CN113904707B

CN113904707B - Human body detection method and device, and shutdown method and device of household electrical appliance

Info

Publication number: CN113904707B
Application number: CN202111087740.XA
Authority: CN
Inventors: 宋亚东
Original assignee: Shanghai Meiren Semiconductor Co ltd
Current assignee: Shanghai Meiren Semiconductor Co ltd
Priority date: 2021-09-16
Filing date: 2021-09-16
Publication date: 2023-11-10
Anticipated expiration: 2041-09-16
Also published as: CN113904707A

Abstract

The disclosure provides a human body detection method and device, and a shutdown method and device of household appliances, and relates to the field of wireless sensing. The method comprises the steps of sampling Channel State Information (CSI) of each subcarrier of Wi-Fi communication in a target environment where a first household appliance is located for multiple times; based on the sampled CSI, generating a CSI amplitude time sequence matrix, wherein each line of the CSI amplitude time sequence matrix is the CSI amplitude value of each subcarrier acquired at the same sampling time; acquiring the mean variance of the CSI amplitude time sequence matrix; human body detection is carried out on a target environment according to the mean variance of a plurality of CSI amplitude time sequence matrixes which are continuous in time; and controlling the first household appliance to be closed in response to detecting that no person exists in the target environment. The application can detect whether the target environment is someone in real time, does not need to add special detection hardware or carry out training or off-line calibration in advance, has low complexity and wide detection range, and can control the shutdown of the household appliance under the condition of no detection person, thereby saving resources.

Description

Human body detection method and device, and shutdown method and device of household electrical appliance

Technical Field

The present application relates to the field of wireless sensing, and in particular, to a human body detection method and apparatus, and a shutdown method and apparatus for home appliances.

Background

At present, more and more home appliances integrate Wi-Fi modules, and indoor personnel detection is performed by utilizing Wi-Fi channel state information (channel state information, CSI), so that the Wi-Fi module can be applied to various scenes such as intelligent shutdown and intrusion detection of home appliances, and the technology related to indoor personnel detection by utilizing Wi-Fi channel state information has the problems of large calculated amount, small detection range, low practicability and the like.

Disclosure of Invention

The present application aims to solve at least one of the technical problems in the related art to some extent.

Therefore, an object of the present application is to provide a human body detection method, which samples channel state information CSI of each subcarrier of Wi-Fi communication in an environment to be detected for a plurality of times; generating a CSI amplitude time sequence matrix based on the sampled CSI, wherein each line of the CSI amplitude time sequence matrix is a CSI amplitude value of each subcarrier acquired at the same sampling moment; acquiring the mean variance of the CSI amplitude time sequence matrix; and detecting the human body of the environment to be detected according to the mean variance of the time series matrixes of the CSI amplitude and the time series.

A second object of the present application is to provide a method for powering off a home appliance.

A third object of the present application is to provide a human body detecting device.

A fourth object of the present application is to provide a shutdown device for a home appliance.

A fifth object of the present application is to propose a receiving device with human detection.

A sixth object of the present application is to provide a human body detection system.

A seventh object of the present application is to propose a home appliance.

An eighth object of the present application is to propose a non-transitory computer readable storage medium.

A ninth object of the application is to propose a computer programme product.

To achieve the above object, an embodiment of a first aspect of the present application provides a human body detection method, including: sampling Channel State Information (CSI) of each subcarrier of Wi-Fi communication in an environment to be detected for multiple times; generating an amplitude time sequence matrix based on the sampled CSI, wherein each line of the CSI amplitude time sequence matrix is a CSI amplitude value of each subcarrier acquired at the same sampling moment; acquiring the mean variance of the CSI amplitude time sequence matrix; and detecting the human body of the environment to be detected according to the mean variance of the time series matrixes of the CSI amplitude and the time series.

The personnel detection method of the embodiment can detect whether a person exists in the current environment in a non-invasive and real-time mode, does not need to add special detection hardware or train in advance or calibrate in an off-line mode, is low in calculation complexity and wide in detection range, greatly improves the application range and practicality, can achieve all-weather real-time monitoring, improves the detection success rate, and improves user experience.

To achieve the above object, an embodiment of a second aspect of the present application provides a shutdown method of a home appliance, including: sampling Channel State Information (CSI) of each subcarrier of Wi-Fi communication in a target environment where the first household appliance is located for multiple times; generating a CSI amplitude time sequence matrix based on the sampled CSI, wherein each line of the CSI amplitude time sequence matrix is a CSI amplitude value of each subcarrier acquired at the same sampling moment; acquiring the mean variance of the CSI amplitude time sequence matrix; human body detection is carried out on the target environment according to the mean variance of a plurality of CSI amplitude time sequence matrixes which are continuous in time; and controlling the first household appliance to be closed in response to detecting that no person exists in the target environment.

The embodiment of the application can detect whether a person exists in the target environment in a non-invasive and real-time manner, does not need to add special detection hardware or perform advanced training or off-line calibration, has low calculation complexity and wide detection range, greatly improves the application range and the practicability, can control the shutdown of the household appliance under the condition of unmanned detection, and is convenient to use and saves resources.

To achieve the above object, an embodiment of a third aspect of the present application provides a human body detection device, including: the sampling module is used for sampling Channel State Information (CSI) of each subcarrier of Wi-Fi communication in the environment to be detected for a plurality of times; the generation module is used for generating a CSI amplitude time sequence matrix based on the sampled CSI, wherein each line of the CSI amplitude time sequence matrix is the CSI amplitude value of each subcarrier acquired at the same sampling time; the acquisition module is used for acquiring the mean variance of the CSI amplitude time sequence matrix; and the detection module is used for detecting the human body of the environment to be detected according to the mean variance of the time series matrixes of the CSI amplitude and the time series.

To achieve the above object, a fourth aspect of the present application provides a shutdown device of a home appliance, including: the sampling module is used for sampling Channel State Information (CSI) of each subcarrier of Wi-Fi communication in a target environment where the first household appliance is located for a plurality of times; the generation module is used for generating a CSI amplitude time sequence matrix based on the sampled CSI, wherein each line of the CSI amplitude time sequence matrix is the CSI amplitude value of each subcarrier acquired at the same sampling time; the acquisition module is used for acquiring the mean variance of the CSI amplitude time sequence matrix; the detection module is used for detecting the human body of the target environment according to the mean variance of the time series matrixes of the CSI amplitude in time; and the control module is used for controlling the first household appliance to be closed in response to detecting that no person exists in the target environment.

