CN103364761B - Method for positioning system to position indoor sound source - Google Patents

Abstract

The invention discloses a positioning system of an indoor sound source and a method using the positioning system to position the indoor sound source, and relates to the sound source positioning technology. The positioning system of the indoor sound source and the method using the positioning system to position the indoor sound source are used for solving the problems that a positioning dead zone is formed due to blocking or the direction of sound production of a sound source target, the requirements for positioning accuracy on different application backgrounds are different, complexity of system operation cannot be adjusted in a self-adaptation mode, and accuracy is low. Due to the fact that sound sensors are reasonably distributed in a three-dimensional space, correlation operation is carried out on the time sequence of signals captured by the sound sensors to obtain arrival time differences, the obtained arrival time differences are used for solving a sound source positioning model to obtain the result of three-dimensional positioning of the sound source, and dead zones of sound source positioning can be well prevented due to reasonable distribution of the sound sensors. The positioning system of the indoor sound source and the method using the positioning system to position the indoor sound source are applicable to positioning of sound source targets of a plurality of indoor places.

Description

A kind of method of sound source in indoor sonic location system orientation room

Technical field

The present invention relates to a kind of indoor sonic location system and adopt the method for sound source in this system orientation room.

Background technology

In recent years, auditory localization technology has become a new study hotspot, is with a wide range of applications and practical significance.Particularly for the auditory localization of indoor, application demand is vigorous.Such as wired home amusement/service robot needs when carrying out man-machine interaction to position voice, robot by the visual field that the detection of voice, location changed to oneself towards, can be located it accurately to carry out mutual object by the find a view detection and Identification of laggard pedestrian's face of camera, finally realize voice, video mutual.

For existing auditory localization technology, exist due to block or acoustic target sounding towards the problem of the blind area, location of bringing, to the different requirements of positioning precision under different application background, can not self-adapted adjustment system computing complexity and make that positions calculations amount is excessive, real-time is not high, positioning calculation algorithm is simple and make positioning precision not high, and positioning system physical arrangement is complicated or harsh, be unfavorable for the problems such as practical layout.

Summary of the invention

The present invention be in order to solve existing due to block or acoustic target sounding towards the problem of the blind area, location of bringing, under different application background to the different requirements of positioning precision can not self-adapted adjustment system computing complexity and make that positions calculations amount is excessive, real-time is not high, positioning calculation algorithm is simple and make positioning precision not high, positioning system physical arrangement is complicated or harsh, be unfavorable for the problems such as practical layout, provides a kind of indoor sonic location system and adopts the time difference and the localization method of this system.

The indoor sonic location system of one of the present invention, it comprises first sound sound sensor, second sound sound sensor, 3rd sound transducer, fourth sound sound sensor, fifth sound sound sensor, first signal conditioning circuit, secondary signal modulate circuit, 3rd signal conditioning circuit, 4th signal conditioning circuit, 5th signal conditioning circuit, one A/D modular converter, 2nd A/D modular converter, 3rd A/D modular converter, 4th A/D modular converter, 5th A/D modular converter, sound Sources Detection module, time difference module, locating module, state modulator module and gain control module,

The signal output part of first sound sound sensor connects the signal input part of the first signal conditioning circuit, and the signal output part of described first signal conditioning circuit connects the signal input part of an A/D modular converter; The signal output part of second sound sound sensor connects the signal input part of secondary signal modulate circuit, and the signal output part of described secondary signal modulate circuit connects the signal input part of the 2nd A/D modular converter; The signal output part of the 3rd sound transducer connects the signal input part of the 3rd signal conditioning circuit, and the signal output part of described 3rd signal conditioning circuit connects the signal input part of the 3rd A/D modular converter; The signal output part of fourth sound sound sensor connects the signal input part of the 4th signal conditioning circuit, and the signal output part of described 4th signal conditioning circuit connects the signal input part of the 4th A/D modular converter; The signal output part of fifth sound sound sensor connects the signal input part of the 5th signal conditioning circuit, and the signal output part of described 5th signal conditioning circuit connects the signal input part of the 5th A/D modular converter;

One A/D modular converter, 2nd A/D modular converter, 3rd A/D modular converter, 4th A/D modular converter and the signal output part of the 5th A/D modular converter are connected five sound detection signals input ends of sound Sources Detection module respectively, a sound detection control signal output terminal of described sound Sources Detection module connects a signal input part of time difference module, another sound detection control signal output terminal of sound Sources Detection module connects the control signal input end of gain control module, the control signal output terminal of gain control module connects the gain control signal input end of the first signal conditioning circuit simultaneously, the gain control signal input end of secondary signal modulate circuit, the gain control signal input end of the 3rd signal conditioning circuit, the gain control signal input end of the 4th signal conditioning circuit and the gain control signal input end of the 5th signal conditioning circuit, the signal output part of described time difference module connects a signal input part of locating module,

A signal output part of state modulator module connects another signal input part of time difference module, and another signal output part of state modulator module connects another signal input part of locating module.

Above-mentioned sound Sources Detection module, time difference module, locating module, state modulator module and gain control module can adopt software simulating, such as: can the mode of programming be adopted to be embedded in fpga chip above-mentioned five modules and realize.

A method for sound source in indoor sonic location system orientation room,

Step one, by first sound sound sensor, second sound sound sensor, 3rd sound transducer, fourth sound sound sensor and fifth sound sound sensor are positioned on 5 summits of the indoor virtual rectangular parallelepiped in sound source place to be measured, wherein: first sound sound sensor, second sound sound sensor and the 3rd sound transducer lay respectively at three summit O of rectangular parallelepiped end face, A and B place, these three summit O, A and B forms a right-angle triangle, the summit of first sound sound sensor place summit O corresponding to the right angle of this right-angle triangle, the rectangular parallelepiped bottom surface summit C being positioned at same limit with summit A places fourth sound sound sensor, the rectangular parallelepiped bottom surface summit F being positioned at same limit with summit B places fifth sound sound sensor,

Using summit O as the line of true origin, summit O and summit B as X-axis, using the line of summit O and summit A as Y-axis, form right-handed helix rectangular coordinate system in space, the length of OA is D01, the length of the length of OB to be the length of D02, AC be D13, BF is D24, D01 is greater than 0 and is less than 20m, D02 is greater than 0 and is less than 20m, and D13 is greater than 0 and is less than 5m, and D24 is greater than 0 and is less than 5m and D13=D24;

Step 2, time difference generalized correlation method is utilized to obtain four delay inequalities after same sound-source signal different delay according to the voice signal of five sound transducer collections, described four delay inequalities are: the voice signal that first sound sound sensor is caught and the voice signal that second sound sound sensor is caught carry out the delay inequality of related operation acquisition, the voice signal that first sound sound sensor is caught and the voice signal that the 3rd sound transducer is caught carry out the delay inequality of related operation acquisition, the voice signal that second sound sound sensor is caught and the voice signal that fourth sound sound sensor is caught carry out the delay inequality of related operation acquisition, the voice signal that 3rd sound transducer is caught and the voice signal that fifth sound sound sensor is caught carry out the delay inequality of auto-correlation computation acquisition,

Time stepped parameter is obtained, by time stepping Parameter transfer to time difference module after the positioning precision parameter that state modulator module sets calculates; Iteration error parameter and the iterations parameter of the setting of state modulator module carry out calculating iteration error threshold value and iterations threshold value, and iteration error threshold value and iterations threshold value are passed to locating module; The space position parameter of state modulator module setting passes to locating module;

