JPH07218614A - Sound source position calculation method and apparatus thereof - Google Patents

Abstract

(57) [Summary] [Purpose] The sound source position can be accurately calculated regardless of the length of time the sound is generated and the sound generation range, and the accuracy of calculation of the sound source position is improved. [Structure] Non-directional first mounted on the three-dimensional coordinate axes
~ The fifth microphones 1 to 5, the arrival time of the sound received by each of the microphones 1 to 5, the sound velocity at this time, the storage unit 31 for storing the coordinate data of each of the microphones 1 to 5, and each of the microphones 1 to 5. Calculation of the sound source position and the sound generation time obtained by randomly selecting four equations from the five sound propagation distance equations related to the sound source position and sound generation time of the received sound twice for each selected group An expression storage unit 32 that stores an expression for each selection group, and a sound source position and a sound generation time are calculated for each selection group based on the calculation expression of the expression storage unit 32, and the values obtained for each selection group are close to each other. The calculation unit 29 determines and specifies the value as the sound source position and the sound generation time.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a sound source position calculation method and apparatus, and particularly to a sound source position calculation suitable for grasping the surrounding situation of an autonomous vehicle based on the calculated sound source position. A method and an apparatus thereof.

[0002]

2. Description of the Related Art Conventionally, when estimating the position of a sound source (place of sound generation), a method of estimating the sound source position by a predetermined method such as the sound intensity method has been adopted. That is, in the conventional method, it is common to estimate that the sound source exists in the direction in which the reception level of the microphone becomes maximum when the microphone having strong directivity is shaken. As this type of technology, for example, the technology described in Japanese Patent Laid-Open No. 4-72525 is proposed.

[0003]

However, in the above-mentioned prior art, when estimating the position of the sound source, a method of estimating that the sound source is in the direction in which the reception level detected by the microphone is maximum is generally used. However, it is extremely difficult to estimate the sound source position when a sound is generated instantaneously or to separate the sound source position when a sound is generated at a plurality of places over a wide range. .

[0004]

It is an object of the present invention to improve the inconveniences of the above-mentioned conventional examples, and in particular, to enable accurate calculation of the sound source position regardless of the length of time the sound is generated and the sound generation range. It is an object of the present invention to provide a sound source position calculation method and apparatus for improving the sound source position calculation accuracy.

[0005]

According to a first aspect of the present invention, at least five omnidirectional microphones are provided in the three-dimensional coordinate space at the origin and at a position separated from the origin by a predetermined distance. The arrival time (t ₀ , t ₁ , t ₂ , t ₃ , t ₄ ) of the received sound is stored, and the sound velocity at this time is V, and the sound source position (x, y, z) and at least five sound propagation distance equations for sound generation time T are provided, and
By arbitrarily selecting four equations from these five sound propagation distance equations and performing simultaneous solution, the sound source position (x, y,
z) and the sound generation time T are calculated. This is intended to achieve the above-mentioned object.

According to the present invention of claim 2, in the three-dimensional coordinate space,
At least five omnidirectional microphones are provided at the origin and at a position separated from the origin by a predetermined distance, and arrival times (t ₀ , t ₁ , t ₂ , t ₃ ,
t ₄ ) is stored, and the sound velocity at this time is V, and the sound source position (x, y,
z) and at least five sound propagation distance equations related to the sound generation time T are provided, and four equations are arbitrarily selected a plurality of times from the five sound propagation distance equations, and the sound source position (x, y, z) and sound generation time T
X is calculated separately for each selected group, x, y,
Among the values of z and T, the close values of the sound source position (x, y,
z) and the sound generation time T. This is intended to achieve the above-mentioned object.

The present invention according to claim 3 provides the microphone according to claim 1 or 2, wherein the microphones are mounted on a three-dimensional coordinate axis and at a position at an equal distance L from the origin.

According to a fourth aspect of the present invention, in the three-dimensional coordinate space,
At least five omnidirectional microphones are provided at the origin and at a position separated from the origin by a predetermined distance, and arrival times (t ₀ , t ₁ , t ₂ ,
t ₃ , t ₄ ), the sound velocity V at this time, and the coordinate position data of each of the microphones are provided, and the sound source position (x, y, z) of the sound received by each of the microphones and At least five sound propagation distance equations related to the sound generation time T are set, and four equations are arbitrarily selected from these five sound propagation distance equations to obtain a sound source position (x, y, z) and a formula storage unit for storing a calculation formula of the sound generation time T, and based on the calculation formula stored in the formula storage unit, the sound source position (x, y,
z) and a sound generation time T are provided. This is intended to achieve the above-mentioned object.

According to a fifth aspect of the present invention, in the three-dimensional coordinate space,
At least five omnidirectional microphones are provided at the origin and at a position separated from the origin by a predetermined distance, and arrival times (t ₀ , t ₁ , t ₂ ,
t ₃ , t ₄ ), the sound velocity V at this time, and the coordinate position data of each of the microphones are provided, and the sound source position (x, y, z) of the sound received by each of the microphones and It is obtained by setting at least five sound propagation distance equations related to the sound generation time T, arbitrarily selecting four equations from the five sound propagation distance equations a plurality of times, and performing simultaneous solution for each selected group. Sound source position (x,
y, z) and a formula storage unit for storing the calculation formula of the sound generation time T for each selected group, and based on the calculation formula stored in the formula storage unit, the sound source position (x, y, z) and A calculation unit for calculating the sound generation time T for each selected group is provided, and this calculation unit determines a close value among the values of x, y, z, and T obtained for each selected group as the sound source position (x , Y, z) and sound generation time T, and has a sound source position specifying function for specifying and specifying. This is intended to achieve the above-mentioned object.

