JP7162192B2 - Bionumber estimation device, bionumber estimation method, and program - Google Patents

Description

The present disclosure relates to an estimating device, a living body number estimating device, an estimating method, and a program, and in particular, an estimating device and a living body number estimating device that estimate at least one of the position and number of living bodies using radio signals. , an estimation method, and a program.

A technique for detecting a detection target using a signal transmitted wirelessly has been developed (see Patent Document 1, for example).

Patent Literature 1 discloses a technology capable of knowing the number and positions of persons to be detected by analyzing the eigenvalues of the components including the Doppler shift using the Fourier transform of a signal received wirelessly. ing.

JP 2015-117972 A JP 2014-228291 A Japanese Patent No. 5047002 Japanese Patent No. 5025170

However, with the technology disclosed in Patent Document 1, when the number of living organisms to be detected is large, and the difference in the magnitude of the eigenvalue corresponding to the living organisms is small, the estimation accuracy of the number of living organisms to be detected is is reduced, and the accuracy of estimating the position of the living body is also reduced.

The present disclosure has been made in view of the above circumstances, and provides an estimating device, a living body number estimating device, an estimating method, and a program that can accurately estimate at least one of the position and number of living organisms using radio signals. intended to provide

In order to achieve the above object, an estimating device or the like according to an aspect of the present disclosure provides components corresponding to one or more living organisms present in a predetermined space from a received signal transmitted in a predetermined space. a biometric information extraction unit that extracts biometric information, an eigenvector calculation unit that calculates one or more eigenvectors of a biometric correlation matrix obtained from the biometric information, and a predetermined position estimation method using the biometric correlation matrix, a first position estimating unit for estimating a position including a virtual image in the one or more living bodies; A second steering vector output unit that extracts and outputs the first steering vector corresponding to each position including the virtual image of the living body as a second steering vector, and using the eigenvector and the second steering vector, the and a second position estimator that estimates at least one of the positions and number of one or more living bodies.

According to the estimation device and the like of the present disclosure, it is possible to accurately estimate at least one of the positions and the number of living bodies using radio signals.

FIG. 1 is a block diagram showing the configuration of a sensor according to Embodiment 1. FIG. FIG. 2 is a diagram showing an example of the arrangement of transmitters and receivers and the concept of a method of estimating the position of an estimating device. FIG. 3 is a diagram showing an example of the result of calculating the correlation between the second steering vector and the eigenvector and the evaluation function according to the first embodiment. FIG. 4 is a flow chart showing a sensor population estimation process according to the first embodiment. 5 is a block diagram showing a configuration of a first position estimating unit according to a modification of Embodiment 1. FIG. FIG. 6 is a block diagram showing the configuration of a composite sensor according to Embodiment 2. FIG. 7A is a conceptual diagram of an estimated position distribution calculated from the positions and the number of persons estimated by one sensor according to Embodiment 2. FIG. 7B is a conceptual diagram of an estimated position distribution calculated from the positions and the number of persons estimated by one sensor according to Embodiment 2. FIG. 8 is a conceptual diagram showing an example of an estimated position distribution calculated by a living body number estimating unit according to Embodiment 2. FIG. FIG. 9 is a flow chart showing a process of estimating the number of people of the composite sensor according to the second embodiment. FIG. 10 is a diagram showing an experimental environment using the estimation method according to the second embodiment. FIG. 11A is a diagram showing an estimated position distribution by the Capon method. FIG. 11B is a diagram showing estimated position distribution in the example. FIG. 12 is a spectrum showing the variance of the estimated position distribution shown in FIG. 11B.

(Findings on which this disclosure is based)
Techniques for detecting a detection target using a signal transmitted wirelessly have been developed (see Patent Documents 1 to 4, for example).

For example, Patent Literature 1 discloses a technique of estimating the number and positions of persons to be detected by analyzing eigenvalues of components including Doppler shift using Fourier transform. Specifically, a Fourier transform is performed on the received signal, an autocorrelation matrix is obtained for a waveform obtained by extracting a specific frequency component, and eigenvalue decomposition is performed on the autocorrelation matrix to obtain eigenvalues. In general, each of the eigenvalues and eigenvectors represents a propagation path of radio waves from a transmitting antenna to a receiving antenna, that is, one path. Originally, there are various paths such as direct waves or reflections from fixed objects such as walls, and each path corresponds to each eigenvalue and eigenvector. However, since the technique of Patent Document 1 removes components that do not contain biological information, only paths reflected by the biological body and paths corresponding to noise appear in the eigenvalues and eigenvectors. Here, since the eigenvalue corresponding to noise is smaller than the eigenvalue corresponding to living organisms, the number of living organisms can be estimated by counting the number of eigenvalues larger than a predetermined threshold among the eigenvalues.

However, in the technique disclosed in Patent Document 1, when the target living body is far away or when there are many living bodies, the difference between the eigenvalue corresponding to the living body and the eigenvalue corresponding to noise is reduced, and the accuracy of the number estimation is reduced. There is a problem of lowering This is because when the Doppler effect is very weak, the Doppler shift is influenced by the internal noise of the receiver, interference waves coming from outside the detection target, and objects that cause Doppler shift outside the detection target. This is because it becomes difficult to detect a weak signal that is

Patent Literature 2 discloses a technique of estimating the position of an object using a direction estimation algorithm such as the MUSIC (MUltiple Signal Classification) method. Specifically, a receiving station that receives a signal emitted by a transmitting station performs a Fourier transform on the received signal, obtains an autocorrelation matrix for a waveform that extracts a specific frequency component, and uses the direction of the MUSIC method or the like. Apply an estimation algorithm. This makes it possible to estimate the direction of the living body with high precision. However, since the MUSIC method used in Patent Document 2 is a method for estimating the direction of a living body position by using the number of living bodies to be detected as known in advance, the number of living bodies cannot be estimated.

Further, for example, in Patent Document 3, the number of arriving waves, that is, the number of transmitters such as mobile phones, is calculated from the correlation between the eigenvectors of received signals received by a plurality of antennas and the steering vectors in the range where radio waves may arrive. , is disclosed.

Further, for example, in Patent Document 4, various arrival wave numbers are assumed for received signals received by a plurality of antennas, an evaluation function using a steering vector is calculated for each, and the arrival that maximizes the evaluation function Techniques for estimating the number of waves as the true number of arriving waves have been disclosed.

However, the techniques disclosed in Patent Documents 3 and 4 are techniques for estimating the number of transmitters that emit radio waves, and cannot estimate the number of living organisms.

In view of the above, the inventors have made an estimation that can accurately estimate at least one of the position and number of living organisms using radio signals without making the target living organism possess special equipment such as a transmitter. This led to the idea of devices (sensors) and the like.

That is, the estimating device according to one aspect of the present disclosure extracts biometric information, which is a component corresponding to one or more living organisms present in the predetermined space, from a received signal that is received from a signal transmitted in the predetermined space. a biological information extraction unit, an eigenvector calculation unit that calculates one or more eigenvectors of a biological correlation matrix obtained from the biological information, and a predetermined position estimation method using the biological correlation matrix, in the one or more living organisms, a first position estimating unit for estimating a position including a virtual image; and one or more virtual images of the living body estimated by the first position estimating unit among a plurality of first steering vectors pre-stored in a storage unit. a second steering vector output unit for extracting and outputting first steering vectors corresponding to respective positions as second steering vectors; a second position estimator for estimating at least one of the numbers.