With time stepped parameter obtained above, iteration error threshold value and iterations threshold value, the voice signal that first sound sound sensor and second sound sound sensor are caught is carried out related operation and obtains correlation, search maximum related value, associated time lengths corresponding to this maximum related value is exactly the delay inequality between first sound sound sensor and second sound sound sensor; The voice signal that first sound sound sensor and the 3rd sound transducer are caught carries out related operation and obtains correlation, search maximum related value, associated time lengths corresponding to this maximum related value is exactly the delay inequality between first sound sound sensor and the 3rd sound transducer; The voice signal that second sound sound sensor and fourth sound sound sensor are caught carries out related operation and obtains correlation, search maximum related value, associated time lengths corresponding to this maximum related value is exactly the delay inequality between second sound sound sensor and fourth sound sound sensor; The voice signal that 3rd sound transducer and fifth sound sound sensor are caught carries out related operation and obtains correlation, search maximum related value, associated time lengths corresponding to this maximum related value is exactly the delay inequality between the 3rd sound transducer and fifth sound sound sensor;

Step 3, utilize quasi-Newton method to obtain according to step 2 same sound-source signal different delay after the locus distribution of four delay inequalities, voice signal that five sound transducers gather and five sound transducers obtain sound source position,

The mistiming arriving each sensor according to the locus distribution of sensor and sound source forms positioning equation group, by iteration, obtains positioning solution, calculates the Euclidean distance of the positioning result of front and back twice;

When the positioning result Euclidean distance of twice is less than the threshold value of Operation system setting, iterative solution has converged to sound source position; Otherwise when iterations is greater than predetermined threshold value, the voice signal of Resurvey five sound transducers, repeat step 2, until obtain sound source position.

The present invention can solve due to block or acoustic target sounding towards the problem of the blind area, location of bringing; The present invention can according to the complexity under different application background, the difference of positioning precision being required to self-adapted adjustment system computing, makes the speed of positioning calculation reach the fastest and system power dissipation reaches minimum when positioning performance meets practical application request; The three-dimensional localization solution to acoustic target can be obtained; In voice collecting process, automatically can adjust the gain of each road signal, both remain the positional information be hidden in the signal amplitude of each road, turn avoid the too low positioning precision loss brought of signal amplitude.

Accompanying drawing explanation

Fig. 1 is one-piece construction schematic diagram of the present invention; Fig. 2 is the inner structure schematic diagram of signal conditioning circuit of the present invention;

Fig. 3 is sound transducer space structure figure of the present invention; Fig. 4 is software flow pattern of the present invention.

Embodiment

Embodiment one: present embodiment is described below in conjunction with Fig. 1, a kind of indoor sonic location system described in present embodiment, it comprises first sound sound sensor S ₀, second sound sound sensor S ₁, the 3rd sound transducer S ₂, fourth sound sound sensor S ₃, fifth sound sound sensor S ₄, the first signal conditioning circuit 1, secondary signal modulate circuit 2, the 3rd signal conditioning circuit 3, the 4th signal conditioning circuit 4, the 5th signal conditioning circuit 5, an A/D modular converter 6, the 2nd A/D modular converter 7, the 3rd A/D modular converter 8, the 4th A/D modular converter 9, the 5th A/D modular converter 10, sound Sources Detection module 11, time difference module 12, locating module 13, state modulator module 14 and gain control module 15;

First sound sound sensor S ₀signal output part connect the signal input part of the first signal conditioning circuit 1, the signal output part of described first signal conditioning circuit 1 connects the signal input part of an A/D modular converter 6; Second sound sound sensor S ₁signal output part connect the signal input part of secondary signal modulate circuit 2, the signal output part of described secondary signal modulate circuit 2 connects the signal input part of the 2nd A/D modular converter 7; 3rd sound transducer S ₂signal output part connect the signal input part of the 3rd signal conditioning circuit 3, the signal output part of described 3rd signal conditioning circuit 3 connects the signal input part of the 3rd A/D modular converter 8; Fourth sound sound sensor S ₃signal output part connect the signal input part of the 4th signal conditioning circuit 4, the signal output part of described 4th signal conditioning circuit 4 connects the signal input part of the 4th A/D modular converter 9; Fifth sound sound sensor S ₄signal output part connect the signal input part of the 5th signal conditioning circuit 5, the signal output part of described 5th signal conditioning circuit 5 connects the signal input part of the 5th A/D modular converter 10;

One A/D modular converter 6, 2nd A/D modular converter 7, 3rd A/D modular converter 8, 4th A/D modular converter 9 and the signal output part of the 5th A/D modular converter 10 are connected five sound detection signals input ends of sound Sources Detection module 11 respectively, a sound detection control signal output terminal of described sound Sources Detection module 11 connects a signal input part of time difference module 12, another sound detection control signal output terminal of sound Sources Detection module 11 connects the control signal input end of gain control module 15, the control signal output terminal of gain control module 15 connects the gain control signal input end of the first signal conditioning circuit 1 simultaneously, the gain control signal input end of secondary signal modulate circuit 2, the gain control signal input end of the 3rd signal conditioning circuit 3, the gain control signal input end of the 4th signal conditioning circuit 4 and the gain control signal input end of the 5th signal conditioning circuit 5, the signal output part of described time difference module 12 connects a signal input part of locating module 13,

A signal output part of state modulator module 14 connects another signal input part of time difference module 12, and another signal output part of state modulator module 14 connects another signal input part of locating module 13.

Embodiment two: present embodiment is described below in conjunction with Fig. 2, present embodiment is to the further restriction of a kind of indoor sonic location system described in embodiment one, in present embodiment, described secondary signal modulate circuit 2, the 3rd signal conditioning circuit 3, the 4th signal conditioning circuit 4, the 5th signal conditioning circuit 5 are identical with the internal circuit configuration of the first signal conditioning circuit 1, and wherein the first signal conditioning circuit 1 is made up of filtering circuit, PGA circuit and signal holding circuit;

The signal input part of filtering circuit is the signal input part of the first signal conditioning circuit 1, the signal output part of described filtering circuit connects the signal input part of PGA circuit, the signal input part of the signal output part connection signal holding circuit of described PGA circuit, the signal output part of described signal holding circuit is the signal output part of the first signal conditioning circuit 1.

Embodiment three: present embodiment is to the further restriction of a kind of indoor sonic location system described in embodiment two, and in present embodiment, described PGA circuit is programme-controlled gain amplifying circuit.

Sound Sources Detection module 11 determines the parameter of PGA circuit being carried out to programming Control by the signal level analyzing every road signal, guarantee that the enlargement factor of every road signal is the same and the signal amplitude of signal Qiang Na road signal reaches the maximal value of AD conversion, this remains the positional information of the sound source in the amplitude of lying in.

Embodiment four: composition graphs 3 and Fig. 4 illustrate present embodiment, present embodiment to the method for sound source in indoor sonic location system orientation room a kind of described in any one embodiment of embodiment one to three, step one, by first sound sound sensor S ₀, second sound sound sensor S ₁, the 3rd sound transducer S ₂, fourth sound sound sensor S ₃with fifth sound sound sensor S ₄be positioned on 5 summits of the indoor virtual rectangular parallelepiped Cube in sound source place to be measured, wherein: first sound sound sensor S ₀, second sound sound sensor S ₁with the 3rd sound transducer S ₂lay respectively at O, A and B place, three summits of rectangular parallelepiped Cube end face, these three summits O, A and B form a right-angle triangle, first sound sound sensor S ₀the summit of summit, place O corresponding to the right angle of this right-angle triangle, the rectangular parallelepiped Cube bottom surface summit C being positioned at same limit with summit A places fourth sound sound sensor S ₃, the rectangular parallelepiped Cube bottom surface summit F being positioned at same limit with summit B places fifth sound sound sensor S ₄;

Using summit O as the line of true origin, summit O and summit B as X-axis, using the line of summit O and summit A as Y-axis, form right-handed helix rectangular coordinate system in space, the length of OA is D01, the length of the length of OB to be the length of D02, AC be D13, BF is D24, D01 is greater than 0 and is less than 20m, D02 is greater than 0 and is less than 20m, and D13 is greater than 0 and is less than 5m, and D24 is greater than 0 and is less than 5m and D13=D24;