According to a sixth aspect of the present invention, in the fourth or fifth aspect, each of the microphones is mounted on the three-dimensional coordinate axis and at a position at an equal distance L from the origin.

[0011]

According to the present invention of claim 1, the arrival times of the sounds received by the origin and at least five omnidirectional microphones provided at positions separated by a predetermined distance from the origin in the three-dimensional coordinate space are determined. At least five sound propagation distance equations related to the sound source position and sound generation time that are stored and stored with respect to the origin as a reference for each microphone are provided, and four equations are arbitrarily selected from the five sound propagation distance equations and are simultaneous. Since the sound source position and sound generation time are calculated by the solution, the length of time the sound is generated is compared to the conventional case where the sound source is estimated to be in the direction in which the reception level of the microphone is maximum. It is possible to accurately calculate the sound source position regardless of the sound generation range.

According to the present invention of claim 2, further, four equations are arbitrarily selected a plurality of times from the above-mentioned five sound propagation distance equations, and the sound source position and the sound generation time are separately set for each selected group. In order to calculate the sound source position and sound generation time, the close values among the values obtained for each selection group are calculated.
It is possible to add objectivity to the obtained sound source position,
As a result, it is possible to improve the calculation accuracy of the sound source position.

According to the present invention of claim 3, since the microphones of claim 1 or 2 are mounted on the three-dimensional coordinate axis and at the origin and at positions equidistant L from the origin, the sound source position and the sound generation are generated. The calculation when calculating the time can be simplified, and the sound source position and the sound generation time can be calculated quickly. Therefore, when the present invention is applied to, for example, an autonomous vehicle, a sound generated around the vehicle (for example, engine sound of another vehicle / footsteps of humans / sound caused by living noise such as crying of animals, sound of river flow / Since the situation around the vehicle can be accurately grasped based on the position of the sound source (such as a sound caused by a natural phenomenon such as thunder), it is possible to properly operate the autonomous vehicle.

According to the present invention of claim 4, at least five omnidirectional microphones provided at the origin and a position separated by a predetermined distance from the origin in the three-dimensional coordinate space, and the sound received by each microphone. Arrival time, sound velocity at this time,
A storage unit that stores coordinate position data of each microphone,
Calculation formula of sound source position and sound generation time obtained by arbitrarily selecting four equations from at least five sound propagation distance equations related to sound source position and sound generation time of sound received by each microphone and performing simultaneous solution Since a formula storage unit that stores the sound source and a calculation unit that calculates the sound source position and the sound generation time based on the calculation formula stored in the formula storage unit are provided, the length of time the sound is generated and the sound The sound source position can be accurately calculated regardless of the generation range.

According to the present invention of claim 5, the sound source position obtained by further selecting four equations arbitrarily from the above-mentioned five sound propagation distance equations a plurality of times and performing simultaneous solution for each selected group. And a formula storage unit that stores the calculation formula of the sound generation time for each selection group, and a calculation unit that calculates the sound source position and the sound generation time for each selection group based on the calculation formula stored in the formula storage unit. , The calculation unit has a sound source position specifying function that determines and determines the close value among the values obtained for each selected group as the sound source position and the sound generation time, so that the specified sound source position has objectivity. As a result, it is possible to improve the accuracy of calculation of the sound source position.

According to the present invention of claim 6, since the microphones of claim 4 or 5 are mounted on the three-dimensional coordinate axis and at the origin and at positions equidistant from the origin, the sound source position and the sound generation. The calculation when calculating the time can be simplified, and the sound source position and the sound generation time can be calculated quickly. Therefore, when the present invention is applied to, for example, an autonomous vehicle, a sound generated around the vehicle (for example, engine sound of another vehicle / footsteps of humans / sound caused by living noise such as crying of animals, sound of river flow / Since the situation around the vehicle can be accurately grasped based on the position of the sound source (such as a sound caused by a natural phenomenon such as thunder), it is possible to properly operate the autonomous vehicle.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment in which the sound source position calculating method and apparatus of the present invention are applied to an autonomous vehicle will be described below with reference to the drawings.

First, the structure of the autonomous traveling vehicle according to the present embodiment will be explained with reference to FIG. 2. For example, five omnidirectional first to fifth non-directional vehicles are provided above the autonomous traveling vehicle 11 which is traveling by remote control by radio. 5th microphone 1, 2, 3,
A sound collecting mechanism 6 composed of 4, 5 and the like is mounted via a platform 7 and a platform driving motor 8 and an outside air temperature sensor 9 for detecting the outside air temperature. On the other hand, inside the autonomous traveling vehicle 11, five reception amplifiers 17 to be described later, which are respectively provided corresponding to the first to fifth microphones 1 to 5, are provided.
A control box 10 accommodating 18 and 5 low-pass filters 22 to 26, a multiplexer 27, an A / D converter 28, and a computer 33 is provided.

