CN103064061B - Sound source localization method of three-dimensional space - Google Patents

Abstract

The three-dimensional sound source localization method is a three-dimensional sound source localization method using a movable small microphone array and sound source localization technology based on time delay estimation. A small microphone array is used to collect target sound source signals for a period of time, and a cross-correlation algorithm is used. Calculate the time difference between each of the microphones, and substitute the time difference into the calculation formula of azimuth angle and elevation angle to obtain the azimuth angle and elevation angle of the target sound source, and then move the small microphone array along a certain azimuth angle for a certain distance, repeat the above operation , get the azimuth angle and elevation angle of the target sound source at this time, and calculate the distance of the sound source target by measuring the azimuth angle and elevation angle of the target sound source twice. The method for measuring the sound source target adopted by the method of the present invention is a passive method, and the small-sized microphone array is moved in the measurement process, which overcomes the low accuracy of measuring the sound source target distance in the prior art and adopts the active method to measure the cost of the sound source target distance High and unsafe disadvantages.

Description

Three dimensions sound localization method

Technical field

Technical scheme of the present invention relates to application sound wave by the device of determining that the cooperation of multiple directions is located, specifically three dimensions sound localization method.

Background technology

Now, along with the continuous expansion of bionics techniques application, the Auditory Perception technology based on microphone array becomes the important topic of numerous research fields such as Mobile Robotics Navigation, voice signal enhancing and submarine target perception gradually.Can say, the sense of hearing is one of important symbol of New Generation of Intelligent robot, is to realize " people-machine-environment " mutual important means.Because sound has the characteristic of cut-through thing, in many information acquisition systems, the sense of hearing can match with vision, thereby it is limited and can not be through the limitation of non-printing opacity barrier to make up the visual field of vision.In addition, can not only localization of sound source target in " auditory scene ", can also obtain more valuable information by modern signal processing technology.Therefore, design high-precision sound source locating device and there is important theory significance and using value in medical treatment, service and military field.

Microphone sound source locating device of the prior art and method can only localization of sound source orientation angle, accurately orientation distances.For example, CN201010191634.1 disclosed " a kind of sound source locating device ", the method of the location spatial sound source adopting is: gather one section of acoustic target signal, then the mistiming between each microphone be can obtain by said apparatus and its method for calculating and locating, deflection, the elevation angle and distance calculated according to mistiming and array geometry model.The deflection that the method calculates and elevation accuracy are very high, but the precision of distance is just poor.The measuring distance method using in this patented technology is passive means, and in the method entirety measuring process, microphone array is not moved.The shortcoming of this prior art is that the precision of orientation distance is not high enough.The sound localization method that document " based on the object locating system of orthopyramid battle array " was mentioned in (the 25th the 5th phase of volume of Southeast China University's journal) is: gather one section of acoustic target signal, then can obtain the mistiming between each microphone by said system and algorithm, then can only calculate deflection and the elevation angle according to mistiming and array geometry model, and now not calculate distance.And the measuring method of this section of article middle distance is: in array center, the device that can launch sound is set, after audio emission is gone out, while arriving acoustic target, can be reflected back, now array received is to sound, when sound and receive mistiming in these two moment while reflecting sound, calculate the distance of acoustic target according to transmitting.The measuring distance method of this section of article use is active method, and in the process of measuring distance, array is not also moved.The shortcoming of the method is to carry out measuring distance with active method, so just needs the equipment of extra additional emission sound, has increased the cost of system.Also have potential safety hazard if be used in addition military aspect, for example, be used in submarine detection, while at this moment using active method, need to launch voice signal, this signal is easily found by the other side and is received, thereby exposed oneself, generation potential safety hazard.

Summary of the invention

Technical matters to be solved by this invention is: three dimensions sound localization method is provided, and is to adopt movable small microphone array and the auditory localization technology based on time delay estimation to carry out three dimensions sound localization method.The method of the measurement acoustic target that the inventive method adopts is passive means, the middle-size and small-size microphone array of measuring process moves, and has overcome the measurement acoustic target range accuracy of prior art low and adopt active method to measure the high and unsafe shortcoming of acoustic target distance costs.

The present invention solves this technical problem adopted technical scheme: three dimensions sound localization method is to adopt movable small microphone array and the auditory localization technology based on time delay estimation to carry out three dimensions sound localization method,

A. the method device used

Comprise minitype microphone array, power supply conditioning device, data collecting card and host computer, four summits that wherein minitype microphone array lays respectively at positive tetrahedron by four independences and the identical microphone of characteristic form, and host computer comprises that time-delay calculation model, position angle calculate model, elevation angle computation model and apart from computation model; Each microphone needs the data line of a BNC connector to be connected with power supply conditioning device, power supply conditioning device is connected and thinks that the latter powers with whole microphone array by the data line of 4 BNC connectors, power supply conditioning device is also connected with data collecting card by the data line of 4 BNC connectors, and data collecting card is connected by a usb data line with host computer;

B. by the step that said apparatus carries out three dimensions sound localization method be:

Three dimensions auditory localization comprises that position angle B, elevation angle F, the acoustic target of determining acoustic target are from the horizontal range d of the central point of microphone array bottom surface ₂with the distance B of acoustic target to the central point of minitype microphone array bottom surface,

The first step, measures the position angle A of acoustic target before minitype microphone array moves and the elevation angle E before minitype microphone array moves

Minitype microphone array is positive tetrahedron, and the center of establishing positive tetrahedron bottom surface is true origin O, and the microphone of position, true origin O dead ahead is S1, the microphone of true origin O location right is S2, the microphone of true origin O left position is S3, and the microphone of true origin O top position is S4

(1) carrying out by the time-delay calculation model in host computer the time delay that in minitype microphone array, each microphone is relative estimates

One according to measure the selected position of environmental baseline, with minitype microphone array to gather a period of time be 10ms～30ms target sound signal, voice signal passes to host computer by data collecting card, first host computer calculates voice signal, and to arrive relative time between four microphones on four summits that lay respectively at positive tetrahedron poor, be that voice signal arrives the time delay value between the moment of microphone S2, microphone S3 and microphone S4 and the moment of voice signal arrival microphone S1, concrete grammar is as follows:

The coordinate of supposing the discrete event signal model of two microphones reception voice signals is:

x ₁(t)＝a ₁s(t)+n ₁(t)，x ₂(t)＝a ₂s(t-τ ₁₂)+n ₂(t) （1）

In above formula, α _ifor the attenuation coefficient of sound-source signal, s (t) is acoustic target signal, x _i(t) voice signal gathering for microphone, n _i(t) be the additional noise signal of sound source, τ ₁₂be two time delays that microphone picks up voice signal, i.e. time delay,

By the voice signal x gathering _i(t), i=1,2 by Fourier transform, changes into frequency domain signal X by time domain _i(ω), its cross-power spectrum function is:

$G_{X_1{X}_2}\left(\mathrm{\&omega;}\right)=X_1\left(\mathrm{\&omega;}\right){{\mathrm{<figure-callout\; id="2"\; label="X"\; filenames="130105094101.png,130105094110.png"\; state="\{\{state\}\}">X</figure-callout>}}_2}^{*}\left(\mathrm{\&omega;}\right)---\left(2\right)$

Its cross correlation function is:

$R_{x_1{x}_2}\left(\mathrm{\&tau;}\right)=\underset{0}{\overset{\mathrm{\&pi;}}{\&Integral;}}{G}_{X_1{X}_2}\left(\mathrm{\&omega;}\right)e^{\mathrm{j\&omega;\&tau;}}\mathrm{d\&omega;}---\left(3\right)$

Finally carry out peak value detection, the point of the horizontal ordinate that the peak value of cross correlation function is corresponding is exactly time delay value t ₂₁, use the same method and can calculate time delay value t ₃₁and t ₄₁, finally show that the time delay value between microphone S2, microphone S3 and microphone S4 and microphone S1 is respectively: t ₂₁, t ₃₁, t ₄₁;

(2) calculate the position angle A of acoustic target before minitype microphone array moves and the elevation angle E before minitype microphone array moves by position angle calculating model and elevation angle computation model in host computer

If Q is acoustic target point, coordinate is Q (x, y, z), true origin O is r to the distance of acoustic target point Q, and OQ is projected as OQ ' XOY plane, definition OQ ' is α with the angle of X-axis, the angle of OQ and Z axis is β, supposes that S1 is a to the distance of true origin O, and the coordinate of four microphones is respectively: S ₁=(a, 0,0), ${\mathrm{<figure-callout\; id="2"\; label="S"\; filenames="130105094101.png,130105094110.png"\; state="\{\{state\}\}">S</figure-callout>}}_2=(-a/2,\sqrt{3}a/\mathrm{2,0}),$ ${\mathrm{<figure-callout\; id="3"\; label="S"\; filenames="130105094110.png"\; state="\{\{state\}\}">S</figure-callout>}}_3=(-a/2,-\sqrt{3}a/\mathrm{2,0}),$ $S_4=(\mathrm{0,0},\sqrt{2}a),$

The time delay value obtaining between the moment of voice signal arrival microphone S2, microphone S3 and microphone S4 and the moment of voice signal arrival microphone S1 from above-mentioned (1) is respectively: t ₂₁, t ₃₁, t ₄₁,

At this moment the position angle formula that draws sound source is:

$\mathrm{\&alpha;}\&ap;\arctan \sqrt{3}\frac{t_{31}-t_{21}}{t_{21}+t_{31}}---\left(4\right)$

The angle of the α that above-mentioned formula (4) calculates is not necessarily in regulation azimuth coverage, because the problem of quadrant, the range of results that formula (4) Arctan calculates is that-90 degree are to 90 degree, and needed regulation position angle range of results is that-180 degree are to 180 degree, this just need to divide quadrant processing, via following position angle quadrant processing procedure, show that measuring for the first time the position angle A of acoustic target before minitype microphone array moves is:

In the time of result of calculation α > 0, and t ₃₁when >0, position angle A=α,

In the time of result of calculation α > 0, and t ₃₁when <0, position angle A=-180+ α,

In the time of result of calculation α < 0, and t ₂₁when >0, position angle A=α,

In the time of result of calculation α < 0, and t ₂₁when <0, position angle A=180+ α,

At this moment the elevation angle formula that draws sound source is:

$\mathrm{\&beta;}\&ap;\mathrm{arccot}\frac{t_{21}+t_{31}-{3t}_{41}}{2\sqrt{2}\sqrt{t_{21}^2+t_{31}^2-t_{21}{t}_{31}}}---\left(5\right),$

Above-mentioned formula (5) calculates the angle of β not necessarily in regulation elevation coverage, because the problem of quadrant, the range of results that formula (5) Arccot calculates is that-90 degree are to 90 degree, and the regulation elevation angle range of results needing is that 0 degree is to 180 degree, according to the geometric model calculating, here divide quadrant processing, via following elevation angle quadrant processing procedure, show that measuring for the first time the elevation angle E of acoustic target before minitype microphone array moves is:

In the time of result of calculation β > 0, elevation angle E=β,

In the time of result of calculation β < 0, elevation angle E=180+ β,

Second step, mobile minitype microphone array

After the first step completes, clockwise rotate the realization of χ-A degree by mobile robot original place minitype microphone array is clockwise rotated to χ-A degree, 0 degree < χ < 180 spends, then minitype microphone array is spent to direction along array orientation angle 0 is also that mobile robot dead ahead moves forward distance L;

The 3rd step, measures the position angle B of acoustic target after minitype microphone array moves and the elevation angle F after minitype microphone array moves

After second step has operated, on the position arriving after minitype microphone array moves, repeat with operation and the calculating of the first step, result is,

(1) carrying out by the time-delay calculation model in host computer the time delay that in minitype microphone array, each microphone is relative estimates

Finally show that the time delay value between microphone S2, microphone S3 and microphone S4 and microphone S1 is respectively: t ₂₁, t ₃₁, t ₄₁;

(2) calculate the position angle B of acoustic target after minitype microphone array moves and the elevation angle F after minitype microphone array moves by position angle calculating model and elevation angle computation model in host computer

Show that measuring for the second time the position angle B of acoustic target after minitype microphone array moves is:

In the time of result of calculation α > 0, and t ₃₁when >0, position angle B=α,

In the time of result of calculation α > 0, and t ₃₁when <0, position angle B=-180+ α,

In the time of result of calculation α < 0, and t ₂₁when >0, position angle B=α,

In the time of result of calculation α < 0, and t ₂₁when <0, position angle B=180+ α,

Show that measuring for the second time the elevation angle F of acoustic target after minitype microphone array moves is:

In the time of result of calculation β > 0, elevation angle F=β,

In the time of result of calculation β < 0, elevation angle F=180+ β,

The 4th step, calculates the horizontal range of acoustic target to the central point of microphone array bottom surface

Calculate the horizontal range of acoustic target to the central point of microphone array bottom surface by the distance computation model in host computer, computing formula is:

d ₂＝L*sin(χ)/sin(180-χ-δ) （6）

D ₂for acoustic target is to the horizontal range of the central point of microphone array bottom surface, the opposite direction O of the moving direction of minitype microphone array 2 ₂o ₁the direction O of the acoustic target 3 after moving with minitype microphone array 2 ₂the angle of Z is δ, is also δ=180-B;

The 5th step, calculates the distance of acoustic target to the central point of minitype microphone array bottom surface

Calculate the distance of acoustic target to the central point of minitype microphone array bottom surface by the distance computation model in host computer, computing formula is:

D＝d ₂/sin(F) （7）

D is the distance of acoustic target to the central point of minitype microphone array bottom surface;

The 6th step, the demonstration output of three dimensions auditory localization data

By the peripheral hardware of computing machine, display shows or outputs on other computer the elevation angle F, acoustic target of the position angle B, the acoustic target that show output acoustic target to the horizontal range d of the central point of microphone array bottom surface by network interface card ₂with the distance B of acoustic target to the central point of minitype microphone array bottom surface, complete thus three dimensions auditory localization.