Step 2, utilize time difference generalized correlation method to obtain four delay inequalities after same sound-source signal different delay according to the voice signal of five sound transducer collections, described four delay inequalities are: first sound sound sensor S ₀the voice signal of catching and second sound sound sensor S ₁the voice signal of catching carries out delay inequality, the first sound sound sensor S of related operation acquisition ₀the voice signal of catching and the 3rd sound transducer S ₂the voice signal of catching carries out delay inequality, the second sound sound sensor S of related operation acquisition ₁the voice signal of catching and fourth sound sound sensor S ₃the voice signal of catching carries out delay inequality, the 3rd sound transducer S of related operation acquisition ₂the voice signal of catching and fifth sound sound sensor S ₄the voice signal of catching carries out the delay inequality of auto-correlation computation acquisition;

Time stepped parameter is obtained, by time stepping Parameter transfer to time difference module 12 after the positioning precision parameter that state modulator module 14 sets calculates; The iteration error parameter that state modulator module 14 sets and iterations parameter carry out calculating iteration error threshold value and iterations threshold value, and iteration error threshold value and iterations threshold value are passed to locating module 13; The space position parameter that state modulator module 14 sets passes to locating module 13;

With time stepped parameter obtained above, iteration error threshold value and iterations threshold value, by first sound sound sensor S ₀with second sound sound sensor S ₁the voice signal of catching carries out related operation and obtains correlation, and search maximum related value, associated time lengths corresponding to this maximum related value is exactly first sound sound sensor S ₀with second sound sound sensor S ₁between delay inequality; First sound sound sensor S ₀with the 3rd sound transducer S ₂the voice signal of catching carries out related operation and obtains correlation, and search maximum related value, associated time lengths corresponding to this maximum related value is exactly first sound sound sensor S ₀with the 3rd sound transducer S ₂between delay inequality; Second sound sound sensor S ₁with fourth sound sound sensor S ₃the voice signal of catching carries out related operation and obtains correlation, and search maximum related value, associated time lengths corresponding to this maximum related value is exactly second sound sound sensor S ₁with fourth sound sound sensor S ₃between delay inequality; 3rd sound transducer S ₂with fifth sound sound sensor S ₄the voice signal of catching carries out related operation and obtains correlation, and search maximum related value, associated time lengths corresponding to this maximum related value is exactly the 3rd sound transducer S ₂with fifth sound sound sensor S ₄between delay inequality;

Step 3, utilize quasi-Newton method to obtain according to step 2 same sound-source signal different delay after the locus distribution of four delay inequalities, voice signal that five sound transducers gather and five sound transducers obtain sound source position,

The mistiming arriving each sensor according to the locus distribution of sensor and sound source forms positioning equation group, by iteration, obtains positioning solution, calculates the Euclidean distance of the positioning result of front and back twice;

When the positioning result Euclidean distance of twice is less than the iteration error threshold value described in step 2, iterative solution has converged to sound source position; Otherwise when iterations is greater than the iterations threshold value described in step 2, the voice signal of Resurvey five sound transducers, repeats step 2, until obtain sound source position.

Embodiment five: present embodiment is to the further restriction described in embodiment four, the method of sound source in a kind of indoor sonic location system orientation room described in present embodiment, described in step 2 utilize time difference generalized correlation method to obtain same sound-source signal different delay according to the voice signal of five sound transducer collections after the detailed process of four delay inequalities be:

By sound transducer S _ireceived signal strength r _i(t) and sensor S _jreceived signal strength r _jt the signal of () adopts formula (0.1) to represent:

$\left\{\begin{array}{c}{r}_i\left(t\right)=A_{i}s(t-{\mathrm{\τ}}_i)+w_i\left(t\right)\\ r_j\left(t\right)=A_{j}s(t-{\mathrm{\τ}}_j)+w_j\left(t\right)\end{array}\right.---\left(0.1\right)$

In formula (0.1), the voice signal that s (t) sends for sound source; The fourier function that S (ω) is s (t), A _ifor sound transducer S _ithe decay of propagation channel, A _jfor sensor S _jthe decay of propagation channel; w _it () is room background and S _ithe circuit noise sum of road signal, w _jt () is room background and S _jthe circuit noise sum of road signal, i, j=0,1,2,3,4 and i and j is unequal;

By signal r _i(t) and signal r _jt () carries out related operation, obtain r _i(t) and r _j(t) correlation result

$\begin{array}{c}{R}_{r_i{r}_j}\left(\mathrm{\τ}\right)=E[r_i\left(t\right)\·r_j^{*}(t+\mathrm{\τ})]\\ ={\∫}_{-\∞}^{+\∞}{r}_i\left(t\right)\·r_j^{*}(t+\mathrm{\τ})\mathrm{dt}\\ =A_i{A}_j\·R_{\mathrm{ss}}[\mathrm{\τ}-({\mathrm{\τ}}_j-{\mathrm{\τ}}_i)]+A_i\·R_{w_{i}s}(\mathrm{\τ}-{\mathrm{\τ}}_j)+A_i\·R_{{\mathrm{sw}}_j}(\mathrm{\τ}+{\mathrm{\τ}}_i)+R_{w_i{w}_j}\left(\mathrm{\τ}\right)\end{array}---(0.2)$

In formula (0.2), τ is the associated time lengths of related operation, τ _i, τ _jbe respectively S _iroad signal and S _jthe propagation delay time of road signal path, i, j=0,1,2,3,4 and i and j is unequal, wherein $\left\{\begin{array}{c}{R}_{w_{i}s}(\mathrm{\τ}-{\mathrm{\τ}}_j)=0\\ R_{{\mathrm{sw}}_j}(\mathrm{\τ}+{\mathrm{\τ}}_i)=0\\ R_{w_i{w}_j}\left(\mathrm{\τ}\right)=0\end{array}\right.,$ Then have:

$R_{r_i{r}_j}\left(\mathrm{\τ}\right)=A_i{A}_j\·R_{\mathrm{ss}}[\mathrm{\τ}-({\mathrm{\τ}}_j-{\mathrm{\τ}}_i)]---\left(0.3\right)$

R in formula (0.3) _ss[τ-(τ _j-τ _i)] obtained by following formula:

$\begin{array}{c}{R}_{\mathrm{ss}}[\mathrm{\τ}-({\mathrm{\τ}}_j-{\mathrm{\τ}}_i)]={\∫}_{-\∞}^{+\∞}s\left(t\right)\·s^{*}(t+\mathrm{\τ}-({\mathrm{\τ}}_j-{\mathrm{\τ}}_i))\mathrm{dt}\\ ={\∫}_{-\∞}^{+\∞}\left(\frac{1}{2\mathrm{\π}}{\∫}_{-\∞}^{+\∞}S\left(\mathrm{\ω}\right)e^{\mathrm{j\ωt}}\mathrm{d\ω}\right)\·s^{*}(t+\mathrm{\τ}-{\mathrm{\τ}}_j+{\mathrm{\τ}}_i)\mathrm{dt}\\ =\frac{1}{2\mathrm{\π}}{\∫}_{-\∞}^{+\∞}S\left(\mathrm{\ω}\right){\left({\∫}_{-\∞}^{+\∞}s(t+\mathrm{\τ}-{\mathrm{\τ}}_j+{\mathrm{\τ}}_i)e^{-\mathrm{j\ωt}}\mathrm{dt}\right)}^{*}\mathrm{d\ω}\\ =\frac{1}{2\mathrm{\π}}{\∫}_{-\∞}^{+\∞}S\left(\mathrm{\ω}\right)S^{*}\left(\mathrm{\ω}\right)e^{\mathrm{j\ω}({\mathrm{\τ}}_j-{\mathrm{\τ}}_i-\mathrm{\τ})}\mathrm{d\ω}\\ =\frac{1}{2\mathrm{\π}}{\∫}_{-\∞}^{+\∞}{\left|S\left(\mathrm{\ω}\right)\right|}^2\·e^{\mathrm{j\ω}({\mathrm{\τ}}_j-{\mathrm{\τ}}_i-\mathrm{\τ})}\mathrm{d\ω}\\ \≤\frac{1}{2\mathrm{\π}}{\∫}_{-\∞}^{+\∞}{\left|S\left(\mathrm{\ω}\right)\right|}^2\·\left|e^{\mathrm{j\ω}({\mathrm{\τ}}_j-{\mathrm{\τ}}_i-\mathrm{\τ})}\right|\mathrm{d\ω}\\ =\frac{1}{2\mathrm{\π}}{\∫}_{-\∞}^{+\∞}{\left|S\left(\mathrm{\ω}\right)\right|}^2\mathrm{d\ω}\end{array}---\left(0.4\right)$

As τ=τ _j-τ _itime, R _ss[τ-(τ _j-τ _i)] get maximal value, namely get maximal value; Adjustment associated time lengths obtains corresponding correlation, and search maximum related value, associated time lengths τ corresponding to this maximum related value is exactly signal r _i(t) and signal r _jthe delay inequality of (t).

