CN103399293A - Method and device for realizing sound source direction finding - Google Patents

Abstract

The invention provides a method and a device for realizing sound source direction finding. Based on the advantages of passive cross locating and typical time difference locating, in combination with judgment of an approximate direction and analysis of a hyperbola, the direction locating of an accurate direction is finally realized by a robot. By adopting the scheme, not only can the realization difficulty of a system be reduced, but also the higher accuracy can be guaranteed.

Description

A kind of method and device of realizing the sound source direction finding

Technical field

The present invention relates to a kind of method and device of realizing the sound source direction finding, particularly relate to a kind of method and device that is applicable to realize robot sound source direction finding.

Background technology

The passive location technology is an important technology in the direction finding means, in engineering and military affairs, have a wide range of applications, and along with the upgrading of technology, research and all progressively upgradings of application to passive location, particularly, aspect voice application, a large amount of patent and documents about shotgun microphone, directed osophone aspect have been emerged.We wish in machine person to person's reciprocal process, allow the behavior pattern of robot more close to the people, corresponding feedback is all done in action, expression and position etc. to the people, and the identification of Sounnd source direction when the people is spoken, in the time of can allowing the machine person to person exchange, all the time recognize speaker's direction, but because robot is self-existent, we also have no idea to realize cross bearing by the mode of establishing a plurality of observation point, therefore general passive location technology as: intersecting passive location method and multistation time difference passive location method all can't use.

Passive cross bearing is a kind of traditional localization method, utilize the azimuthal measurement of a plurality of measurement points to target, obtain the coordinate position of point of crossing, in order to make mathematical model more clear, usually be reduced to plane problem and solve, describe by the localization method of 2 intersections here.

As shown in Figure 1, the coordinate of establishing two measurement points is respectively (x ₁, y ₁) and (x ₂, y ₂), record azimuth of target and calculate and be respectively θ from the X-axis forward ₁And θ ₂, the equation of two direction finding lines is respectively:

(x-x _i)×sinθ _i＝(y-y _i)×cosθ _i（i=1,2）

Work as θ ₁≠ θ ₂The time, have intersecting point coordinate to be:

$\left\{\begin{array}{c}{x}_0=x_1+l\cos {\mathrm{\θ}}_1\\ y_0=y_1+l\sin {\mathrm{\θ}}_1\end{array}\right.$

Wherein, l is impact point (x ₀, y ₀) to measurement point (x ₁, y ₁) distance, l can be expressed as:

$l=\frac{\mathrm{\Δx}\×\sin {\mathrm{\θ}}_2-\mathrm{\Δy}\×\cos {\mathrm{\θ}}_2}{\sin ({\mathrm{\θ}}_2-{\mathrm{\θ}}_1)}$

Wherein, Δ x=x ₂-x ₁, Δ y=y ₂-y ₁.

And the principle of utilizing time difference measurement can obtain the equation of two direction finding lines, article one, time difference line needs two measurement points, article two, time difference line needs three measurement points, so by the detection of three measurement points to the impact point signal, can set up two groups of Hyperbolic Equations, utilize hyp intersection point, obtain the position of impact point.

This patent, by extracting the advantage of classical time difference positioning method, has provided a kind of sound source direction method and the device that can realize based on robot self, the method and and device not only reduced realizing difficulty but also having guaranteed higher precision of system.

Summary of the invention

The technical problem to be solved in the present invention is to provide a kind of method and device that is applicable to realize robot sound source direction finding, advantage based on passive cross bearing and classical time difference location, in conjunction with the judgement to general orientation with to hyp parsing, realize that finally robot is to the direction in accurate orientation location, this scheme of the invention had not only reduced realizing difficulty but also having guaranteed higher precision of system.

The technical solution used in the present invention is as follows:

A kind of method that realizes the sound source direction finding, its concrete grammar is: one, in robot, establish two measurement points with it; Two, take two measurement points as focal length, obtain Hyperbolic Equation; Three, by the impact point source sound, propagate into the time order and function of two measurement points, reach the division of robot front and back, detect the quadrant area of sound source; Four, resolve hyperbolic curve and draw the angle between impact point and measurement point.

As preferably, the concrete grammar of described step 2 is: by in passive location, obtaining the mode of time difference line, utilizing corresponding fixing distance of the fixing time difference, is 2 times of hyperbolic curve real axis from 2 range differences, and the distance of two measurement points is namely focal length; If the coordinate of two measurement points is respectively (x _a, y _a) and (x _b, y _b), impact point is D to this range difference of 2, thereby can draw hyp equation.

As preferably, the method for obtaining the time difference of two measurement points is: with two-way independently, interrupt processing two paths of signals, and independently timer is set calculates the two paths of signals time difference.

As preferably, measure two paths of signals and calculate the concrete grammar step of the time difference and be: A, time difference counter original state is set; B, judge whether to interrupt first and triggered, be, entered step C, no, continue to wait for; C, unlatching timer first are also closed the interruption first; Whether D, judgement are interrupted second and are triggered, and are, enter step e, no, return steps A after entering blocking time; E, stop the timer first, close interruption second, and open timer second; F, report time difference data to host computer; G, return steps A after entering blocking time.