Above-mentioned three dimensions sound localization method, the bottom surface circumradius of the positive tetrahedron of described minitype microphone array is 10 centimetres.

Above-mentioned three dimensions sound localization method, in described host computer, using software is matlab.

Above-mentioned three dimensions sound localization method, the scope of described L is preferably 0.5～1.5m.

Above-mentioned three dimensions sound localization method, the scope of described χ is preferably 45 degree～135 degree.

Above-mentioned three dimensions sound localization method, described mobile minitype microphone array is by manually moving with minitype microphone array or being moved with minitype microphone array by mobile robot.

Above-mentioned three dimensions sound localization method, microphone used is the MPA201 microphone that Beijing Sheng Wang Acoustic-Electric (BSWA) Technology Co., Ltd. produces, and forms MPA201 microphone array by four these microphones; Power supply conditioning device is the microphone power supply conditioning device MC104 that Beijing popularity company produces; Data collecting card is the NI9215A data collecting card that American National instrument and equipment (NI) is produced, host computer is general PC, host computer has been installed after NIDAQ driving, just can write Matlab program, read the data that NI data collecting card collects, can also use the filter function of matlab to realize filtering.

The invention has the beneficial effects as follows: compared with prior art, the outstanding substantive distinguishing features of three dimensions sound localization method of the present invention is to adopt movable small microphone array and the auditory localization technology based on time delay estimation to carry out three dimensions auditory localization, adopt minitype microphone array to gather the acoustic target signal that a period of time is 10ms～30ms, carry out the orientation angles for the first time and the elevation angle that after mistiming calculating, draw acoustic target, after this this minitype microphone array moves a segment distance along certain orientation, and then to gather a period of time be 10ms～30ms acoustic target signal, carry out the orientation angles for the second time and the elevation angle that after mistiming calculating, draw again acoustic target, the position angle of moving by the position angle of twice of acoustic target and the measurement at the elevation angle and this minitype microphone array and mobile distance just can calculate the distance of acoustic target, thereby complete three dimensions auditory localization.

Compared with prior art, significant progressive being of three dimensions sound localization method of the present invention, calculated amount is little, precision is high, can be in three dimensions localization of sound source target accurately, the method of the measurement acoustic target that the inventive method adopts is passive means, and the middle-size and small-size microphone array of measuring process moves, and has overcome prior art and has adopted active method to measure the high also unsafe shortcoming of acoustic target distance costs.

Brief description of the drawings

Below in conjunction with drawings and Examples, the present invention is further described.

Fig. 1 is the formation of the inventive method device used and forms each several part connected mode schematic block diagram.

Fig. 2 is that the minitype microphone array in the inventive method device used forms schematic diagram.

Fig. 3 is the algorithm principle figure of position angle, the elevation angle and the distance of the calculating acoustic target of the inventive method.

Fig. 4 is that the calculating acoustic target of the inventive method is to the schematic diagram calculation of the horizontal range of the central point of minitype microphone array bottom surface.

In figure, 1. microphone, 2. minitype microphone array, 3. acoustic target.

Embodiment

Embodiment illustrated in fig. 1 showing, the inventive method device used comprises minitype microphone array, power supply conditioning device, data collecting card and host computer, and four summits that wherein minitype microphone array lays respectively at positive tetrahedron by four independences and the identical microphone of characteristic form; Power supply conditioning device is connected with whole microphone array by the data line of 4 BNC connectors, power supply conditioning device is also connected with data collecting card by the data line of 4 BNC connectors, data collecting card is connected by a usb data line with host computer, and power supply conditioning device is connected with 220v AC power by wire.

Embodiment illustrated in fig. 2 showing, the minitype microphone array 2 in three dimensions sound localization method of the present invention device used is that four summits that lay respectively at positive tetrahedron by four independences and the identical microphone 1 of characteristic form.

Embodiment illustrated in fig. 3 showing, the algorithm principle of calculating acoustic target orientation angles, the elevation angle and the distance of the inventive method is:

Minitype microphone array is positive tetrahedron, and S1, S2, S3, S4 are respectively four microphones, and O had been both true origin, also be the center of positive tetrahedron bottom surface, establishing Q is acoustic target point simultaneously, and coordinate is Q (x, y, z), true origin O is r to the distance of acoustic target point Q, and OQ is projected as OQ ' XOY plane, definition OQ ' is α with the angle of X-axis, the angle of OQ and Z axis is β, supposes that S1 is a to the distance of true origin O, and the coordinate of four microphones is respectively: S ₁=(a, 0,0), ${\mathrm{<figure-callout\; id="2"\; label="S"\; filenames="130105094101.png,130105094110.png"\; state="\{\{state\}\}">S</figure-callout>}}_2=(-a/2,\sqrt{3}a/\mathrm{2,0}),$ ${\mathrm{<figure-callout\; id="3"\; label="S"\; filenames="130105094110.png"\; state="\{\{state\}\}">S</figure-callout>}}_3=(-a/2,-\sqrt{3}a/\mathrm{2,0}),$ $S_4=(\mathrm{0,0},\sqrt{2}a),$

The time delay value that voice signal arrived between the moment of microphone S2, S3 and S4 and the moment of voice signal arrival microphone S1 is respectively: t ₂₁, t ₃₁, t ₄₁, at this moment show that the position angle formula of sound source is:

$\mathrm{\&alpha;}\&ap;\arctan \sqrt{3}\frac{t_{31}-t_{21}}{t_{21}+t_{31}}$

Owing to will carrying out the twice azimuthal measurement of minitype microphone array before and after moving in the inventive method, each measurement is all to have used position angle publicity above to carry out computer azimuth angle, calculate for the first time the position angle A of acoustic target before minitype microphone array moves, calculate for the second time the position angle B of acoustic target after minitype microphone array moves.α is azimuthal general designation.The relation of α and position angle A or B is as follows:

In the time of result of calculation α > 0, and t ₃₁when >0, position angle A or B=α

In the time of result of calculation α > 0, and t ₃₁when <0, position angle A or B=-180+ α

In the time of result of calculation α < 0, and t ₂₁when >0, position angle A or B=α

In the time of result of calculation α < 0, and t ₂₁when <0, position angle A or B=180+ α

At this moment the elevation angle formula that draws sound source is:

$\mathrm{\&beta;}\&ap;\mathrm{arccot}\frac{t_{21}+t_{31}-{3t}_{41}}{2\sqrt{2}\sqrt{t_{21}^2+t_{31}^2-t_{21}{t}_{31}}}$

Owing to will carrying out the measurement at twice elevation angle of minitype microphone array before and after moving in the inventive method, each measurement is all to have used elevation angle publicity above to calculate the elevation angle, calculate for the first time the elevation angle E of acoustic target before minitype microphone array moves, calculate for the second time the elevation angle F of acoustic target after minitype microphone array moves.β is the general designation at the elevation angle.The relation of β and elevation angle E or F is as follows:

In the time of result of calculation β > 0, elevation angle E or elevation angle F=β.

