JP2013213739A - Sound source position estimating device, sound source position estimating method and program therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a sound source position estimating technique which allows time and labor for microphone position measurement and wiring to be saved by transmitting information necessary for sound source position estimation via visible light communication, and which allows easy alteration and expansion of an area covered by a microphone array and advantageous in estimating a sound source direction.SOLUTION: A sound source position estimating device comprises a plurality of sensor units and a receiver. Each of the sensor unit includes: a soundwave direction estimating section for estimating a strength by soundwave receiving direction that is an aggregation of soundwave strengths for each direction, by using acoustic signals; and a transmitting section for encoding the strength by soundwave receiving direction and transmitting it via visible light communication by making a light emitter flash. The receiver includes: a decoding section which decodes the flashing of the light emitter on the image data to acquire the strength by soundwave receiving direction, and further estimates a sensor unit position from a flashing position of the light emitter on the image data using a position of a light receiving section as a reference point; and a sound source position reconfiguration section for estimating a sound source position using the strength by soundwave receiving direction and the sensor unit position.

Description

  The present invention relates to a sound source position estimation apparatus, a sound source position estimation method, and a program thereof that estimate the position of a sound source using a microphone.

  Various research and development have been conducted on the estimation of sound source direction and position using a microphone array composed of a plurality of microphones. For the target of sound source direction and position estimation, two pieces of information such as a microphone position and measurement information are usually required. In general, the arrangement area of the microphone array is large, and the larger the number of microphones, the more advantageous for sound source position estimation.

  For this reason, a method is employed in which the position of the microphone is measured each time or a microphone array having a predetermined positional relationship is created (see Non-Patent Document 1).

  However, in the former, there is a problem that if the number of microphone channels is increased, there is a problem that the labor of position measurement and wiring increases.

H. Krim, M. Viberg, “Two decades of array signal processing research”, IEEE Signal Processing Magazine, July 1996, pp.67-94

In response to this problem, if the configuration is such that information from the microphone array is acquired using an image sensor by visible light communication (see Reference 1), position information and information on the microphone array can be acquired simultaneously, and wiring can be simplified. (See Reference 2).
(Reference 1) Masao Nakagawa, “The World of Visible Light Communication”, Industrial Research Committee, February 1, 2006, p3-21, p159-170
(Reference 2) Toru Nagura, Takaya YAMAZATO, Masaaki KATAYAMA, Tomohiro YENDO, Toshiaki FUJII, Hiraku OKADA, “Improved Decoding Methods of Visible Light Communication System for ITS Using LED Array and High-Speed Camera”, Vehicular Technology Conference (VTC 2010 -Spring), 2010 IEEE 71st, IEEE, 16-19 May 2010, pp.1-5

However, when information from a microphone array is simply transmitted by visible light communication, there is a problem that a communication band is insufficient to transmit all of the acquired acoustic signals when the number of channels of the microphone array is increased (see Reference 3). ).
(Reference 3) Masanori Ishida, Shinichiro Haruyama, Masao Nakagawa, “Examination of Communication Speed Limit in Parallel Visible Light Wireless Communication System”, IEICE Technical Report CS Communication System, The Institute of Electronics, Information and Communication Engineers, 2007 1 May 4, Vol. 106, no. 450, pp. 37-41

  In the present invention, instead of transmitting all acoustic signals, a solution for extracting only information necessary for sound source position estimation within a sensor unit or between a plurality of sensor units to some extent and then transmitting only the information is the visible light communication described above. Used in combination with the above solution.

  The present invention eliminates the trouble of microphone position measurement and wiring by transmitting information necessary for sound source position estimation by visible light communication, and covers the area of the microphone array that works advantageously for sound source direction estimation. An object of the present invention is to provide a sound source position estimation technique that can easily change or increase the number of sound sources.

  In order to solve the above-described problem, according to the first aspect of the present invention, the sound source position estimating device includes a plurality of sensor units and a receiving device. The sensor unit converts a sound wave into an electrical signal, generates a sound signal, and uses the sound signal to estimate a sound wave direction estimation unit that estimates a strength for each sound wave reception direction, which is a set of sound wave strengths in each direction. A transmitter that encodes the intensity for each sound wave reception direction, blinks the light emitting element, and transmits the light by visible light communication. The receiving device is composed of an image sensor, receives a flashing light emitting element as image data, decodes the flashing light emitting element on the image data to obtain the intensity for each sound wave receiving direction, and further receives the position of the light receiving unit. A decoding unit that estimates the position of the sensor unit from the blinking position of the light emitting element on the image data, and a sound source position reconstruction unit that estimates the sound source position using the intensity for each sound wave reception direction and the position of the sensor unit; Have.