As preferably, described blocking time greater than impact point to two measurement points apart from the quotient of difference D divided by sound propagation velocity.

As preferably, described blocking time is 2 seconds.

As preferably, described step 4 is resolved hyp concrete grammar and is: the method that adopts asymptotic line to approach solves.

Based on the described system that realizes the sound source direction-finding method, it is characterized in that: comprise the sound collection unit, operation amplifier unit, pulse transmitter, time difference counter and the host computer that are connected successively.

Compared with prior art, the invention has the beneficial effects as follows: combine the passive location technology, utilize the interrupt handling capacity of single-chip microcomputer, realized to the detection of audio direction and to the control of integral body, not only reduced realizing difficulty but also having guaranteed higher precision of system, enriched the application of intellectual technology, also for follow-up development in this field, laid a good foundation.

The accompanying drawing explanation

Fig. 1 is two straight line cross bearing principle schematic.

Fig. 2 is a wherein examples measure two paths of signals calculate the process flow diagram of temporal difference method of the present invention.

Fig. 3 is the wherein apparatus structure block diagram of an embodiment of the present invention.

Embodiment

In order to make purpose of the present invention, technical scheme and advantage clearer, below in conjunction with drawings and Examples, the present invention is further elaborated.Should be appreciated that specific embodiment described herein, only in order to explain the present invention, is not intended to limit the present invention.

Disclosed all features in this instructions, except the feature of mutual eliminating, all can make up by any way.

Disclosed arbitrary feature in this instructions (comprising any accessory claim, summary and accompanying drawing), unless special narration all can be replaced by other equivalences or the alternative features with similar purpose.That is, unless special narration, each feature is an example in a series of equivalences or similar characteristics.

In robot application, because robot is self-existent, we can't set up outside measurement point and come auxiliary robot to complete location, therefore the mode of cross bearing is also not exclusively applicable, but concerning robot, do not need to know the range information of impact point and robot, only need to pay close attention to directional information and just can complete direction finding, thus we only need in robot, establish two measurement points can extract directional data.

A kind of method that realizes the sound source direction finding, its concrete grammar is: one, in robot, establish two measurement points with it; Two, take two measurement points as focal length, obtain Hyperbolic Equation; Three, by the impact point source sound, propagate into the time order and function of two measurement points, reach the division of robot front and back, detect the quadrant area of sound source; Four, resolve hyperbolic curve and draw the angle between impact point and measurement point.

The concrete grammar of described step 2 is: by in passive location, obtaining the mode of time difference line, utilize corresponding fixing distance of the fixing time difference, and when from 2 range differences, being fixed value, this value is 2 times of hyperbolic curve real axis, and the distance of two measurement points is namely focal length; If the coordinate of two measurement points is respectively (x _a, y _a) and (x _b, y _b), impact point is D to this range difference of 2, thereby can draw hyp equation.

If the coordinate of two measurement points is respectively (x _a, y _a) and (x _b, y _b), impact point is D to this range difference of 2, so corresponding hyp equation is:

$(\mathrm{\Δ}{x}^2-D^2){(x-\frac{\mathrm{\Σx}}{2})}^2+2\mathrm{\Δx\Δy}(x-\frac{\mathrm{\Σx}}{2})(y-\frac{\mathrm{\Σy}}{2})+(\mathrm{\Δ}{y}^2-D^2){(y-\frac{\mathrm{\Σy}}{2})}^2$

$={\left(\frac{D}{2}\right)}^2[D^2-(\mathrm{\Δ}{x}^2+\mathrm{\Δ}{y}^2)]$

Wherein, Δ x=x _b-x _a, Δ y=y _b-y _a; Σ x=x _b+ x _a, Σ y=y _b+ y _a.

By equation, can find out, when we recorded D, point source of sound must be on this hyperbolic curve, but still faced two problems: the one, and there are 4 quadrants in hyperbolic curve, and these 4 quadrants are on all four concerning robot, and we need artificial distinguishing; The 2nd, hyp parsing is comparatively complicated, and single-chip microcomputer carries out the poor ability of floating-point operation, and along with the change of distance, the angle that hyperbolic curve is corresponding is also different.

The settling mode of first problem is to judge by the time order and function that detects two measurement points, the propagation of sound is slower, two measurement points the time difference must occur to the acceptance of sound, when one of them measurement point is first received signal, sound source must be in a corresponding side, so just can exclude 2 quadrants, then by the division to the robot front and back, can exclude again a quadrant, by such mode, just detect the approximate region of sound source.

The method of obtaining the time difference of two measurement points is: with two-way independently, interrupt processing two paths of signals, and independently timer is set calculates the two paths of signals time difference.