In the time of result of calculation β < 0, elevation angle E or elevation angle F=180+ β.

Apart from computing formula:

r＝OQ′/sin(β)

R is the general designation of acoustic target to the distance of the central point of microphone array bottom surface.Move owing to will carrying out minitype microphone array in the inventive method, before and after have twice acoustic target distance to the horizontal range of the central point of minitype microphone array bottom surface and twice acoustic target to the central point of minitype microphone array bottom surface, wherein, after minitype microphone array moves acoustic target to the horizontal range d of the central point of minitype microphone array bottom surface ₂represent, after minitype microphone array moves, acoustic target represents with alphabetical D to the distance of the central point of minitype microphone array bottom surface.

Draw thus the central point O of the bottom surface of measuring orientation angles α, the elevation angle β of acoustic target 3 and the positive tetrahedron of the minitype microphone array after minitype microphone array 2 moves 2 ₂and the distance of acoustic target 3 between the projection Z of the place of positive tetrahedron bottom surface plane is d ₂, the central point O of the bottom surface of the positive tetrahedron of the minitype microphone array 2 after minitype microphone array 2 moves ₂and the distance between acoustic target 3 is D.

Embodiment illustrated in fig. 4 showing, the calculating acoustic target of the inventive method to the Computing Principle of the horizontal range of the central point of minitype microphone array bottom surface is:

O in this figure ₁the central point of the bottom surface of the positive tetrahedron of the minitype microphone array 2 of minitype microphone array 2 while not moving, O ₂for the central point of the bottom surface of the positive tetrahedron of minitype microphone array 2 after minitype microphone array 2 displacement L, Z is the projection of acoustic target 3 in the place of positive tetrahedron bottom surface plane.Black arrow represents the direction that the position angle 0 of minitype microphone array 2 is spent, O ₁place's black arrow and O ₁the angle of Z is A, O ₂place's black arrow and O ₂the angle of Z is B.

The central point O of the bottom surface of the positive tetrahedron of the minitype microphone array 2 in the time that minitype microphone array 2 does not move ₁and the distance of acoustic target 3 between the projection Z of the place of positive tetrahedron bottom surface plane is d ₁, the central point O of the bottom surface of the positive tetrahedron of the minitype microphone array 2 after minitype microphone array 2 moves ₂and the distance of acoustic target 3 between the projection Z of the place of positive tetrahedron bottom surface plane is d ₂, the distance that minitype microphone array 2 moves is L, the moving direction O of minitype microphone array 2 ₁o ₂the direction O of the acoustic target 3 before not mobile with minitype microphone array 2 ₁the angle of Z is χ, the opposite direction O of the moving direction of minitype microphone array 2 ₂o ₁the direction O of the acoustic target 3 after moving with minitype microphone array 2 ₂the angle of Z is δ, calculates the central point O of the bottom surface of the positive tetrahedron of the minitype microphone array 2 after moving by following computing formula ₂and the distance d of acoustic target 3 between the projection Z of the place of positive tetrahedron bottom surface plane ₂for:

d ₂＝L*sin(χ)/sin(180-χ-δ)

Embodiment 1

A. the present embodiment device used

Comprise minitype microphone array MPA201 microphone array, power supply conditioning device MC104, NI9215A data collecting card and the PC after NIDAQ driving has been installed, this minitype microphone array is that bottom surface circumradius is the positive tetrahedron of 10 centimetres, microphone used is the MPA201 microphone that Beijing Sheng Wang Acoustic-Electric (BSWA) Technology Co., Ltd. produces, in host computer, use software matlab, in host computer, mainly comprise time-delay calculation model, orientation angles computation model, elevation angle computation model and apart from computation model; Each microphone needs the data line of a BNC connector to be connected with power supply conditioning device, power supply conditioning device is connected and thinks that the latter powers with whole microphone array by the data line of 4 BNC connectors, power supply conditioning device is also connected with data collecting card by the data line of 4 BNC connectors, and data collecting card is connected by a usb data line with host computer.

B. this enforcement by the step that said apparatus carries out three dimensions sound localization method is:

Three dimensions auditory localization comprises that position angle B, elevation angle F, the acoustic target of determining acoustic target are from the horizontal range d of the central point of microphone array bottom surface ₂with the distance B of acoustic target to the central point of minitype microphone array bottom surface,

The first step, measures the position angle A of acoustic target before minitype microphone array moves and the elevation angle E before minitype microphone array moves

Minitype microphone array is positive tetrahedron, and the center of establishing positive tetrahedron bottom surface is true origin O, and the microphone of position, true origin O dead ahead is S1, the microphone of true origin O location right is S2, the microphone of true origin O left position is S3, and the microphone of true origin O top position is S4

(1) carrying out by the time-delay calculation model in host computer the time delay that in minitype microphone array, each microphone is relative estimates

One according to measure the selected position of environmental baseline, with minitype microphone array to gather a period of time be 10ms target sound signal, voice signal passes to host computer by data collecting card, first host computer calculates voice signal, and to arrive relative time between four microphones on four summits that lay respectively at positive tetrahedron poor, be that voice signal arrives the time delay value between the moment of microphone S2, microphone S3 and microphone S4 and the moment of voice signal arrival microphone S1, concrete grammar is as follows:

The coordinate of supposing the discrete event signal model of two microphones reception voice signals is:

x ₁(t)＝a ₁s(t)+n ₁(t)，x ₂(t)＝a ₂s(t-τ ₁₂)+n ₂(t) （1）

In above formula, α _ifor the attenuation coefficient of sound-source signal, s (t) is acoustic target signal, x _i(t) voice signal gathering for microphone, n _i(t) be the additional noise signal of sound source, τ ₁₂be two time delays that microphone picks up voice signal, i.e. time delay,

By the voice signal x gathering _i(t), i=1,2 by Fourier transform, changes into frequency domain signal X by time domain _i(ω), its cross-power spectrum function is:

$G_{X_1{X}_2}\left(\mathrm{\&omega;}\right)=X_1\left(\mathrm{\&omega;}\right){{\mathrm{<figure-callout\; id="2"\; label="X"\; filenames="130105094101.png,130105094110.png"\; state="\{\{state\}\}">X</figure-callout>}}_2}^{*}\left(\mathrm{\&omega;}\right)---\left(2\right)$

Its cross correlation function is:

$R_{x_1{x}_2}\left(\mathrm{\&tau;}\right)=\underset{0}{\overset{\mathrm{\&pi;}}{\&Integral;}}{G}_{X_1{X}_2}\left(\mathrm{\&omega;}\right)e^{\mathrm{j\&omega;\&tau;}}\mathrm{d\&omega;}---\left(3\right)$

Finally carry out peak value detection, the point of the horizontal ordinate that the peak value of cross correlation function is corresponding is exactly time delay value t ₂₁, use the same method and can calculate time delay value t ₃₁and t ₄₁;

The result that the time delay that gained is relative is estimated is: microphone S2 is 29 with respect to the mistiming of microphone S1, and microphone S3 is 49 with respect to the mistiming of microphone S1, and microphone S4 is 28 with respect to the mistiming of microphone 1.