The sound collecting mechanism 6 includes a first microphone 1 and a second microphone 2.
Microphone 2, third microphone 3, fourth microphone 4, fifth microphone 5 and connecting shafts 12, 13, 14, 15 for connecting first microphone 1 and second to fifth microphones 2, 3, 4, 5 respectively. It is configured to include and. Further, the sound collecting mechanism 6 is supported by the platform 7 via a biaxial support shaft 16 connected to the base ends of the connecting shafts 12 to 15. The platform 7 is driven by a platform drive motor 8 which is operated based on a drive control signal from the control device 9 to change the attitude of the sound collecting mechanism 6.

Inside the connecting shafts 12 to 15 connecting the first microphone 1 and the second to fifth microphones 2 to 5, respectively, and the support shaft 16 connecting the first microphone 1 to the platform 7, A signal line for transmitting an audio signal received by the fifth microphones 1 to 5 to receiving amplifiers 17 to 21 described later is incorporated.

As shown in FIG. 3, the first to fifth microphones 1 to 5 are arranged in such a manner that the first microphone 1 (coordinates (0,0,0)) is arranged at the origin of the three-dimensional coordinate axes, and the first microphone is arranged. The second microphone 2 (coordinates (L, 0, 0)) is arranged at a position spaced from the first microphone 1 in the X-axis direction, and the third microphone 2 is arranged at a position spaced from the first microphone 1 in the Y-axis direction at the third position.
The microphone 3 (coordinates (0, L, 0)) is placed, and the first
A fourth microphone 4 (coordinates (0, 0, L)) is arranged at a position spaced from the microphone 1 in the Z-axis positive direction by a distance L,
The fifth microphone 5 (coordinates (0, 0, -L)) is arranged at a position spaced from the first microphone 1 in the negative Z-axis direction by a distance L. In this case, in the present embodiment, the first to fifth microphones 1 to 5 are mounted on the three-dimensional coordinate axes to simplify the calculation when calculating the sound source position and the sound arrival time.

Next, the configuration of the sound source position calculation control system in this embodiment will be described with reference to FIG. 1. The sound source position calculation control system includes a first microphone 1, a second microphone 2, and
The third microphone 3, the fourth microphone 4, the fifth microphone 5, and the reception amplifiers 17, 18, 19, 20,
21, a low-pass filter (low-pass analog filter) 22, 23, 24, 25, 26, a multiplexer 27, an A / D converter 28, a calculation unit 29, a posture control unit 30, a storage unit 31, and an expression storage unit 32. And a computer 33 having

More specifically, the first to fifth microphones 1 to 5 supply respective audio signals based on the sound received from the outside of the autonomous vehicle 11 to the reception amplifiers 17 to 21, and the reception amplifiers 17 to 21 respectively. , Each audio signal is amplified and supplied to the low-pass filters 22-26,
26 is a multiplexer 2 for converting the low-frequency component signal of each audio signal.
It is designed to supply to 7.

The multiplexer 27 sequentially selects each output signal of the low-pass filters 22 to 26 and supplies a signal corresponding to each output signal to the A / D converter 28, so that A
The / D converter 28 converts the output signal of the multiplexer 27 from analog to digital and supplies it to the arithmetic unit 29 of the computer 33. Further, the outside air temperature sensor 9 detects the outside air temperature and supplies an outside air temperature signal to the arithmetic unit 29 of the computer 33.

The calculation unit 29 of the computer 33 uses the calculation formula stored in the formula storage unit 32 to calculate the sound source position (x,
y, z) and a sound generation time T for each selected group, which will be described later, a sound source position (x, y, z) of the sound received by the first to fifth microphones 1 to 5, and a sound generation time. From the five sound propagation distance equations (described later) concerning T, four equations are arbitrarily selected a plurality of times (for example, twice in this embodiment), and each value of x, y, z, and T obtained for each selected group. A sound source position specifying function for determining and specifying a close value of the sound source position (x, y, z) and the sound generation time T is provided. Further, the calculation unit 29 has a sound source recognition function for recognizing what the sound source is based on the characteristics of the sound, and the outside air temperature sensor 9
The sound velocity correction function that corrects the sound velocity based on the outside air temperature detected by and the posture control function that outputs a posture control command signal that changes the posture of the sound collecting mechanism 6 to the posture control unit 30 are provided. In this case, the speed of sound at 0 degrees C in air is about 3
Since it is 31 [m / sec], the sound velocity when the outside air temperature is α degrees C is 331 + 0.6 × α [m / sec] (the temperature coefficient of the sound velocity at 0.6: 0 degrees C).

The attitude control unit 30 drives and controls the pan head driving motor 8 based on the attitude control command signal output from the arithmetic unit 29, thereby rotating the pan head by a predetermined angle and by vertically moving the support shaft 16 in a predetermined direction. Angle operation, first to fifth
The posture of the sound collecting mechanism 6 including the microphones 1 to 5 is controlled.

The storage unit 31 includes the first to fifth microphones 1.
Arrival times (t ₀ , t ₁ , t ₂ , t ₃ ,
t ₄ ), the sound velocity V at this time, and the coordinate position data of the first to fifth microphones 1 to 5 are stored. Further, the storage unit 31 is corrected based on the characteristic pattern of the waveforms received by the first to fifth microphones 1 to 5 extracted by the calculation unit 29, the outside air temperature detected by the outside air temperature sensor 9, and the outside air temperature. It stores various data such as the speed of sound.

