by other researchers who chose different weighting factors to alter the aperture of the main lobe and the level of the secondary lobe of the directivity diagram. It should be noted that in these processes, the sensor's directivity diagram is not used. [0020] If the antenna is to receive a wide band acoustic signal, ie a signal containing frequencies as low as 20 Hz, this conventional process necessarily requires a large number of sensors in the antenna and the dimensions of the antenna We encounter two difficulties, getting bigger. Thus, conventional processing results in expensive and bulky results. 05-05-2019 5 [0021] As a variant, so-called "superdirective" antenna processing is proposed, in which the directivity factor is optimized. In this subject, Y. T. Low and S. W. A book entitled "Antenant Handbook", edited by Lee in 1993, Volume 2 II, Chapter 11 entitled "Array Theory", particularly pages 11-61 to 11-79 of this Chapter 11 You can refer to According to the research content of the present application described above, the maximization of the directivity factor (relation 5) for far-field sources (α equals all 1) is expressed starting from relations 4 and 5 according to the following equations ing. Set a conversion function equal to the unit in the direction of the input signal represented by [0022] By this processing, the distance between the sensors can be reduced to be shorter than the wavelength. Therefore, good spatial selectivity can be obtained with a small size antenna. The disadvantage of this superdirective antenna is that it is less robust, that is, there is a rapid degradation in performance if the optimization is not complete or the optimum conditions for use are deviated; amplification of incoherent noise; information is information It is in the performance fall when not coming from the vertical direction. [0023] Among the recent achievements associated with vertical acoustic antennas, the IEEE Journal on Acoustic Speech and Signal Processing, vol. The article entitled "Practical supergain" described in Cox et al., Pp. 393-398 can be cited. This superdirective antenna is still optimized for the far field since no coefficients are used. Moreover, there are no linear constraints, and the directivity of the sensor is still not considered. Weighting is conditioned only on a selected number on gain for uncorrelated white noise. [0024] Attempts have been made again to improve performance by using fitness algorithms that allow field approximations to be followed to follow the changes. The result is three conditions: (1) the number of sources must be small compared to the number of sensors; (2) ambient noise has 05-05-2019 6 more energy than the indirect path of the source And (3) it is sufficient if it is satisfied that the fluctuations in the field are not too early. If the first condition is not met, it is difficult to analyze the place due to the ambiguity. The second condition needs to be such as not to perturb the perturbation signal which is minimized in the incoming signal. The third condition is necessary to allow the algorithm to follow with an application step small enough to avoid unstable operation. [0025] Starting with all these basic processes such as conventional and hyper-directional processing and processing with fitness algorithms, the development of lobe formation by delay / weighting / summed near-field focusing is sought, all starting in the far field It was being done. Instead of equalizing the delay for the direction, the delay for the near field point is equalized. However, while the known processing method presented above is well understood as the directivity diagram can be represented by a weighted Fourier transform, satisfactory results for near-field focusing have hardly been published. [0026] "The Measures of the IEEE ICASSP", 1997, pp. 363-366. G. Lian and R. A. In the article entitled "Near-field beamforming for microphone arrays" by Golan, the 1 / R term is taken into account for attenuation, so the coefficients of the signal are used. A linear, uniformly spaced conventional antenna geometry is again used. However, the sensor's directivity diagram is not integrated. Moreover, as will become apparent, functions dependent on the signal to be processed are optimized and additional linear constraints are not integrated. [0027] In fact, on the one hand, the speech signal to be processed belongs to a wide band frequency spectrum occupying a large number of octaves, for example 100 to 8000 Hz, and on the other hand the hypothesis of propagation of speech waves by plane waves is not proved The processing method described so far due to the presence of the field sound source does not solve some of the drawbacks. In particular, small conventional antennas can not be made selective at low frequencies. 