  In order to solve the above problem, according to the second aspect of the present invention, the sound source position estimation method uses a plurality of sensor units and a receiving device. The sound source position estimation method includes a sound collection step in which a microphone of a sensor unit converts a sound wave into an electrical signal and generates an acoustic signal, and a sound wave direction estimation unit of the sensor unit uses the sound signal to generate a sound wave for each direction. A sound wave direction estimating step for estimating the intensity for each sound wave reception direction, which is a set of strengths, and a transmission in which the transmission unit of the sensor unit encodes the intensity for each sound wave reception direction, blinks the light emitting element, and transmits it by visible light communication. A receiving step in which the light receiving unit of the receiving device receives the blinking of the light emitting element as image data; and a decoding unit of the receiving device decodes the blinking of the light emitting element on the image data to obtain the intensity for each sound wave receiving direction. And a decoding step of estimating the position of the sensor unit from the blinking position of the light emitting element on the image data with the position of the light receiving unit as a base point, Source position reconstruction unit, using the position of the wave reception direction-strength and the sensor unit includes a sound source position reconstruction step of estimating the sound source position.

  According to the present invention, by transmitting information necessary for sound source position estimation by visible light communication, it is possible to save the trouble of microphone position measurement and wiring, and it is advantageous for sound source direction estimation. The area to be changed can be easily changed or increased.

The functional block diagram of the sound source position estimation apparatus 100 which concerns on 1st embodiment. The figure which shows the processing flow of the sound source position estimation apparatus 100 which concerns on 1st embodiment. FIG. 3A is an arrangement example of two sensor units and a receiving device, FIG. 3B is an example of image data obtained by a light receiving unit of the receiving device, and FIG. 3C is a diagram showing a relationship between maximum intensity and overlapping portions. 4A is a front view of the sensor unit, FIG. 4B is a plan view thereof, and FIG. 4C is a rear view thereof. The figure which shows the example of the image data obtained by the light-receiving part of the receiver which concerns on the modification of 1st embodiment. The figure which shows the example at the time of arrange | positioning the sound collection area of a microphone three-dimensionally. The figure which shows the example which arrange | positions three sensor units. The figure which shows the example of arrangement | positioning of three sensor units and a receiver. The figure which shows the example of the image data in the case of FIG. The figure which shows the relationship of the maximum intensity | strength and overlapping part in the case of FIG. FIG. 11A is a front view of the sensor unit when the LED of the transmission unit is arranged on the front surface of the substrate, and FIG. 11B is a plan view thereof. The figure which shows the example of arrangement | positioning with the sensor unit at the time of arrange | positioning LED of a transmission part on the front of a board | substrate, and a receiver. The figure which shows the example of arrangement | positioning of a sensor unit and a receiver at the time of installing two sensor units in a different direction. The functional block diagram of the sound source position estimation apparatus 200. FIG. 15A is a front view of the sensor unit according to the second embodiment, FIG. 15B is a plan view thereof, and FIG. 15C is a rear view thereof. In the sensor unit according to the second embodiment, it shows the relationship between the acoustic signals _x 1 _{(t), x} 2 (t) and _x 3 (t) and wave receiving direction-strength. The figure for demonstrating the relationship between the intensity | strength according to a sound wave receiving direction, and an arrival axis. The figure which shows the intersection of an arrival axis.

  Hereinafter, embodiments of the present invention will be described. In the drawings used for the following description, constituent parts having the same function and steps for performing the same process are denoted by the same reference numerals, and redundant description is omitted. Further, the processing performed for each element of a vector or matrix is applied to all elements of the vector or matrix unless otherwise specified.

<First embodiment>
FIG. 1 is a functional block diagram of the sound source position estimating apparatus 100, and FIG.

  The sound source position estimation device 100 includes two sensor units 110-1 and 110-2 and a reception device 120. 3A shows an arrangement example of the two sensor units 110-1 and 110-2 and the receiving device 120, and FIG. 3B shows an example of image data obtained by the light receiving unit 121 of the receiving device 120. In the present embodiment, the case where the sound source direction estimation is limited to the horizontal direction is shown for simplicity. Therefore, it is assumed that the sensor units 110-1 and 110-2 are horizontally installed at the same height when viewed from the receiving device 120 (the light receiving unit 121).

[Sensor units 110-1 and 110-2]
4A is a front view of the sensor unit 110-1, FIG. 4B is a plan view thereof, and FIG. 4C is a rear view thereof. The sensor unit 110-1 includes microphones 111 _R , 111 _F, and 111 _L , a sound wave direction estimation unit 113, and a transmission unit 115 (see FIG. 1). The sensor unit 110-2 has a similar configuration.

  The sensor unit 110-1 estimates the arrival direction of the sound wave at the installed position, and transmits information regarding the arrival direction of the sound wave to the reception device using visible light communication.