To achieve the above object, a fifth aspect of the present application provides a receiving apparatus with human body detection, comprising: the receiver is used for receiving Wi-Fi signals sent by Wi-Fi communication on all subcarriers; the controller is used for acquiring Channel State Information (CSI) of the subcarriers according to the Wi-Fi signals, generating a CSI amplitude time sequence matrix based on the sampled CSI, wherein the CSI amplitude values of the subcarriers acquired at the same sampling time in each row of the CSI amplitude time sequence matrix are acquired, the mean variance of the CSI amplitude time sequence matrix is acquired, and human body detection is carried out on the target environment according to the mean variances of a plurality of continuous time CSI amplitude time sequence matrices.

To achieve the above object, a sixth aspect of the present application provides a human body detection system, including: a receiving device and a transmitting device with human body detection as described in the fifth aspect embodiment; the transmitting device is used for transmitting Wi-Fi signals through all sub-carriers of Wi-Fi communication.

To achieve the above object, an embodiment of a seventh aspect of the present application provides an electric home appliance, including: at least one processor; and a memory communicatively coupled to the at least one processor; the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the human body detection method according to the first aspect embodiment and the power-off method of the home appliance according to the second aspect embodiment.

To achieve the above object, an eighth aspect of the present application provides a non-transitory computer-readable storage medium storing computer instructions for causing the computer to execute the human body detection method according to the first aspect of the embodiment and the power-off method of the home appliance according to the second aspect of the embodiment.

To achieve the above object, an embodiment of a ninth aspect of the present application proposes a computer program product, comprising a computer program which, when executed by a processor, implements a human body detection method according to the embodiment of the first aspect and a shutdown method of a home appliance according to the embodiment of the second aspect.

Drawings

Fig. 1 is an exemplary implementation of a human body detection method according to an embodiment of the present application.

Fig. 2 is a schematic diagram of the mean variance of the acquired CSI amplitude time series matrix according to an embodiment of the present application.

Fig. 3 is a schematic diagram of human detection of an environment to be detected according to an embodiment of the present application.

Fig. 4 is a schematic diagram of human body detection of an environment to be detected according to another embodiment of the present application.

Fig. 5 is a schematic diagram of human detection of an environment to be detected according to an embodiment of the present application.

Fig. 6 is a general flow chart of a human body detection method according to an embodiment of the present application.

Fig. 7 is a schematic diagram of a shutdown method of an electric home appliance according to an embodiment of the present application.

Fig. 8 is a schematic diagram of an electrical home device receiving Wi-Fi signals according to one embodiment of the present application.

Fig. 9 is a schematic diagram of controlling the first home device to be turned off according to an embodiment of the present application.

Fig. 10 is a schematic diagram of a plurality of home devices receiving Wi-Fi signals according to one embodiment of the application.

Fig. 11 is a general schematic diagram of a shutdown method of an electric home appliance according to an embodiment of the present application.

Fig. 12 is a schematic view of a human body detection device according to an embodiment of the present application.

Fig. 13 is a schematic diagram of a shutdown device of an electric home appliance according to an embodiment of the present application.

Fig. 14 is a schematic diagram of a receiving device with human detection according to an embodiment of the present application.

Fig. 15 is a schematic diagram of a human detection system according to an embodiment of the present application.

Fig. 16 is a schematic diagram of an electric home appliance according to an embodiment of the present application.

Detailed Description

Embodiments of the present application are described in detail below, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to like or similar elements or elements having like or similar functions throughout. The embodiments described below by referring to the drawings are illustrative and intended to explain the present application and should not be construed as limiting the application.

Fig. 1 is an exemplary embodiment of a human body detection method according to the present application, as shown in fig. 1, comprising the steps of:

s101, sampling channel state information CSI of all subcarriers of Wi-Fi communication in an environment to be detected for a plurality of times.

Wi-Fi is the most widely used wireless local area network transmission technology today, which is to convert information to be transmitted into wireless signals for receiving signals by devices such as computers, mobile phones, home appliances and the like supporting the technology.

Channel state information (Channel State Information, CSI) refers to the channel properties of a communication link in the field of wireless communications. It describes the attenuation factor of the signal on each transmission path, i.e. the value of each element in the channel gain matrix H, such as signal Scattering (Scattering), multipath fading or shadow fading (multipath fading or shadowing fading), distance-induced power attenuation (power decay of distance), etc. The CSI feedback may adapt the communication system to the current channel conditions, and provides a guarantee for high reliability and high rate communication in a multi-antenna system.

In the embodiment of the application, in order to detect whether personnel exist in certain environments, wi-Fi channel state information in the environment to be detected needs to be acquired, and optionally, the environment to be detected can be a family house, an industrial house and the like. Assuming Wi-Fi contains N valid subcarriers, then the sampling instant t _i The CSI vector of (c) may be expressed as:

H(t _i )＝[H(f ₁ ，t _i )，H(f ₂ ，t _i )，…，H(f _N ，t _i )]

wherein,represents the CSI amplitude of the nth subcarrier, +.H (f) _n ，t _i ) The CSI phase of the nth subcarrier is represented.

Sampling CSI for each subcarrier of Wi-Fi communication within an environment to be detected multiple times, with a sampling time interval Δt=t _i+1 -t _i For example, the time interval may be selected to be 0.2s, and the CSI vector of each sampling time is obtained separately.

S102, based on the sampled CSI, a CSI amplitude time sequence matrix is generated, wherein each line of the CSI amplitude time sequence matrix is the CSI amplitude value of each subcarrier acquired at the same sampling time.

Based on the CSI vector at each sampling time obtained by the above-mentioned multiple sampling, a CSI amplitude time series matrix may be generated. Assuming that the number of times of the multiple sampling is K times, the sampling point CSI amplitude time sequence of the continuous K times of sampling can form a CSI amplitude time sequence matrix, which is expressed as A _K×N ＝[a ₁ ，a ₂ ，…，a _N ]Wherein a is _j ＝(||H(f _j ，t ₁ )||，||H(f _j ，t ₂ )||，…，||H(f _j ，t _K ) I|) j=1, 2, …, N, that is, CSI amplitude values of each subcarrier acquired at the same sampling time in each of the CSI amplitude time series matrices. Alternatively, the value of K may be set by the practitioner itself, for example, the value of K may be 16.