Because capture card is 100k, therefore here 29 representative real times be 29 sampling periods, be 29*10 ^-5second; The real time of 49 representatives is here 49 sampling periods, is 49*10 ^-5second; The real time of 28 representatives is here 28 sampling periods, is 28*10 ^-5second.

The time delay value being derived as between microphone S2, microphone S3 and microphone S4 and microphone S1 is respectively: t ₂₁=29, t ₃₁=49, t ₄₁=28;

(2) calculate the position angle A of acoustic target before minitype microphone array moves and the elevation angle E before minitype microphone array moves by position angle calculating model and elevation angle computation model in host computer

Microphone array is classified positive tetrahedron as, establishes S1, S2, S3, S4 are respectively four microphones, and O had been both true origin, also be the center of positive tetrahedron bottom surface, establishing Q is acoustic target point simultaneously, and coordinate is Q (x, y, z), true origin O is r to the distance of acoustic target point Q, and OQ is projected as OQ ' XOY plane, definition OQ ' is α with the angle of X-axis, the angle of OQ and Z axis is β, supposes that S1 is a to the distance of true origin O, and the coordinate of four microphones is respectively: S ₁=(a, 0,0), ${\mathrm{<figure-callout\; id="2"\; label="S"\; filenames="130105094101.png,130105094110.png"\; state="\{\{state\}\}">S</figure-callout>}}_2=(-a/2,\sqrt{3}a/\mathrm{2,0}),$ ${\mathrm{<figure-callout\; id="3"\; label="S"\; filenames="130105094110.png"\; state="\{\{state\}\}">S</figure-callout>}}_3=(-a/2,-\sqrt{3}a/\mathrm{2,0}),$ $S_4=(\mathrm{0,0},\sqrt{2}a),$

The time delay value obtaining between the moment of voice signal arrival microphone S2, microphone S3 and microphone S4 and the moment of voice signal arrival microphone S1 from above-mentioned (1) is respectively: t ₂₁, t ₃₁, t ₄₁;

At this moment the position angle formula that draws sound source is:

$\mathrm{\&alpha;}\&ap;\arctan \sqrt{3}\frac{t_{31}-t_{21}}{t_{21}+t_{31}}---\left(4\right)$

Calculate α=29 degree, according to dividing quadrant problem to obtain about position angle: A=29 degree above.

At this moment show that elevation angle formula is:

$\mathrm{\&beta;}\&ap;\mathrm{arccot}\frac{t_{21}+t_{31}-{3t}_{41}}{2\sqrt{2}\sqrt{t_{21}^2+t_{31}^2-t_{21}{t}_{31}}}---\left(5\right)$

Calculate β=-87 degree, according to above divide quadrant problem to obtain about the elevation angle: E=180+(-87)=93 degree, elevation angle E is 93 degree.

Draw thus and measure for the first time the position angle A=29 degree of acoustic target that acoustic target calculates before minitype microphone array moves and calculate the elevation angle E=93 degree of acoustic target before minitype microphone array moves;

Second step, mobile minitype microphone array

After the first step completes, clockwise rotate the realization of 90-29 degree by mobile robot original place minitype microphone array is clockwise rotated to 90-29 degree, χ=90 degree, then minitype microphone array is spent to direction Ye Ji robot dead ahead along array orientation angle 0 move forward distance L=1 meter;

The 3rd step, measures the position angle B of acoustic target after minitype microphone array moves and the elevation angle F after minitype microphone array moves

After second step has operated, on the position arriving at minitype microphone array, repeat the operation of the first step,

(1) carrying out by the time-delay calculation model in host computer the time delay that in minitype microphone array, each microphone is relative estimates

The result that the time delay that gained is relative is estimated is: microphone S2 is-49 with respect to the mistiming of microphone S1, and microphone S3 is-13 with respect to the mistiming of microphone S1, and microphone S4 is-20 with respect to the mistiming of microphone S1.

Because capture card is 100k, therefore here-49 representative real times be-49 sampling periods, be-49*10 ^-5second; The real time of-13 representatives is here-13 sampling periods, is-13*10 ^-5second; The real time of-20 representatives is here-20 sampling periods, is-20*10 ^-5second.

Show that the time delay value between microphone S2, microphone S3 and microphone S4 and microphone S1 is respectively: t ₂₁=-49, t ₃₁=-13, t ₄₁=-20;

(2) calculate the position angle B of acoustic target after minitype microphone array moves and the elevation angle F after minitype microphone array moves by position angle calculating model and elevation angle computation model in host computer

Show that the result of measuring for the second time acoustic target is:

α=-45.1630 degree, the position angle B=180+(-45.1630 of acoustic target after minitype microphone array moves)=-134.8370 degree;

β=-88.1577 degree, the elevation angle F=180+(-88.1577 of acoustic target after minitype microphone array moves)=91.8423 degree.

The 4th step, calculates the horizontal range of acoustic target from the central point of microphone array bottom surface

Computing formula is:

d ₂＝L*sin(χ)/sin(180-χ-δ) （6）

D ₂for acoustic target is to the horizontal range of the central point of microphone array bottom surface, the opposite direction O of the moving direction of minitype microphone array 2 ₂o ₁the direction O of the acoustic target 3 after moving with minitype microphone array 2 ₂the angle of Z is δ, is also δ=180-B, and the above-mentioned the data obtained of substitution draws the horizontal range d of the central point of sound source distance microphone array bottom surface ₂for:

d ₂＝1*sin(90)/sin(180-90-(180-134.837))=1.4183；

The 5th step, calculates the distance B of acoustic target to the central point of minitype microphone array bottom surface

Computing formula is:

D＝d ₂/sin(F) （7）

The above-mentioned the data obtained of substitution show that acoustic target to the distance B of the central point of minitype microphone array bottom surface is:

D=1.4183/sin(91.8423)=1.4187

The 6th step, the demonstration output of three dimensions auditory localization data

By the peripheral hardware of computing machine, be that display shows or outputed to and on other computer, shown that the concrete outcome of output the present embodiment three dimensions auditory localization is: the position angle B=-134.8370 degree of acoustic target by network interface card, the elevation angle F=91.8423 degree of acoustic target, acoustic target is from the horizontal range d of the central point of microphone array bottom surface ₂=1.4183 and acoustic target to distance B=1.4187 of the central point of minitype microphone array bottom surface, complete thus three dimensions auditory localization.