With this configuration, two different estimation methods, ie, eigenvectors instead of eigenvalues and a predetermined position estimation method such as the Capon method, can be used together. At least one of them can be estimated with high accuracy. In this specification, a living body whose position is estimated is called an image, an image that does not actually exist is defined as a virtual image, and an image that actually exists is defined as a real image.

Here, for example, the second position estimation unit calculates the product or correlation between the eigenvector and the second steering vector, and calculates the eigenvector or the first The position pointed to by the two steering vectors is assumed as the position.

This makes it possible to accurately estimate the position of a living body using radio signals.

Also, for example, the second position estimation unit estimates, as the number, the number of the eigenvectors or the second steering vectors that take a value equal to or greater than a threshold among the calculated products or correlations.

This makes it possible to accurately estimate the number of living organisms using radio signals.

Here, for example, the predetermined position estimation method may be the Capon method, or the predetermined position estimation method may be the beamformer method.

As a result, position estimation methods such as the Capon method, which can be used even if the number of people is unknown, have poor position estimation accuracy. At least one of can be estimated with high accuracy.

Further, for example, the first position estimating unit further includes a prior number of people estimating unit that estimates the number of people including the virtual image in the one or more living organisms using the eigenvalues of the biometric correlation matrix, and the predetermined position estimating unit The method may be the MUSIC method using the number of people estimated by the prior population estimation unit.

As a result, even if the number of people is inaccurate including a virtual image, by using the position estimation method by the MUSIC method in which the number of people estimated for the time being using the eigenvalues is known in advance and the method using the eigenvalues, Radio signals can be used to accurately estimate at least one of the location and number of living organisms.

Here, for example, a transmitter having N (N is a natural number of 2 or more) transmitting antenna elements, M (M is a natural number of 2 or more) receiving antenna elements, and each of the M receiving antenna elements a transfer function calculator that calculates a plurality of complex transfer functions representing propagation characteristics between the N transmitting antenna elements and the M receiving antenna elements from the received signals received for a predetermined period in and the biological information extraction unit extracts, from the plurality of complex transfer functions calculated by the transfer function calculation unit, fluctuation components due to the influence of the living body and fluctuations in each of the M receiving antenna elements. components are extracted as the biometric information, and the eigenvector calculation unit calculates the biocorrelation matrix from the variation components in each of the M receiving antenna elements extracted by the biometric information extraction unit. Eigenvectors of the biometric correlation matrix are calculated.

Furthermore, the living body count estimation device according to an aspect of the present disclosure further includes a plurality of estimation devices according to the above aspects, and the positions of one or more living bodies existing in each of the predetermined spaces estimated by the plurality of estimation devices. based on the estimated position distribution calculation unit that calculates the estimated position distribution of the entire space consisting of the predetermined space that partially overlaps; and a living body number estimating unit for estimating the number of living bodies, which is the number of one or more living bodies existing in the entire space, using each of the numbers of living bodies.

With this configuration, by using a plurality of estimating devices, a plurality of transmitting stations and receiving stations can be linked, so that the range in which the number of people can be estimated can be expanded.

Here, for example, the living body number estimating unit may estimate, as the living body number, the number of local minimum values of variance of the positions estimated by the plurality of estimation devices. For example, the living body number estimating unit may The number of maximum values of the density of positions estimated by the plurality of estimation devices may be estimated as the number of living organisms.

With this configuration, the same living body that is redundantly estimated by a plurality of estimating devices can be counted as one living body, so that the range that can be estimated with high accuracy can be expanded.

In addition, an estimation method according to an aspect of the present disclosure extracts biometric information, which is a component corresponding to one or more living organisms present in a predetermined space, from a received signal that is a signal transmitted in a predetermined space. A biological information extraction step, an eigenvector calculation step of calculating one or more eigenvectors of a biological correlation matrix obtained from the biological information, and a predetermined position estimation method using the biological correlation matrix, in the one or more living organisms, a first position estimating step of estimating a position including a virtual image; and one or more virtual images of the living body estimated in the first position estimating step among a plurality of first steering vectors pre-stored in a storage unit. a second steering vector output step of extracting and outputting a first steering vector corresponding to each position as a second steering vector; and a second position estimation step of estimating at least one of the numbers.

Further, a program according to an aspect of the present disclosure extracts biological information, which is a component corresponding to one or more living organisms present in the predetermined space, from a received signal transmitted in the predetermined space. An information extraction step, an eigenvector calculation step of calculating one or more eigenvectors of a biometric correlation matrix obtained from the biometric information, and a predetermined position estimation method using the biometric correlation matrix to generate a virtual image of the one or more biometric objects. a first position estimation step of estimating a position including a virtual image of the one or more living bodies estimated in the first position estimation step among a plurality of first steering vectors pre-stored in a storage unit a second steering vector output step of extracting and outputting the corresponding first steering vectors as second steering vectors; and a second position estimation step of estimating at least one of the above.

It should be noted that the present disclosure can be implemented not only as a device, but also as an integrated circuit including processing means included in such a device, as a method using processing means that constitute the device as steps, or as steps can be realized as a program to be executed by a computer, or as information, data or a signal indicating the program. These programs, information, data and signals may be distributed via recording media such as CD-ROMs and communication media such as the Internet.

Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. It should be noted that each of the embodiments described below is a preferred specific example of the present disclosure. Numerical values, shapes, materials, components, arrangement positions and connection forms of components, steps, order of steps, and the like shown in the following embodiments are examples, and are not intended to limit the present disclosure. In addition, among the constituent elements in the following embodiments, constituent elements that are not described in independent claims representing the highest concept of the present disclosure will be described as arbitrary constituent elements that constitute more preferred embodiments. Further, in the present specification and drawings, constituent elements having substantially the same functional configuration are denoted by the same reference numerals, thereby omitting redundant description.

(Embodiment 1)
Below, the estimation method of the sensor 100 in Embodiment 1, etc. are demonstrated, referring drawings.

[Configuration of sensor 100]
FIG. 1 is a block diagram showing the configuration of sensor 100 according to Embodiment 1. As shown in FIG. FIG. 2 is a diagram showing an example of arrangement of the transmitter 10 and the receiver 20 and the concept of the position estimation method of the estimation device 1 according to the first embodiment. Sensor 100 shown in FIG. A detailed configuration will be described below.

[Transmitter 10]
The transmitter 10 has N (N is a natural number equal to or greater than 2) transmitting antenna elements. Specifically, as shown in FIG. 1, the transmitter 10 is composed of a transmission section 11 and a transmission antenna section 12 having an _MT element array antenna. Here, M _T corresponds to N transmit antenna elements (N is a natural number of 2 or more) of M T elements, and M _T is a natural number of 2 or more.

<Transmitting antenna unit 12>
The transmission antenna unit 12 is composed of a transmission array antenna of _MT elements. Here, the transmission array antenna of M _T elements corresponds to N transmission antenna elements, and M _T is a natural number of 2 or more.

The transmitting antenna section 12 transmits the high-frequency signal generated by the transmitting section 11 .