Therefore the time difference value of two measurement points is parameters most crucial in scheme, and the precision of this value directly affects the precision of direction, with two-way independently, interrupts processing this two paths of signals, and independently timer is set calculates the time difference.

As shown in Figure 2, measure two paths of signals calculate the concrete grammar step of the time difference and be: A, time difference counter original state is set; B, judge whether to interrupt first and triggered, be, entered step C, no, continue to wait for; C, unlatching timer first are also closed the interruption first; Whether D, judgement are interrupted second and are triggered, and are, enter step e, no, return steps A after entering blocking time; E, stop the timer first, close interruption second, and open timer second; F, report time difference data to host computer; G, return steps A after entering blocking time.

In this specific embodiment, interrupt first and interrupt not strict differentiation of second, but convenient the description distinguished name afterwards according to the elder generation of down trigger.Also can interrupt 1 by called after, interrupt 2, same, timer is similarly random state, according to successively distinguishing name.

Described blocking time greater than impact point to two measurement points apart from the quotient of difference D divided by sound propagation velocity.

Described blocking time is 2 seconds.

A blocking time is set, is conducive to the stable of system, be convenient to voice signal is changed to point value.

Described step 4 is resolved hyp concrete grammar: the method that adopts asymptotic line to approach solves.

The method that Second Problem adopts asymptotic line to approach solves, when point source of sound from measurement point away from the time, the tangent line y that puts on hyperbolic curve will overlap with asymptotic line, when the distance value of measurement point and impact point during much larger than focal length, the distance of two measurement points namely, asymptotic line and hyperbolic error can be ignored.Therefore we,, hyp parsing is converted into asymptotic parsing, have greatly reduced operand.Below derive asymptotic equation:

$\left\{\begin{array}{c}(x-\frac{\mathrm{\Σx}}{2})=\frac{\mathrm{\Δ}{y}^2-D^2}{\mathrm{\Δx\Δy}+\sqrt{D^2(\mathrm{\Δ}{x}^2+\mathrm{\Δ}{y}^2-D^2)}}(y-\frac{\mathrm{\Σy}}{2})\\ (y-\frac{\mathrm{\Σy}}{2})=\frac{\mathrm{\Δ}{x}^2-D^2}{\mathrm{\Δx\Δy}+\sqrt{D^2(\mathrm{\Δ}{x}^2+\mathrm{\Δ}{y}^2-D^2)}}(x-\frac{\mathrm{\Σx}}{2})\end{array}\right.$

Only need to solve the hyperbolic curve asymptotic line, namely can draw the angle between impact point and measurement point.

As shown in Figure 3, based on the described system that realizes the sound source direction-finding method, it is characterized in that: comprise the sound collection unit, operation amplifier unit, pulse transmitter, time difference counter and the host computer that are connected successively.

Sound collection unit is responsible for voice signal is converted to electric signal, transfers to operational amplifier and realizes the amplification of signal and stablize and produce pulse, and time difference counter calculates the time difference of two-way pulse, finally gives host computer and completes positioning calculation process and control.

Claims (8)

1. method that realizes the sound source direction finding, its concrete grammar is: one, in robot, establish two measurement points with it; Two, take two measurement points as focal length, obtain Hyperbolic Equation; Three, by the impact point source sound, propagate into the time order and function of two measurement points, reach the division of robot front and back, detect the quadrant area of sound source; Four, resolve hyperbolic curve and draw the angle between impact point and measurement point.

2. method according to claim 1, the concrete grammar of described step 2 is: by in passive location, obtaining the mode of time difference line, utilizing corresponding fixing distance of the fixing time difference, is 2 times of hyperbolic curve real axis from 2 range differences, and the distance of two measurement points is namely focal length; If the coordinate of two measurement points is respectively (x _a, y _a) and (x _b, y _b), impact point is D to this range difference of 2, thereby draws hyp equation.

3. method according to claim 2, the method for obtaining the time difference of two measurement points is: with two-way independently, interrupt processing two paths of signals, and independently timer is set calculates the two paths of signals time difference.

4. method according to claim 3, measure two paths of signals and calculate the concrete grammar step of the time difference and be: A, time difference counter original state is set; B, judge whether to interrupt first and triggered, be, entered step C, no, continue to wait for; C, unlatching timer first are also closed the interruption first; Whether D, judgement are interrupted second and are triggered, and are, enter step e, no, return steps A after entering blocking time; E, stop the timer first, close interruption second, and open timer second; F, report time difference data to host computer; G, return steps A after entering blocking time.

5. method according to claim 4, described blocking time greater than impact point to two measurement points apart from the quotient of difference D divided by sound propagation velocity.

6. method according to claim 5, described blocking time is 2 seconds.

7. method according to claim 1, described step 4 are resolved hyp concrete grammar and be: the method that adopts asymptotic line to approach solves.

8. based on the device of the described method of one of claim 1 to 7, it is characterized in that: comprise the sound collection unit, operation amplifier unit, pulse transmitter, time difference counter and the host computer that are connected successively.

Images