Alternatively, the sampling device itself may be used for environmental reasons or Abnormal values can occur in the CSI amplitude time sequence matrix, so that whether a human body exists in an environment to be detected or not can be judged more accurately, and the abnormal values in the CSI amplitude values need to be removed. Optionally, filtering may be selected to remove outliers in the CSI amplitude values. Illustrating: each CSI amplitude time series a in the CSI amplitude time series matrix may be removed using a Hampel filter _j J=1, 2, …, N, and after removing the outliers in the CSI amplitude values, the CSI amplitude values are reconstructed. Wherein, alternatively, the median calculation window length of the Hampel filter may be 7, and the standard deviation threshold may be 3sigma.

S103, obtaining the mean variance of the CSI amplitude time sequence matrix.

According to the obtained CSI amplitude time sequence matrix, the variance corresponding to the CSI amplitude value of each column of the CSI amplitude time sequence matrix can be obtained, so that a variance vector is formed, and the obtained variance vector is subjected to weighted average, so that the mean variance of the CSI amplitude time sequence matrix can be obtained.

And S104, detecting the human body of the environment to be detected according to the mean variance of the time series matrixes of the plurality of CSI amplitude in time.

The mean variance of the CSI amplitude time sequence matrix corresponding to a plurality of continuous sampling points in time is obtained, and the mean variance of the set number is compared with a preset threshold value or a number threshold value, so that whether a human body exists in the environment to be detected is judged. If the detected environment is judged to be the human body after being compared with the preset threshold value or the quantity threshold value, the detected environment is considered to be the human body; if the detected environment is not human after being compared with the preset threshold value or the quantity threshold value, static personnel detection is carried out on the detected environment based on the CSI amplitude time sequence matrix so as to determine whether a human body exists in the detected environment.

The embodiment of the application provides a human body detection method, which is characterized in that Channel State Information (CSI) of subcarriers of Wi-Fi communication in an environment to be detected is sampled for a plurality of times; based on the sampled CSI, generating a CSI amplitude time sequence matrix, wherein each line of the CSI amplitude time sequence matrix is the CSI amplitude value of each subcarrier acquired at the same sampling time; acquiring the mean variance of the CSI amplitude time sequence matrix; and carrying out human body detection on the environment to be detected according to the mean variance of the time series matrixes of the plurality of CSI amplitude in time. The personnel detection method of the embodiment can detect whether a person exists in the current environment in a non-invasive and real-time mode, does not need to add special detection hardware or train in advance or calibrate in an off-line mode, is low in calculation complexity and wide in detection range, greatly improves the application range and practicality, can achieve all-weather real-time monitoring, improves the detection success rate, and improves user experience.

Fig. 2 is an exemplary implementation manner of a human body detection method according to the present application, as shown in fig. 2, based on the foregoing embodiment, the method for obtaining the mean variance of the CSI amplitude time series matrix includes the following steps:

s201, based on the CSI amplitude value on each column of the CSI amplitude time sequence matrix, acquiring a first variance of each column to form a first variance vector corresponding to the CSI amplitude time sequence matrix.

Obtaining a first variance of each CSI amplitude time sequence according to the CSI amplitude value of each CSI amplitude time sequence of the obtained CSI amplitude time sequence matrix, wherein each column of the CSI amplitude time sequence matrix can be expressed as a _j ＝(||H(f _j ，t ₁ )||，||H(f _j ，t ₂ )||，…，||H(f _j ，t _K ) I) j=1, 2, …, N, and since j=1, 2, …, N, respectively, the first variances corresponding to the N columns of CSI amplitude time sequences can be obtained, thereby obtaining a CSI amplitude time sequence matrix a composed of N first variances _K×N Corresponding first variance vector, the first variance vector is marked as [ var ] ₁ ，var ₂ ，…，var _N ]。

S202, carrying out weighted average on the first variance in the first variance vector to obtain the mean variance of the CSI amplitude time sequence matrix.

For the first direction difference vector [ var ] obtained above ₁ ，var ₂ ，…，var _N ]The N first differences in the set are weighted and averaged, and the weighted average formula is that General get->Thereby obtaining the mean variance of the CSI amplitude time sequence matrix>

According to the embodiment of the application, the mean variance of the CSI amplitude time sequence matrix is obtained, so that human body detection can be conveniently carried out on the environment to be detected according to the mean variance of a plurality of CSI amplitude time sequence matrices which are continuous in time.

Fig. 3 is an exemplary implementation manner of a human body detection method according to the present application, as shown in fig. 3, based on the foregoing embodiment, according to a mean variance of a plurality of CSI amplitude time series matrices that are continuous in time, human body detection is performed on an environment to be detected, including the following steps:

s301, generating a total mean variance vector based on a plurality of mean variances which are continuous in time.

Because the sampling is performed for a plurality of times, when each sampling time is updated, the CSI amplitude time sequence matrix of the last sampling time is required to be updated based on the CSI acquired at the current sampling time, so as to generate the CSI amplitude time sequence matrix of the current sampling time. Optionally, when updating the CSI amplitude time sequence matrix at the previous sampling time, the first row of matrix elements in the CSI amplitude time sequence matrix corresponding to the previous sampling time may be shifted out, each remaining row in the CSI amplitude time sequence matrix is sequentially shifted forward by one row, then based on the CSI acquired at the current sampling time, the last row of matrix elements in the CSI amplitude time sequence matrix is generated, and written into the CSI time sequence matrix, so as to generate the CSI amplitude time sequence matrix at the current sampling time. By repeating the method, CSI amplitude time sequences respectively corresponding to a plurality of sampling moments with continuous time can be obtained The matrix is used for calculating the mean variance of the CSI amplitude time sequence matrix corresponding to each sampling moment based on the CSI amplitude time sequence matrix corresponding to each sampling moment so as to obtain a plurality of mean variances which are continuous in time, and the mean variances can be respectively marked as if the sampling is performed for m timesGenerating a total mean variance vector by using a plurality of mean variances which are continuous in time, and marking the total mean variance vector as +.>Alternatively, m may take on a value of 64.