05-05-2019 7 [0028] SUMMARY OF THE INVENTION One object of the present invention begins with the processing of superdirectivity classes in which the coefficients are processed so as not to introduce distortion of the incoming signal coming from a near-field acoustic source. It makes it possible to improve existing conventional processing and consists in providing antenna processing that meets a number of constraints. [0029] Another object of the present invention is a prior art antenna consisting of a plurality of acoustic sensors, the output signals of which are processed and the output signal of the process being close to the input acoustic source in terms of quality. It consists of providing an antenna that is superior to the output signal by. [0030] Another object of the invention is to provide an antenna, the processing of which is to provide good selectivity at low frequencies. [0031] Another object of the invention consists of providing an antenna with a high directivity factor, a low distortion input signal, and a high incoherent noise reduction. [0032] SUMMARY OF THE INVENTION According to one aspect of the present invention, there is provided an antenna formed by a plurality of acoustic sensors, wherein the sensor output signal has fixed coefficient constraints and incoherent noise rejection. The basic equations of these constraints are as follows, where the first transformation is that the overall transformation function is a pure delay τ The second constraint specifies that the limit is fixed for incoherent noise reduction. [0033] According to another feature, the processing of the antenna also indicates other constraints representing the presence of one or more zeros in the directivity diagram, for example in one or more predetermined directions, ie here Where C (f) is a matrix of propagation vectors and p (f) is a composite gain vector for each propagation vector. 05-05-2019 8 [0034] According to another feature, said processing is realized by a so-called superdirectivity / coefficient / phase mathematical operator or SDMP flow chart, the input data of antenna geometry and propagation model data, weighting data and the aforementioned It is data related to constraints, and its output data is the coefficients of a number of digital filters as well as acoustic sensors in the frequency domain. [0035] According to another feature, an antenna formed by a plurality of acoustic sensors is provided, the first portion of which is placed against a near wanted source and is aligned with the first row And a second portion of the first row behind the first source associated with the near-field source comprises the sensors aligned in at least the second row. [0036] According to another feature, the common direction of the rows of sensors in the first part and in the second part is transverse to the middle direction of the incoming acoustic wave. [0037] According to another feature, the common direction of the rows of sensors in the first part and the second part is slightly oblique to the middle direction of the incoming acoustic wave. [0038] According to another feature, the sensors of the first part are symmetrically distributed in a logarithmic manner around the intermediate sensor. [0039] According to another feature, the sensors of the first part are selectively assigned to a number of sub-antennas, each sub-antenna being associated with a predetermined frequency band, and the sensors being processed in a conventional process Selectively assigned to this sub-antenna transmitting the signal, this frequency band is continuous and practically not more than 1 kHz in total, each process being configured with specific filtering and each specific The filter's