[Microphones 111 _R , 111 _F and 111 _L ]
Each of the microphones 111 _R , 111 _F, and 111 _L has a sharp directivity and a different sound collection area. In the present embodiment, the microphones 111 _R , 111 _F, and 111 _L are installed on the front surface of the substrate 117 so as to best capture sound waves coming from the front, 45 ° left and right (see FIG. 4). The microphones 111 _R , 111 _F, and 111 _L pick up sound waves and convert them into electric signals, generate acoustic signals x _R (t), x _F (t), and x _L (t), and estimate sound wave directions. The data is transmitted to the unit 113 (s1). Here, t represents a frame number or a time corresponding to the frame number.

[Sound wave direction estimation unit 113]
The sound wave direction estimation unit 113 includes a CPU, a memory, a processing circuit, and the like. The sound wave direction estimation unit 113 receives the acoustic signals x _R (t), x _F (t), and x _L (t), and uses these values to determine the strength of the sound wave in each direction that is the strength of the sound wave in each direction. Is estimated (s2) and transmitted to the transmitter.

  Here, the intensity by sound wave reception direction is a set of intensities by sound wave traveling direction at the measurement position (installation position of the sensor unit), and if the information on the measurement position is given, the traveling direction of the sound wave can be estimated. Information that can be done. Note that the intensity of the traveling wave exists only in the direction of the sound source when measuring in an ideal sound source and infinite space.

For example, the sound wave direction estimating unit 113 obtains the short-time integration of the acoustic signals x _R (t), x _F (t), and x _L (t) generated from the microphones 111 _R , 111 _F, and 111 _L as its intensity. The intensity for each sound wave receiving direction, which is the set, is estimated. For example, the intensity of the sound wave coming from each direction is calculated for each period T1. The intensity φ _F (n) of the sound wave from the front direction calculated for the nth time is expressed as x _F (t),

It is estimated to be. T2 represents the integration time. Similarly, sound wave strengths φ _L (n) and φ _R (n) regarding 45 ° to the left and 45 ° to the right can be obtained from x _L (t) and x _R (t). The obtained intensity for each sound wave receiving direction φ (n) = [φ _F (n), φ _L (n), φ _R (n)] is output to the transmitter 115 every period T1.

[Transmitter 115]
The transmission unit 115 includes a circuit or program that performs encoding, a light emitting element driver (such as an LED driver), and a light emitting element (such as an LED). In this embodiment, the LED is disposed on the back surface of the substrate 117 (see FIG. 4). The transmitter 115 receives the intensity φ (n) = [φ _F (n), φ _L (n), φ _R (n)] for each sound wave receiving direction, encodes it for visible light communication, and blinks the LED. Then, it transmits by visible light communication (s3). The blinking by visible light is output toward the light receiving unit 121 of the receiving device 120 described later.

  In the present embodiment, for example, it is only necessary that the intensity φ (n) for each sound wave reception direction obtained at the nth time can be sent to the receiving device 120 within T1 time. For example, the communication method described in Reference 3 may be used. .

In the following description, the intensity for each sound wave reception direction output from the sensor unit 110-1 is φ ₁ (n) = [φ _1F (n), φ _1L (n), φ _1R (n)], and the sensor unit 110-2. The intensity for each sound wave reception direction output from the signal is φ ₂ (n) = [φ _2F (n), φ _2L (n), φ _2R (n)].

  In other words, in the present embodiment, the acoustic signal of the microphone and its feature amount are not transmitted, but only the information necessary for sound source estimation (the intensity for each sound wave reception direction) is extracted in advance by the sensor unit alone, and then the information is extracted. Only transmit with visible light communication. The intensity for each sound wave reception direction is the intensity for each sound wave direction calculated by a single sensor unit.

[Receiving device 120]
The receiving device 120 includes a light receiving unit 121, a decoding unit 123, and a sound source position reconstruction unit 125 (see FIG. 1). The intensity for each sound wave receiving direction emitted from a plurality of sensor units is acquired, and the position of the sound source is reconstructed.

[Light receiving unit 121]
The light receiving unit 121 includes an image sensor. The blinking LED is received as image data (s4), and the received image data is output to the decoding unit 123. FIG. 3B shows an example of image data.

[Decoding unit 123]
The decoding unit 123 receives the image data, decodes the blinking of the LED of the image data, and acquires the intensity φ ₁ (n), φ ₂ (n) for each sound wave reception direction (s5). Further, the positions of the sensor units 110-1 and 110-2 are estimated from the blinking positions of the LEDs on the image data with the position of the light receiving unit 121 as a base point. The acquired strengths φ ₁ (n) and φ ₂ (n) for each sound wave reception direction and the estimated position of the sensor unit are output to the sound source position reconstruction unit 125. With such a configuration, the position information of the sensor unit and the measurement information of the sensor unit can be obtained by a single measurement.

  In the present embodiment, for example, the blinking transmitted by the sensor units 110-1 and 110-2 is decoded as described in Reference 3 corresponding to the encoding method.