S302, acquiring a first number of first mean variances from the total mean variance vector according to a time sequence from late to early.

Taking a first number of first mean variances from the total mean variance vector V in time sequence from late to early, and assuming that the first number is set to be k, obtaining k first mean variances respectively expressed asAlternatively, k may take on a value of 8.

S303, generating a first mean variance vector based on the first mean variance.

And generating a first mean variance vector according to the obtained first number of first mean variances. Assuming that the first number is set to k, the first mean variance vector is composed of k first mean variances obtained as described above, and the first mean variance vector is noted as Alternatively, k may take on a value of 8.

S304, determining that the movable human body exists in the environment to be detected in response to the first mean variance vector meeting a first set condition.

And counting the number of which all the first mean variances in the first mean variance vectors are larger than a first threshold value beta as a second number. If statistically obtained firstThe first mean variance of the first mean variance vector with the latest sampling time, the second number exceeding the first number threshold uGreater than a second threshold g _abs And determining that the first mean variance vector meets a first set condition and that the movable human body exists in the environment to be detected. Alternatively, k may take a value of 8,u and 2 g _abs The value is 0.25.

According to the embodiment of the application, human body detection is carried out on the environment to be detected according to the mean variance of a plurality of CSI amplitude time sequence matrixes which are continuous in time, the calculation complexity is low, the detection range is wide, and the application range and the practicability are greatly improved.

Fig. 4 is an exemplary implementation manner of a human body detection method according to the present application, as shown in fig. 4, based on the foregoing embodiment, according to a mean variance of a plurality of CSI amplitude time series matrices that are continuous in time, human body detection is performed on an environment to be detected, including the following steps:

S401, generating a total mean variance vector based on a plurality of mean variances which are continuous in time.

In step S401, the above embodiments have been specifically described, and will not be described herein.

S402, acquiring a first number of first mean variances from the total mean variance vector according to a time sequence from late to early.

In step S402, the above embodiments have been specifically described, and will not be described herein.

S403, generating a first mean variance vector based on the first mean variance.

In step S403, the above embodiments have been specifically described, and will not be described herein.

And S404, determining that no active human body exists in the environment to be detected in response to the first mean variance vector not meeting the first setting condition.

All the first mean variances in the first mean variance vectors are counted for a second number that is greater than the first threshold value beta. If it is countedThe second number does not exceed the first number threshold u, or the first mean variance of the first mean variance vector with the latest sampling timeLess than or equal to the second threshold value g _abs And if the first mean variance vector does not meet the first setting condition, determining that no active human body exists in the environment to be detected. Alternatively, k may take a value of 8,u and 2 g _abs The value is 0.25.

S405, obtaining a third number of second mean square values from the total mean variance vectors according to the time sequence from late to early.

The first mean variance vector is compared with a first set condition, the first mean variance vector does not meet the first set condition, no moving human body exists in the environment to be detected, and in order to avoid errors of a detection result that no moving human body exists in the environment to be detected, further judgment on the environment to be detected is needed to obtain a more accurate result. Similarly to the above, a third number of second mean square values are obtained from the total mean variance vector in time order from late to early. Assume that the third number is set toThen get +.>Second mean square values of +.>Optionally, ->The value is 10.

And S406, generating a second mean variance vector based on the second mean square difference value.

Generating a second mean variance vector based on the third number of second mean squared values obtained above, the second mean variance vector being inscribed as

And S407, determining that no person exists in the environment to be detected in response to the second mean variance vector meeting a second set condition.

Obtaining a second mean variance vectorIf the second mean variance of the second mean variance vector with the latest sampling time is +. >Less than a third threshold value h _abs And the second variance is smaller than the variance threshold sigma _th And determining that no person exists in the environment to be detected. Alternatively, h _abs 、σ _th The magnitude of the value can be set according to the actual situation.

According to the embodiment of the application, human body detection is carried out on the environment to be detected according to the mean variance of a plurality of CSI amplitude time sequence matrixes which are continuous in time, so that errors caused by comparison with only the first set condition are avoided, the calculation complexity is low, the detection range is wide, and the application range and the practicability are greatly improved.

Fig. 5 is an exemplary implementation manner of a human body detection method according to the present application, as shown in fig. 5, based on the foregoing embodiment, according to mean variance of a plurality of CSI amplitude time series matrices that are continuous in time, human body detection is performed on an environment to be detected, including the following steps:

s501, generating a total mean variance vector based on a plurality of mean variances which are continuous in time.

In step S501, the above embodiments have been specifically described, and will not be described herein.

S502, acquiring a first number of first mean variances from the total mean variance vector according to a time sequence from late to early.

In step S502, the above embodiments have been specifically described, and will not be described herein.

S503, generating a first mean variance vector based on the first mean variance.

In step S503, the above embodiment has been specifically described, and will not be described herein.

And S504, determining that no active human body exists in the environment to be detected in response to the first mean variance vector not meeting the first setting condition.

In step S504, the above embodiments have been specifically described, and will not be described herein.

S505, obtaining a third number of second mean square values from the total mean variance vectors according to the time sequence from late to early.

In step S505, the above embodiments have been specifically described, and will not be described herein.

S506, generating a second mean variance vector based on the second mean square value.

In step S506, the above embodiments have been specifically described, and will not be described herein.

S507, performing singular spectrum analysis on the CSI amplitude time sequence matrix to detect candidate respiratory frequencies of the stationary human body in response to the second mean variance vector not meeting the second set condition.

The second mean variance vector is compared with a second set condition, and the second mean variance vector does not meet the second set condition, so that no moving human body exists in the environment to be detected, optionally, in order to avoid error of the detection result that no moving human body exists in the environment to be detected, in order to obtain a more accurate result, singular spectrum analysis is required to be performed on the CSI amplitude time sequence matrix, so as to detect candidate respiratory rate of the stationary human body, and further judge the environment to be detected.