output signal is summed. [0040] According to another feature, in the antenna, all the following processing is performed: SDMP 05-05-2019 9 algorithm for low frequencies, division into frequency bands by logarithmic antenna method and conventional channel formation for frequencies not processed by SDMP algorithm Filters each sensor output signal. [0041] According to another feature, a propagation model is used. [0042] According to another feature, measurement of the propagation vector is used. [0043] Like the other features, the features of the invention described above will become more apparent from reading the following description of exemplary embodiments, which description will be made in connection with the attached drawings. . [0044] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT FIG. 1 shows a set 11 containing digital data related to the topographical layout of the antenna sensor and the required source, a set 12 containing data related to the linear constraints, related to the spatial weighting A set 13 containing certain data, a set 14 containing data relating to selected non-coherent noise reduction constraints, and an SDMP flow chart table receiving input data from set 15 containing data related to sub-antenna definition 10 is shown symbolically. SDMP Flow Chart Table 10 carries the output data to set 16, which is related to the set of M digital filter coefficients in the frequency domain, where M is equal to the number of antenna sensors. [0045] The arrangement of the SDMP flow diagram of the present invention that implements the mathematical operators described above is shown in the appendix at the end of the specification. 05-05-2019 10 The flow diagram is described in the MATLAB language, which is well known to those skilled in the art. [0046] If a set of M filters is provided in the frequency domain, then filtering in the frequency band with multiplication can be implemented, or for the purpose of obtaining a set of filters in the time domain, a “generalized least squares” form The transformation performed by a conventional filter design algorithm can be implemented by taking the algorithm of. [0047] In FIG. 2, the antenna is formed by two acoustic sensors or microphones 21, 22 placed behind each other with respect to the speaker or input acoustic source 23. The sensors 21 and 22 and the input acoustic source 23 are aligned. The distance d between the sensors is, for example, 30 cm, which is equal to the distance from the sensor 21 to the input acoustic source 23. Thus, this very simple antenna symbolizes the near-field sound pick up. Moreover, it is still assumed that for the purpose of simplification, the two sensors comprise omnidirectional directivity diagrams. [0048] The outputs of the sensors 21, 22 are respectively connected to the inputs of low pass filters 24, 25 whose outputs are connected at 27 to the input of a summer 26 which produces an antenna output signal. [0049] By the conventional process of "equalization of delays due to propagation, then summing" at very 05-05-2019 11 low frequencies, coherent perturbations from all directions are summed in a phase fashion, which is given by equation (2) above. Quadruple the power. [0050] The incoming signal is also summed in phase fashion, but the amplitude of the signal on sensor 22 is half of the magnitude on sensor 21, which makes the amplification of the incoming signal power equal. Directing factor-make equation (3) above equal. [0051] If subtraction is performed instead of summing, as in conventional processing, this subtraction is [0052] Therefore, if the frequency tends to go to 0, the directivity factor tends to go to indeterminate. On the other hand, since the incoming signal is weak at this output, its processing is susceptible. The amplification of the signal amplifies the unequal numbers in the two sensors 21 and 22, ie the power added incoherent noise 12 + 12 = 2, which is incoherent compared to the incoming signal It means amplification of noise. [0053] This amplification remains small when compared to the indefinite directivity factor. It has been shown that the process of the invention can be found to compensate between the directivity factor and the amplification of the incoherent noise. [0054] 