<Transmitting unit 11>
The transmitting unit 11 generates a high-frequency signal used for estimating the presence/absence, position, and/or number of people of the living body 200 . For example, the transmission unit 11 generates a CW (Continuous Wave) and transmits the generated CW from the transmission antenna unit 12 as a transmission wave. The signal to be generated is not limited to a sine wave signal such as CW, and may be a modulated signal (modulated wave signal).

[Receiver 20]
The receiver 20 includes M receiving antenna elements (M is a natural number of 2 or more) and a complex transfer function calculator 23 . Specifically, the receiver 20 includes a receiving antenna section 21, a receiving section 22, and a complex transfer function calculating section 23, as shown in FIG.

<Receiving antenna unit 21>
The receiving antenna section 21 is composed of a receiving array antenna of _MR elements. Here, the receiving array antenna of MR elements corresponds to _M receiving antenna elements, and _MR is a natural number of 2 or more. The receiving antenna unit 21 is a receiving array antenna and receives high frequency signals.

<Receiver 22>
The receiving unit 22 converts the high-frequency signal received by the receiving antenna unit 21 into a low-frequency signal that can be processed using, for example, a down converter. The receiver 22 transmits the converted low-frequency signal to the complex transfer function calculator 23 .

Although the transmitter 10 and the receiver 20 are configured adjacent to each other in FIG. 1, the present invention is not limited to this. For example, as shown in FIG. 2, the transmitter 10 and the receiver 20 may be arranged at distant locations.

In addition, in FIG. 1, the transmitting array antenna used by the transmitter 10 and the receiving array antenna used by the receiver 20 are different and arranged at different positions, but the present invention is not limited to this. The transmitting array antenna and the receiving array antenna used by the transmitter 10 and the receiver 20 may be shared. Further, transmitter 10 and receiver 20 may be shared with wireless device hardware, such as a Wi-Fi® router or handset.

<Complex transfer function calculator 23>
The complex transfer function calculation unit 23 is an example of a transfer function calculation unit. A plurality of complex transfer functions representing propagation characteristics between and are calculated.

Specifically, the complex transfer function calculation unit 23 calculates the complex transfer function representing the propagation characteristics between the reception array antenna of the reception antenna unit 21 and the transmission array antenna of the transmission antenna unit 12 from the signal observed by the reception array antenna of the reception antenna unit 21 . Calculate the transfer function. In the present embodiment, the complex transfer function calculation unit 23 calculates M _T transmission antenna elements of the transmission antenna unit 12 and M _R reception antenna elements of the reception antenna unit from the low-frequency signal transmitted by the reception unit 22. A complex transfer function representing the propagation characteristics between elements is calculated.

The complex transfer function calculated by the complex transfer function calculation unit 23 includes reflected waves and/or scattered waves, which are signals in which part of the transmission waves transmitted from the transmission antenna unit 12 are reflected and scattered by the living body 200. may contain. In addition, the complex transfer function calculated by the complex transfer function calculation unit 23 includes reflected waves that do not pass through the living body 200, such as direct waves from the transmission antenna unit 12 and reflected waves originating from fixed objects. In addition, the amplitude and phase of the signal reflected or scattered by the living body 200, that is, the reflected wave and the scattered wave passing through the living body 200, constantly fluctuate according to the living body's activities such as respiration and heartbeat. In the following description, it is assumed that the complex transfer function calculated by the complex transfer function calculation unit 23 includes reflected waves and scattered waves, which are signals reflected or scattered by the living body 200 .

[Estimation device 1]
The estimation device 1 uses a position estimation method such as the Capon method, which can be used even if the number of people is unknown, and a position estimation method using eigenvalues in combination to obtain at least one of the position and number of living bodies using radio signals. Estimate with good accuracy. In the present embodiment, as shown in FIG. 1, the estimation device 1 includes a biological information extraction unit 30, a first position estimation unit 40, a second steering vector output unit 60, an eigenvector calculation unit 70, a second and a position estimator 80 . Note that the estimation device 1 may include a storage unit 50 . Also, the estimating device 1 does not have to include the transmitter 10 and the receiver 20, but may include them. A detailed configuration of the estimation device 1 will be described below.

<Biological information extraction unit 30>
The biometric information extraction unit 30 extracts biometric information, which is a component corresponding to one or more living organisms existing in a predetermined space, from a received signal that has been transmitted to a predetermined space. More specifically, the biometric information extraction unit 30 extracts, from the plurality of complex transfer functions calculated by the complex transfer function calculation unit 23, fluctuation components due to the influence of the living body, which are fluctuation components in each of the M receiving antenna elements. is extracted as biometric information.

In the present embodiment, the biological information extraction unit 30 extracts signals from signals observed by the receiving array antenna of the receiving antenna unit 21, which are transmitted from the transmitting array antenna of the transmitting antenna unit 12 and reflected or reflected by one or more living bodies 200. A biological component, which is a scattered signal component, is extracted.

More specifically, the biological information extraction unit 30 records the complex transfer functions calculated by the complex transfer function calculation unit 23 in chronological order in which the signals were observed. Then, the biological information extraction unit 30 extracts the fluctuation component due to the influence of the living body 200 from the changes in the complex transfer function recorded in time series. Fluctuation components of the complex transfer function due to the influence of the living body 200 are hereinafter referred to as biological components or biological information.

As a method of extracting the biological component, for example, a method of extracting only the biological component after conversion to the frequency domain by Fourier transform, etc., or a method of extracting by calculating the difference between the complex transfer functions at two different times. be. By these methods, the components of the direct wave and the reflected wave that pass through the fixed object are removed, leaving only the biological component that passes through the living body 200 .

In the present embodiment, there are M _T transmitting antenna elements constituting the transmitting array antenna, and M _R receiving antenna elements constituting the receiving array antenna, that is, a plurality of them. The biological component of the function via the biological body 200 also becomes plural. Hereinafter, a matrix that summarizes these biocomponents will be referred to as a biocomponent channel matrix H(t).

In this manner, the biological information extraction unit 30 calculates the biological component channel matrix H(t) from the time-series complex transfer function. The biological information extraction unit 30 also outputs the biological component channel matrix H(t) obtained from the extracted biological components to the first position estimation unit 40 and the eigenvector calculation unit 70 .

<Eigenvector calculator 70>
The eigenvector calculator 70 calculates one or more eigenvectors of the biometric correlation matrix obtained from the biometric information. More specifically, the eigenvector calculating unit 70 calculates a biocorrelation matrix from the biometric information, which is the variation component of each of the M receiving antenna elements, extracted by the bioinformation extracting unit 30, and calculates the biocorrelation matrix Calculate the eigenvectors of

In the present embodiment, the eigenvector calculator 70 calculates the correlation matrix R from the biological component channel matrix H(t) received from the biological information extractor 30 according to (Equation 1). In (Formula 1), E[·] represents an ensemble average, and {·} ^H represents a complex conjugate transposition operation. The correlation matrix R is an example of the biometric correlation matrix described above.

The eigenvector calculation unit 70 also transfers the eigenvectors obtained by performing eigenvalue decomposition of the calculated correlation matrix R to the second position estimation unit 80 . Here, when the correlation matrix R of the biological component channel matrix H(t) is eigenvalue decomposed, the following (formula 2) to (formula 4) can be written. Note that u ₁ to u _N in (Equation 3) represent eigenvectors having N elements. In (Formula 4), λ ₁ to λ _N represent eigenvalues corresponding to eigenvectors.