When the CSI amplitude time sequence matrix is subjected to singular spectrum analysis to detect the candidate respiratory frequency of a static human body, the CSI amplitude time sequence matrix A is firstly subjected to _K×N ＝[a ₁ ，a ₂ ，…，a _N ]Performing embedded transformation of N variable with window length of L to form a target matrix X _{L×((K-L+1)·N)} ＝[X ₁ ，X ₂ ，…，X _(K-L+1)·N] L is more than or equal to 2 and K is more than or equal to K. Wherein for vector a _j ＝(a _1，j ，a _2，j ，…，a _K，j ) ' j is greater than or equal to 1 and less than or equal to N, and matrix X is formed after the aperture length L is subjected to the Embedding transformation _{L×((K-L+1)·N)} Is respectively: x is X _{(K-L+1)·(j-1)+1} ，X _{(K-L+1)·(j-1)+2} ,…,X _(K-L+1)·j Wherein X is _{(K-L+1)·(j-1)+i} ＝(a _i，j ，a _i+1，j ，…，a _i+L-1，j ) ' i is more than or equal to 1 and less than or equal to K-L+1. For the target matrix X _{L×((K-L+1)·N)} Singular value decomposition (Singular Value Decomposition, SVD) is performed to obtain a plurality of decomposition matrices, the decomposition process being denoted as x=x ₁ +…+X _d . Wherein the matrix is decomposedTo decompose matrix X _i Is the characteristic value of U _i And V _i Respectively a decomposition matrix X _i Left and right eigenvectors of (a) will be (λ) _i ,U _i ，V _i ) As a feature triplet.

According to characteristic values in characteristic tripletsSorting feature triples, for example, if +.>Arranged in order of from big to small as +.>The corresponding feature triplets (lambda _i ，U _i ，V _i ) And also the ordering. The first r feature triples can be set as target feature triples, and when the first r feature triples are screened, the first sum of feature values in the selected feature triples is obtained >Obtaining the second sum of eigenvalues of all decomposition matrices +.>The first sum value to be obtainedAnd a second sum value->Dividing to obtain a first ratio->Likewise, the feature value lambda of the next feature triplet to be selected is acquired _r+1 And a second sum value->Second ratio>If the first ratio isGreater than or equal to a second set threshold η and a second ratio +.>And if the feature triplet is smaller than the third set threshold epsilon, determining that the currently selected feature triplet meets the third set condition, stopping backward continuous selection, and taking the currently selected feature triplet as the target feature triplet.

After the target feature triplet is screened from the feature triples, the left feature vector U of the target feature triplet _i Or right eigenvector V _i Performing fast Fourier transform to obtain left feature vector U in target feature triplet _i Or right eigenvector V _i And identifying the peak position of the spectrogram by the corresponding spectrogram, and determining the frequency corresponding to the peak position as the candidate respiratory frequency.

S508, determining that a stationary human body exists in the environment to be detected in response to the candidate respiratory frequency being in the respiratory frequency range.

Judging the frequency of the candidate respiratory rate, and if the candidate respiratory rate is in the respiratory rate range, determining that a stationary human body exists in the environment to be detected. For example, the normal adult respiratory rate is 12-20 times per minute, and if the candidate respiratory rate is 17 times per minute, it is determined that a stationary human body exists in the environment to be detected.

And S509, determining that no person exists in the environment to be detected in response to the candidate respiratory rate not being in the respiratory rate range.

And judging the frequency of the candidate respiratory rate, and if the candidate respiratory rate is not in the respiratory rate range, determining that no person exists in the environment to be detected. For example, the normal adult respiratory rate is 12-20 times per minute, and if the candidate respiratory rate is 2 times per minute, no person in the environment to be detected is determined.

According to the embodiment of the application, human body detection is carried out on the environment to be detected according to the mean variance of a plurality of CSI amplitude time sequence matrixes which are continuous in time, errors caused by comparison with the first setting condition and the second setting condition are avoided, the calculation complexity is low, the detection range is wide, and the application range and the practicability are greatly improved.

Fig. 6 is a schematic flow chart of an exemplary embodiment of a human body detection method according to the present application, as shown in fig. 6, comprising the following steps:

s601, sampling channel state information CSI of each subcarrier of Wi-Fi communication in an environment to be detected for multiple times.

S602, based on the sampled CSI, a CSI amplitude time sequence matrix is generated, wherein each line of the CSI amplitude time sequence matrix is the CSI amplitude value of each subcarrier acquired at the same sampling time.

S603, based on the CSI amplitude value on each column of the CSI amplitude time sequence matrix, acquiring a first variance of each column to form a first variance vector corresponding to the CSI amplitude time sequence matrix.

S604, carrying out weighted average on the first variance in the first variance vector to obtain the mean variance of the CSI amplitude time sequence matrix.

The steps S601 to S604 are specifically described in the above embodiments, and are not described herein.

S605, generating a total mean variance vector based on a plurality of mean variances which are continuous in time.

S606, acquiring a first number of first mean variances from the total mean variance vector according to the time sequence from late to early.

S607, generating a first mean variance vector based on the first mean variance.

S608, determining that the movable human body exists in the environment to be detected in response to the first mean variance vector meeting the first setting condition.

S609, determining that no active human body exists in the environment to be detected in response to the first mean variance vector not meeting the first setting condition.

S610, obtaining a third number of second mean square values from the total mean variance vectors in a time sequence from late to early.

S611, based on the second mean square value, a second mean variance vector is generated.

And S612, determining that no person exists in the environment to be detected in response to the second mean variance vector meeting a second set condition.

And S613, performing singular spectrum analysis on the CSI amplitude time sequence matrix to detect candidate respiratory frequencies of the static human body in response to the second mean variance vector not meeting the second set condition.

S614, in response to the candidate respiratory rate being within the respiratory rate range, determining that a stationary human body exists in the environment to be detected.

And S615, determining that no person exists in the environment to be detected in response to the candidate respiratory rate not being in the respiratory rate range.

The steps S605 to S615 are specifically described in the above embodiments, and are not described herein.

Fig. 7 is an exemplary embodiment of a shutdown method of a home appliance according to the present application, as shown in fig. 7, the shutdown method of a home appliance includes the following steps:

s701, channel state information CSI of all sub-carriers of Wi-Fi communication in a target environment where a first household appliance is located is sampled for a plurality of times.