The expression storage unit 32 has five sound propagation distance equations (described later) for sound source positions (x, y, z) of sounds received by the first to fifth microphones 1 to 5 and sound generation time T.
Calculation of the sound source position (x, y, z) and the sound generation time T obtained by arbitrarily selecting four equations from the above multiple times (for example, twice in this embodiment) and performing simultaneous solution for each selected group. The formula is stored for each selected group.

Next, the first method for calculating the arrival time of the sound generated at a certain position in the first to fifth microphones 1 to 5 in this embodiment will be described. FIG. 4 shows received waves W1 to W5 when a certain sound generated outside the autonomous vehicle 11 is received by the first to fifth microphones 1 to 5, and the calculation unit 29 of the computer 33 is shown in FIG. By wavelet transforming the received wave of each frequency (for example, 50 [Hz], 100 [Hz], 200 [Hz], 40
A time-amplitude graph of 0 [Hz], 800 [Hz]) is created. FIG. 5 shows a time-amplitude graph relating to the received wave W1 of the first microphone 1, and FIG.
3 shows a time-amplitude graph related to a received wave W2 of the microphone 2.

The arithmetic unit 29 is adapted to store in the storage unit 31 the sound pattern of each frequency and the arrival time (reception time) received by the first microphone 1 within a predetermined time (time Ta to time Ta + ΔT). There is. The same applies to the second to fifth microphones 2 to 5. In the example of FIG. 5, it can be seen that the sound of frequency 200 [Hz] is received at time Ta. The arithmetic unit 29 also compares the sound patterns of the respective frequencies received by the first to fifth microphones 1 to 5 stored in the storage unit 31, and arrives at the matching sound pattern of the respective frequencies. Is selected from the storage unit 31.

Further, as shown in FIG. 6, when the sound of frequency 200 [Hz] is received at different times as shown in FIG. 6, the arithmetic unit 29 arrives at the pattern corresponding to the pattern shown in FIG. The time is selected from the storage unit 31. This is at other frequencies (50, 100, 40
The same applies to 0,800 [Hz]). Through the above-described processing, the calculation unit 29 selects from the storage unit 31 the arrival time at which the sound generated at a certain position reaches the first to fifth microphones 1 to 5. In this case, the first microphone 1 and the second to fifth microphones 2 to
Since the difference in the arrival time of the sound is limited according to the interval with 5, the arrival time of the sound may be calculated within a time section as necessary.

Next, the first to fifth sounds generated at a certain position
A second method of calculating the arrival times at the microphones 1 to 5 will be described. In the second method, instead of the low pass filters 22 to 26 of FIG.
38 (see FIG. 7), the bandpass filter 34-
The first zero-cross point of the waveform whose amplitude has reached a predetermined level in the waveform output from 38 is set as the arrival time of the sound generated at a certain position. In this case, the common configuration between FIG. 7 and FIG.
In the example of FIG. 8, when the amplitude of the waveform Wf output from the bandpass filter exceeds a preset threshold value, the first zero-crossing point Pz of the waveform Wf becomes the arrival time of the sound generated at a certain position. It is supposed to do.

Next, the first to fifth sounds generated at a certain position
A third method of calculating the arrival times at the microphones 1 to 5 will be described. In the third method, instead of the low pass filters 22 to 26 of FIG.
38 (see FIG. 7), the bandpass filter 34-
The arrival time (arrival time difference) of the sound generated at a certain position is calculated by measuring the phase difference between the waveforms output from 38. For example, when the frequency of the sound passing through the bandpass filter is 1 [Kz], the wavelength of the sound is 3
Since it is 4 [cm], the first to fifth microphones 1 to 5
The first to fifth microphones 1 to 5 are arranged so that the maximum distance is within 34 [cm], and for example, the positions of the reception waveforms of the first to fifth microphones 1 to 5 with the first microphone 1 as a reference. The arrival time of the sound is calculated by measuring the phase difference. In the example of FIG. 3, the fourth and fifth
Since the distance 2L between the microphones 4 and 5 is the maximum, L
May be 17 [cm] or less.

Next, the first to fifth sounds generated at a certain position
A fourth method for calculating the arrival times at the microphones 1 to 5 will be described. In the fourth method, the arrival time of the sound generated at a certain position is calculated by correlating the reception waveforms of the first to fifth microphones 1 to 5 in a certain section. That is, for example, with reference to the reception waveform of the first microphone 1, the reception waveforms of the second to fifth microphones 2 to 5 are shifted in time in a certain section and the time axis is shifted.
It is determined whether or not the respective received waveform patterns match by superimposing on the received waveforms of No. 2, and when the received waveform patterns match, the second to fifth microphones 2 to 5 with the received waveform of the first microphone 1 as a reference. The arrival time of the sound is calculated based on the time when the received waveform of is shifted.

Next, the method of calculating the sound source position in this embodiment will be described. In this case, the processing of calculating the sound source position by obtaining the solution of the simultaneous equations shown below is performed by the arithmetic unit 29 of the computer 33.