Here, the concept of the calculated eigenvectors will be described with reference to FIG. The eigenvector calculation unit 70 calculates eigenvectors 1011-A to 1011-D from the biological component channel matrix H(t) obtained in the space where the living body 200-A and the living body 200-B exist, for example, as shown in FIG. Suppose we calculated In this case, the calculated eigenvectors 1011-A and 1011-C point in the direction of living body 200-B as seen from the receiving antenna of receiver 20 and the transmitting antenna of transmitter . Also, the calculated eigenvectors 1011-B and 1011-D indicate the direction of the living body 200-A viewed from the receiving antenna of the receiver 20 and the transmitting antenna of the transmitter . This is because each eigenvector calculated by the eigenvector calculation unit 70 represents one path from the transmitting antenna to the receiving antenna, and therefore corresponds to a vector indicating the position of the living body to be detected as viewed from the transmitting antenna and the receiving antenna. It will be.

<Storage unit 50>
The storage unit 50 pre-stores a plurality of pre-calculated first steering vectors. The storage unit 50 is accessed by the first position estimation unit 40 and the second steering vector output unit 60, which will be described later, and a plurality of stored first steering vectors are referenced.

Here, a method for calculating a plurality of first steering vectors will be described.

First, according to (Equation 5), a transmitting-side first steering vector a _t (θ) representing the phase difference between each antenna element of the transmitting array antenna is calculated at predetermined intervals within a predetermined range. For example, according to (Equation 5), the front direction of the transmitting array antenna is assumed to be 0 degrees, and the first steering vector on the transmitting side can be calculated from -45 degrees to 45 degrees in increments of 1 degree.

In (Formula 5), λ is the wavelength, N is the number of antenna elements constituting the transmission array antenna, j is the imaginary unit, and e is the base of natural logarithms.

Next, similarly, the receiving side first steering vector a _r (θ) representing the phase difference between each antenna element of the receiving array antenna is obtained with respect to a predetermined θ in the transmitting side first steering vector a _t (θ) as calculate.

In (Equation 6), λ is the wavelength, M is the number of antenna elements forming the receiving array antenna, j is the imaginary unit, and e is the base of natural logarithms.

Then, the transmitting-side first steering vector a _t (θ) and the corresponding receiving-side first steering vector a _r (θ) are stored in the storage unit 50 as first steering vectors.

<First position estimation unit 40>
The first position estimating unit 40 uses a biometric correlation matrix obtained from biometric information to estimate a position, including a virtual image, of one or more living bodies by a predetermined position estimation method. Here, the predetermined position estimation method is the Capon method, but may be the beamformer method.

In the present embodiment, the first position estimating unit 40 estimates the position candidates of the living body 200 using the biological component channel matrix H(t) obtained from the biological components extracted by the biological information extracting unit 30 . The first position estimating unit 40 estimates the position candidate of the living body 200 using a position estimating method that can be used even if the number of living bodies is unknown, such as the beamformer method or the Capon method. The beamformer method and the Capon method are less accurate than a position estimation method such as the MUSIC method, which is performed when the number of living organisms is known.

In the following, a case of estimating candidates for the position of the living body 200 from the living body component channel matrix H(t) using the Capon method as an example of a position estimation method that can be used even if the number of living bodies is unknown will be described.

In the Capon method, first, an evaluation function P _Capon corresponding to the power when the directivity main lobe of the receiving array antenna or the transmitting array antenna is directed in a certain direction θ is calculated for every θ (referred to as the Capon spectrum). is done). Then, the direction θ in which the Capon spectrum peaks is estimated as the target direction.

The evaluation function P _Capon used in the Capon method is shown in (Equation 7) below.

More specifically, position estimation by the Capon method is performed from both the transmitting array antenna of the transmitting antenna unit 12 and the receiving array antenna of the receiving antenna unit 21 to create a two-dimensional Capon spectrum, and the detected peak Estimate location. Note that when detecting the peak, the Capon spectrum may be smoothed in order to reduce the influence of noise. In the following description, for example, _NP peaks are detected.

Here, the concept of the body position calculated by the Capon method will be described with reference to FIG. FIG. 2 shows a diagram in which the peak positions of the Capon spectrum estimated by the Capon method by the first position estimating unit 40 in the space where the living body 200-A and the living body 200-B exist are superimposed on the coordinate system. . That is, the peak positions 1001, 1002, and 1003 indicated by plus signs represent the positions of the peaks estimated by the first position estimator 40. FIG. Note that the peak position 1001 is the position of the peak (hereinafter also referred to as the peak position) detected even though the living body does not actually exist, and is a virtual image.

<Second Steering Vector Output Unit 60>
The second steering vector output unit 60 corresponds to each position including one or more virtual images of the living body estimated by the first position estimation unit 40 among the plurality of first steering vectors pre-stored in the storage unit 50. The first steering vector is extracted and output as the second steering vector.

In the present embodiment, the second steering vector output unit 60 outputs N _p peak positions estimated by the first position estimation unit 40 among the first steering vectors calculated in advance and stored in the storage unit 50. (the position of the living body including the virtual image) is extracted. Specifically, the second steering vector output unit 60 outputs the first steering vector corresponding to the direction θ of the peak position received from the first position estimation unit 40 as viewed from the transmission array antenna, using (Equation 5) and ( N _p pieces are output according to equation 6). In the present embodiment, the first steering vector extracted by the second steering vector output unit 60 is called a second steering vector in order to distinguish it from the first steering vector stored in the storage unit 50. Denote the vector a _peak .

<Second Position Estimation Unit 80>
The second position estimator 80 estimates at least one of the positions and number of one or more living bodies using the calculated eigenvectors and the output second steering vector. Specifically, the second position estimating unit 80 calculates the product or correlation between the calculated eigenvector and the output second steering vector, and the eigenvector or the second The position indicated by the two steering vectors may be estimated as the position of the living body. Also, the second position estimator may estimate the number of eigenvectors or second steering vectors having a value equal to or greater than a threshold among the calculated products or correlations as the number of living organisms. Although the threshold value is 0.5, for example, it may be appropriately determined between 0 and less than 1.

In the present embodiment, the second position estimation unit 80 calculates the correlation between the second steering vector a _peak output by the second steering vector output unit 60 and the eigenvector u calculated by the eigenvector calculation unit 70, Estimate the number and location of organisms.

Here, for example, the second position estimator 80 may estimate the number of living bodies using an evaluation function that calculates the correlation between the second steering vector a _peak and the eigenvector u shown in (Equation 7).

In (Expression 7), the correlation between the second steering vector a _peak and the eigenvector u is calculated by calculating the inner product of the second steering vector a _peak and the eigenvector u. Therefore, if the eigenvector u has a high correlation with the second steering vector a _peak , the value is close to 1, and if there is no high correlation, the value is close to 0.

In addition, in (Formula 7), an operation for obtaining the maximum value of correlation is performed, but any operation that can determine the presence or absence of a term having a correlation value close to 1, such as a sum, can be substituted.

An example of calculating the correlation between the second steering vector a _peak and the eigenvector u by calculating the inner product will be described below with reference to FIG.

FIG. 3 is a diagram showing an example of the result of calculating the correlation between the second steering vector a _peak and the eigenvector u and the evaluation function according to the first embodiment.