05-05-2019 12 The three processes according to the invention were investigated in different hypothetical situations. -In the case of hypothesis (a), there is no constraint on amplification of incoherent noise,-in the case of hypothesis (b), amplification of incoherent noise between 0 and 5 dB is accepted, andhypothesis (c In the case of), non-interference noise reduction equal to conventional solutions is asked, ie [0055] Under assumption (a), low pass filters 24 and 25 are used, for which a diagram of the coefficients as a function of frequency is shown in FIG. 3 respectively. It can be seen that for f = 0, the amplitudes of the two coefficients are equal, which is outside the equation given above. Above 400 Hz, this amplitude decreases from -4 dB to be substantially -12 dB for filter 24 and 18 dB for filter 25. [0056] Furthermore, under assumption (a), the diagram of the phase difference as a function of frequency in FIG. It shows that it is antiphase for = 0, but in fact has the same value over 400 Hz. [0057] The schematic diagram of FIG. 6 shows an exemplary embodiment of the process-filtering and summing process at the output of the sensors 21, 22 in the time domain. The outputs of the sensors 21, 22 are respectively connected to the inputs of the microphone amplifiers 28, 29, the outputs of which are respectively connected to the inputs of the analog to digital converters 30, 31 and the outputs of the converters are each shifted, for example 32 cells. It is connected to the inputs of the memories 32, 33 configured by the register. 05-05-2019 13 The lateral output of the cells of memory 30 associated with sensor 24 is gate 34.1. n are connected to one input, whose second input is the coefficient signal h. 1. Receive n. The lateral output of the cells of memory 31 associated with sensor 25 is gate 34.2. n are connected to one input, whose second input is the coefficient signal h. 2. Receive n. The parameter n mentioned above varies individually from 1 to 32 according to the rank of the cells in the shift register. Gate 34.1. n and 34.2. The outputs of n are connected to the corresponding inputs of the digital summer 26, whose output carries the antenna signal at 17. [0058] In FIG. 5, the variation in directivity factor as a function of frequency in Assumption (a) is shown by curve 1a, which decreases from 25 dB to 5 dB below 100 Hz, and the low frequency performance is conventional as shown by curve 1d. It shows that it improves compared with the case of an antenna. Curve 2a shows the variation in the decrease. [0059] Still in FIG. 5, under the assumption (b) that amplification of incoherent noise between 0 and 5 dB is accepted, curve 1b shows that the low frequency performance is improved to 5 dB, ie It shows that conventional solutions do not work well. Curve 2b corresponds to the set minimum decrease variation. 05-05-2019 14 [0060] Finally, under the assumption (c) that incoherent noise reduction equal to the conventional solution is taken, curve 1c shows that between 2 dB for low frequencies and 0.6 dB for high frequencies can be obtained . The straight line 2c, which is equal to the straight line 2d, corresponds to the set maximum reduction fluctuation. [0061] Under these three assumptions, the greater the incoherent noise reduction, the lower the directivity of the antenna, and the algorithm of the invention compares the conventional solutions 1d and 1d as compared to the curves 1c and 1d. It can be noted that showing better results than 2d and that the directivity factor can be higher for low frequencies. [0062] Thus, one can choose the compensation between incoherent noise reduction and directivity factor. [0063] FIG. 7 shows 13 sensors 101, which in the embodiment described above opposite to the source 100, are sensors that have a heart-shaped directional view forward, ie towards the area containing the source 100 with respect to the antenna. Through 113 are schematically represented. The first nine sensors 101 to 109 are symmetrically aligned around the sensor 105 on the first straight line D1, and the next two sensors 110 and 111 are disposed on the second straight line D2, The two sensors 112 and 113 are disposed on the third straight line D3. The straight lines D1, D2 and D3 are parallel and perpendicular to the straight line D4 through which the sensor 105 is mounted and to which the source 100 is mounted. 