For example, the time when a sound is generated at a position outside the autonomous vehicle 11 is T, the three-dimensional coordinates of the sound source position are (x, y, z), and the speed of sound is corrected based on the detection data of the outside air temperature sensor 9. Is Vs, the intervals between the first microphone 1 and the second to fifth microphones 2 to 5 are L, and the reception waves W1 to W5 by the first to fifth microphones 1 to 5 are the waveforms shown in FIG. The arrival times of the sounds generated at 1 to 5th microphones 1 to 5 are t ₀ , t ₁ , t ₂ , t ₃ , t _{4 respectively.}
If

[0037]

[Equation 1]

The following simultaneous equations (sound propagation distance equations) hold for the first to fifth microphones 1 to 5. If you square the left and right sides of the simultaneous equations,

[0039]

[Equation 2]

The following simultaneous equations (sound propagation distance equations) hold for the first to fifth microphones 1 to 5. Furthermore,
In order to more accurately calculate the solution x, y, z (three-dimensional coordinates (x, y, z) of the sound source position) of the simultaneous equations,

[0041]

[Equation 3]

The simultaneous equations A for the first to fourth microphones 1 to 4

[0043]

[Equation 4]

The simultaneous equations B for the first to third microphones 1 to 3 and the fifth microphone 5 are decomposed. First, when obtaining the solutions x, y, and z of the simultaneous equations A for the first to fourth microphones 1 to 4,

[0045]

[Equation 5]

A _x as _{[0046], b x, a y, b} y, a z, replaced with b _z, further,

[0047]

[Equation 6]

Substituting a _t , b _t , and c _t like
The sound generation time T is

[0049]

[Equation 7]

It can be represented by the following equation. in this case,
The sound generation time T calculated as a negative value in the above equation has its absolute value as a solution. That is, there are two values of the sound generation time T. Therefore, the coordinates (x,
y, z) can be expressed by _{x = a x × T + b} x y = a y × T + b y z = a z × T + b z becomes equation. That is, there are two solutions x, y, z of the simultaneous equations A described above.

Similarly, the solution x, of the simultaneous equations B for the first to third microphones 1 to 3 and the fifth microphone,
When determining y and z,

[0052]

[Equation 8]

A _x as _{[0053], b x, a y, b} y, a z, replaced with b _z, further,

[0054]

[Equation 9]

Substituting a _t , b _t , and c _t like
The sound generation time T is

[0056]

[Equation 10]

It can be calculated by the following equation. In this case, the sound generation time T calculated as a negative value in the above equation
Takes the absolute value as the solution. That is, there are two values of the sound generation time T. Thus, sound source position coordinates (x, y, z) can be expressed by _{x = a x × T + b} x y = a y × T + b y z = a z × T + b z becomes equation. That is, there are two solutions x, y, z of the simultaneous equations B described above.

Since there are two solutions x, y, z of the simultaneous equations A and two solutions x, y, z of the simultaneous equations B, there are two solutions before the decomposition into the simultaneous equations A and B. There are four solutions x, y, z of the simultaneous equations. Therefore, of the four solutions x, y, and z (coordinates (x, y, z) of the sound source position), the one that seems to be close is specified as the sound source position.

In the sound source position calculation procedure described above, there may be a case where the above solution cannot be obtained depending on the error in the arrival time of the sound from the sound source to the first to fifth microphones 1 to 5, but in this case, the sound The solution can be obtained by recalculating the simultaneous equations described above in consideration of the arrival time error.

In this case, in this embodiment, the simultaneous equations A for the first to fourth microphones 1 to 4 and the first to third microphones are used.
The sound source position is calculated based on the simultaneous equations B for the microphones 1 to 3 and the fifth microphone 5.
The combination of the individual microphones is not limited to this.

As described above, according to the present embodiment, the origin and the omnidirectional first to fifth microphones 1 to 5 are provided on the three-dimensional coordinate axis and at positions equidistant from the origin, and each microphone 1 5 to 5 are stored in the arrival time of the received sound, and five sound propagation distance equations relating to the sound source position and the sound generation time are provided for each of the microphones 1 to 5 and the five sound propagation distances are provided. From the equations, four equations are arbitrarily selected twice, the sound source position and the sound generation time are calculated separately for each selection group, and the close values among the values obtained for each selection group are set as the sound source position and the sound source position. Since it is specified as the sound generation time, it is possible to add objectivity to the calculated sound source position compared to the case where it is estimated that there is a sound source in the direction in which the microphone reception level becomes maximum as in the past. Sound source position It is possible to improve the constant accuracy, the irrespective and source position even if the SN ratio is small sound generation range which Occurring Time and sounds of the sound can be calculated accurately.

Further, according to the present embodiment, the first microphone 1 is arranged at the origin on the three-dimensional coordinate axis, and the first microphone 1 is arranged at the positions at equal intervals L in the X, Y, and Z axis directions. 2nd microphone 2, 3rd microphone 3,
Since the fourth microphone 5 and the fifth microphone 5 are arranged, the calculation at the time of calculating the position of the sound source (coordinates (x, y, z)) described above can be simplified, and the sound source position can be quickly determined. It becomes possible to calculate. As a result, sounds generated around the autonomous vehicle (for example, engine sounds of other vehicles, human footsteps, sounds of animal noises and other living noises, river sounds, thunder, etc. Since the situation around the vehicle can be accurately grasped based on the sound source position of the generated sound), it is possible to properly operate the autonomously traveling vehicle.

