Determining the direction of origin of sound signals

Abstract

Method for determining the direction of origin of sound signals using a large number of microphones, comprising the following steps: - Providing (S1) a plurality of microphones (11, 12, 13, 14) attached to an object, in particular a vehicle (1), each directed in different directions; - Grouping (S2) the microphones such that the plurality of microphones (11, 12, 13, 14) is divided into at least two microphone groups, wherein the microphones (11, 12, 13, 14) of each microphone group are directed into a directional area; - Detection (S3) of a sound signal by means of microphones of a microphone group and determination of a most probable direction of origin of the sound signal, wherein the most probable direction of origin lies in the direction range of the microphone group; - Repeat (S4) the step of acquiring the sound signal and determining the most probable direction of origin for each microphone group, such that this step has been performed at least once for each microphone group; - Determining (S5) an actual direction of origin from the most probable direction of origin of the microphone groups by determining a direction-related level of the respective sound signal for each most probable direction of origin and comparing the direction-related levels of each most probable direction of origin.

Description

FIELD OF THE INVENTION

[0001] The present invention relates to a method for determining an origin direction of sound signals using a plurality of microphones.

[0002] US 2021 / 0125494 A1 discloses how the direction from which an emergency is coming can be detected. This involves determining a peak value of an audio signal for each microphone pair and comparing this peak value with the respective peak values ​​of the other microphone pairs.

[0003] US 2020 / 0213728 A1 discloses acoustic detection and tracking of an acoustic source. The direction of motion of the source relative to a microphone array is determined using Doppler shift.

[0004] US 2021 / 0233554 A1 discloses a method for detecting and classifying siren sounds and for locating sources of the siren sounds. SUMMARY OF THE INVENTION

[0005] Accordingly, it is provided: - a method for determining a direction of origin of sound signals using a plurality of microphones, comprising the following steps: providing a plurality of microphones that are attached to an object, in particular a vehicle, and are each oriented in different directions; grouping the microphones such that the plurality of microphones is divided into at least two microphone groups, wherein the microphones of each microphone group are oriented in a directional range; detecting a sound signal using microphones of a microphone group and determining a most probable direction of origin of the sound signal, wherein the most probable direction of origin lies in the directional range of the microphone group; repeating the step of detecting the sound signal and determining the most probable direction of origin for each microphone group such that this step is performed at least once for each microphone group;Determining an actual origin direction from the most probable origin direction of the microphone groups by determining a directional level for each most probable origin direction and comparing the directional levels of each most probable origin direction.;

[0006] A microphone is a sound transducer that converts airborne sound, as alternating sound pressure vibrations, into corresponding electrical voltage changes as a microphone signal. A directional microphone primarily records sound arriving from the front and thus has a directional characteristic. Sound from other directions is recorded in an attenuated manner. The recorded sound is converted into electrical signals.

[0007] In this application, a (directional) microphone is oriented in one direction if the microphone captures sound from different directions differently. This can be due, for example, to structural conditions, such as shadows caused by objects, or to the directional characteristics of a directional microphone.

[0008] Directional microphones are typically aimed at a narrow directional range that is smaller than the directional range of a microphone group.

[0009] Microphone groups comprise microphones that are conceptually grouped together. For example, it is conceivable to group microphones spatially into front / back, right / left, top / bottom. Location specifications refer to the relationship between the microphone and the object to which it is attached. The directional range of a microphone group corresponds to the summary of all directional ranges of the microphones grouped in the microphone group.

[0010] The directional level of a sound signal is the level in a specific direction, in this application the most probable direction of origin.

[0011] Beamforming is a method for determining the position of sources in wave fields, such as sound fields. Corresponding devices are also called acoustic cameras, microphone arrays, or acoustic antennas. The measurement principle is based on the microphone array being focused on the various measuring points on the measurement object. This is achieved by a time shift of the sound signals recorded by each microphone, which corresponds to the travel time from the measuring point to that microphone. Sound travel times can be converted into distances by multiplying them by the speed of sound. The time-corrected signals from all microphones are summed, resulting in a time signal assigned to the respective measuring point—i.e., a signal curve over time.The sound from sources at other positions is attenuated because their signals are no longer completely time-corrected and partially overlap destructively; however, the sound emitted from the respective measuring point is amplified.

[0012] In the following, the terms beamformer and beamforming algorithm are used synonymously.