Embodiment six: present embodiment is the further restriction to embodiment four, the method of sound source in a kind of indoor sonic location system orientation room described in present embodiment, the detailed process that the locus distribution of four delay inequalities after the same sound-source signal different delay utilizing quasi-Newton method to obtain according to step 2 described in step 3, the voice signal of five sound transducer collections and five sound transducers obtains sound source position is:

Gather sound transducer S ₀~ S ₄the voice signal of catching, supposes sound transducer S ₀~ S ₂and S ₃or S ₄gather and all acoustic target detected, comparing fourth sound sound sensor S ₃with fifth sound sound sensor S ₄the amplitude of the voice signal of catching;

Work as S ₃the amplitude of the voice signal of catching is greater than S ₄the amplitude of the voice signal of catching;

Positioning equation group (1.1) is set up to the locus of sound source;

$\left\{\begin{array}{c}\sqrt{{(x-x_1)}^2+{(y-y_1)}^2+{(z-z_1)}^2}-\sqrt{{(x-x_0)}^2+{(y-y_0)}^2+{(z-z_0)}^2}=v\·\mathrm{\Δ}{T}_{01}\\ \sqrt{{(x-x_2)}^2+{(y-y_2)}^2+{(z-z_2)}^2}-\sqrt{{(x-x_0)}^2+{(y-y_0)}^2+{(z-z_0)}^2}=v\·\mathrm{\Δ}{T}_{02}\\ \sqrt{{(x-x_3)}^2+{(y-y_3)}^2+{(z-z_3)}^2}-\sqrt{{(x-x_1)}^2+{(y-y_1)}^2+{(z-z_1)}^2}=v\·\mathrm{\Δ}{T}_{13}\end{array}\right.---\left(1.1\right)$

(x, y, z) represents the locus coordinate of acoustic target, (x _i, y _i, z _i) represent S _ilocus coordinate, i=0,1,2,3,4, △ T ₀₁for S ₀and S ₁the delay inequality that the voice signal related operation of catching obtains, τ ₀₁=△ T ₀₁=τ ₁-τ ₀, △ T ₀₂for S ₀and S ₂the delay inequality that the voice signal related operation of catching obtains, τ ₀₂=△ T ₀₂=τ ₂-τ ₀, △ T ₁₃for S ₁and S ₃the delay inequality that the voice signal related operation of catching obtains, τ ₁₃=△ T ₁₃=τ ₃-τ ₁, △ T ₂₄for S ₂and S ₄the delay inequality that the voice signal related operation of catching obtains, τ ₂₄=△ T ₂₄=τ ₄-τ ₂, v is sound propagation velocity, and meets the known conditions shown in formula (1.2);

$\left\{\begin{array}{c}(x_0,y_0,z_0)=\left(\mathrm{0,0,0}\right)\\ (x_1,y_1,z_1)=(0,y_1,0)\\ (x_2,y_2,z_2)=(x_2,\mathrm{0,0})\\ (x_3,y_3,z_3)=(0,y_3,z_3)\end{array}\right.---\left(1.2\right)$

Positioning equation group (1.1) is Nonlinear System of Equations,

Defined function is as shown in (1.3);

$\left\{\begin{array}{c}{f}_1=\sqrt{{(x-x_1)}^2+{(y-y_1)}^2+{(z-z_1)}^2}-\sqrt{{(x-x_0)}^2+{(y-y_0)}^2+{(z-z_0)}^2}-v\·\mathrm{\Δ}{T}_{01}\\ f_2=\sqrt{{(x-x_1)}^2+{(y-y_1)}^2+{(z-z_1)}^2}-\sqrt{{(x-x_0)}^2+{(y-y_0)}^2+{(z-z_0)}^2}-v\·\mathrm{\Δ}{T}_{02}\\ f_3=\sqrt{{(x-x_3)}^2+{(y-y_3)}^2+{(z-z_3)}^2}-\sqrt{{(x-x_1)}^2+{(y-y_1)}^2+{(z-z_1)}^2}-v\·\mathrm{\Δ}{T}_{13}\end{array}\right.---\left(1.3\right)$

Build vectorial F (x)=(f ₁, f ₂, f ₃) ^t, x=(x, y, z) ^t, and this vector meets equation shown in formula (1.4);

F(x)＝(f ₁,f ₂,f ₃) ^T＝0 (1.4)

The Jacobi matrix of formula (1.4) is (1.5);

$A_i=F^{\′}\left(x^i\right)=\left[\begin{array}{ccc}\frac{\∂f_1}{\∂x}& \frac{\∂f_1}{\∂y}& \frac{\∂f_1}{\∂z}\\ \frac{\∂f_2}{\∂x}& \frac{\∂f_2}{\∂y}& \frac{\∂f_2}{\∂z}\\ \frac{\∂f_3}{\∂x}& \frac{\∂f_3}{\∂y}& \frac{\∂f_3}{\∂z}\end{array}\right]---\left(1.5\right)$

Then the iterative formula of solving equations (1.1) is formula (1.6);

$x^{i+1}=x^i-A_i^{-1}F\left(x^i\right)---\left(1.6\right)$

Iteration error threshold value E described in step 2 _th, as the Euclidean distance D of the positioning result of front and back twice _e=|| x ⁱ⁺¹-x ⁱ|| ₂<E _thtime iteration exit, obtain positioning solution; Otherwise when iterations is greater than the iterations threshold value T described in step 2 _th, the voice signal of Resurvey five sound transducers, repeats step 2, until obtain sound source position;

The amplitude of the voice signal of catching as S3 is less than S ₄the amplitude of the voice signal of catching;

According to the rectangular coordinate system in space that the present invention sets up, positioning equation group (2.1) can be set up to the locus of sound source;

$\left\{\begin{array}{c}\sqrt{{(x-x_1)}^2+{(y-y_1)}^2+{(z-z_1)}^2}-\sqrt{{(x-x_0)}^2+{(y-y_0)}^2+{(z-z_0)}^2}=v\&CenterDot;\mathrm{\&Delta;}{T}_{01}\\ \sqrt{{(x-x_2)}^2+{(y-y_2)}^2+{(z-z_2)}^2}-\sqrt{{(x-x_0)}^2+{(y-y_0)}^2+{(z-z_0)}^2}=v\&CenterDot;\mathrm{\&Delta;}{T}_{02}\\ \sqrt{{(x-x_4)}^2+{(y-y_4)}^2+{(z-z_4)}^2}-\sqrt{{(x-x_2)}^2+{(y-y_2)}^2+{(z-z_2)}^2}=v\&CenterDot;\mathrm{\&Delta;}{T}_{24}\end{array}\right.---\left(2.1\right)$

And meet the known conditions shown in formula (2.2);