Further, according to the present embodiment, the first to fifth microphones 1 to 5 are based on the time-frequency-amplitude graph of the sound created by wavelet transforming the received waves of the first to fifth microphones 1 to 5. When the arrival time of the sound with respect to 5 is obtained, it becomes possible to classify the positions of a plurality of sound sources and to recognize what the sound source is.

Furthermore, according to the present embodiment, the on-vehicle camera (visual sensor) is directed toward the sound source based on the position information of the sound source estimated during the operation of the autonomous vehicle 11, and the surroundings of the sound source are photographed. Autonomous vehicle 1 based on shooting information
When 1 is remotely controlled, the autonomous vehicle 11 can be accurately driven.

Next, the sound source recognition method in this embodiment will be described. To recognize what the sound source is,
For example, various sounds that are expected to be generated outdoors operating the autonomous vehicle 11 (for example, engine noise of other vehicles, human footsteps, sounds of animal noises and other living noises, river sounds, and thunderstorms). Data corresponding to a template (sound waveform template) in which the characteristics of a natural phenomenon such as a sound) are extracted is stored in the storage unit 31 of the computer 33 in advance.

When the autonomous vehicle 11 is operated outdoors, the arithmetic unit 29 of the computer 33 corresponds to the waveform of the sound received by the first to fifth microphones 1 to 5 supplied from the A / D converter 28. The data is compared with the data corresponding to the model of the waveform of the sound stored in advance in the storage unit 31 to determine what the sound source is.

In this case, the first to fifth microphones 1 to 5
When the arrival time of a sound is calculated by wavelet transforming the waveform received by the above as described above, the “arrival time” when a certain sound arrives at the first to fifth microphones 1 to 5,
It is possible to recognize what the sound source is by using a method similar to known image recognition using a three-dimensional pattern composed of “frequency” and “amplitude”.

Next, a modified example of this embodiment will be described.

FIG. 10 and FIG. 11 are diagrams showing the arrangement of microphones and speakers in the modified example. The first microphone 1 (coordinates (0,0,0)) is arranged at the origin of the three-dimensional coordinates, and A speaker 40 is arranged in the vicinity of the microphone 1, and a second microphone 2 (coordinates (L, 0,0)) and a third microphone are provided at positions spaced apart from the first microphone 1 in the X-, Y-, and Z-axis directions, respectively. 3 (coordinates (0,
L, 0)), the fourth microphone 4 (coordinates (0, 0,
L)), the fifth microphone 5 (coordinates (0,0, -L))
Is arranged. In this case, by installing the first to fifth microphones 1 to 5 on the three-dimensional coordinate axis, the calculation when calculating the sound source position and the sound arrival time is simplified.

FIG. 12 is a block diagram of an obstacle position calculation control system mounted on an autonomous vehicle according to a modification. The point different from the above embodiment is that a speaker control command signal output from the arithmetic unit 29 is used. The point is that a speaker 40 for a sound source whose sound generation is controlled by a speaker control unit 39 that operates based on the above is provided. The same configurations as those of the modified example and the above-described embodiment are designated by the same reference numerals and the description thereof will be omitted.

That is, in the modification, the autonomous traveling vehicle receives the speaker 40 that emits sound to the outside of the vehicle and the sounds that are returned from the obstacles such as walls around the vehicle and returned. Equipped with the fifth microphones 1 to 5, the distance to the obstacle is calculated based on the time until the sound emitted from the speaker 40 is reflected from the obstacle and returns to the first to fifth microphones 1 to 5. By doing so, the position of the obstacle is estimated. Therefore, it is possible to drive the autonomous vehicle while avoiding the obstacle based on the estimated position of the obstacle.

In this case, in the modification, after the sound generation time by the speaker 40 is calculated based on the received signal of the first microphone 1, the position of the obstacle is calculated by the same method as the sound source position calculation method described above. be able to.

When the sound source position is calculated using the first to fifth microphones 1 to 5 in the above embodiment as a passive sensor, the speaker 40 and the first to fifth microphones 1 to 1 in the modified example are used. The case where the sound source position is calculated using 5 can be called usage as an active sensor.

In this case, in the present embodiment and the modified example, the first to fifth microphones 1 to 5 are arranged on the three-dimensional coordinate axes as shown in FIGS. 3 and 10, but four or more microphones are used. The first to fifth microphones 1 to 5 may be arranged at arbitrary positions in the three-dimensional coordinate space unless they are arranged on a straight line and the five microphones are not arranged on the same plane.