As shown in FIG. 3, the correlation between the second steering vector a _peak1 and the eigenvector u2 indicated by reference numeral 1101 is 0.9, which exceeds the threshold value of 0.5 and is close to 1. Also, the correlation between the second steering vector a _peak2 and the eigenvector u1 indicated by reference numeral 1102 is 0.83, which exceeds the threshold value of 0.5 and is close to 1.

Also, from the calculation results of the evaluation functions P _peaki for the second steering vectors a _peak1 to a _peakNp , the two evaluation functions P _peaki indicated by reference numerals 1103 and 1104 exceed the threshold value of 0.5. As a result, the estimated number of living organisms is two.

That is, as shown in FIG. 3, the evaluation function P _peaki is calculated for the N _p second steering vectors a _peak1 to a _peakNp output by the second steering vector output unit 60 . Then, the number of living organisms larger than the threshold value of 0.5 is counted and output as the estimated number of living organisms. Also, the peak positions corresponding to those exceeding the threshold value of 0.5 may be output as the estimated position of the living body.

3 corresponds to FIG. 2, the second steering vector a _peak1 corresponding to the peak position 1001 has a value close to 0 because there is no eigenvector having a high correlation. On the other hand, the second steering vector a _peak2 corresponding to the peak position 1002 has a value close to 1 because there are eigenvectors 1011-A and 1011-C with high correlation. The second steering vector a _peak3 corresponding to the peak position 1003 also has a value close to 1 because there are eigenvectors 1011-B and 1011-D with high correlation. Therefore, there are two that exceed the threshold, the number of living bodies is estimated to be two, and the peak positions 1002 and 1003 corresponding to two that exceed the threshold are estimated to be the positions of living bodies.

[Operation of sensor 100]
Processing for estimating the number of living bodies by the sensor 100 configured as described above will be described.

FIG. 4 is a flow chart showing the living body count estimation process of the sensor 100 according to the first embodiment.

First, as shown in FIG. 4, the sensor 100 observes a received signal in the receiver 20 for a predetermined period (S10), and calculates a complex transfer function from the observed received signal (S20).

Next, the sensor 100 records each of the calculated complex transfer functions in time series, and calculates a biological component channel matrix from the recorded time-series complex transfer functions (S30). In other words, the sensor 100 extracts biometric information, which is a component corresponding to one or more living organisms present in a given space, from a received signal that has been transmitted to the given space.

Next, the sensor 100 calculates eigenvectors from the calculated biological component channel matrix (S35). In other words, the sensor 100 calculates one or more eigenvectors of the biometric correlation matrix obtained from the biometric information.

Next, the sensor 100 estimates biometric position candidates from the calculated biocomponent channel matrix using a predetermined position estimation method such as the Capon method that can be used even if the number of people is unknown (S40). In other words, the sensor 100 uses the calculated bio-correlation matrix to estimate the position of one or more living bodies, including the virtual image, by a predetermined position estimation method.

Next, the sensor 100 calculates _NP second steering vectors corresponding to the estimated biological position candidates (S50). In other words, the sensor 100 converts the first steering vectors corresponding to the positions including the estimated one or more virtual images of the living body from among the plurality of first steering vectors pre-stored in the storage unit 50 to the second steering vectors. Extract and output as

Next, the sensor 100 calculates the correlations between the N _p second steering vectors and the eigenvectors (S60), and estimates the number of living bodies or the positions of living bodies based on the number of correlations above the threshold among the correlations calculated in S60 ( S70). In other words, the sensor 100 uses the eigenvectors and the second steering vector to estimate at least one of the position and number of one or more living organisms.

[Effects, etc.]
According to the estimation device 1 and the like of the present embodiment, it is possible to accurately estimate the number of living organisms present in a predetermined space using radio signals. Further, according to the estimation apparatus 1 and the like of the present embodiment, it is possible to estimate the positions of the corresponding living bodies at the same time as estimating the number of living bodies.

More specifically, according to the estimation device 1 and the like of the present embodiment, a position estimation method using eigenvectors corresponding to eigenvalues and a predetermined position estimation method such as the Capon method that can be used even if the number of people is unknown By using two different methods together, the accuracy of population estimation can be improved.

In the existing method of estimating the number of living organisms using eigenvalues, when the signal-to-noise ratio is small, such as when the living organisms are far away, it is difficult to distinguish between the eigenvalues corresponding to the living organisms and the eigenvalues corresponding to noise using a threshold. Become. The direction of the eigenvector corresponding to the eigenvalue is less susceptible to noise compared to the eigenvalue, but since sorting by the corresponding eigenvalue is required to use it for population estimation, it is not possible to estimate the population using the eigenvector alone. In other words, the eigenvalues are susceptible to noise, and when the signal-to-noise ratio is small, the accuracy of population estimation decreases.

On the other hand, in a method using a position estimation method such as the Capon method, which can be used even if the number of people is unknown, virtual images are likely to occur around the antenna, so there is a high possibility of erroneously estimating the number of people. In other words, in a method using a position estimation method that can be used even if the number of people is unknown, such as the Capon method, the position estimation method has poor position estimation accuracy, and it is estimated that the living body is actually in a position where there is no living body (that is, a virtual image is estimated). ), it is not possible to estimate the number of people with high accuracy by itself.

Therefore, in the present embodiment, two methods, a position estimation method using eigenvectors corresponding to eigenvalues and a position estimation method such as the Capon method, are used together to improve the accuracy of population estimation. That is, the estimated position excluding the virtual image can be estimated by associating the calculated eigenvector with the vector (steering vector) associated with the estimated position including the virtual image calculated by a predetermined position estimation method. In this way, the information from the eigenvectors can be used to improve the accuracy of the estimated position calculated by the predetermined position estimation method. As a result, it is possible to overcome the problems of both existing methods of estimating the number of living organisms using eigenvalues and position estimation methods that can be used even if the number of people is unknown, such as the Capon method. It is possible to accurately estimate the number and positions of living organisms.

In addition, in the present disclosure, in order to discriminate the virtual image included in the estimated position calculated by the predetermined position estimation method, the correlation between the eigenvector and the steering vector is calculated as an index for quantitatively evaluating which image is likely and how likely. and eliminate virtual images. This makes it possible to implement an estimation method that is resistant to noise and has a wide detection range.

(Modification)
In the above embodiment, a position estimation method such as the Capon method that can be used even if the number of people is unknown is used as the predetermined position estimation method, but the method is not limited to this. A position estimation method based on the MUSIC method in which the number of people is known may be used. Hereinafter, this case will be described as a modified example, focusing on points different from the first embodiment.

FIG. 5 is a block diagram showing the configuration of the first position estimator 40A according to the modification of the first embodiment. 40 A of 1st position estimation parts are provided with the advance population estimation part 401 and the candidate position estimation part 402, as shown in FIG.

The prior population estimation unit 401 estimates the number of people including virtual images in one or more living organisms using the eigenvalues of the biometric correlation matrix. In this modification, prior population estimation section 401 uses eigenvalues obtained by eigenvalue decomposition of correlation matrix R by eigenvector calculation section 70 before candidate position estimation section 402 estimates the position of a living body to be detected. to estimate the number of living organisms to be detected. Here, the number of people estimated by the a priori number-of-people estimation unit 401 is the number of people tentatively estimated using the eigenvalue, and is an inaccurate number of people including a virtual image.