$\left\{\begin{array}{c}(x_0,y_0,z_0)=\left(\mathrm{0,0,0}\right)\\ (x_1,y_1,z_1)=(0,y_1,0)\\ (x_2,y_2,z_2)=(x_2,\mathrm{0,0})\\ (x_4,y_4,z_4)=(x_4,0,z_4)\end{array}\right.---(.22)$

Positioning equation group (2.1) is Nonlinear System of Equations,

Positioning equation group is rewritten into vector form:

Defined function is as shown in (2.3);

$\left\{\begin{array}{c}{f}_1=\sqrt{{(x-x_1)}^2+{(y-y_1)}^2+{(z-z_1)}^2}-\sqrt{{(x-x_0)}^2+{(y-y_0)}^2+{(z-z_0)}^2}-v\&CenterDot;\mathrm{\&Delta;}{T}_{01}\\ f_2=\sqrt{{(x-x_1)}^2+{(y-y_1)}^2+{(z-z_1)}^2}-\sqrt{{(x-x_0)}^2+{(y-y_0)}^2+{(z-z_0)}^2}-v\&CenterDot;\mathrm{\&Delta;}{T}_{02}\\ f_3=\sqrt{{(x-x_4)}^2+{(y-y_4)}^2+{(z-z_4)}^2}-\sqrt{{(x-x_2)}^2+{(y-y_2)}^2+{(z-z_2)}^2}-v\&CenterDot;\mathrm{\&Delta;}{T}_{24}\end{array}\right.---(.23)$

Build vectorial F (x)=(f ₁, f ₂, f ₃) ^t, x=(x, y, z) ^t, and this vector meets equation shown in formula (2.4);

F(x)＝(f ₁,f ₂,f ₃) ^T＝0 (2.4)

The Jacobi matrix of formula (2.4) is (2.5);

$A_i=F^{\&prime;}\left(x^i\right)=\left[\begin{array}{ccc}\frac{\&PartialD;f_1}{\&PartialD;x}& \frac{\&PartialD;f_1}{\&PartialD;y}& \frac{\&PartialD;f_1}{\&PartialD;z}\\ \frac{\&PartialD;f_2}{\&PartialD;x}& \frac{\&PartialD;f_2}{\&PartialD;y}& \frac{\&PartialD;f_2}{\&PartialD;z}\\ \frac{\&PartialD;f_3}{\&PartialD;x}& \frac{\&PartialD;f_3}{\&PartialD;y}& \frac{\&PartialD;f_3}{\&PartialD;z}\end{array}\right]---\left(2.5\right)$

Then the iterative formula of solving equations (2.1) is formula (2.6);

$x^{i+1}=x^i-A_i^{-1}F\left(x^i\right)---(2.6)$

Iteration error threshold value E described in step 2 _th, as the Euclidean distance D of the positioning result of front and back twice _e=|| x ⁱ⁺¹-x ⁱ|| ₂<E _thtime iteration exit, obtain positioning solution; Otherwise when iterations is greater than the iterations threshold value T described in step 2 _th, the voice signal of Resurvey five sound transducers, repeats step 2, until obtain sound source position.

The plan Newton process that said process adopts is the mistiming formation positioning equation group arriving each sensor according to the distribution of the locus of sensor and sound source target utterance, intend Newton process to solve positioning equation group, key step has: initial position elects the position at the sensor place of amplitude maximum in 5 collecting sensor signal results as; By iteration, obtain positioning solution, the Euclidean distance of the positioning result that calculating front and back are twice; When the positioning result Euclidean distance of twice is less than the iteration error threshold value described in step 2, think that iterative solution has converged to sound source position; Otherwise when iterations is greater than the iterations threshold value described in step 2, the voice signal of Resurvey five sound transducers, repeats step 2, until obtain sound source position.

The principal character intending the sound localization method that Newton process proposes comprises: obtain time stepped parameter, by time stepping Parameter transfer to time difference module 12 after the positioning precision parameter that state modulator module 14 sets calculates; The iteration error parameter that state modulator module 14 sets and iterations parameter carry out calculating iteration error threshold value and iterations threshold value, and iteration error threshold value and iterations threshold value are passed to locating module 13; The space position parameter that state modulator module 14 sets passes to locating module 13; Stepped parameter is less, time difference operation precision higher, auditory localization precision is higher, and system operations amount is larger; Stepped parameter is larger, time difference operation precision lower, auditory localization precision is lower, and system operations amount is less; According to the requirement of practical application to positioning precision, stepped parameter is rationally set and can makes this algorithm on computational complexity, reach minimum; When quasi_Kantorovich operator solves positioning equation group, according to the requirement of practical application to positioning precision, the positioning result distance threshold rationally arranging front and back twice can make algorithm on computational complexity, reach minimum.

The present invention's " embodiment " part symbol definition used is as follows:

S _i, S _j; Sound transducer numbering i, j=0,1,2,3,4;

S (t); The signal model of sound source;

R _i(t); Sound transducer S _ithe signal received;

R _j(t); Sound transducer S _jthe signal received;

W _i(t); Room background and S _ithe circuit noise sum of road signal;

W _j(t); Room background and S _jthe circuit noise sum of road signal;

A _i; S _iroad signal propagation channel decay;

A _j; S _jroad signal propagation channel decay;

r _i(t) and r _j(t) correlation result;

w _i(t) and s (t) correlation result;

s (t) and w _j(t) correlation result;

w _i(t) and w _j(t) correlation result;

τ; The associated time lengths of related operation;

τ _i, τ _j; S _iroad signal and S _jroad signal path propagation delay time

i,j＝0,1,2,3,4；

△ T ₀₁; S ₀and S ₁difference estimated value time of arrival that related operation obtains;

△ T ₀₂; S ₀and S ₂difference estimated value time of arrival that related operation obtains;

△ T ₁₃; S ₁and S ₃difference estimated value time of arrival that related operation obtains;

△ T ₂₄; S ₂and S ₄difference estimated value time of arrival that related operation obtains;

(x, y, z); Represent the locus coordinate of acoustic target;

(x _i, y _i, z _i); Represent S _ilocus coordinate i=0,1,2,3,4;

V; Sound propagation velocity;

S (ω); The fourier function of s (t);

D _e; The Euclidean distance of the positioning result that front and back are twice;

E _th; Iteration error threshold value;

T _th; Iterations threshold value.

Claims (3)

1. the method for sound source in indoor sonic location system orientation room, and adopt and realize with lower device, this device comprises first sound sound sensor (S ₀), second sound sound sensor (S ₁), the 3rd sound transducer (S ₂), fourth sound sound sensor (S ₃), fifth sound sound sensor (S ₄), first signal conditioning circuit (1), secondary signal modulate circuit (2), 3rd signal conditioning circuit (3), 4th signal conditioning circuit (4), 5th signal conditioning circuit (5), one A/D modular converter (6), 2nd A/D modular converter (7), 3rd A/D modular converter (8), 4th A/D modular converter (9), 5th A/D modular converter (10), sound Sources Detection module (11), time difference module (12), locating module (13), state modulator module (14) and gain control module (15),

First sound sound sensor (S ₀) signal output part connect the signal input part of the first signal conditioning circuit (1), the signal output part of described first signal conditioning circuit (1) connects the signal input part of an A/D modular converter (6); Second sound sound sensor (S ₁) signal output part connect the signal input part of secondary signal modulate circuit (2), the signal output part of described secondary signal modulate circuit (2) connects the signal input part of the 2nd A/D modular converter (7); 3rd sound transducer (S ₂) signal output part connect the signal input part of the 3rd signal conditioning circuit (3), the signal output part of described 3rd signal conditioning circuit (3) connects the signal input part of the 3rd A/D modular converter (8); Fourth sound sound sensor (S ₃) signal output part connect the signal input part of the 4th signal conditioning circuit (4), the signal output part of described 4th signal conditioning circuit (4) connects the signal input part of the 4th A/D modular converter (9); Fifth sound sound sensor (S ₄) signal output part connect the signal input part of the 5th signal conditioning circuit (5), the signal output part of described 5th signal conditioning circuit (5) connects the signal input part of the 5th A/D modular converter (10);