The sound source position calculation method when the first to fifth microphones 1 to 5 are arranged at arbitrary positions in the three-dimensional coordinate space is as follows. That is, T is the time when a sound is generated at a certain position outside the autonomous vehicle 11, the three-dimensional coordinate of the sound source position is (x, y, z), and the sound velocity corrected based on the detection data of the outside air temperature sensor 9 is Vs, The intervals between the first microphone and the second microphone in the X-axis, Y-axis, and Z-axis directions are L ₁ , L ₂ , and L ₃ , respectively, and the X-axis and the Y-axis between the first microphone and the third microphone, The intervals in the Z-axis direction are L ₄ , L ₅ , and L ₆ , respectively, and the intervals in the X-axis, Y-axis, and Z-axis directions between the first microphone and the fourth microphone are L, respectively.
₇ , L ₈ , L ₉ , the distances between the first and fifth microphones in the X-axis, Y-axis, and Z-axis directions are L ₁₀ , respectively.
When L ₁₁ and L ₁₂ and arrival times of sounds generated at a certain position to the fifth to fifth microphones are t ₀ , t ₁ , t ₂ , t ₃ and t ₄ , respectively,

[0076]

[Equation 11]

The following simultaneous equations (sound propagation distance equations) hold for the first to fifth microphones. After that, if the simultaneous equations are solved by the same procedure as in the above embodiment,
It is possible to calculate the sound generation time T and the coordinates (x, y, z) of the sound source position. In this case, the calculation is slightly complicated as compared with the above embodiment, but the initial purpose can be achieved.

Further, in the present embodiment and the modified examples, the case where the sound source position calculation method and the obstacle position calculation method are applied to the autonomous traveling vehicle has been taken as an example, but the invention is not limited to the autonomous traveling vehicle, and it is not limited to the autonomous traveling vehicle. It can be applied to various fields. in this case,
The sound source position calculation method can be applied to, for example, a crime prevention device in a building or a house.

Further, in the present embodiment, the position of the acoustic lighthouse is calculated based on the reception information of the sound transmitted from the acoustic lighthouse installed in advance at a predetermined location where the autonomous vehicle is operated, and the calculation is performed. It is also possible to calculate the current position of the autonomous vehicle based on the position of the acoustic lighthouse.

[0080]

As described above, according to the present invention of claim 1, at least five omnidirectional microphones mounted on the origin in the three-dimensional coordinate space and at a position separated by a predetermined distance from the origin. The arrival time of the received sound is stored, and at least five sound propagation distance equations relating to the sound source position and sound generation time obtained with respect to the origin for each microphone are provided. Since the sound source position and the sound generation time are calculated by arbitrarily selecting the equations and performing the simultaneous solution method, the sound source is compared with the case where the sound source is estimated to be in the direction in which the microphone reception level is maximum as in the past. It is possible to accurately calculate the position of the sound source regardless of the length of time the sound is generated and the sound generation range.

According to the present invention of claim 2, further, four equations are arbitrarily selected a plurality of times from the above-mentioned five sound propagation distance equations, and the sound source position and the sound generation time are separately set for each selected group. In order to calculate the sound source position and sound generation time, the close values among the values obtained for each selection group are calculated.
It is possible to add objectivity to the obtained sound source position,
As a result, it is possible to improve the calculation accuracy of the sound source position.

According to the present invention of claim 3, since the microphones of claim 1 or 2 are mounted on the three-dimensional coordinate axis and at the origin and at positions equidistant L from the origin, the sound source position and the sound generation are generated. The calculation when calculating the time can be simplified, and the sound source position and the sound generation time can be calculated quickly. Therefore, when the present invention is applied to, for example, an autonomous vehicle, a sound generated around the vehicle (for example, engine sound of another vehicle / footsteps of humans / sound caused by living noise such as crying of animals, sound of river flow / The situation around the vehicle can be accurately grasped based on the position of the sound source such as a sound caused by a natural phenomenon such as thunder, etc., so that it is possible to accurately operate the autonomous vehicle. Play.

According to the present invention of claim 4, at least five omnidirectional microphones provided at the origin and a position separated by a predetermined distance from the origin in the three-dimensional coordinate space, and the sound received by each microphone. Arrival time, sound velocity at this time,
A storage unit that stores coordinate position data of each microphone,
Calculation formula of sound source position and sound generation time obtained by arbitrarily selecting four equations from at least five sound propagation distance equations related to sound source position and sound generation time of sound received by each microphone and performing simultaneous solution Since a formula storage unit that stores the sound source and a calculation unit that calculates the sound source position and the sound generation time based on the calculation formula stored in the formula storage unit are provided, the length of time the sound is generated and the sound It is possible to accurately calculate the sound source position regardless of the generation range.

According to the present invention of claim 5, a sound source position obtained by arbitrarily selecting four equations from the above-mentioned five sound propagation distance equations a plurality of times and performing simultaneous solution for each selected group. And a formula storage unit that stores the calculation formula of the sound generation time for each selection group, and a calculation unit that calculates the sound source position and the sound generation time for each selection group based on the calculation formula stored in the formula storage unit. , The calculation unit has a sound source position specifying function that determines and determines the close value among the values obtained for each selected group as the sound source position and the sound generation time, so that the specified sound source position has objectivity. As a result, it is possible to improve the calculation accuracy of the sound source position.

According to the present invention of claim 6, since the microphones of claim 4 or 5 are mounted on the three-dimensional coordinate axis and at the origin and at positions equidistant L from the origin, the sound source position and the sound generation are generated. The calculation when calculating the time can be simplified, and the sound source position and the sound generation time can be calculated quickly. Therefore, when the present invention is applied to, for example, an autonomous vehicle, a sound generated around the vehicle (for example, engine sound of another vehicle / footsteps of humans / sound caused by living noise such as crying of animals, sound of river flow / The situation around the vehicle can be accurately grasped based on the position of the sound source such as a sound caused by a natural phenomenon such as thunder, etc., so that it is possible to accurately operate the autonomous vehicle. Play.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a sound source position calculation control system of an autonomous traveling vehicle in the present embodiment to which the present invention is applied.