The embodiment of the application is applied to the first household appliance, and a Wi-Fi module is arranged in the first household appliance. As shown in fig. 8, the first home device may receive, by using a Wi-Fi module, a Wi-Fi signal sent by a Wi-Fi transmitter, to obtain channel state information CSI, and a micro control unit (Microcontroller Unit, MCU) in the Wi-Fi module runs an indoor personnel detection algorithm, so as to determine whether a person exists in the target environment. Illustratively, the first home device may be an air conditioner, a television, a sweeping robot, etc., and the Wi-Fi transmitter may be a wireless router.

The foregoing embodiments have been specifically described for sampling the channel state information CSI of each subcarrier of Wi-Fi communication in the target environment where the first home device is located for multiple times, and will not be described herein.

S702, based on the sampled CSI, a CSI amplitude time sequence matrix is generated, wherein each line of the CSI amplitude time sequence matrix is the CSI amplitude value of each subcarrier acquired at the same sampling time.

In step S702, the above embodiments have been specifically described, and will not be described herein.

S703, obtaining the mean variance of the CSI amplitude time sequence matrix.

In step S703, the above embodiments have been specifically described, and will not be described herein.

S704, detecting a human body of the target environment according to the mean variance of the time series matrixes of the plurality of CSI amplitude in time.

In step S704, the above embodiment has been specifically described, and will not be described herein.

And S705, controlling the first household appliance to be closed in response to detecting that no person exists in the target environment.

And taking each home appliance which detects the target environment as a first home appliance, and acquiring detection results of the first home appliance and all other home appliances which are accessed into the Wi-Fi network on whether people exist in the target environment. Optionally, according to the detection result of the first home device and all other home devices accessing the Wi-Fi network on whether the target environment is occupied, whether the target environment is occupied or not may be finally determined by voting.

As an achievable manner, if it is finally determined that no person exists in the target environment, the first home device may send a close instruction to itself, so as to control the first home device to close itself.

As another implementation manner, if no person is finally determined in the target environment, the first home device may send a prompt message about whether to shutdown to the terminal device of the user, and the user may confirm whether to shutdown the first home device, if the user indicates to shutdown, the first home device is turned off, and if the user indicates not to shutdown or does not indicate, the first home device is not turned off.

As another implementation manner, if it is finally determined that a person exists in the target environment, the home appliance repeats steps S701 to S704.

The embodiment of the application provides a shutdown method of home appliances, which is implemented by sampling Channel State Information (CSI) of each subcarrier of Wi-Fi communication in a target environment where a first home appliance is located for a plurality of times; based on the sampled CSI, generating a CSI amplitude time sequence matrix, wherein each line of the CSI amplitude time sequence matrix is the CSI amplitude value of each subcarrier acquired at the same sampling time; acquiring the mean variance of the CSI amplitude time sequence matrix; human body detection is carried out on a target environment according to the mean variance of a plurality of CSI amplitude time sequence matrixes which are continuous in time; and controlling the first household appliance to be closed in response to detecting that no person exists in the target environment. The embodiment of the application can detect whether a person exists in the target environment in a non-invasive and real-time manner, does not need to add special detection hardware or perform advanced training or off-line calibration, has low calculation complexity and wide detection range, greatly improves the application range and the practicability, can control the shutdown of the household appliance under the condition of unmanned detection, and is convenient to use and saves resources.

Fig. 9 is an exemplary implementation manner of a power-off method of a home appliance according to the present application, as shown in fig. 9, based on the foregoing embodiment, in response to detecting that no person is in a target environment, the method includes the following steps:

and S901, receiving a detection result of a second household appliance accessed into the Wi-Fi network, wherein the detection result is used for indicating whether a person exists in the target environment.

In the embodiment of the application, each home appliance internally provided with the Wi-Fi module can be used as a first home appliance needing to confirm whether to perform shutdown operation, and when any home appliance internally provided with the Wi-Fi module is the first home appliance, all the other home appliances internally provided with the Wi-Fi module can be regarded as second home appliances. Fig. 10 is a schematic diagram of a plurality of home devices having Wi-Fi modules disposed therein, receiving Wi-Fi signals from Wi-Fi transmitters, taking fig. 10 as an example, when the home device 1 itself determines whether a person is in the target environment, the home device 1 itself may be regarded as a first home device, and all home devices having Wi-Fi modules disposed therein except the home device 1 may be regarded as second home devices. The first home appliance needs to judge whether a person exists in the target environment or not, and also needs to receive a detection result of the second home appliance connected to the Wi-Fi network, wherein the detection result is used for indicating whether the person exists in the target environment or not. Similarly, each first home device sends its detection result to a second home device in the Wi-Fi network, so that the second home device corresponding to the first home device can also receive the detection results of home devices other than the first home device.

S902, acquiring the total equipment number of the household appliances in the Wi-Fi network.

The total number of home appliances connected to the Wi-Fi network is obtained, for example, if 10 home appliances with Wi-Fi modules are built in a certain home, and the 10 home appliances are all in operation at present, the built-in Wi-Fi modules can judge whether someone is in the target environment, and then the total number of home appliances connected to the Wi-Fi network can be considered to be 10. Optionally, if a user of a home appliance has manually turned off the home appliance, or the user sets a home appliance to a no-disturb stage, so that the Wi-Fi module built in the home appliance does not operate, the home appliance cannot be considered as a home appliance accessing the Wi-Fi network.

S903, obtaining the number of unmanned devices in the target environment indicated by the detection result in the second household appliance.

Taking the total number of home devices accessing to the Wi-Fi network as 10 as an example, taking any one of the home devices as a first home device, obtaining the number of unmanned devices in the target environment indicated by the detection results in the remaining 9 second home devices, for example, 8 second home devices accessing to the Wi-Fi network indicate unmanned devices in the target environment, and considering that the number of unmanned devices in the target environment indicated by the detection results in the second home devices is 8.

S904, voting is carried out according to the number of the devices and the total device data, and a voting result is obtained.

And combining the detection result of the first household appliance with the detection result of the second household appliance, and voting whether someone exists in the target environment. Continuing taking the total number of the household appliances connected into the Wi-Fi network as 10 as an example, if the detection result of the first household appliance indicates that no person exists in the target environment and the detection result of the second household appliance indicates that the number of the unmanned appliances in the target environment is 8, the method indicates that among the 10 household appliances connected into the Wi-Fi network, 9 household appliances vote to the unmanned appliances in the target environment and 1 household appliance vote to the unmanned appliances in the target environment.