One A/D modular converter (6), 2nd A/D modular converter (7), 3rd A/D modular converter (8), 4th A/D modular converter (9) is connected five sound detection signals input ends of sound Sources Detection module (11) respectively with the signal output part of the 5th A/D modular converter (10), a sound detection control signal output terminal of described sound Sources Detection module (11) connects a signal input part of time difference module (12), another sound detection control signal output terminal of sound Sources Detection module (11) connects the control signal input end of gain control module (15), the control signal output terminal of gain control module (15) connects the gain control signal input end of the first signal conditioning circuit (1) simultaneously, the gain control signal input end of secondary signal modulate circuit (2), the gain control signal input end of the 3rd signal conditioning circuit (3), the gain control signal input end of the 4th signal conditioning circuit (4) and the gain control signal input end of the 5th signal conditioning circuit (5), the signal output part of described time difference module (12) connects a signal input part of locating module (13),

A signal output part of state modulator module (14) connects another signal input part of time difference module (12), and another signal output part of state modulator module (14) connects another signal input part of locating module (13);

Secondary signal modulate circuit (2), the 3rd signal conditioning circuit (3), the 4th signal conditioning circuit (4), the 5th signal conditioning circuit (5) are identical with the internal circuit configuration of the first signal conditioning circuit (1), and wherein the first signal conditioning circuit (1) is made up of filtering circuit, PGA circuit and signal holding circuit;

The signal input part of filtering circuit is the signal input part of the first signal conditioning circuit (1), the signal output part of described filtering circuit connects the signal input part of PGA circuit, the signal input part of the signal output part connection signal holding circuit of described PGA circuit, the signal output part of described signal holding circuit is the signal output part of the first signal conditioning circuit (1), it is characterized in that:

Step one, by first sound sound sensor (S ₀), second sound sound sensor (S ₁), the 3rd sound transducer (S ₂), fourth sound sound sensor (S ₃) and fifth sound sound sensor (S ₄) be positioned on 5 summits of the indoor virtual rectangular parallelepiped (Cube) in sound source place to be measured, wherein: first sound sound sensor (S ₀), second sound sound sensor (S ₁) and the 3rd sound transducer (S ₂) laying respectively at O, A and B place, three summits of rectangular parallelepiped (Cube) end face, these three summits O, A and B form a right-angle triangle, first sound sound sensor (S ₀) summit of summit, place O corresponding to the right angle of this right-angle triangle, rectangular parallelepiped (Cube) summit, the bottom surface C being positioned at same limit with summit A places fourth sound sound sensor (S ₃), rectangular parallelepiped (Cube) summit, the bottom surface F being positioned at same limit with summit B places fifth sound sound sensor (S ₄);

Using summit O as the line of true origin, summit O and summit B as X-axis, using the line of summit O and summit A as Y-axis, form right-handed helix rectangular coordinate system in space, the length of OA is D01, the length of the length of OB to be the length of D02, AC be D13, BF is D24, D01 is greater than 0 and is less than 20m, D02 is greater than 0 and is less than 20m, and D13 is greater than 0 and is less than 5m, and D24 is greater than 0 and is less than 5m and D13=D24;

Step 2, utilize time difference generalized correlation method to obtain four delay inequalities after same sound-source signal different delay according to the voice signal of five sound transducer collections, described four delay inequalities are: first sound sound sensor (S ₀) voice signal of catching and second sound sound sensor (S ₁) voice signal of catching carries out delay inequality, the first sound sound sensor (S of related operation acquisition ₀) voice signal of catching and the 3rd sound transducer (S ₂) voice signal of catching carries out delay inequality, the second sound sound sensor (S of related operation acquisition ₁) voice signal of catching and fourth sound sound sensor (S ₃) voice signal of catching carries out delay inequality, the 3rd sound transducer (S of related operation acquisition ₂) voice signal of catching and fifth sound sound sensor (S ₄) voice signal of catching carries out the delay inequality of related operation acquisition;

Time stepped parameter is obtained, by time stepping Parameter transfer to time difference module (12) after the positioning precision parameter that state modulator module (14) sets calculates; The iteration error parameter that state modulator module (14) sets and iterations parameter carry out calculating iteration error threshold value and iterations threshold value, and iteration error threshold value and iterations threshold value are passed to locating module (13); The space position parameter that state modulator module (14) sets passes to locating module (13);

With time stepped parameter obtained above, iteration error threshold value and iterations threshold value, by first sound sound sensor (S ₀) and second sound sound sensor (S ₁) voice signal of catching carries out related operation and obtain correlation, search maximum related value, associated time lengths corresponding to this maximum related value is exactly first sound sound sensor (S ₀) and second sound sound sensor (S ₁) between delay inequality; First sound sound sensor (S ₀) and the 3rd sound transducer (S ₂) voice signal of catching carries out related operation and obtain correlation, search maximum related value, associated time lengths corresponding to this maximum related value is exactly first sound sound sensor (S ₀) and the 3rd sound transducer (S ₂) between delay inequality; Second sound sound sensor (S ₁) and fourth sound sound sensor (S ₃) voice signal of catching carries out related operation and obtain correlation, search maximum related value, associated time lengths corresponding to this maximum related value is exactly second sound sound sensor (S ₁) and fourth sound sound sensor (S ₃) between delay inequality; 3rd sound transducer (S ₂) and fifth sound sound sensor (S ₄) voice signal of catching carries out related operation and obtain correlation, search maximum related value, associated time lengths corresponding to this maximum related value is exactly the 3rd sound transducer (S ₂) and fifth sound sound sensor (S ₄) between delay inequality;

Step 3, utilize quasi-Newton method to obtain according to step 2 same sound-source signal different delay after the locus distribution of four delay inequalities, voice signal that five sound transducers gather and five sound transducers obtain sound source position,

The mistiming arriving each sensor according to the locus distribution of sensor and sound source forms positioning equation group, by iteration, obtains positioning solution, calculates the Euclidean distance of the positioning result of front and back twice;

When the positioning result Euclidean distance of twice is less than the iteration error threshold value described in step 2, iterative solution has converged to sound source position; Otherwise when iterations is greater than the iterations threshold value described in step 2, the voice signal of Resurvey five sound transducers, repeats step 2, until obtain sound source position.

2. the method for sound source in a kind of indoor sonic location system orientation room according to claim 1, is characterized in that: described in step 2 utilize time difference generalized correlation method to obtain same sound-source signal different delay according to the voice signal of five sound transducer collections after the detailed process of four delay inequalities be:

By sound transducer S _ireceived signal strength r _i(t) and sensor S _jreceived signal strength r _jt the signal of () adopts formula (0.1) to represent:

$\left\{\begin{array}{c}{r}_i\left(t\right)=A_{i}s(t-{\mathrm{\&tau;}}_i)+w_i\left(t\right)\\ r_j\left(t\right)=A_{j}s(t-{\mathrm{\&tau;}}_j)+w_j\left(t\right)\end{array}\right.---\left(0.1\right)$

In formula (0.1), the voice signal that s (t) sends for sound source; The fourier function that S (ω) is s (t), A _ifor sound transducer S _ithe decay of propagation channel, A _jfor sensor S _jthe decay of propagation channel; w _it () is room background and S _ithe circuit noise sum of road signal, w _jt circuit noise sum that () is room background and Sj road signal, i, j=0,1,2,3,4 and i and j is unequal;

By signal r _i(t) and signal r _jt () carries out related operation, obtain r _i(t) and r _j(t) correlation result