  At the same time, the actual sensor position is estimated from the blinking positions of the LEDs on the image data of the sensor units 110-1 and 110-2. Assume that the distance from the surface D formed by the sensor units 110-1 and 110-2 to the receiving device 120 is given as z0 (see FIG. 3A). Then, the decoding unit 123 determines the positions of the sensor units 110-1 and 110-2 based on the blinking positions of the LEDs on the image data, and the vector positions r (n) based on the position of the light receiving unit 121 of the receiving device 120, respectively. , 1) and r (n, 2) can be estimated using known techniques. With such a configuration, even if the positions of the sensor units 110-1 and 110-2 move on the plane D, the vector positions r (n) are based on the blinking positions of the LEDs on the image data. , 1), r (n, 2) can be easily estimated.

The decoding unit 123 outputs information [r (n, 1), φ ₁ (n)] and [r (n, 2), φ ₂ (n)] that are all of these to the sound source position reconstruction unit 125. .

[Sound source position reconstruction unit 125]
The sound source position reconstruction unit 125 receives the intensity φ ₁ (n), φ ₂ (n) for each sound wave reception direction, and the position r (n, 1), r (n, 2) of the sensor unit, and these values. Is used to estimate the sound source position (s6), and the estimated sound source position is output as an output value of the sound source position estimating apparatus 100.

In the present embodiment, the sound source position reconstruction unit 125 collects the microphones 111 _R , 111 _F, and 111 _L from the positions r (n, 1) and r (n, 2) of the sensor units 110-1 and 110-2. Estimate the sound area. The sound collection area is specified by the positions and directions of the sensor units 110-1 and 110-2. As described above, the position is estimated based on the blinking position of the LED on the image data. It shall be given in advance.

Furthermore, the sound source position reconstructing unit 125 has the maximum intensity φ ₁ (n) = [φ _1F (n), φ _1L (n), φ _1R (n)] for each sound wave reception direction of the sensor unit 110-1. The microphone (any one of the microphones 111 _R , 111 _F, and 111 _L ) corresponding to the intensity (any one of φ _1F (n), φ _1L (n), φ _1R (n)) is selected, and similarly, the sensor Intensities φ ₂ (n) = [φ _2F (n), φ _2L (n), φ _2R (n)] for each sound wave reception direction of the unit 110-2, the maximum intensity (φ _2F (n), φ _2L (N) or any one of φ _2R (n)) (one of the microphones 111 _R , 111 _F, and 111 _L ) is selected. Furthermore, the sound source position reconstruction unit 125 estimates that there is a sound source in an overlapping portion of the sound pickup area of the selected microphone. FIG. 3C shows the relationship between maximum intensity and overlap. For example, when φ _1L (n) and φ _2F (n) are maximum values, it is estimated that a sound source exists in area (3) in FIG. 3A.

<Effect>
With such a configuration, since the position of the sensor unit and the intensity for each sound wave reception direction generated by the sensor unit are simultaneously acquired as image data, it does not have to be performed each time the arrangement or position measurement of the sensor unit is changed. .

  The target sensor unit may not be fixed. As long as the movement in the period T1 is negligible on the image sensor, the movement on the surface D may be performed.

  Further, the measurement in the sensor unit does not need to be limited to the band for visible light communication to the image sensor. Therefore, a sufficient measurement band of the microphone in the sensor unit can be secured. Therefore, the estimation accuracy of the sound source direction in each sensor unit is unlikely to deteriorate. For example, the intensity for each sound wave reception direction measured at 40 kHz may be transmitted at a video rate of image data (for example, 30 Hz to 40 Hz).

  Wiring on the sensor unit side requires only a power source. Therefore, it becomes easy to arrange a large amount of sensor units over a wide range. In addition, since information measured and processed by each sensor unit is acquired by an image sensor, wiring work is reduced.

  Further, since the sensor unit can be easily installed in a wide range, it is advantageous in estimating the sound source direction and the sound source position.

<Modification>
The sensor unit 110-1 and the sensor unit 110-2 do not have to be installed horizontally at the same height. For example, the sensor units may be two-dimensionally arranged at different heights when viewed from the position of the light receiving unit 121 of the receiving device 120. An example of image data obtained by the light receiving unit 121 of the receiving apparatus 120 in this case is shown in FIG. Even with such a configuration, the position of the sensor unit relative to the receiving device 120 can be estimated from the blinking position of the LED on the image data and the distance z0.

Further, as shown in FIG. 6, the microphones 111 _RH , 111 _RL , 111 _LL , and 111 _LH surround the four sides around the microphone 111 _F, and the intensity of sound waves from 45 ° left and right and up and down 45 ° can be measured. The sound collection area may be three-dimensionally arranged. With such a configuration, visible light communication and sound source position estimation can be performed in a form in which information in the vertical direction is added in addition to the horizontal direction.