Embodiment 2

A. the method device used

With embodiment 1.

B. this enforcement by the step that said apparatus carries out three dimensions sound localization method is:

Herein with embodiment 1.

The first step, measures the position angle A of acoustic target before minitype microphone array moves and the elevation angle E before minitype microphone array moves

Herein with embodiment 1.

(1) except gathering the target sound signal that a period of time is 20ms with minitype microphone array, other are with embodiment 1.

The result that the time delay that gained is relative is estimated is: microphone S2 is 36 with respect to the mistiming of microphone S1, and microphone S3 is 49 with respect to the mistiming of microphone S1, and microphone S4 is 28 with respect to the mistiming of microphone S1.

Because capture card is 100k, therefore here 36 representative real times be 36 sampling periods, be 36*10 ^-5second; The real time of 49 representatives is here 49 sampling periods, is 49*10 ^-5second; The real time of 28 representatives is here 28 sampling periods, is 28*10 ^-5second.

Show that the time delay value between microphone S2, microphone S3 and microphone S4 and microphone S1 is respectively: t ₂₁=36, t ₃₁=49, t ₄₁=28

(2) calculate the position angle A of acoustic target before minitype microphone array moves and the elevation angle E before minitype microphone array moves by position angle calculating model and elevation angle computation model in host computer

Herein with embodiment 1.

Calculate α=14.8370 degree, calculate β=89.0786 degree;

Draw thus and measure for the first time the position angle A=14.8370 degree of acoustic target that acoustic target calculates before minitype microphone array moves and calculate the elevation angle E=89.0786 degree of acoustic target before minitype microphone array moves;

Second step, mobile minitype microphone array

After the first step completes, clockwise rotate the realization of 45-14.8370 degree by mobile robot original place minitype microphone array is clockwise rotated to 45-14.8370 degree, χ=45 degree, then minitype microphone array is spent to direction Ye Ji robot dead ahead along array orientation angle 0 move forward distance L=0.5 meter;

The 3rd step, measures the position angle B of acoustic target after minitype microphone array moves and the elevation angle F after minitype microphone array moves

After second step has operated, on the position arriving at minitype microphone array, repeat the operation of the first step,

(1) result that the time delay that gained is relative is estimated is: microphone S2 is-45 with respect to the mistiming of microphone S1, and microphone S3 is-3 with respect to the mistiming of microphone S1, and microphone S4 is-17 with respect to the mistiming of microphone S1.

Because capture card is 100k, therefore here-45 representative real times be-45 sampling periods, be-45*10 ^-5second; The real time of-3 representatives is here-3 sampling periods, is-3*10 ^-5second; The real time of-17 representatives is here-17 sampling periods, is-17*10 ^-5second.

The time delay value being derived as between microphone S2, microphone S3 and microphone S4 and microphone S1 is respectively: t ₂₁=-45, t ₃₁=-3, t ₄₁=-17;

(2) calculate the position angle B of acoustic target after minitype microphone array moves and the elevation angle F after minitype microphone array moves by position angle calculating model and elevation angle computation model in host computer

Show that the result of measuring for the second time acoustic target is:

α=-56.5820 degree, the position angle B=123.4180 degree of acoustic target after minitype microphone array moves;

β=88.6057 degree, the elevation angle F=88.6057 degree of acoustic target after minitype microphone array moves.

The 4th step, calculates the horizontal range of acoustic target from the central point of microphone array bottom surface

Computing formula is:

d ₂＝L*sin(χ)/sin(180-χ-δ) （6）

D ₂for acoustic target is to the horizontal range of the central point of microphone array bottom surface, the opposite direction O of the moving direction of minitype microphone array 2 ₂o ₁the direction O of the acoustic target 3 after moving with minitype microphone array 2 ₂the angle of Z is δ, is also δ=180-B, the horizontal range d of the central point of the above-mentioned the data obtained sound source distance microphone array of substitution bottom surface ₂for:

d ₂＝0.5*sin(45)/sin(180-45-(180-123.4180))=0.3609；

The 5th step, calculates the distance B of acoustic target to the central point of minitype microphone array bottom surface

Computing formula is:

D＝d ₂/sin(F) （7）

The above-mentioned the data obtained of substitution show that acoustic target to the distance B of the central point of minitype microphone array bottom surface is:

D=0.3609/sin(88.6057)=0.3610

The 6th step, the demonstration output of three dimensions auditory localization data

By the peripheral hardware of computing machine, be that display shows or outputed to and on other computer, shown that the concrete outcome of output the present embodiment three dimensions auditory localization is: the position angle B=123.4180 degree of acoustic target by network interface card, the elevation angle F=88.6057 degree of acoustic target, acoustic target is from the horizontal range d of the central point of microphone array bottom surface ₂=0.3609 and acoustic target to distance B=0.3610 of the central point of minitype microphone array bottom surface, complete thus three dimensions auditory localization.

Embodiment 3

B. this enforcement by the step that said apparatus carries out three dimensions sound localization method is:

Herein with embodiment 1.

The first step, measures the position angle A of acoustic target before minitype microphone array moves and the elevation angle E before minitype microphone array moves

Herein with embodiment 1.

(1) except gathering the target sound signal that a period of time is 30ms with minitype microphone array, other are with embodiment 1.

The result that the time delay that gained is relative is estimated is: microphone S2 is 39 with respect to the mistiming of microphone S1, and microphone S3 is 47 with respect to the mistiming of microphone S1, and microphone S4 is 28 with respect to the mistiming of microphone S1.

Because capture card is 100k, therefore here 39 representative real times be 39 sampling periods, be 39*10 ^-5second; The real time of 47 representatives is here 47 sampling periods, is 47*10 ^-5second; The real time of 28 representatives is here 28 sampling periods, is 28*10 ^-5second.

The time delay value being derived as between microphone S2, microphone S3 and microphone S4 and microphone S1 is respectively: t ₂₁=39, t ₃₁=47, t ₄₁=28

(2) calculate the position angle A of acoustic target before minitype microphone array moves and the elevation angle E before minitype microphone array moves by position angle calculating model and elevation angle computation model in host computer

Herein with embodiment 1.