$\begin{array}{c}{R}_{r_i{r}_j}\left(\mathrm{\&tau;}\right)=E[r_i\left(t\right)\&CenterDot;r_j^{*}(t+\mathrm{\&tau;})]\\ ={\&Integral;}_{-\&infin;}^{+\&infin;}{r}_i\left(t\right)\&CenterDot;r_j^{*}(t+\mathrm{\&tau;})\mathrm{dt}\\ =A_i{A}_j\&CenterDot;R_{\mathrm{ss}}[\mathrm{\&tau;}-({\mathrm{\&tau;}}_j-{\mathrm{\&tau;}}_i)]+A_i\&CenterDot;R_{w_{i}s}(\mathrm{\&tau;}-{\mathrm{\&tau;}}_j)+A_i\&CenterDot;R_{{\mathrm{sw}}_j}(\mathrm{\&tau;}+{\mathrm{\&tau;}}_i)+R_{w_i{w}_j}\left(\mathrm{\&tau;}\right)\end{array}---\left(0.2\right)$

In formula (0.2), τ is the associated time lengths of related operation, τ _i, τ _jbe respectively S _iroad signal and S _jthe propagation delay time of road signal path, wherein $\left\{\begin{array}{c}{R}_{w_{i}s}(\mathrm{\&tau;}-{\mathrm{\&tau;}}_j)=0,\\ R_{{\mathrm{sw}}_j}(\mathrm{\&tau;}+{\mathrm{\&tau;}}_i)=0,\\ R_{w_i{w}_j}\left(\mathrm{\&tau;}\right)=0\end{array}\right.$ Then have:

$R_{r_i{r}_j}\left(\mathrm{\&tau;}\right)=A_i{A}_j\&CenterDot;R_{\mathrm{ss}}[\mathrm{\&tau;}-({\mathrm{\&tau;}}_j-{\mathrm{\&tau;}}_i)]---\left(0.3\right)$

R in formula (0.3) _ss[τ-(τ _j-τ _i)] obtained by following formula:

$\begin{array}{c}{R}_{\mathrm{ss}}[\mathrm{\&tau;}-({\mathrm{\&tau;}}_j-{\mathrm{\&tau;}}_i)]={\&Integral;}_{-\&infin;}^{+\&infin;}s\left(t\right)\&CenterDot;s^{*}(t+\mathrm{\&tau;}-({\mathrm{\&tau;}}_j-{\mathrm{\&tau;}}_i))\mathrm{dt}\\ ={\&Integral;}_{-\&infin;}^{+\&infin;}\left(\frac{1}{2\mathrm{\&pi;}}{\&Integral;}_{-\&infin;}^{+\&infin;}S\left(\mathrm{\&omega;}\right)e^{\mathrm{j\&omega;t}}\mathrm{d\&omega;}\right)\&CenterDot;s^{*}(t+\mathrm{\&tau;}-{\mathrm{\&tau;}}_j+{\mathrm{\&tau;}}_i)\mathrm{dt}\\ =\frac{1}{2\mathrm{\&pi;}}{\&Integral;}_{-\&infin;}^{+\&infin;}S\left(\mathrm{\&omega;}\right){\left({\&Integral;}_{-\&infin;}^{+\&infin;}s(t+\mathrm{\&tau;}-{\mathrm{\&tau;}}_j+{\mathrm{\&tau;}}_i)e^{-\mathrm{j\&omega;t}}\mathrm{dt}\right)}^{*}\mathrm{d\&omega;}\\ =\frac{1}{2\mathrm{\&pi;}}{\&Integral;}_{-\&infin;}^{+\&infin;}S\left(\mathrm{\&omega;}\right)S^{*}\left(\mathrm{\&omega;}\right)e^{\mathrm{j\&omega;}({\mathrm{\&tau;}}_j-{\mathrm{\&tau;}}_i-\mathrm{\&tau;})}\mathrm{d\&omega;}\\ =\frac{1}{2\mathrm{\&pi;}}{\&Integral;}_{-\&infin;}^{+\&infin;}{\left|S\left(\mathrm{\&omega;}\right)\right|}^2\&CenterDot;e^{\mathrm{j\&omega;}({\mathrm{\&tau;}}_j-{\mathrm{\&tau;}}_i-\mathrm{\&tau;})}\mathrm{d\&omega;}\\ \&le;\frac{1}{2\mathrm{\&pi;}}{\&Integral;}_{-\&infin;}^{+\&infin;}{\left|S\left(\mathrm{\&omega;}\right)\right|}^2\&CenterDot;\left|e^{\mathrm{j\&omega;}({\mathrm{\&tau;}}_j-{\mathrm{\&tau;}}_i-\mathrm{\&tau;})}\right|\mathrm{d\&omega;}\\ =\frac{1}{2\mathrm{\&pi;}}{\&Integral;}_{-\&infin;}^{+\&infin;}{\left|S\left(\mathrm{\&omega;}\right)\right|}^2\mathrm{d\&omega;}\end{array}---\left(0.4\right)$

As τ=τ _j-τ _itime, R _ss[τ-(τ _j-τ _i)] get maximal value, namely get maximal value; Adjustment associated time lengths obtains corresponding correlation, and search maximum related value, associated time lengths τ corresponding to this maximum related value is exactly signal r _i(t) and signal r _jthe delay inequality of (t).

3. the method for sound source in a kind of indoor sonic location system orientation room according to claim 1, is characterized in that: the detailed process that four delay inequalities after the same sound-source signal different delay utilizing quasi-Newton method to obtain according to step 2 described in step 3, the voice signal of five sound transducer collections and the locus distribution of five sound transducers obtain sound source position is:

Gather first sound sound sensor (S ₀), second sound sound sensor (S ₁), the 3rd sound transducer (S ₂), fourth sound sound sensor (S ₃) and fifth sound sound sensor (S ₄) voice signal of catching, suppose first sound sound sensor (S ₀), second sound sound sensor (S ₁), the 3rd sound transducer (S ₂) and fourth sound sound sensor (S ₃) or fifth sound sound sensor (S ₄) gather and all acoustic target detected, compare fourth sound sound sensor (S ₃) and fifth sound sound sensor (S ₄) amplitude of voice signal of catching;

As fourth sound sound sensor (S ₃) amplitude of voice signal of catching is greater than fifth sound sound sensor (S ₄) amplitude of voice signal of catching;

Positioning equation group (1.1) is set up to the locus of sound source;

$\left\{\begin{array}{c}\sqrt{{(x-x_1)}^2+{(y-y_1)}^2+{(z-z_1)}^2}-\sqrt{{(x-x_0)}^2+{(y-y_0)}^2+{(z-z_0)}^2}=v\&CenterDot;\mathrm{\&Delta;}{T}_{01}\\ \sqrt{{(x-x_2)}^2+{(y-y_2)}^2+{(z-z_2)}^2}-\sqrt{{(x-x_0)}^2+{(y-y_0)}^2+{(z-z_0)}^2}=v\&CenterDot;{\mathrm{\&Delta;T}}_{02}\\ \sqrt{{(x-x_3)}^2+{(y-y_3)}^2+{(z-z_3)}^2}-\sqrt{{(x-x_1)}^2+{(y-y_1)}^2+{(z-z_1)}^2}=v\&CenterDot;\mathrm{\&Delta;}{T}_{13}\end{array}\right.---\left(1.1\right)$

(x, y, z) represents the locus coordinate of acoustic target, (x _i, y _i, z _i) represent sensor S _ilocus coordinate, i=0,1,2,3,4, △ T ₀₁for first sound sound sensor (S ₀) and second sound sound sensor (S ₁) delay inequality that obtains of the voice signal related operation of catching, τ ₀₁=△ T ₀₁=τ ₁-τ ₀, △ T ₀₂for first sound sound sensor (S ₀) and the 3rd sound transducer (S ₂) delay inequality that obtains of the voice signal related operation of catching, τ ₀₂=△ T ₀₂=τ ₂-τ ₀, △ T ₁₃for second sound sound sensor (S ₁) and fourth sound sound sensor (S ₃) delay inequality that obtains of the voice signal related operation of catching, τ ₁₃=△ T ₁₃=τ ₃-τ ₁, △ T ₂₄be the 3rd sound transducer (S ₂) and fifth sound sound sensor (S ₄) delay inequality that obtains of the voice signal related operation of catching, τ ₂₄=△ T ₂₄=τ ₄-τ ₂, v is sound propagation velocity, and meets the known conditions shown in formula (1.2);