  In addition, the number of sensor units is not necessarily two, and a plurality of sensor units may be used. For example, an example in which three sensor units 110-1, 110-2, and 110-3 are arranged is shown in FIG. 8 shows an arrangement example of the three sensor units 110-1, 110-2 and 110-3 and the receiving device 120, FIG. 9 shows an example of image data in the case of FIG. 8, and FIG. 10 shows a case of FIG. Shows the relationship between maximum intensity and overlap. Thus, by arranging a plurality of sensor units, it is possible to estimate the sound source position in more detail.

  Moreover, you may arrange | position LED of the transmission part 115 on the front of the board | substrate 117 (refer FIG. 11A and FIG. 11B). In that case, the receiving device 120 is disposed on the front side of the substrate 117 (FIG. 12).

A clock may be shared between the sensor units. With such a configuration, [φ _1F (n), φ _1L (n), φ _1R (n)], [φ _2F (n), φ _2L (n), φ _2R (n) transmitted every T1. ] And the transmission timing are synchronized between the units.

  It is desirable for T1 to be set to a time when the sensor unit can be regarded as being substantially stationary in order to specify the position of the sensor unit.

  There may be a plurality of receiving apparatuses 120. For example, a known technique such as a stereo camera may be used by using the two receiving devices 120. With such a configuration, it is not necessary to limit the position of the sensor unit on the surface D, and the sensor unit may be arranged three-dimensionally when viewed from the position of the light receiving unit 121 of the receiving device 120. Further, it is possible to save the trouble of measuring the value of z0 in advance.

  In the present embodiment, the two sensor units 110-1 and 110-2 are installed in the same direction, but may be installed in different directions (see FIG. 13). With such a configuration, the sound source position can be estimated in more detail. However, since the direction (orientation) of the sensor unit cannot be acquired only by blinking the LED, it is given to the receiving device 120 in advance.

<Second embodiment>
Only parts different from the first embodiment will be described.

  FIG. 14 shows a functional block diagram of the sound source position estimating apparatus 200.

  The sound source position estimating apparatus 200 includes P sensor units 210-p and a receiving apparatus 220. However, P is an integer greater than or equal to 3, p is a positive integer showing the index of a sensor unit, and p = 1, 2, ..., P.

  The P sensor units 210-p are connected in cascade, and the acoustic signals of the microphones are mutually transmitted and exchanged between adjacent sensor units. For the sake of simplicity, the case where both sides of the sensor unit connected in the horizontal direction are connected will be described.

[Sensor unit 210-p]
The sensor unit 210-p includes an omnidirectional microphone 211, a sound wave direction estimation unit 213, and a transmission unit 115.

  The sensor unit 210-p estimates the sound source arrival direction at the installed position, and transmits the information to the receiving device by visible light communication. It is assumed that the wired connection is made to the closest sensor unit. Without losing generality, the p-th sensor unit 110-p has (p−1) th sensor unit 210- (p−1) on the left side and (p + 1) th on the right side as viewed from the receiving device 120. The sensor unit 210- (p + 1) may be wired.

[Microphone 211]
The microphone 211 is a general microphone and preferably has no directivity, and is installed on the front surface of the substrate 117 (see FIG. 15). The microphone 211 picks up the sound wave and converts it into an electric signal, generates an acoustic signal x (t), and transmits it to the sound wave direction estimation unit 113 (s1). Note that an acoustic signal generated by the microphone 211 of the sensor unit 210-p is represented as x _p (t).

[Sound wave direction estimation unit 213]
Wave direction estimation section 213, a sound signal _x p (t) that are generated from the microphone 211, the acoustic signal _{x p} generated from the microphone 211 of the adjacent sensor units 210- (p-1) and 210- (p + 1) ₋₁ (t) and x _{p + 1} (t) are received, and the intensity for each sound wave reception direction that arrives in the vicinity of the sensor unit 210-p is calculated from the information. The calculated intensity for each sound wave reception direction is output to the transmitter 115. The sound wave direction estimation unit 213 outputs the acoustic signal x _p (t) generated by the microphone 211 to the adjacent sensor units 210- (p−1) and 210- (p + 1).

In the present embodiment, a part of the intensity φ _p (n) for each sound wave reception direction is calculated for each period T1 as the correlation function φ _p (n, a, b) of the integration time T2 based on the following equation (FIG. 15). A set consisting of a correlation function between x _p (t) and x _p-1 (t) and a correlation function between x _p (t) and x _{p + 1} (t) is defined as the intensity φ _p (n) for each sound wave reception direction. presume. For example, the correlation function φ _p (n, a, b) calculated for the nth time is

Calculate as Here, a is an integer index that distinguishes adjacent sensor units, and takes -1 and 1 in this embodiment. τ is a constant representing the difference in time delay, b is an index representing the time delay, and b is an integer from 0 to B. FIG. 16 shows the relationship between the acoustic signals x ₁ (t), x ₂ (t) and x ₃ (t) and the intensity for each sound wave reception direction in the sensor unit 210-2. The correlation function φ ₂ (n, a, b) is calculated using the acoustic signals x ₁ (t), x ₂ (t) and x ₃ (t) corresponding to the integration time T2 while shifting by the difference τ.