Calculate α=9.1529 degree, calculate β=88.1403 degree,

Draw thus and measure for the first time the position angle A=9.1529 degree of acoustic target that acoustic target calculates before minitype microphone array moves and calculate the elevation angle E=88.1403 degree of acoustic target before minitype microphone array moves;

Second step, mobile minitype microphone array

After the first step completes, clockwise rotate the realization of 135-9.1529 degree by mobile robot original place minitype microphone array is clockwise rotated to 135-9.1529 degree, χ=135 degree, then minitype microphone array is spent to direction Ye Ji robot dead ahead along array orientation angle 0 move forward distance L=1.5 meter;

The 3rd step, measures the position angle B of acoustic target after minitype microphone array moves and the elevation angle F after minitype microphone array moves

After second step has operated, on the position arriving at minitype microphone array, repeat the operation of the first step,

(1) result that the time delay that gained is relative is estimated is: microphone S2 is-13 with respect to the mistiming of microphone S1, and microphone S3 is-9 with respect to the mistiming of microphone S1, and microphone S4 is-8 with respect to the mistiming of microphone S1.

Because capture card is 100k, therefore here-13 representative real times be-13 sampling periods, be-13*10 ^-5second; The real time of-9 representatives is here-9 sampling periods, is-9*10 ^-5second; The real time of-8 representatives is here-8 sampling periods, is-8*10 ^-5second.

The time delay value being derived as between microphone S2, microphone S3 and microphone S4 and microphone S1 is respectively: t ₂₁=-13, t ₃₁=-9, t ₄₁=-8

(2) calculate the position angle B of acoustic target after minitype microphone array moves and the elevation angle F after minitype microphone array moves by position angle calculating model and elevation angle computation model in host computer

Show that the result of measuring for the second time acoustic target is:

α=-17.4802 degree, the position angle B=162.5198 degree of acoustic target after minitype microphone array moves;

β=86.4914 degree, the elevation angle F=86.4914 degree of acoustic target after minitype microphone array moves.

The 4th step, calculates the horizontal range of acoustic target from the central point of microphone array bottom surface

Computing formula is:

d ₂＝L*sin(χ)/sin(180-χ-δ) （6）

D ₂for acoustic target is to the horizontal range of the central point of microphone array bottom surface, the opposite direction O of the moving direction of minitype microphone array 2 ₂o ₁the direction O of the acoustic target 3 after moving with minitype microphone array 2 ₂the angle of Z is δ, is also δ=180-B, the horizontal range d of the central point of the above-mentioned the data obtained sound source distance microphone array of substitution bottom surface ₂for:

d ₂＝1.5*sin(135)/sin(180-135-(180-162.5198))=2.2955；

The 5th step, calculates the distance B of acoustic target to the central point of minitype microphone array bottom surface

Computing formula is:

D＝d ₂/sin(F) （7）

The above-mentioned the data obtained of substitution show that acoustic target to the distance B of the central point of minitype microphone array bottom surface is:

D=2.2955/sin(86.4914)=2.2998

The 6th step, the demonstration output of three dimensions auditory localization data

By the peripheral hardware of computing machine, be that display shows or outputed to and on other computer, shown that the concrete outcome of output the present embodiment three dimensions auditory localization is: the position angle B=162.5198 degree of acoustic target by network interface card, the elevation angle F=86.4914 degree of acoustic target, acoustic target is from the horizontal range d of the central point of microphone array bottom surface ₂=2.2955 and acoustic target to distance B=2.2998 of the central point of minitype microphone array bottom surface, complete thus three dimensions auditory localization.

Components and parts used in above-mentioned all embodiment are all by commercially available.

Claims (5)