05-05-2019 15 As an example, the distance from the source 100 to the straight line D1 is 60 cm, and the straight lines D2 and D3 are placed behind the straight line D1 at 15 cm and 30 cm, respectively. Sensors 111 and 112 are aligned behind sensor 101, and sensors 111 and 113 are aligned behind sensor 109 to form U legs. [0064] On the straight line D1, the interval between the sensors 105, 104, 103, 102 and 101 increases and fluctuates in a logarithmic manner and symmetrically as the interval between the sensors 105, 106, 107, 108 and 109. [0065] The interval between 105 and 104 is 2.5 cm; the interval between 104 and 103 is 2.5 cm; the interval between 103 and 102 is 5 cm; the interval between 102 and 101 is 10 cm It is. The sensor 110 is placed 15 cm behind the sensor 101, and similarly, the sensor 111 is placed behind the sensor 109, the sensor 112 is placed 15 cm behind the sensor 110, and similarly the sensor 113 is of the sensor 112 It is installed behind. [0066] The schematic diagram of FIG. 8 shows that the filtering of the output signals of the sensors 101 to 113 of FIG. 7 is carried out frequently. The sensor 101 is an analog to digital converter following a circuit C01 operating according to a fast Fourier transform algorithm (RFT with zero padding) whose output is connected to the serial input of the filter D01 connected to the corresponding input of the adder SOM. The sensor 101 supplies an amplifier A01 following B01. The parallel input of filter D01 receives the set of coefficients computed in the SDMP flow chart for this filter. [0067] 05-05-2019 16 FIG. 8 shows a sensor feeding an amplifier A13 following an analog to digital converter B13 following a circuit C13 operating in the same manner as the circuit C01 whose output is connected to the series input of the filter D13 connected to the input to the adder SOM 113 is shown. The parallel input of filter D13 also receives the set of coefficients computed in the SDMP flow chart. [0068] The output of the adder SOM is connected to a circuit E operating according to an inverse rapid Fourier transform algorithm (IRFT with overlap addition) followed by a digital to analog converter F carrying the antenna output signal. [0069] In fact, this algorithm can be implemented in real time using DSP (C50 from Texas Instruments). [0070] In fact, for processing, the antenna of FIG. 7 is divided into four sub-antennas, and the first three sub-antennas are three sub-sensors in which the sensors 101 to 109 of the straight line D1 play a role. The fourth sub-antenna is used for the purpose of covering high frequency octaves, and the sensors 101 to 113 play all roles, for the purpose of covering the low frequency of 0 to 1 kHz. [0071] As mentioned above, on the straight line D1, the sensors 101 to 109 are distributed symmetrically in a logarithmic manner, which can reduce the number of sensors, in this case 9 in this example, as known per se. It becomes. The figure of 5 sensors per octave band has proven to be sufficient. The sensors 103 to 107 constituting the first sub-antenna are used for the 4 to 7 kHz band; the sensors 102, 103, 105, 107 and 108 constituting the second sub-antenna are for the band 2 to 4 kHz The sensors 101, 102, 105, 108 and 109 which are used; constituting the third sub-antenna 05-05-2019 17 are used for the band 1 to 2 kHz. [0072] In the fourth sub-antenna, its processing uses the algorithm of the present invention, ie, in a manner similar to the processing described above for the antenna of FIG. All sensors 101 to 113 are included to take account of the differences. [0073] Thus, the process according to the invention is useful for a wide band of frequencies, such as, for example, speech, bands ranging from 20 Hz to 7 kHz. [0074] In FIG. 9, the variant of the antenna of FIG. 6 comprises thirteen sensors 101 to 113 with a heartshaped directivity diagram as opposed to the input source 200. The first nine sensors 201 to 209 are symmetrically aligned around the sensor 205 on the first straight line D1, the next two sensors 210 and 211 are arranged on the second straight line D2, and the last two