And S905, controlling the first household appliance to be closed in response to the voting result indicating no person in the target environment.

Setting a voting threshold, for example, the voting threshold may be set to 70%, i.e., if greater than or equal to 70% of the home devices indicate no people in the target environment, then no people in the target environment are considered; i.e., if greater than or equal to 70% of the home devices indicate a person in the target environment, then the person in the target environment is considered to be present.

When the household electrical appliance with the duration being more than or equal to 70% indicates that no person exists in the target environment, the duration of the unmanned person is continuously detected in the target environment, if the duration reaches the preset duration, optionally, the first household electrical appliance sends a closing instruction to the first household electrical appliance to control the first household electrical appliance to close the first household electrical appliance, or optionally, the first household electrical appliance sends a prompt message of whether the first household electrical appliance is closed to the terminal device of the user, and the user operation confirms whether the first household electrical appliance is closed. For example, if the preset duration is set to 30 minutes, and when more than or equal to 70% of the home devices indicate no person in the target environment and the duration of the continuous detection of the no person in the target environment is greater than or equal to 30 minutes, optionally, the first home device sends a closing instruction to itself to control itself to close, or alternatively, the first home appliance sends a prompt message of whether to shut down to the terminal device of the user, the user operates to confirm whether the first home appliance is shut down, if the user indicates to shut down, the first home appliance is shut down, and if the user indicates not to shut down or does not indicate, the first home appliance is not shut down.

According to the embodiment of the application, when no person is detected in the target environment, the first household appliance is controlled to be closed, whether a person exists in the current environment or not can be detected in a non-invasive way in real time, special detection hardware is not required to be added, and advanced training or offline calibration is not required.

Fig. 11 is an exemplary embodiment of a shutdown method of a home appliance according to the present application, as shown in fig. 11, the shutdown method of a home appliance includes the following steps:

s1101, sampling channel state information CSI of each subcarrier of Wi-Fi communication in the target environment where the first home device is located multiple times.

S1102, based on the sampled CSI, a CSI amplitude time sequence matrix is generated, wherein each line of the CSI amplitude time sequence matrix is the CSI amplitude value of each subcarrier acquired at the same sampling time.

S1103, the mean variance of the CSI amplitude time sequence matrix is obtained.

S1104, detecting the human body of the target environment according to the mean variance of the time series matrixes of the plurality of CSI amplitude in time.

The above embodiments have been specifically described in relation to steps S1101 to S1104, and will not be described herein.

S1105, receiving a detection result of a second household appliance accessed into the Wi-Fi network, wherein the detection result is used for indicating whether a person exists in the target environment.

S1106, acquiring the total equipment number of the household appliances in the Wi-Fi network.

S1107, acquiring the detection result in the second household appliance to indicate the number of unmanned appliances in the target environment.

S1108, voting is carried out according to the number of the devices and the total device data, and voting results are obtained.

S1109, controlling the first household appliance to be closed in response to the voting result indicating no person in the target environment.

The above embodiments have been specifically described in relation to steps S1105 to S1109, and will not be described here again.

Fig. 12 is a schematic diagram of a human body detection device according to the present application, and as shown in fig. 12, the human body detection device 1200 includes: a sampling module 1201, a generating module 1202, an obtaining module 1203 and a detecting module 1204, wherein:

the sampling module 1201 is configured to sample channel state information CSI of each subcarrier of Wi-Fi communication in an environment to be detected multiple times.

The generating module 1202 is configured to generate a CSI amplitude time sequence matrix based on the sampled CSI, where each of the CSI amplitude time sequence matrices is a CSI amplitude value of each subcarrier acquired at the same sampling time.

An obtaining module 1203 is configured to obtain a mean variance of the CSI amplitude time series matrix.

The detection module 1204 is configured to perform human body detection on an environment to be detected according to mean variances of a plurality of CSI amplitude time sequence matrices that are continuous in time.

Further, the generating module 1202 is further configured to: updating the CSI amplitude time sequence matrix of the last sampling time based on the CSI acquired at the current sampling time to generate the CSI amplitude time sequence matrix of the current sampling time; and acquiring the CSI amplitude time sequence matrixes of a plurality of sampling moments which are continuous in time, and acquiring the mean variance of each CSI amplitude time sequence matrix to acquire a plurality of mean variances which are continuous in time.

Further, the generating module 1202 is further configured to: shifting out a first row of matrix elements in the CSI amplitude time sequence matrix corresponding to the last sampling moment, and sequentially shifting forward each remaining row in the CSI amplitude time sequence matrix by one row; based on the CSI acquired at the current sampling moment, generating the last row of matrix elements of the CSI amplitude time sequence matrix, writing the last row of matrix elements into the CSI time sequence matrix, and generating the CSI amplitude time sequence matrix at the current sampling moment.

Further, the obtaining module 1203 is further configured to: based on the CSI amplitude value on each column of the CSI amplitude time sequence matrix, acquiring a first variance of each column to form a first variance vector corresponding to the CSI amplitude time sequence matrix; and carrying out weighted average on the first variance in the first variance vector to obtain the mean variance of the CSI amplitude time sequence matrix.

Further, the detecting module 1204 is further configured to: generating a total mean variance vector based on the plurality of mean variances that are temporally continuous; acquiring a first number of first mean variances from the total mean variance vector in a time sequence from late to early; generating a first mean variance vector based on the first mean variance; and responding to the first mean variance vector to meet a first set condition, and determining that the movable human body exists in the environment to be detected.

Further, the detecting module 1204 is further configured to: counting a second number of first mean variances larger than a first threshold value in the first mean variance vector; and in response to the second number exceeding the first number threshold and the first mean variance of the first mean variance vector at the latest sampling time being greater than the second threshold, determining that the first mean variance vector meets a first set condition, and determining that a movable human body exists in the environment to be detected.

Further, the detecting module 1204 is further configured to: determining that no movable human body exists in the environment to be detected in response to the first mean variance vector not meeting the first setting condition; obtaining a third number of second mean square values from the total mean variance vectors in a time sequence from late to early; generating a second mean variance vector based on the second mean variance value; and responding to the second mean variance vector to meet a second set condition, and determining that no person exists in the environment to be detected.