Therefore, for example, when B = 2, a = 1, or a = −1, the intensity φ _p (n) for each sound wave reception direction sent to the transmission unit 115 for the tenth time is φ _p (n) = [φ _{p (10, -1,0), φ} p (10,1,0), φ p (10, -1,1), φ p (10,1,1), φ p (10, -1,2 ), Φ _p (10, 1, 2)].

The transmission unit 115 has the same configuration as that of the first embodiment. In this embodiment, the intensity φ _p (n) (in other words, 2 × (B + 1) correlation functions φ within the period T1 time. _{It is} only necessary that _p (n, a, b)) can be transmitted to the receiving device 120.

When there is no adjacent sensor unit (sensor units 210-1 and 210-P at both ends), since a part of the correlation function φ _p (n, a, b) is not calculated, “null” or blank information is used. Instead, it is sent to the transmission unit 115.

  With such a configuration, only information necessary for sound source estimation (intensity by sound wave reception direction) is extracted in advance between sensor units, and only the information is transmitted by visible light communication.

[Receiver 220]
In the present embodiment, the receiving device 220 includes a light receiving unit 121, a decoding unit 223, and a sound source position reconstruction unit 225 (see FIG. 14).

[Decoding unit 223]
The decoding unit 223 receives the image data, decodes the blinking of the LED of the image data, and acquires _p sound wave reception direction-specific intensities φ _p (n). Further, the position r (n, p) of the P sensor units 210-p is estimated from the blinking position of the LED on the image data with the position of the light receiving unit 121 as a base point. The acquired P sound wave reception direction-specific intensities φ _p (n) and the P estimated sensor unit positions r (n, p) are output to the sound source position reconstruction unit 225.

[Sound source position reconstruction unit 225]
The sound source position reconstructing unit 225 receives the P sound wave receiving direction-specific intensities φ _p (n) and the P sensor unit positions r (n, p), and uses these values to determine the sound source position. The estimated sound source position is output as an output value of the sound source position estimating apparatus 200.

In the present embodiment, the position r (n, p) of the sensor unit is used to correct the intensity φ _p (n) for each sound wave reception direction to reflect the actual sound wave direction. First, the sound source position reconstruction unit 225 obtains the relative distance between adjacent sensor units using the position of the position r (n, p) of the P sensor units. For example, the distance between the sensor units 210-p and 210- (p + a) in the n-th measurement is
d (n, p, p + a) = | r (n, p) -r (n, p + a) |
Is required.

Next, the arrival direction of the sound wave is predicted using the P sound wave reception direction-specific intensities φ _p (n). For example, the actual angle θ _{p (n, a, b)} and 2 × (B + 1) correlation functions φ _p (n, a, b) can be linked by the following expression. The correlation function φ _p (n, a, b) is a strength that can be regarded as a sound wave coming from the following angle θ _{p (n, a, b) when} viewed from the midpoint of the sensor units 210-p and 210- (p + 1). It can be regarded as representing the certainty.

Here, sign (a) represents a function representing the sign of a, sin ⁻¹ () represents an arc tangent function, and c represents sound velocity.

In other words, when a part (half) of the intensity φ _p (n) and φ _{p + 1} (n) according to the sound wave reception direction transmitted from the sensor units 210-p and 210- (p + 1) is brought to the middle point, respectively. An axis (hereinafter referred to as “a (n, p) + r (n, p + 1)) / 2” as viewed in the direction of angles θ _{p (n, 1, b)} and θ _{p + 1 (n, −1, b)} The intensity of sound waves from the “arrival axis”) can be summarized on the diagram as φ _{p (n, 1, b)} and φ _{p + 1 (n, −1, b)} , respectively (see FIG. 17).

Therefore, the sound source position reconstruction unit 225 calculates one of the correlation functions included in the intensity of each sound wave reception direction of the p-th sensor unit and the p′-th (where p ′ = 1, 2,... P, p ≠ p Calculate the average of all the combinations with one of the correlation functions included in the intensity of the sound wave reception direction of the sensor unit of ') for all the combinations, and estimate that the sound source is at the intersection of the arrival axes where the sum of the average is the maximum (Fig. 18). For example, in the case of FIG. 18, the addition average (φ _p (n, 1,1) + φ _p−1 (n, 1,1)) ÷ 2 of the intersection 4 closest to the sound source 3 is maximized, and it is estimated that there is a sound source. Is done. The position of the intersection of the arrival axes is obtained from the midpoint of adjacent sensor units obtained from the positions of the P sensor units, the sound speed c, the constant τ, and the relative distance d (n, p, p + a).