The candidate position estimating unit 402 uses the MUSIC method using the number of people estimated by the prior population estimating unit 401 as a predetermined position estimating method to estimate candidate positions including virtual images on one or more living bodies. In this modification, the candidate position estimation unit 402 estimates candidates for the position of the living body 200 using the MUSIC method using the inaccurate number of people estimated by the prior population estimation unit 401 as known.

As a result, the estimation apparatus and the like of this modified example use not only a position estimation method such as the Capon method that can be used even if the number of people is unknown, but also a position estimation method based on the MUSIC method in which the number of people is known. One can be estimated with high accuracy.

(Embodiment 2)
In the first embodiment, a case has been described in which at least one of the positions and the number of living bodies is estimated using a single sensor 100 . In the second embodiment, two or more sensors 100 are used to estimate at least one of the positions and the number of living bodies.

A sensor comprising two or more sensors 100 is hereinafter referred to as a composite sensor 101 .

[Configuration of composite sensor 101]
FIG. 6 is a block diagram showing the configuration of compound sensor 101 according to the second embodiment. Elements similar to those in FIG. 1 are denoted by the same reference numerals, and detailed description thereof is omitted.

The composite sensor 101 is an example of a bio-number estimating device, and includes a plurality of sensors 100 described in the first embodiment. The compound sensor 101 uses the estimation results of the plurality of sensors 100 to estimate the positions and number of living bodies in a wider range and with higher accuracy. In the example shown in FIG. 6, the composite sensor 101 includes sensors 100-A to 100-D, each of which is the sensor 100, an estimated position distribution calculator 120, and a living body count estimator . Note that in FIG. 6, the composite sensor 101 includes four sensors 100 in order to cover a room (that is, a predetermined space) in which a plurality of living bodies 200 (living bodies 200-A and 200-B) exist. It is not limited to this. The number of sensors 100 included in the composite sensor 101 is not limited as long as it is two or more. A detailed configuration will be described below.

[Sensor 100-A to Sensor 100-D]
Sensors 100-A to 100-D are the sensors 100 described in Embodiment 1, respectively, and between the transmitter 10-A and the receiver 20-A to the transmitter 10-D and the receiver 20-D Radio signals transmitted and received between the bodies are used to estimate the location and number of living bodies. In this embodiment, sensors 100-A to 100-D estimate the number of living bodies 200 and the positions of living bodies 200, respectively. Sensors 100 -A to 100 -D then output the estimated number of living bodies 200 and the estimated positions of living bodies 200 to estimated position distribution calculating section 120 .

[Estimated position distribution calculator 120]
Based on the positions of one or more living bodies 200 existing in each predetermined space estimated by a plurality of sensors 100, the estimated position distribution calculator 120 calculates the estimated position distribution of the entire space consisting of partially overlapping predetermined spaces. Calculate

In the present embodiment, estimated position distribution calculating section 120 calculates the number of living bodies 200 estimated by sensors 100-A to 100-D (referred to as estimated number of people) and the positions of living bodies 200 (referred to as estimated positions). receive. The estimated position distribution calculation unit 120 calculates the estimated position distribution by plotting the received estimated number of people and estimated positions on a plane or space.

More specifically, the estimated position distribution calculation unit 120 first calculates the space (referred to as the entire space) as the measurement range based on the predetermined space to be measured by each of the sensors 100-A to 100-D. Define a coordinate system. Here, the predetermined spaces to be measured by the sensors 100-A to 100-D partially overlap to constitute the entire space. Then, estimated position distribution calculating section 120 converts the estimated positions output from sensors 100-A to 100-D into the coordinate system of the entire space, which is the measurement range, and plots them. Estimated position distribution calculating section 120 superimposes the time-series data indicating the estimated position estimated for a period of about several seconds when living body 200 does not move, which are output from sensors 100-A to 100-D. can be plotted.

The concept of the estimated position distribution calculated by the estimated position distribution calculator 120 will be described below with reference to the drawings. 7A and 7B are conceptual diagrams of the estimated position distribution calculated from the positions and the number of people estimated by one sensor 100 according to the second embodiment. FIG. 8 is a conceptual diagram showing an example of the estimated position distribution calculated by the living body count estimation unit 130 according to the second embodiment.

As shown in FIG. 7A, the estimated position distribution calculation unit 120 calculates the estimated positions of the living body 200-A and the living body 200-C in a predetermined space 1201-A, which is the measurement target range, output from the sensor 100-A, for example. , are converted into the coordinate system of the entire space 1201 of the measurement range and plotted. In the example shown in FIG. 7A, the predetermined space 1201-A constitutes a partial area (left area) of the total space 1201. In the example shown in FIG. Similarly, as shown in FIG. 7B, the estimated position distribution calculation unit 120 calculates the position of the living body 200-A and the living body 200-B in the predetermined space 1201-B, which is the measurement target range, output from the sensor 100-B, for example. The estimated position is transformed into the coordinate system of the entire space 1201 of the measurement range and plotted. In the example shown in FIG. 7B, the predetermined space 1201-B constitutes a partial area (lower area) of the total space 1201. In the example shown in FIG.

Then, estimated position distribution calculating section 120 calculates the estimated positions of living bodies 200-A and 200-C in predetermined space 1201-A shown in FIG. 7A and living bodies 200-A and 200 in predetermined space 1201-B shown in FIG. -B's estimated positions are overlaid and plotted as shown in FIG. In the example shown in FIG. 8, the predetermined space 1201-A and the predetermined space 1201-B partially overlap to form the entire space 1201. In FIG.

Note that, like the living body 200-A shown in FIG. 8, for living bodies that appear redundantly in the output from the sensor 100-A and the output from the sensor 100-B, the living body number estimation unit 130 mathematically removes the overlap. .

[Bionumber estimation unit 130]
The living body number estimating unit 130 estimates the number of living bodies, which is the number of one or more living bodies existing in the entire space, by using the number of one or more living bodies existing in each predetermined space estimated by the plurality of sensors 100. do. Here, the living body count estimating unit 130 may estimate the number of local minimum values of the variance of the positions estimated by the plurality of sensors 100 as the living body count. In addition, the living body count estimating unit 130 may estimate the number of maximum values of the density of the positions estimated by the plurality of sensors 100 as the living body count.

In the present embodiment, the living body count estimation unit 130 estimates the number of people using the estimated location distribution calculated by the estimated location distribution calculation unit 120 . More specifically, the living body number estimating unit 130 searches for points where the density of the estimated position distribution takes a maximum value, and estimates the number as the number of living bodies 200 . In addition, the living body count estimating unit 130 may estimate the searched point of maximum value as the position of the living body 200 . In this way, the living body number estimating unit 130 mathematically removes the number of living bodies 200 plotted overlapping the estimated position distribution calculated by the estimated position distribution calculating unit 120, 200 numbers can be estimated.

As a method for calculating the density of the estimated position distribution, the space of the measurement range may be divided into grids of appropriate sizes, and the number of estimated positions plotted on each grid may be counted. A method using the reciprocal of the variance of a predetermined number of plots close to . However, among the local maximum values of the density of the estimated position distribution, it is preferable to select those having a value equal to or greater than a predetermined threshold. As a result, erroneous estimation can be prevented when there is no living body 200 in the measurement range, that is, when no one is present.