$\left\{\begin{array}{c}(x_0,y_0,z_0)=\left(\mathrm{0,0,0}\right)\\ (x_1,y_1,z_1)=(0,y_1,0)\\ (x_2,y_2,z_2)=(x_2,\mathrm{0,0})\\ (x_3,y_3,z_3)=(0,y_3,z_3)\end{array}\right.---\left(1.2\right)$

Positioning equation group (1.1) is Nonlinear System of Equations,

Defined function is as shown in (1.3);

$\left\{\begin{array}{c}{f}_1=\sqrt{{(x-x_1)}^2+{(y-y_1)}^2+{(z-z_1)}^2}-\sqrt{{(x-x_0)}^2+{(y-y_0)}^2+{(z-z_0)}^2}-v\&CenterDot;\mathrm{\&Delta;}{T}_{01}\\ f_2=\sqrt{{(x-x_2)}^2+{(y-y_2)}^2+{(z-z_2)}^2}-\sqrt{{(x-x_0)}^2+{(y-y_0)}^2+{(z-z_0)}^2}-v\&CenterDot;{\mathrm{\&Delta;T}}_{02}\\ f_3=\sqrt{{(x-x_3)}^2+{(y-y_3)}^2+{(z-z_3)}^2}-\sqrt{{(x-x_1)}^2+{(y-y_1)}^2+{(z-z_1)}^2}-v\&CenterDot;\mathrm{\&Delta;}{T}_{13}\end{array}\right.---(1.3)$

Build vectorial F (x)=(f ₁, f ₂, f ₃) ^t, x=(x, y, z) ^t, and this vector meets equation shown in formula (1.4);

F(x)＝(f ₁,f ₂,f ₃) ^T＝0 (1.4)

The Jacobi matrix of formula (1.4) is (1.5);

$A_i=F^{\&prime;}\left(x^i\right)=\left[\begin{array}{ccc}\frac{{\&PartialD;f}_1}{\&PartialD;x}& \frac{{\&PartialD;f}_1}{\&PartialD;y}& \frac{{\&PartialD;f}_1}{\&PartialD;z}\\ \frac{{\&PartialD;f}_2}{\&PartialD;x}& \frac{{\&PartialD;f}_2}{\&PartialD;y}& \frac{{\&PartialD;f}_2}{\&PartialD;z}\\ \frac{{\&PartialD;f}_3}{\&PartialD;x}& \frac{{\&PartialD;f}_3}{\&PartialD;y}& \frac{{\&PartialD;f}_3}{\&PartialD;z}\end{array}\right]---(1.5)$

Then the iterative formula of solving equations (1.1) is formula (1.6);

$x^{i+1}=x^i-A_i^{-1}F\left(x^i\right)---\left(1.6\right)$

Iteration error threshold value E described in step 2 _th, as the Euclidean distance D of the positioning result of front and back twice _e=|| x ⁱ⁺¹-x ⁱ|| ₂<E _thtime iteration exit, obtain positioning solution; Otherwise when iterations is greater than the iterations threshold value T described in step 2 _th, the voice signal of Resurvey five sound transducers, repeats step 2, until obtain sound source position;

The amplitude of the voice signal of catching when fourth sound sound sensor (S3) is less than fifth sound sound sensor (S ₄) amplitude of voice signal of catching;

According to the rectangular coordinate system in space that the present invention sets up, positioning equation group (2.1) can be set up to the locus of sound source;

$\left\{\begin{array}{c}\sqrt{{(x-x_1)}^2+{(y-y_1)}^2+{(z-z_1)}^2}-\sqrt{{(x-x_0)}^2+{(y-y_0)}^2+{(z-z_0)}^2}=v\&CenterDot;\mathrm{\&Delta;}{T}_{01}\\ \sqrt{{(x-x_2)}^2+{(y-y_2)}^2+{(z-z_2)}^2}-\sqrt{{(x-x_0)}^2+{(y-y_0)}^2+{(z-z_0)}^2}=v\&CenterDot;{\mathrm{\&Delta;T}}_{02}\\ \sqrt{{(x-x_4)}^2+{(y-y_4)}^2+{(z-z_4)}^2}-\sqrt{{(x-x_2)}^2+{(y-y_2)}^2+{(z-z_2)}^2}=v\&CenterDot;\mathrm{\&Delta;}{T}_{24}\end{array}\right.---\left(2.1\right)$

And meet the known conditions shown in formula (2.2);

$\left\{\begin{array}{c}(x_0,y_0,z_0)=\left(\mathrm{0,0,0}\right)\\ (x_1,y_1,z_1)=(0,y_1,0)\\ (x_2,y_2,z_2)=(x_2,\mathrm{0,0})\\ (x_4,y_4,z_4)=(x_4,0,z_4)\end{array}\right.---\left(2.2\right)$

Positioning equation group (2.1) is Nonlinear System of Equations,

Positioning equation group is rewritten into vector form:

Defined function is as shown in (2.3);

$\left\{\begin{array}{c}{f}_1=\sqrt{{(x-x_1)}^2+{(y-y_1)}^2+{(z-z_1)}^2}-\sqrt{{(x-x_0)}^2+{(y-y_0)}^2+{(z-z_0)}^2}=v\&CenterDot;\mathrm{\&Delta;}{T}_{01}\\ f_2=\sqrt{{(x-x_2)}^2+{(y-y_2)}^2+{(z-z_2)}^2}-\sqrt{{(x-x_0)}^2+{(y-y_0)}^2+{(z-z_0)}^2}=v\&CenterDot;{\mathrm{\&Delta;T}}_{02}\\ f_3=\sqrt{{(x-x_4)}^2+{(y-y_4)}^2+{(z-z_4)}^2}-\sqrt{{(x-x_2)}^2+{(y-y_2)}^2+{(z-z_2)}^2}=v\&CenterDot;\mathrm{\&Delta;}{T}_{24}\end{array}\right.---(2.3)$

Build vectorial F (x)=(f ₁, f ₂, f ₃) ^t, x=(x, y, z) ^t, and this vector meets equation shown in formula (2.4);

F(x)＝(f ₁,f ₂,f ₃) ^T＝0 (2.4)

The Jacobi matrix of formula (2.4) is (2.5);

$A_i=F^{\&prime;}\left(x^i\right)=\left[\begin{array}{ccc}\frac{{\&PartialD;f}_1}{\&PartialD;x}& \frac{{\&PartialD;f}_1}{\&PartialD;f}& \frac{{\&PartialD;f}_1}{\&PartialD;z}\\ \frac{{\&PartialD;f}_2}{\&PartialD;x}& \frac{{\&PartialD;f}_2}{\&PartialD;y}& \frac{{\&PartialD;f}_2}{\&PartialD;z}\\ \frac{{\&PartialD;f}_3}{\&PartialD;x}& \frac{{\&PartialD;f}_3}{\&PartialD;y}& \frac{{\&PartialD;f}_3}{\&PartialD;z}\end{array}\right]---\left(2.5\right)$

Then the iterative formula of solving equations (2.1) is formula (2.6);

x ⁱ⁺¹＝x ⁱ-A _i ^-1F(x ⁱ) (2.6)

Iteration error threshold value E described in step 2 _th, as the Euclidean distance D of the positioning result of front and back twice _e=|| x ⁱ⁺¹-x ⁱ|| ₂<E _thtime iteration exit, obtain positioning solution; Otherwise when iterations is greater than the iterations threshold value T described in step 2 _th, the voice signal of Resurvey five sound transducers, repeats step 2, until obtain sound source position.