Note that θ _{p (a, b) in} the sound source position reconstruction unit 225 may be calculated only once if the position of the sensor unit is fixed.

<Effect>
With such a configuration, the same effect as that of the first embodiment can be obtained. In the present embodiment, only a communication line between the power source and the adjacent unit is required, and a large number of sensor units can be easily arranged over a wide range.

<Modification>
A modification similar to the first embodiment can be considered.

As is clear from FIG. 17, the correlation functions φ _p (n, 1, 0) and φ _{p + 1} (n, −1, 0) have the same value. That is, the correlation function when b = 0 is the same value. Therefore, in order to avoid duplication of transmission data, the sensor unit 210-p does not send φ _p (n, 1, 0) or the sensor unit 210- (p + 1) does not send φ _{p + 1} (n, 1, 0). It is good also as a structure. In short, the overlapping correlation function may be configured such that the sound wave direction estimating unit 213 does not calculate or the transmitting unit 115 does not transmit. With such a configuration, the calculation amount and the transmission amount can be reduced.

When the final sound source position is specified, the sound source position reconstruction unit 125 calculates all the intersections of the arrival axes, and further calculates the addition average of the correlation function φ _p (n, a, b) at that position. . As the number of sensor units increases, there is a problem that the calculation amount increases rapidly. Therefore, the following processing may be performed.

At the stage where the sensor units 210-p and 210- (p + 1) are gathered from the midpoint position (see FIG. 17), the correlation function φ _p (n, a, b) in which the subsequent calculations are gathered at the midpoint position. This is limited to the direction of the arrival axis having the maximum value. That is, the sound source position reconstructing unit 225 includes 2 × (B + 1) correlation functions φ _p (included in the intensity φ _p (n) for each sound wave receiving direction estimated by the sound wave direction estimating unit 213 of each sensor unit 210-p. The maximum correlation function is selected from n, a, b). Further, the maximum correlation function is selected from each of the plurality of sensor units. Only the addition average of a plurality of selected maximum correlation functions is obtained, and it is estimated that there is a sound source at the intersection of the arrival axes that maximizes the addition average.

Referring to the example of FIG. 17, the intensities according to the sound wave receiving directions recombined based on the midpoint position (r (n, p) + r (n, p + 1)) / 2 of the sensor units 210-p and 210- (p + 1). φ _p (n) = [φ _p (n, 1, 0), φ _p (n, 1, 1), φ _p (n, 1, 2), φ _{p + 1} (n, 1, 1), φ _{p + 1} ( n, 1, 2)]. When this maximum value is φ _p (n, 1, 1), the arrival axis from the midpoint position used in the subsequent calculations is considered only with respect to the arrival axis of the angle θ _{p (n, 1, 1)} . With such a configuration, it is possible to greatly reduce the amount of calculation while ensuring the accuracy of sound source position estimation to some extent.

In the present embodiment, a correlation function between x _p (t) and x _p-1 (t) and a correlation function between x _p (t) and x _{p + 1} (t) are used as the similarity. Other values may be used as the similarity. For example, the correlation between adjacent sensor units normalized by the energy of the acoustic signal may be used as the similarity. At this time, the similarity (information to be transmitted) is, for example,

Is calculated.

  In this embodiment, a = 1 or a = −1 and a correlation function with one adjacent sensor unit is obtained, but a correlation function with two or more adjacent sensor units may be provided. When obtaining correlation functions with A adjacent sensor units, 2A × (B + 1) correlation functions may be obtained. With such a configuration, it is possible to estimate the sound source position more finely.

<Other variations>
The present invention is not limited to the above-described embodiments and modifications. For example, the various processes described above are not only executed in time series according to the description, but may also be executed in parallel or individually as required by the processing capability of the apparatus that executes the processes. In addition, it can change suitably in the range which does not deviate from the meaning of this invention.

<Program and recording medium>
The above-described sound source position estimation apparatus can be functioned by a computer. More specifically, each sensor unit and receiving device constituting the sound source position estimating device can be caused to function by a computer. In this case, each process of a program for causing a computer to function as a target device (a device having the functional configuration shown in the drawings in various embodiments) or a process procedure (shown in each embodiment) is processed by the computer. A program to be executed by the computer may be downloaded from a recording medium such as a CD-ROM, a magnetic disk, or a semiconductor storage device or via a communication line into the computer, and the program may be executed.