Note that the living body number estimation unit 130 may search for points where the variance of the estimated position distribution takes a minimum value, and estimate the number as the number of living bodies 200 .

[Operation of composite sensor 101]
Processing for estimating the number of living organisms and the like by the composite sensor 101 configured as described above will be described.

FIG. 9 is a flow chart showing the living body count estimation process of the composite sensor 101 according to the second embodiment.

First, as shown in FIG. 9, the composite sensor 101 causes each sensor 100 (sensor 100-A to sensor 100-D) to estimate the position and number of living bodies for a predetermined period (S110).

Next, the composite sensor 101 calculates an estimated position distribution using the estimation results of each sensor 100 (S120). More specifically, the composite sensor 101 calculates the estimated position distribution by plotting the estimated number of people and positions output by each sensor 100 over the entire space that is the measurement range.

Next, the composite sensor 101 searches for the maximum value of the density of the estimated position distribution calculated in S120 (S130). Note that the composite sensor 101 may search for the minimum value of the variance of the estimated position distribution calculated in S120.

Finally, the composite sensor 101 estimates the number or positions of living bodies by counting the number of maximum values found in S130 that are equal to or greater than a threshold value (S140). More specifically, the composite sensor 101 estimates the number of maximum values found in S130 that are equal to or greater than the threshold as the number of living organisms. In addition, the living body number estimating unit 130 estimates the point having the maximum value searched in S130 as the position of the living body. Note that the composite sensor 101 may estimate the number of living bodies or the positions of living bodies by counting the number of minimum values found in S130.

[Effects, etc.]
According to the composite sensor 101 and the like of Embodiment 2, the number of living organisms present in the measurement range can be estimated with high accuracy using radio signals. More specifically, the composite sensor 101 of Embodiment 2 can link a plurality of sensors 100 in addition to the effects obtained by the sensor 100 and the like of Embodiment 1. Estimates of the position and number of living organisms can be made in In other words, the composite sensor 101 of the second embodiment highly accurately estimates the number of living organisms in a measurement range composed of a plurality of predetermined ranges, which is wider than the predetermined space that is the measurement range of the individual sensors 100. can. For this reason, the composite sensor 101 of the second embodiment has a metal obstacle in the measurement range of a certain sensor 100, and even in a situation where it is difficult for radio waves to reach, the other sensors 100 are linked to each other so that the measurement range can be expanded. It is possible to reduce blind spots in the body and estimate the number of living organisms with high accuracy.

(Example)
In order to confirm the effects of the second embodiment, an evaluation was conducted by experiments, and the results will be described as examples below.

FIG. 10 is a diagram showing an environment in which an experiment was conducted using the estimation method according to Embodiment 2. In FIG.

This experiment was conducted in a room of 6 m wide by 7 m long, in which living organisms stood at three positions (0.5, 2), (2, 2), and (3.5, 2). Further, in this experiment, element linear array antennas were used in the transmitting/receiving stations indicated by TRx1 to TRx3 in FIG. Note that TRx1 to TRx3 correspond to the transmitting/receiving array antennas forming the transmitting antenna section 12 and the receiving antenna section 21 forming the sensor 100, respectively. In this experiment, the transmitting/receiving stations TRx1 to TRx3 are arranged at coordinates (0,0), (2,0), and (4,0). In addition, the distance between the transmitting and receiving array elements constituting the element linear array antenna used in the transmitting and receiving stations TRx1 to TRx3 is 0.5 wavelength, the operating frequency is 2.47125 GHz, the sampling frequency is 100 Hz, and the frequency range of biological activity to be extracted is 0.3 to 3. .3 Hz.

FIG. 11A is a diagram showing the estimated position distribution by the Capon method, and FIG. 11B is a diagram showing the estimated position distribution in this embodiment. In FIGS. 11A and 11B, a picture of a person is drawn at the position of the real image where the living body actually exists.

FIG. 11A shows the result of estimating the position of a living body in the room shown in FIG. 10 using only the Capon method. In FIG. 11A, the time-series data indicating the estimated position (+ in the figure) estimated during a period of about several seconds when the living body does not move is plotted, but the estimated position estimated only once is plotted. may be In addition, FIG. 11B shows a superposition of the results of estimating the position of a living body in the example using both the Capon method and two different estimation methods for eigenvectors in the room shown in FIG. 10, that is, the estimated position distribution. ing.

As can be seen by comparing FIGS. 11A and 11B, in FIG. 11B, there are fewer estimated positions (+ in the figure) other than those around the positions where people are drawn. In other words, it can be seen that many virtual images generated by the Capon method can be removed by the method of the present embodiment, which uses both a position estimation method such as the Capon method and a method using eigenvalues.

FIG. 12 is a spectrum showing the variance of the estimated position distribution shown in FIG. 11B.

As shown in FIG. 12, in this experiment, the spectrum indicating the variance of the estimated position distribution shown in FIG. 11B was calculated, and the maximum value of the density of the estimated position distribution was searched. As shown in FIG. 12, since the number of maximum values (+ in the figure) is three, the number of living organisms could be correctly estimated to be three. As a result, by searching for the maximum value of the density of the estimated position distribution, it is possible to count the same living body that is redundantly estimated by a plurality of estimating devices as one living body.

INDUSTRIAL APPLICABILITY The present disclosure can be used for an estimation device or the like that estimates at least one of the positions and the number of living organisms using radio signals. It can be used for home electric appliances that perform such functions, monitoring devices that detect the intrusion of living organisms, and the like.

1 estimation device 10, 10-A, 10-B, 10-C, 10-D transmitter 11 transmission unit 12 transmission antenna unit 20, 20-A, 20-B, 20-C, 20-D receiver 21 reception Antenna section 22 Receiving section 23 Complex transfer function calculating section 30 Biological information extracting section 40, 40A First position estimating section 50 Storage section 60 Second steering vector output section 70 Eigenvector calculating section 80 Second position estimating section 100, 100-A, 100-B, 100-C, 100-D Sensor 101 Combined sensor 120 Estimated position distribution calculator 130 Living body count estimator 200, 200-A, 200-B, 200-C Living body 401 Prior population estimator 402 Candidate position estimator 1001, 1002, 1003 Peak position 1011-A, 1011-B, 1011-C, 1011-D Eigenvector 1201 Whole space 1201-A, 1201-B Predetermined space

Claims (12)