[0013] An SRP-PHAT algorithm can be interpreted as a beamforming-based approach that searches for the candidate position that maximizes the output of a steered beamformer.

[0014] The spectrum of a signal is a function of frequency. The cepstrum (as the spectrum of a function in the frequency domain) is the function of a new variable, the transverse frequency. The dimension of the transverse frequency is identical to the dimension of the independent variable of the original function from which the spectrum was formed. For a time-dependent signal, the transverse frequency therefore has the dimension of time. Precisely formulated, the cepstrum is the inverse Fourier transform of the logarithmized, one-sided auto-power spectrum, rendered dimensionless by division by a reference quantity G0. The special feature of the cepstrum function compared to the similar autocorrelation function is its logarithmization in the frequency domain. This transforms multiplications (e.g., of a spectrum with a frequency response) into additions. This makes harmonic components in the signal clearly visible, even if they have relatively small amplitudes.Cepstrum analysis is therefore particularly suitable for separating the effects of source and transmission properties, for example for detecting echoes and possibly suppressing the echo effects by windowing (as a derivative of “filtering” referred to as “liftering”) of the cepstrum in the area not affected by the echo signal.

[0015] The direction of movement of an object is also called the direction of travel for moving objects.

[0016] The term "speed-dependent filter" refers to the speed of the object. Accordingly, a specific filter can be selected for high object speeds.

[0017] Computer program products typically comprise a sequence of instructions that, when the program is loaded, cause the hardware to perform a specific procedure leading to a specific result.

[0018] The basic idea of ​​the invention is to arrange microphones, each oriented in different directions, on an object. The microphones are conceptually grouped into at least two microphone groups, with the microphones of each microphone group oriented in a directional range.

[0019] The invention provides for determining the direction of origin of a sound signal using this setup by determining a most probable direction of origin for the directional range of a first microphone group. Subsequently, a most probable direction of origin is also determined for the directional range of the second microphone group.

[0020] The invention assumes that the actual direction of origin of the sound signal is either the most probable direction of origin determined for the first microphone group or the most probable direction of origin determined for the second microphone group.

[0021] From these possible candidates of the most probable origin directions for the actual origin direction, the most probable origin direction for which the directional level is highest is selected.

[0022] Although the invention has been explained above and below with reference to two microphone groups, it can also be provided that the microphones on the object are grouped into more than two microphone groups. In this case, more than two most probable directions of origin are also determined, and the actual direction of origin is selected from more than two most probable directions of origin.

[0023] The term "most probable direction of origin" is to be understood in the context of patent claim 1. Accordingly, several most probable directions of origin are determined. If the term "most probable direction of origin" were intended to suggest that there is only one most probable direction of origin, this is incorrect.

[0024] Advantageous embodiments and further developments emerge from the further subclaims and from the description with reference to the figures of the drawing.

[0025] For example, the directional range of the microphone groups can be 180° or 90°. If the directional range covers 180°, the microphones can be grouped, for example, into microphones mounted at the front or rear of the object. Alternatively, grouping them into right or left is also conceivable.

[0026] Alternatively, it is also conceivable for the directional range of the microphone groups to cover 90°. Accordingly, the microphones can be grouped into front right, front left, rear right, and rear left.

[0027] It is advisable to group the microphones in such a way that sound signals captured by microphone groups have similar properties.

[0028] For example, the sound signals captured by microphones mounted on a vehicle are essentially symmetrical with respect to whether they are mounted on the right or left side of the vehicle. In contrast, sound signals captured at the front of a vehicle differ systematically from those captured at the rear of the vehicle due to the vehicle's aerodynamic properties. Accordingly, grouping vehicle microphones into those mounted on the rear and those mounted on the front of the vehicle seems sensible.

[0029] Alternatively or additionally, it may also be possible to group the microphones based on the environment, i.e., in real time. Such environment-dependent grouping of the vehicle microphones could be useful, for example, in strong crosswinds or when driving on a (two-lane) highway, where noise emissions from other vehicles come exclusively from the right or left. In these cases, grouping the microphones into microphones mounted on the front left, front right, rear left, or rear right of the vehicle may be appropriate.

[0030] According to a preferred development of the invention, the most probable direction of origin of the microphone groups is determined by means of an SRP-PHAT algorithm or by means of a beamforming algorithm.

[0031] SRP-PHAT determines the direction of origin of a sound signal based on the time delays of individual signal components at a microphone. Incoming sound signals are normalized according to their level, meaning SRP-PHAT does not take the level into account.