The two sensors 212 and 213 are aligned on the third straight line D3. The straight lines D1 to D3 are parallel and perpendicular to the straight line D4 passing through the sensor 205 and the source 200. In the example shown, the mutual distances between the straight lines D1 to D3 and the source 200 are similar to the distances mentioned at the beginning for the antenna of FIG. [0075] On the straight line D1, the mutual distance between the sensors 201 to 209 is the same as the distance between the sensors 101 to 109. [0076] 05-05-2019 18 Sensors 210, 212 are aligned behind the middle of segments 201-202, and sensors 211 and 213 are aligned behind the middle of segments 208-209. In the depth direction, the mutual distance is the same as in FIG. The relative displacement of the sensors 210-213 towards the center of the antenna is indicated with the Pi antenna. [0077] The output signal of the Pi antenna is processed according to the superdirectivity / coefficient / phase flow diagram of the present invention. [0078] In FIG. 10, the other variable of the antenna of FIG. 6 comprises thirteen sensors 301 to 313 with heart-shaped directivity diagrams against the source of input 300. The first nine sensors 301 to 309 are arranged on the straight line D1 in the same manner as the first nine sensors of FIG. [0079] The last four sensors 309-313 are continuously aligned along the same straight line D 4 of FIG. 6 behind 305 to form a T antenna with the sensors 301-309. The distance between the sensors 310-305 is equal to 10 cm as well as between the sensors 311 and 310, 312 and 311 and 313 and 312. [0080] The output signal of the T antenna is processed in accordance with the superdirectivity / coefficient / phase diagram of the present invention. [0081] Instead of providing a linear structure to the U antenna, Pi antenna or T antenna described above 05-05-2019 19 in connection with FIG. 7, FIG. 8 or FIG. 9, as a variant, these antennas may be provided with an oblique structure That is, straight lines D1, D2, D3 are no longer at right angles to straight line D4, but rather at an angle, and the position of the input source is still aligned with straight line D4. [0082] FIG. 1 represents a set 11 containing digital data related to the aerial layout of the antenna and the source sensor. This set 11 also includes data representing the propagation model and / or the measurement of the pulse response as described above. [0083] The following appendix shows an SDMP flow diagram written in the MATLAB language as described above. [0084] %%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%% This file contains two parts:% SDMP part contains % Linear problems for speakers and interference units% Problems (antenna, speaker position, interference unit position)% Non-linear suppression for incoherent noise reduction% Algorithm make G called at the end of the SDMP section% Conventional The antenna part is a delay / weighting / total lobe formation algorithm %%%%%%%%%%%%%%%%%%%%% %%% SDMP antenna part %%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Microphone position, orientation and cardio factor Including am = 1: 13;% Using sensor M = length (am); FocusingPoint = [0.60]; Speaker position → pure delay stop number InterferenceUnitPoint = [10 10 0];% 0 Interfacial unit position → 0% in the required drawing 0 %%%%%%%%%%%%%%% Propagation PropagationModel = 'PropModel';% This The function must be called to obtain delay and attenuation [focdlay focatt] = eval ([PropagationModel '(GeometryFile, am, FocusingPoint, 1)']);% focdlay = focdlay-min (focdlay) for speakers ;% Additive fixed delay removed NormalizationFactor = max (focatt);% Standardized attenuation focatt = focatt / NormalizationFactor; [iudlay iuatt] = eval (PropagationModel '(GeometryFile, am, InterferenceUnitPoint, 0)')); Same as above iuatt = iuatt / NormalizationFactor; iudlay = iudlay-min (iudlay); %%%% Frequency at which the filter is 05-05-2019 20 calculated by the SDMP algorithm FrequencyVector = [0:25:90]; NoOfFreuencies = length (FrequencyVector): SamplingFrequency = 16000; SubAntenna = repmat (am, NoOfFrequencys, 1); %%%%% (of the frequency %) Sdmp → conventional antenna transition IncoherentNoiseRedution = [-2 x ones (1, TransitionFrequency) linspace (-2, 5, No Of Frequencies-TransitionFrequency) %]%;%%%%% Constraint numbers for speakers and interference units constraintMatrixPrefix = 'Cm';% Cm1, Cm2,. . (For all frequencies in the frequency vector) ConstraintVectorPrefix = 'Cv';% Cv1, Cv2, ... for f = FrequencyVector fc = fc + 1; Constraint 1 = (focatt (am). × exp (2i × pi × f ×) focdlay (am));% Conjugation of propagation vector Constraint 2 = (iuatt (am). × exp (2i × pi × f × iudlay (am)); 'int2str (fc)]); eval ([' global Cv 'int2str (fc)]); eval ([' Cm'int2str (fc) '= [Constraint1, Constraint2];']); eval (['Cv' int2 str (fc) '= [1; 0];']); %%%%% Approximate step definition of integration by sum dphi = pi / 25; dtheta = pi / 6; %%%%% of the SDMP algorithm Call G = makeG (GeometryFile, PropagationModel, FrequencyVector, SamplingFrequency, subAntenna, IncoherentNoiseReduction, ConstraintMatrixPrefix, ConstraintVectorPrefix, dphi, dtheta) fr = FrequencyVector;% The frequency of the conventional part is added to this latter frequency vector %%%%%%% Conventional antenna part %%%%%% Conventional use for high frequency Antenna design% Applied to nine microphones forward (3 sub-antennas in 5)% Sub-antenna definition antmic (1, :) = [1 2 5 8 9];% 950-1800 Hz Band antmic (2, :) = [2 3 5 7 8];% 1800-3600 Hz band antmic (3, :) = [3: 7];% 3600-8000 Hz band% definition of sub band limit frequency f min = [950 1800 3600];% lower limit fmax = [1800 3600 8000];% upper limit width = fmax-fmin;% band width% weighting for more or less constant main lobe holes win = [.6; .9; 1; 9; .6] win2 = hamming (5); no_of_pts = 50;% points per band fc = length (fr);% Weighting for 1: ffor already calculated by the super dir algorithm ffor band = 1: 3 bandam = antmic (band, :); [tau0, att0] = PropModel (GeometryFile, am, FocusingPoint, 1); tau0 = tau 0-min (tau0); ctr = 0; forf = fmin (band) + width (band) / no_of_pts: width (band) / no_of_pts: fmax (band) fc = fc + 1; fr (fc) = f; f% more or less Heavily constant main lobe holes weighted smooth = 1-ctr / no_of_pts; b = smooth × win1 + (1-smooth) × win2; b = b / sum (b); cp = b. × exp (2i × pi × f × (tau0 / SamplingFrequency); G (fc, am) = cp. '; Ctr = ctr + 1; end %%%%%%%%%%%%%%%%%%%%%%%%%%% % function G = make G (GeometryFile, PropagationModel, FrequencyVector, SamplingFrequency, SubAntenna, IncoherentNoiseReduction, ConstraintMatrixPrefix, constraintVectorPrefix, dphi, dtheta)% The geometry file contains% of the delay due to the propagation model due to the propagation frequency vector (one, number of frequencies) It is a file that contains the antenna geometry for which attenuation can be calculated. % Includes the frequency at which the filter is computed. The% sub antenna (number of sensors, number of frequencies) indicates which sensor% is used at each frequency. % Non-coherent noise reduction: Minimum required non-coherent noise reduction% Blocking matrix prefix: Prefix to 05-05-2019 21 obtain linear block matrix% Blocking vector prefix: prefix to obtain linear block vector G (number of sensors, number of frequencies) Filter in frequency domain [xm, ym, zm, mictype, xo, yo, zo, mcardio] = readgeo (GeometryFile);% geometry reading M = length (xm);% number of sensors G = zeros (M , length (FrequencyVector)); pr = 0: dphi: (2 × pi-eps);% phi angle (azimuth) vector t = (dtheta / 2): dtheta: (pi-dtheta / 2 + eps); θ angle (elevation angle) vector sr = [logspace (-7, 7, 800)];% Find the parameter for INR%%%%% f = Calculate filter with frequency for frequency vector for f = FrequencyVector f%% Frequency display fc = fc + 1eval (['global Cm'int2str (fc)]);% Stop for this frequency Lix eval (['global Cv' int2str (fc)]);% restraining vector for this frequency [am, Msa] = getam (SubAntenna, fc);% sub antenna r = 1 e4 for this frequency; 10km = far field fac = 2i × pi × f; D = zeros (Msa); %%%%% integration for all directions for theta = trst = sin (theta); for phi = prp = r × [cos (phi) × st sin (phi) × st cos (theta)];% far-field point [dlay (am), att (am)] = eval ([PropagationModel '(GeometryFile, am, p, 0)']); att = att × r d2 = att (am). × exp (-fac × dlay (am)); D−D + d2 ′ × d2 × st; D = D × dphi × dtheta + eps × eye (size (D));% Avoid over-conditions + eps x eyeCm = eval ([ConstraintMatrixPrefix int2str (fc)]); Cv = eval ([ConstraintVectorPrefix int2str (fc)]); %%% Find directional parameters that provide sufficient incoherent noise reduction Loop INR = -Inf; while sc <= length (sr) -1 & INR <IncoherentNoiseReduction (fc) sc = sc + 1; direction = sr (sc); Kic = (D-direction × eye (Msa)) \ Cm b = KiC / (Cm ′ × KiC) × Cv; INR = 10 × log 10 (1 / (b ′ × b)); if sc == length (s r) b = Cm × inv (Cm ′ × Cm) × Cv′warning: Incoherent Noise Reduction impossible 'G (am, fc) = b;% memorize result b for frequency examined in matrix G %%%%%% geometry Shape read %%%%% function [xm, ym, zm, mictype, xo, yo, zo, mcardio] = readgeo (geoname)% function [xm, ym, zm, mictype, xo, yo, zo, mcardio] = readgeo (geoname)% Used to load the antenna geometry stored in geoname:% xm, ym, zm: Sensor position% mictype: Microphone type ('omni', 'cardio', etc)% xo, yo , zo: Microphone orientation% mcardio: in case of cardioid: cardio factor str = ['/ users / cmc / tager / geometries /' geoname];% complete filename fid = fopen (str); if fid <0error (' file not found ')% read microphone format (character string ending with 0) Maxlength = 100; while i <Maxlengthi = i + 1; mictype (i) = fred (fid, 1,' char '); if mictype ( i) == 0 break; endmicrtype = setstr (mictype (1: i-1));% M = fread (fi d, 1, 'short');% Read position xm = fread (fid M, 'float') '; ym = fread (fid, M,' float ')'; zm = fread (fid, M, 'float') %) Read the orientation xo = fread (fid, M, 'float') '; yo = fread (fid, M,' float ')'; zo = fread (fid, M, 'float') ';% mcardio = fread (fid (M, 'float') '; fclose (fid); %%%%% Propagation model %%%%% function [dlay, att] = PropModel (GeometryFile, am, p, read always)% sound wave propagation model% delay = distance / velocity% attenuation = sensor attenuation x distance attenuation global GeometryRead xm ym zm mcardio MO% If still unknown, the geometry is read if ~ exist ('GeometryRead') | always [xm, ym, zm, mictype, xo, yo, zo, mcardio] = readgeo (GeometryFile) MO = [xo; yo; zo]; GeometryRead = 1 tau = []; atten = []; c = 340; (xm);% number of sensors for m = amvec_m_p = p- [xm (m) ym (m) zm (m)];% sound sourcemicrophone m vector dist = norm (vec_m_p);% distance cosangl = vec_m_p × MO (:, m) / (dist x norm (MO (:, m))); dlay (m, 1) = dist / c;% delayed att (m, 1) = (1 + mcardio x cosangl) / (di st × (1 + mcardio));% attenuation dlay = dlay (am); att = att (am); 05-05-2019 22 [0085] Brief description of the drawings [0086] Figure 1 illustrates processing an output signal obtained from an acoustic sensor of the antenna of the invention. [0087] FIG. 2 is a schematic view of a first example antenna according to the present invention. [0088] FIG. 3 shows two coefficient diagrams and two phase difference diagrams for the filter used in the antenna of FIG. 2 respectively. [0089] FIG. 4 shows two coefficient diagrams and two phase difference diagrams for the filter used in the antenna of FIG. 2 respectively. [0090] 5 is a schematic view of a circuit for processing an output signal obtained from the sensor of the antenna of FIG. 2; [0091] Figure 6 schematically represents three response curves as a function of frequency obtained by three different assumptions. [0092] Figure 7 is a schematic view of a second exemplary embodiment of a U antenna according to the invention; [0093] 8 is a schematic view of a circuit for processing an output signal obtained from the sensor of the antenna of FIG. 7; 05-05-2019 23 [0094] FIG. 9 is a schematic view of a third exemplary embodiment of a Pi antenna according to the present invention. [0095] FIG. 10 is a schematic view of a fourth exemplary embodiment of a T antenna according to the invention. [0096] Explanation of sign [0097] DESCRIPTION OF SYMBOLS 10 SDMP flowchart 11 set 12 set 13 set 14 set 15 set 16 set 21 sensor 22 sensor 23 sound source 24 low pass filter 25 low pass filter 26 totalizer 27 antenna output signal 28 microphone amplifier 29 microphone amplifier 30 analog digital conversion 31 Analog-to-digital converter 32 Memory 33 Memory 34 Gate 100 Application source 101 to 113 Sensor 200 Application source 201 to 213 Sensor 301 to 309 Sensor 310 to 313 Sensor A13 Amplifier C01 Circuit C13 Circuit D01 to D13 Straight E Circuit F Digital to analog converter SOM adder 05-05-2019 24

1/--страниц