A living body count estimation device comprising a plurality of estimation devices,
Each of the plurality of estimation devices,
a biometric information extraction unit that extracts biometric information, which is a component corresponding to one or more living organisms present in the predetermined space, from a received signal that is a signal transmitted in the predetermined space;
an eigenvector calculator that calculates one or more eigenvectors of a biometric correlation matrix obtained from the biometric information;
a first position estimator for estimating a position, including a virtual image, of the one or more living bodies by a predetermined position estimation method using the biometric correlation matrix;
Among a plurality of first steering vectors stored in advance in a storage unit, first steering vectors corresponding to respective positions including the one or more virtual images of the living body estimated by the first position estimation unit are used as second steering vectors. a second steering vector output unit for extracting and outputting as a vector;
a second position estimating unit that estimates at least one of the positions and the number of the one or more living organisms using the eigenvector and the second steering vector;
The living body count estimation device further comprises:
Estimated position distribution calculation for calculating the estimated position distribution of the entire space including the predetermined space that partially overlaps based on the positions of one or more living organisms existing in each of the predetermined spaces, which are estimated by the plurality of estimating devices. Department and
Living body count estimation for estimating the number of living bodies, which is the number of one or more living bodies existing in the entire space, using each of the numbers of one or more living bodies existing in each of the predetermined spaces estimated by the plurality of estimating devices. and
The number of living organisms estimating unit estimates the number of local minimum values of variance of the positions estimated by the plurality of estimating devices as the number of living organisms.
number of lives estimation device. A living body count estimation device comprising a plurality of estimation devices,
Each of the plurality of estimation devices,
a biometric information extraction unit that extracts biometric information, which is a component corresponding to one or more living organisms present in the predetermined space, from a received signal that is a signal transmitted in the predetermined space;
an eigenvector calculator that calculates one or more eigenvectors of a biometric correlation matrix obtained from the biometric information;
a first position estimating unit for estimating a position, including a virtual image, of the one or more living bodies by a predetermined position estimation method using the biometric correlation matrix;
Among a plurality of first steering vectors stored in advance in a storage unit, first steering vectors corresponding to respective positions including the one or more virtual images of the living body estimated by the first position estimation unit are used as second steering vectors. a second steering vector output unit for extracting and outputting as a vector;
a second position estimating unit that estimates at least one of the positions and the number of the one or more living bodies using the eigenvector and the second steering vector;
The living body count estimation device further comprises:
Estimated position distribution calculation for calculating the estimated position distribution of the entire space including the predetermined space that partially overlaps based on the positions of one or more living organisms existing in each of the predetermined spaces, which are estimated by the plurality of estimating devices. Department and
Living body count estimation for estimating the number of living bodies, which is the number of one or more living bodies existing in the entire space, using each of the numbers of one or more living bodies existing in each of the predetermined spaces estimated by the plurality of estimating devices. and
The living body number estimating unit estimates, as the living body number, the number of maximum values of the density of positions estimated by the plurality of estimation devices.
Bio-number estimator. The second position estimating unit calculates a product or correlation between the eigenvector and the second steering vector, and indicates the eigenvector or the second steering vector having a value equal to or greater than a threshold among the calculated product or the correlation. estimating a position as said position;
The living body number estimation device according to claim 1 or 2 . The second position estimation unit estimates the number of the eigenvectors or the second steering vectors that take a value equal to or greater than a threshold among the calculated products or correlations as the number.
The living body number estimation device according to claim 3 . The predetermined position estimation method is the Capon method,
A living body number estimation device according to any one of claims 1 to 4 . The predetermined position estimation method is the beamformer method,
A living body number estimation device according to any one of claims 1 to 4 . The first position estimating unit further includes a prior population estimating unit that estimates the number of people including virtual images in the one or more living organisms using the eigenvalues of the biometric correlation matrix,
The predetermined position estimation method is the MUSIC method using the number of people estimated by the prior population estimation unit,
A living body number estimation device according to any one of claims 1 to 4 . a transmitter having N (N is a natural number of 2 or more) transmitting antenna elements;
M (M is a natural number of 2 or more) receiving antenna elements, and the N transmitting antenna elements and the M receiving antenna elements from the received signals received by each of the M receiving antenna elements for a predetermined period. A receiver having a transfer function calculator that calculates a plurality of complex transfer functions representing the propagation characteristics between each of
The biological information extraction unit extracts, from the plurality of complex transfer functions calculated by the transfer function calculation unit, fluctuation components due to the influence of a living body and fluctuation components in each of the M receiving antenna elements as the biological information. extract,
The eigenvector calculation unit calculates the bio-correlation matrix from the variation components in each of the M receiving antenna elements extracted by the bio-information extraction unit, and calculates the eigenvector of the calculated bio-correlation matrix.
A living body number estimation device according to any one of claims 1 to 7 . A method for estimating the number of living organisms for a living organism number estimating apparatus comprising a plurality of estimating apparatuses,
The estimation method performed by each of the plurality of estimation devices includes:
a biometric information extracting step of extracting biometric information, which is a component corresponding to one or more living organisms present in the predetermined space, from a received signal transmitted to the predetermined space;
an eigenvector calculation step of calculating one or more eigenvectors of a biometric correlation matrix obtained from the biometric information;
a first position estimation step of estimating a position including a virtual image in the one or more living bodies by a predetermined position estimation method using the biometric correlation matrix;
Among a plurality of first steering vectors pre-stored in a storage unit, the first steering vectors corresponding to the positions including the one or more virtual images of the living body estimated in the first position estimation step are used as the second steering vectors. a second steering vector output step of extracting and outputting as a vector;
a second position estimation step of estimating at least one of the positions and numbers of the one or more living organisms using the eigenvectors and the second steering vectors;
The method for estimating the number of living organisms,
Based on the positions of one or more living bodies existing in each of the predetermined spaces estimated by the estimation method of the plurality of estimation devices, calculate the estimated position distribution of the entire space consisting of the predetermined spaces that partially overlap. an estimated position distribution calculation step for
The number of living organisms, which is the number of one or more living organisms existing in the entire space, is calculated using the number of one or more living organisms existing in each of the predetermined spaces estimated by the estimation method of the plurality of estimating devices. and a step of estimating the number of living organisms to be estimated,
In the step of estimating the number of living bodies, the estimation method of the plurality of estimation devicesbyestimating the number of local minima of the variance of the estimated position as the number of living organisms;
number of livesestimation method. A method for estimating the number of living organisms for a living organism number estimating device comprising a plurality of estimating devices,
The estimation method performed by each of the plurality of estimation devices includes:
Biometric information extraction for extracting biometric information, which is a component corresponding to one or more living organisms existing in a predetermined space, from a received signal received from a signal transmitted in the predetermined space.stepWhen,
an eigenvector calculation step of calculating one or more eigenvectors of a biometric correlation matrix obtained from the biometric information;
a first position estimation step of estimating a position including a virtual image in the one or more living bodies by a predetermined position estimation method using the biometric correlation matrix;
Among a plurality of first steering vectors pre-stored in a storage unit, the first steering vectors corresponding to the positions including the one or more virtual images of the living body estimated in the first position estimation step are used as the second steering vectors. a second steering vector output step of extracting and outputting as a vector;
a second position estimation step of estimating at least one of the positions and numbers of the one or more living organisms using the eigenvectors and the second steering vectors;
The method for estimating the number of living organisms,
Based on the positions of one or more living bodies existing in each of the predetermined spaces estimated by the estimation method of the plurality of estimation devices, calculate the estimated position distribution of the entire space consisting of the predetermined spaces that partially overlap. an estimated position distribution calculation step for
The number of living organisms, which is the number of one or more living organisms existing in the entire space, is calculated using the number of one or more living organisms existing in each of the predetermined spaces estimated by the estimation method of the plurality of estimating devices. and a step of estimating the number of living organisms to be estimated,
In the step of estimating the number of living organisms, the number of local maximum values estimated by the estimation method of the plurality of estimating devices is estimated as the number of living organisms.
Birth count estimation method. A program that causes a computer to execute the method for estimating the number of living organisms according to claim 9. A program that causes a computer to execute the method for estimating the number of living organisms according to claim 10.

Images