1. The three-dimensional spatial sound source localization method is characterized in that: the sound source localization technique based on movable small microphone array and time delay estimation is used to carry out the three-dimensional spatial sound source localization method, A. Apparatus used in the method Including small microphone array, power supply conditioning equipment, data acquisition card and host computer, in which the small microphone array is composed of four independent microphones with the same characteristics located at the four vertices of the regular tetrahedron, and the host computer includes the delay calculation model, azimuth Angle calculation model, elevation angle calculation model and distance calculation model; each microphone needs to be connected with a data cable of BNC connector to the power supply conditioning device, and the power supply conditioning device is connected to the entire microphone array through 4 data cables of BNC connector to supply power for the latter , the power supply conditioning equipment is also connected to the data acquisition card through 4 data lines with BNC connectors, and the data acquisition card is connected to the host computer through a usb data line; B. The steps of carrying out the three-dimensional sound source localization method with the above-mentioned device are: Three-dimensional sound source localization includes determining the azimuth B of the sound source target, the elevation angle F, the horizontal distance _d2 of the sound source target from the center point of the bottom surface of the microphone array, and the distance D from the sound source target to the center point of the bottom surface of the small microphone array, The first step is to measure the azimuth A of the sound source target before the small microphone array moves and the elevation angle E before the small microphone array moves The small microphone array is a regular tetrahedron, and the center of the bottom surface of the regular tetrahedron is set as the coordinate origin O, the microphone directly in front of the coordinate origin O is S1, the microphone on the right side of the coordinate origin O is S2, and the microphone on the left side of the coordinate origin O is S3 , the microphone above the coordinate origin O is S4, (1) The relative delay estimation of each microphone in the small microphone array is performed by the delay calculation model in the host computer At a location selected according to the measured environmental conditions, a small microphone array is used to collect target sound signals for a period of 10ms to 30ms. The sound signals are transmitted to the host computer through the data acquisition card. The host computer first calculates the arrival of the sound signals on the front four sides The relative time difference between the four microphones of the four vertices of the body, that is, the time delay value between the moment when the sound signal reaches the microphone S2, the microphone S3 and the microphone S4 and the moment when the sound signal reaches the microphone S1, the specific method is as follows: The coordinates of the discrete event signal model assuming that two microphones receive sound signals are: x ₁ (t)=a ₁ s(t)+n ₁ (t), x ₂ (t)=a ₂ s(t-τ ₁₂ )+n ₂ (t) (1) In the above formula, α _i is the attenuation coefficient of the sound source signal, s(t) is the target signal of the sound source, x _i (t) is the sound signal collected by the microphone, n _i (t) is the noise signal added to the sound source, τ ₁₂ is the delay time for two microphones to pick up sound signals, that is delay, The collected sound signal x _i (t), i=1, 2 is converted from time domain to frequency domain signal _Xi (ω) by Fourier transform, and its cross power spectrum function is: ${\mathrm{<span\; class="notranslate">\; G</span>}}_{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; )</span>{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$ Its cross-correlation function is: ${\mathrm{<span\; class="notranslate">\; R</span>}}_{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&tau;</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; \&pi;</span>}}{<span\; class="notranslate">\; \&Integral;</span>}}{\mathrm{<span\; class="notranslate">\; G</span>}}_{{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{\mathrm{<span\; class="notranslate">\; j\&omega;\&tau;</span>}}\mathrm{<span\; class="notranslate">\; d\&omega;</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$ Finally, the peak detection is performed, and the point on the abscissa corresponding to the peak value of the cross-correlation function is the delay value t ₂₁ , and the delay values t ₃₁ and t ₄₁ can be calculated by the same method, and finally the microphone S2, the microphone S3 and the microphone S4 are obtained The time delay values between microphone S1 are: t ₂₁ , t ₃₁ , t ₄₁ ; (2) Calculate the azimuth A of the sound source target before the small microphone array moves and the elevation angle E before the small microphone array moves by the azimuth angle calculation model and the elevation angle calculation model in the host computer Let Q be the sound source target point, the coordinates are Q(x, y, z), the distance from the coordinate origin O to the sound source target point Q is r, the projection of OQ on the XOY plane is OQ′, define the distance between OQ′ and the X axis The included angle is α, the included angle between OQ and the Z axis is β, assuming that the distance from S1 to the coordinate origin O is a, then the coordinates of the four microphones are: S ₁ =(a,0,0), $S_2=(-a/2,\sqrt{3}a/\mathrm{2,0}),$ $S_3=(-a/2,-\sqrt{3}a/\mathrm{2,0}),$ $S_4=(\mathrm{0,0},\sqrt{2}a),$ From the above (1), the time delay values between the moment when the sound signal reaches the microphone S2, the microphone S3 and the microphone S4 and the moment when the sound signal reaches the microphone S1 are respectively: t ₂₁ , t ₃₁ , t ₄₁ , At this time, the formula for the azimuth angle of the sound source is: $\mathrm{<span\; class="notranslate">\; \&alpha;</span>}<span\; class="notranslate">\; \&ap;</span>\mathrm{<span\; class="notranslate">\; arctan</span>}\sqrt{\mathrm{<span\; class="notranslate">\; 3</span>}}\frac{{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; 31</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; one</span>}}}{{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; one</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; 31</span>}}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$ The angle of α calculated by the above formula (4) is not necessarily within the range of the specified azimuth angle, because of the quadrant problem. The range of the result calculated by the formula (4) Arctan is -90 degrees to 90 degrees, and the required specified azimuth angle The angle result ranges from -180 degrees to 180 degrees, which requires sub-quadrant processing. Through the following azimuth quadrant processing, the azimuth A of the first measured sound source target before the small microphone array moves is: When the calculation result α>0, and t ₃₁ >0, the azimuth A=α, When the calculation result α>0, and t ₃₁ <0, the azimuth A=-180+α, When the calculation result α<0, and t ₂₁ >0, the azimuth A=α, When the calculation result α<0, and t ₂₁ <0, the azimuth A=180+α, At this time, the formula for the elevation angle of the sound source is: $\mathrm{<span\; class="notranslate">\; \&beta;</span>}<span\; class="notranslate">\; \&ap;</span>\mathrm{<span\; class="notranslate">\; arccot</span>}\frac{{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; one</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; 31</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; 3</span>}\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; 41</span>}}}{\mathrm{<span\; class="notranslate">\; 2</span>}\sqrt{\mathrm{<span\; class="notranslate">\; 2</span>}}\sqrt{{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; one</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; 31</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; one</span>}}{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; 31</span>}}}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 5</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>$ The angle of β calculated by the above formula (5) is not necessarily within the range of the specified elevation angle, because of the problem of the quadrant. The range of the result calculated by the formula (5) Arccot is -90 degrees to 90 degrees, and the required range of the specified elevation angle is From 0° to 180°, according to the calculated geometric model, sub-quadrant processing is performed here. Through the following elevation angle quadrant processing, the elevation angle E of the sound source target measured for the first time before the small microphone array moves is obtained as: When the calculation result β>0, the elevation angle E=β, When the calculation result β<0, the elevation angle E=180+β, Step two, move the small microphone array After the first step is completed, the small microphone array is rotated clockwise by χ-A degrees by moving the robot in situ by χ-A degrees, 0 degrees<χ<180 degrees, and then the small microphone array is rotated along the array azimuth angle 0 In the direction of degrees, that is, the distance L that the mobile robot moves forward in front of it; The third step is to measure the azimuth B of the sound source target after the small microphone array moves and the elevation angle F after the small microphone array moves After the second step is completed, repeat the same operation and calculation as the first step at the position where the small microphone array has moved, and the result is, (1) The relative delay estimation of each microphone in the small microphone array is performed by the delay calculation model in the host computer Finally, the time delay values between the microphone S2, the microphone S3 and the microphone S4 and the microphone S1 are respectively: t ₂₁ , t ₃₁ , t ₄₁ ; (2) Calculate the azimuth angle B of the sound source target after the small microphone array moves and the elevation angle F after the small microphone array moves by the azimuth angle calculation model and the elevation angle calculation model in the host computer It is obtained that the azimuth B of the second measurement of the sound source target after the small microphone array moves is: When the calculation result α>0, and t ₃₁ >0, the azimuth B=α, When the calculation result α>0, and t ₃₁ <0, the azimuth B=-180+α, When the calculation result α<0, and t ₂₁ >0, the azimuth B=α, When the calculation result α<0, and t ₂₁ <0, the azimuth B=180+α, It is obtained that the elevation angle F of the second measurement sound source target after the small microphone array moves is: When the calculation result β>0, the elevation angle F=β, When the calculation result β<0, the elevation angle F=180+β, The fourth step is to calculate the horizontal distance from the sound source target to the center point of the bottom surface of the microphone array The horizontal distance from the sound source target to the center point of the bottom surface of the microphone array is calculated by the distance calculation model in the host computer, and the calculation formula is: d ₂ ＝L*sin(χ)/sin(180-χ-δ) (6) d ₂ is the horizontal distance from the sound source target to the center point of the bottom surface of the microphone array, the clip between the direction O ₂ O ₁ opposite the moving direction of the small microphone array 2 and the direction O ₂ Z of the sound source target 3 after the small microphone array 2 moves The angle is δ, that is, δ=180-B; The fifth step is to calculate the distance from the sound source target to the center point of the bottom surface of the small microphone array The distance from the sound source target to the center point of the bottom surface of the small microphone array is calculated by the distance calculation model in the host computer, and the calculation formula is: D＝d ₂ /sin(F) (7) D is the distance from the sound source target to the center point of the bottom surface of the small microphone array; The sixth step, display and output of 3D spatial sound source localization data Through the peripherals of the computer, that is, the monitor displays or outputs to other computers through the network card to display the azimuth B of the output sound source target, the elevation angle F of the sound source target, and the horizontal distance d ₂ from the sound source target to the center point of the bottom surface of the microphone array and the distance D from the sound source target to the center point of the bottom surface of the small microphone array, thereby completing three-dimensional sound source localization. 2. The three-dimensional sound source localization method according to claim 1, characterized in that: the radius of the circumscribed circle on the bottom surface of the regular tetrahedron of the small microphone array is 10 cm. 3. The three-dimensional sound source localization method according to claim 1, characterized in that: the software used in the host computer is matlab. 4. The three-dimensional sound source localization method according to claim 1, characterized in that: the range of L is 0.5-1.5m. 5. The three-dimensional sound source localization method according to claim 1, characterized in that: the range of x is 45 degrees to 135 degrees.

Images