[0032] In contrast, the level of a sound signal is taken into account in a beamformer.

[0033] SRP-PHAT is a particularly robust algorithm for determining the direction of origin. Accordingly, the most probable directions of origin, i.e., candidates for an actual direction of origin, can be determined with high reliability using SRP-PHAT.

[0034] According to a preferred development of the invention, the direction-related levels are determined by means of a beamforming algorithm.

[0035] According to a preferred embodiment of the invention, the sound captured by microphones mounted on a moving object in a direction of motion is filtered using a speed-dependent filter. This allows, for example, wind noise, which occurs predominantly at the front (as seen in the direction of motion of the object) and can distort the level of a sound signal, to be filtered out.

[0036] According to a preferred development of the invention, the detected sound signal is filtered with respect to a predetermined signal.

[0037] For example, the method can be designed to determine the direction of origin of specific siren sounds. Siren sounds are country-specific and can be specified as a known target signal. This allows a detected sound signal to be filtered for a predetermined signal.

[0038] It is also useful to determine a cepstrum for the recorded sound signal. Harmonic components of a signal can be clearly identified in a cepstrum, even if their amplitudes are relatively small. Cepstrum analysis can therefore be used to analyze or differentiate the transmission properties of a sound-transmitting medium from those of a sound signal generated at a sound source.

[0039] According to a preferred development of the invention, the most probable direction of origin of a microphone group is averaged over a time interval.

[0040] It is useful to adjust the length of the time interval with respect to a predetermined signal.

[0041] For example, sirens in Europe show a frequency jump that usually occurs every 0.7 seconds.

[0042] Consequently, in order to determine the direction of origin of a sound signal which has regular frequency jumps after predetermined time intervals, it is advantageous to average the most probable directions of origin over a time interval between two frequency jumps, for example 0.7 seconds.

[0043] It is understood that a vehicle with a microphone system for carrying out a method for determining a direction of origin of a sound signal as described above, which has a plurality of microphones, wherein the microphones are attached to the vehicle and are each oriented in different directions and the microphones are grouped such that the plurality of microphones is divided into at least two microphone groups, wherein the microphones of each microphone group are oriented in a directional range, is advantageous.

[0044] A computer program product according to a method of one embodiment of the invention executes the steps of a method according to the preceding description when the computer program product runs on a computer, in particular an in-vehicle computer. When the program in question is used on a computer, the computer program product produces an effect, namely the determination of an origin direction of sound waves. TABLE OF CONTENTS OF THE DRAWINGS

[0045] The present invention will be explained in more detail below with reference to the exemplary embodiments shown in the schematic figures of the drawings. In the drawings: Fig. 1 is a schematic block diagram of an embodiment of the invention. Fig. 2 a schematic diagram of a vehicle according to an embodiment of the invention.

[0046] The accompanying drawings are intended to provide a further understanding of embodiments of the invention. They illustrate embodiments and, in conjunction with the description, serve to explain principles and concepts of the invention. Other embodiments and many of the noted advantages will be apparent upon review of the drawings. Elements of the drawings are not necessarily shown to scale relative to one another.

[0047] In the figures of the drawings, identical, functionally identical and acting elements, features and components are provided with the same reference symbols, unless otherwise stated. DESCRIPTION OF EMBODIMENTS

[0048] The Fig. 1 shows a schematic block diagram of a method for determining a direction of origin of sound signals using a plurality of microphones. In step S1, a plurality of microphones are provided, which are attached to an object, in particular to a vehicle, and are each oriented in different directions. In step S2, the microphones are grouped such that the plurality of microphones is divided into at least two microphone groups, wherein the microphones of each microphone group are oriented in a directional range. In step S3, a sound signal is detected using the microphones of a microphone group, and a most probable direction of origin of the sound signal is determined, wherein the most probable direction of origin lies in the directional range of the microphone group.

[0049] In step S4, the step of detecting the sound signal and determining the most probable direction of origin is repeated for each microphone group, so that the sound signal from each microphone group has been detected once and the most probable direction of origin of a sound signal for the directional range of each microphone group has been determined.

[0050] In step S5, an actual origin direction is determined from the most probable origin directions of the microphone groups by determining a directional level for each most probable origin direction and comparing the directional levels of each most probable origin direction.