100, 200 Sound source position estimation device 110, 210 Sensor unit 111, 211 Microphone 113, 213 Sound wave direction estimation unit 115 Transmission unit 120, 220 Reception device 121 Light reception unit 123, 223 Decoding unit 125, 225 Sound source position reconstruction unit 200 Sound source position Estimating device 210 Sensor unit 210 Each sensor unit 210 Sensor unit 211 Microphone

Claims (8)

A sound source position estimating device including a plurality of sensor units and a receiving device,
The sensor unit is
A microphone that converts sound waves into electrical signals and generates acoustic signals;
Using the acoustic signal, a sound wave direction estimating unit that estimates a strength for each sound wave receiving direction, which is a set of sound wave strengths for each direction,
A transmitter that encodes the intensity for each sound wave receiving direction, blinks the light emitting element, and transmits the visible light communication;
The receiving device is:
A light receiving unit comprising an image sensor and receiving blinking of the light emitting element as image data;
Decoding the blinking of the light emitting element on the image data to obtain the intensity for each sound wave receiving direction, and further, the position of the sensor unit from the blinking position of the light emitting element on the image data based on the position of the light receiving unit A decoding unit for estimating
A sound source position reconstructing unit that estimates a sound source position using the intensity for each sound wave receiving direction and the position of the sensor unit;
Sound source position estimation device. The sound source position estimation apparatus according to claim 1,
The microphone has directivity and is included in plural for one sensor unit,
The plurality of microphones included in one sensor unit each have a different sound collection area,
The sound wave direction estimation unit estimates a set of intensity of the acoustic signal generated from each microphone as the intensity for each sound wave reception direction,
The sound source position reconstruction unit estimates a sound collection area of each microphone from the position of the sensor unit, and selects the microphone corresponding to the maximum intensity from the intensity of each sensor unit in the sound wave reception direction. , It is estimated that there is a sound source in the overlapping portion of the microphone sound collection area selected from each of the plurality of sensor units,
Sound source position estimation device. The sound source position estimation apparatus according to claim 1,
P is an integer of 3 or more, the number of the sensor units, p = 1, 2,..., P, p ′ = 1, 2,..., P, p ≠ p ′,
The sensor unit includes one omnidirectional microphone,
The plurality of sensor units are connected in cascade, and between adjacent sensor units, the acoustic signals of the microphones are transmitted to each other,
The sound wave direction estimating unit calculates a similarity between the acoustic signal generated from each of the microphones and the acoustic signal generated from an adjacent microphone, and sets the similarity to the sound wave receiving direction. Estimated as another strength,
The similarity is the strength of the sound wave from the arrival axis viewed in the direction of the angle determined by the relative distance between the sensor units and the difference in time delay when viewed from the midpoint of the position of the adjacent sensor unit. The position reconstruction unit includes one of the similarities included in the intensity for each sound wave reception direction of the p-th sensor unit, and one of the similarities included in the intensity for each sound wave reception direction of the p'th sensor unit. Is estimated, and it is estimated that there is a sound source at the intersection of the arrival axes where the average is the maximum.
Sound source position estimation device. The sound source position estimation apparatus according to claim 3,
The sensor unit does not transmit the overlapping similarity.
Sound source position estimation device. The sound source position estimation apparatus according to claim 3 or 4,
The sound source position reconstruction unit selects the maximum similarity from the similarities included in the intensity for each sound wave reception direction estimated by the sound wave direction estimation unit of each sensor unit, and a plurality of the sensors Obtaining an average of a plurality of maximum similarities each selected from a unit, and estimating that a sound source exists at the intersection of the arrival axes that maximizes the average of the addition,
Sound source position estimation device. The sound source position estimation apparatus according to claim 2,
The sensor unit is three-dimensionally arranged with respect to the light receiving unit,
The microphone sound collection area is three-dimensionally arranged.
Sound source position estimation device. A sound source position estimation method using a plurality of sensor units and a receiving device,
The microphone of the sensor unit converts a sound wave into an electrical signal and generates an acoustic signal; and
A sound wave direction estimating step in which the sound wave direction estimating unit of the sensor unit uses the acoustic signal to estimate the strength for each sound wave receiving direction, which is a set of sound wave strengths for each direction;
A transmission step in which the transmission unit of the sensor unit encodes the intensity for each sound wave reception direction, blinks the light emitting element, and transmits the visible light communication.
A light receiving unit comprising an image sensor of the receiving device receives the blinking of the light emitting element as image data; and
The decoding unit of the receiving device decodes the blinking of the light emitting element on the image data to acquire the intensity for each sound wave receiving direction, and further, the light emitting element of the light emitting element on the image data on the basis of the position of the light receiving unit. A decoding step of estimating the position of the sensor unit from the blinking position;
A sound source position reconstruction unit of the reception device includes a sound source position reconstruction step of estimating a sound source position using the intensity for each sound wave reception direction and the position of the sensor unit.
Sound source position estimation method.   The program for functioning a computer as an audio | voice position estimation apparatus in any one of Claims 1-6.

Images