FIG. 2 is a schematic diagram of an autonomous traveling vehicle in the present embodiment.

FIG. 3 is an explanatory diagram showing an arrangement state of first to fifth microphones on a three-dimensional coordinate axis in the present embodiment.

FIG. 4 is a diagram showing received waves of first to fifth microphones.

5 (a) is a diagram showing the time-amplitude for each frequency of the received wave of the first microphone, and FIG. 5 (b) is FIG.
It is explanatory drawing which shows the pattern which extracted a part of (a).

FIG. 6 is a diagram showing time-amplitude for each frequency of the received wave of the second microphone.

FIG. 7 is a block diagram showing a configuration of a sound source position calculation control system using a bandpass filter instead of a lowpass filter.

FIG. 8 is a diagram showing a zero-cross point of a waveform that has passed through a bandpass filter.

FIG. 9 is a diagram showing received waves and sound arrival times of the first to fifth microphones.

FIG. 10 is an explanatory diagram showing an arrangement state of a speaker and first to fifth microphones on a three-dimensional coordinate axis in a modified example.

FIG. 11 is an explanatory diagram showing a state where the speaker and the first to fifth microphones shown in FIG. 10 are viewed from the right side.

FIG. 12 is a block diagram showing a configuration of an obstacle position calculation control system of an autonomous traveling vehicle in a modified example.

[Explanation of symbols]

 1 1st microphone 2 2nd microphone 3 3rd microphone 4 4th microphone 5 5th microphone 29 Calculation part 31 Storage part 32 Formula storage part

Claims (6)

[Claims] 1. A three-dimensional coordinate space is provided with at least five omnidirectional microphones at an origin and at a position separated by a predetermined distance from the origin, and arrival times (t ₀ , t) of sounds received by each microphone. ₁ , t ₂ , t ₃ , t ₄ ) are stored, and a sound source position (x, y, z) obtained with reference to the origin for each microphone and sound generation time T The sound source position (x, y, z) and the sound generation time T are calculated by providing five sound propagation distance equations and arbitrarily selecting four equations from these five sound propagation distance equations and performing simultaneous solution. A sound source position calculation method characterized by the above. 2. A three-dimensional coordinate space is equipped with at least five omnidirectional microphones at an origin and at a position separated by a predetermined distance from the origin, and arrival times (t ₀ , t) of sounds received by each microphone. ₁ , t ₂ , t ₃ , t ₄ ) are stored, and a sound source position (x, y, z) obtained with reference to the origin for each microphone and sound generation time T Five sound propagation distance equations are provided, and four equations are arbitrarily selected a plurality of times from these five sound propagation distance equations to determine the sound source position (x, y, z) and the sound generation time T for each selected group. Separately calculated values of x, y, z, and T, which are close to each other and obtained for each selected group, are calculated as the sound source position (x, y, z) and the sound generation time T.
A sound source position calculation method characterized by: 3. The sound source position calculation method according to claim 1, wherein each of the microphones is mounted on the three-dimensional coordinate axis and at the origin and at positions equidistant from the origin. 4. The three-dimensional coordinate space is equipped with an origin and at least five omnidirectional microphones at a position separated from the origin by a predetermined distance, and arrival time (t ₀ , t ₀ ,
_{t 1, t 2, t 3} , t 4 and), and the acoustic velocity V at this time, the storage unit provided for storing the coordinate position data of each microphone, sound source position of the sound received In the respective microphone (x ，
y, z) and at least five sound propagation distance equations related to the sound generation time T, and at the same time, the sound source position obtained by arbitrarily selecting four equations from these five sound propagation distance equations and performing simultaneous solution (X, y, z) and a formula storage unit for storing a calculation formula of the sound generation time T are provided, and the sound source position (x, y, z) and the sound based on the calculation formula stored in the formula storage unit. A sound source position calculation device comprising a calculation unit that calculates a generation time T. 5. The three-dimensional coordinate space is equipped with at least five omnidirectional microphones at an origin and at a position separated by a predetermined distance from the origin, and arrival times (t ₀ , t ₀ ,
_{t 1, t 2, t 3} , t 4 and), and the acoustic velocity V at this time, the storage unit provided for storing the coordinate position data of each microphone, sound source position of the sound received In the respective microphone (x ，
y, z) and at least five sound propagation distance equations for sound generation time T are set, and four equations are arbitrarily selected a plurality of times from these five sound propagation distance equations, and simultaneous solution methods are applied to each selected group. The sound source position (x, y, z) and the sound generation time T obtained by performing the above formula are provided with a formula storage unit for storing each formula group, and the above formula is stored based on the formula stored in the formula storage unit. A calculation unit for calculating the sound source position (x, y, z) and the sound generation time T for each selection group is provided, and the calculation unit obtains x, y, for each selection group.
Among the values of z and T, the close values of the sound source position (x, y,
z) and a sound source position specifying function for determining and specifying the sound generation time T. 6. The sound source position calculation device according to claim 4, wherein each of the microphones is mounted on the three-dimensional coordinate axis and at the origin and at positions equidistant from the origin.