[0051] Fig. 2 shows a vehicle 1 with a microphone system 10. The microphone system 10 comprises 4 microphones 11, 12, 13, 14. The microphones 11 and 12 are mounted at the front of the vehicle 1 and the microphones 13, 14 are mounted at the rear of the vehicle 1. In Fig. 2 schematically illustrates a sound source 15. The sound propagates via sound waves 16 and 17, respectively. The sound waves 17 behind the rear of the vehicle are shadowed by the vehicle and are therefore shown as smaller than the sound waves 16. The sound waves emitted by the sound source 15 propagate from the sound source 15 essentially uniformly, although sound shadows caused by the vehicle, for example, due to rear-view mirrors, also exist at the front and sides of the vehicle.

[0052] If the microphones 11, 12, 13, 14 are divided into two microphone groups, for example front and rear, it is intended to determine a most probable direction of origin of the sound signal using the front microphones 11, 12 and to determine a most probable direction of origin using the rear microphones 13, 14.

[0053] In the following, the directional level of the sound signal at the most likely origin of the front microphone group is compared with the directional level of the rear microphone group. As in Fig. 2, the level at the front microphone group is higher than at the rear microphone group, so the method will determine the most likely origin direction of the front microphone group as the actual origin direction of the sound signal. Reference symbol S1-S5 process steps 1 vehicle. 10 microphone system 11 Microphone 12 Microphone 13 Microphone 14 Microphone 15 Sound source 16 sound wave 17 Sound wave

Claims

[1] Method for determining a direction of origin of sound signals using a plurality of microphones, comprising the following steps: - Providing (S1) a plurality of microphones (11, 12, 13, 14) which are attached to an object, in particular to a vehicle (1), and are each oriented in different directions; - grouping (S2) the microphones such that the plurality of microphones (11, 12, 13, 14) is divided into at least two microphone groups, the microphones (11, 12, 13, 14) of each microphone group being aligned in a directional range; - detecting (S3) a sound signal by means of microphones of a microphone group and determining a most probable direction of origin of the sound signal, wherein the most probable direction of origin lies in the directional range of the microphone group; - repeating (S4) the step of detecting the sound signal and determining the most probable direction of origin for each microphone group such that this step has been performed at least once for each microphone group; - Determining (S5) an actual direction of origin from the most probable direction of origin of the microphone groups by determining a directional level of the respective sound signal for each most probable direction of origin and comparing the directional levels of each most probable direction of origin. [2] Method for determining a direction of origin of sound signals according to claim 1, wherein the directional range of the microphone groups comprises 180° or 90°, wherein the plurality of microphones (11, 12, 13, 14) are grouped in particular in front or rear and / or right or left. [3] Method for determining a direction of origin of sound signals according to one of the preceding claims, wherein the microphones (11, 12, 13, 14) are grouped depending on the environment during a movement of the object. [4] Method for determining a direction of origin of sound signals according to one of the preceding claims, wherein the most probable directions of origin of the microphone groups are determined by means of a PHAT algorithm, in particular by means of an SRP-PHAT or by means of a beamforming algorithm. [5] Method for determining a direction of origin of sound signals according to one of the preceding claims, wherein the direction-related levels are determined by means of a beamforming algorithm. [6] A method for determining a direction of origin of sound signals according to any one of the preceding claims, wherein sound detected by microphones attached to moving objects in a direction of movement is filtered by means of a speed-dependent filter. [7] A method for determining a direction of origin of sound signals according to any one of the preceding claims, wherein the detected sound signal is filtered with respect to a predetermined signal. [8] Method for determining a direction of origin of sound signals according to claim 7, wherein a cepstrum is determined for the detected sound signal. [9] A method for determining a direction of origin of sound signals according to any one of the preceding claims, wherein the most probable direction of origin of a microphone group is averaged over a time interval. [10] A method for determining a direction of origin of sound signals according to claim 9, wherein the length of the time interval is determined with respect to a predetermined signal. [11] Computer program product with program code means for carrying out the method for determining a direction of origin of sound signals according to one of the preceding claims. [12] Vehicle (1) with a microphone system (10) for carrying out a method for determining a direction of origin of sound signals according to one of the preceding claims 1-10, which has a plurality of microphones, wherein the microphones (11, 12, 13, 14) are attached to the vehicle and are each oriented in different directions and the microphones (11, 12, 13, 14) are grouped such that the plurality of microphones (11, 12, 13, 14) are divided into at least two microphone groups, wherein the microphones (11, 12, 13, 14) of each microphone group are oriented in a directional range.

Images