Further, the detecting module 1204 is further configured to: acquiring a second variance of a second mean variance vector; and responding to the fact that the second mean variance of the second mean variance vector, which is the latest in sampling time, is smaller than a third threshold value, wherein the second variance is smaller than a variance threshold value, determining that the second mean variance vector meets a second set condition, and determining that no person exists in the environment to be detected.

Further, the detecting module 1204 is further configured to: and responding to the fact that the second mean variance vector does not meet a second set condition, and performing static personnel detection on the environment to be detected based on the CSI amplitude time sequence matrix.

Further, the detecting module 1204 is further configured to: performing singular spectrum analysis on the CSI amplitude time sequence matrix to detect candidate respiratory frequencies of the static human body; determining that a stationary human body exists in the environment to be detected in response to the candidate respiratory frequency being within the respiratory frequency range; alternatively, in response to the candidate respiratory rate not being within the respiratory rate range, it is determined that no person is within the environment to be detected.

Further, the detecting module 1204 is further configured to: transforming the CSI amplitude time sequence matrix to obtain a transformed target matrix; singular value decomposition is carried out on the target matrix to obtain a plurality of decomposition matrices; acquiring a feature triplet of the decomposition matrix, wherein the feature triplet comprises a feature value, a left feature vector and a right feature vector of the decomposition matrix; screening a target feature triplet from the feature triplet, and acquiring a spectrogram of one of a left feature vector and a right feature vector in the target feature triplet; candidate respiratory frequencies are extracted from the spectrogram.

Further, the detecting module 1204 is further configured to: sorting the feature triples according to the feature values in the feature triples; sequentially selecting feature triples in sequence, and acquiring a first sum of feature values in the selected feature triples; and stopping backward continuous selection in response to the currently acquired first sum value and the characteristic value of the next characteristic triplet to be selected meeting a third setting condition, and taking the currently selected characteristic triplet as a target characteristic triplet.

Further, the detecting module 1204 is further configured to: obtaining second sum values of the eigenvalues of all the decomposition matrixes; acquiring a first ratio of the first sum value to the second sum value and a second ratio of the characteristic value of the next characteristic triplet to be selected to the second sum value; and determining a third setting condition in response to the first ratio being greater than or equal to the second setting threshold and the second ratio being less than the third setting threshold, and stopping continuing to select backwards.

Further, the detecting module 1204 is further configured to: and identifying the peak position of the spectrogram, and determining the frequency corresponding to the peak position as the candidate respiratory frequency.

Fig. 13 is a schematic diagram of a shutdown device of a home appliance according to the present application, as shown in fig. 13, the shutdown device 1300 of the home appliance includes: a sampling module 1301, a generating module 1302, an obtaining module 1303, a detecting module 1304, and a control module 1305, wherein:

The sampling module 1301 is configured to sample, for multiple times, channel state information CSI of each subcarrier of Wi-Fi communication in a target environment where the first home device is located.

The generating module 1302 is configured to generate a CSI amplitude time sequence matrix based on the sampled CSI, where each of the CSI amplitude time sequence matrices is a CSI amplitude value of each subcarrier acquired at the same sampling time.

The obtaining module 1303 is configured to obtain a mean variance of the CSI amplitude time series matrix.

The detection module 1304 is configured to perform human body detection on the target environment according to mean variances of a plurality of CSI amplitude time sequence matrices that are continuous in time.

And the control module 1305 is used for controlling the first household appliance to be closed in response to detecting that no person exists in the target environment.

Further, the detection module 1304 is further configured to: updating the CSI amplitude time sequence matrix of the last sampling time based on the CSI acquired at the current sampling time to generate the CSI amplitude time sequence matrix of the current sampling time; and acquiring the CSI amplitude time sequence matrixes of a plurality of sampling moments which are continuous in time, and acquiring the mean variance of each CSI amplitude time sequence matrix to acquire a plurality of mean variances which are continuous in time.

Further, the detection module 1304 is further configured to: shifting out a first row of matrix elements in the CSI amplitude time sequence matrix corresponding to the last sampling moment, and sequentially shifting forward each remaining row in the CSI amplitude time sequence matrix by one row; based on the CSI acquired at the current sampling moment, generating the last row of matrix elements of the CSI amplitude time sequence matrix, writing the last row of matrix elements into the CSI time sequence matrix, and generating the CSI amplitude time sequence matrix at the current sampling moment.

Further, the obtaining module 1303 is further configured to: based on the CSI amplitude value on each column of the CSI amplitude time sequence matrix, acquiring a first variance of each column to form a first variance vector corresponding to the CSI amplitude time sequence matrix; and carrying out weighted average on the first variance in the first variance vector to obtain the mean variance of the CSI amplitude time sequence matrix.

Further, the detection module 1304 is further configured to: generating a total mean variance vector based on the plurality of mean variances that are temporally continuous; acquiring a first number of first mean variances from the total mean variance vector in a time sequence from late to early; generating a first mean variance vector based on the first mean variance; and determining that an active human body exists in the target environment in response to the first mean variance vector meeting a first set condition.

Further, the detection module 1304 is further configured to: counting a second number of first mean variances larger than a first threshold value in the first mean variance vector; and in response to the second number exceeding the first number threshold and the first mean variance of the first mean variance vector at the latest sampling time being greater than the second threshold, determining that the first mean variance vector meets a first setting condition, and determining that a movable human body exists in the target environment.

Further, the detection module 1304 is further configured to: determining that no moving human body exists in the target environment in response to the first mean variance vector not meeting the first set condition; obtaining a third number of second mean square values from the total mean variance vectors in a time sequence from late to early; generating a second mean variance vector based on the second mean variance value; and determining that no person exists in the target environment in response to the second mean variance vector meeting a second set condition.

Further, the detection module 1304 is further configured to: acquiring a second variance of a second mean variance vector; and responding to the fact that the second mean variance of the second mean variance vector, which is the latest in sampling time, is smaller than a third threshold value, wherein the second variance is smaller than a variance threshold value, determining that the second mean variance vector meets a second setting condition, and determining that no person exists in the target environment.

Further, the detection module 1304 is further configured to: and responding to the second mean variance vector not meeting a second set condition, and performing static personnel detection on the target environment based on the CSI amplitude time sequence matrix.

Further, the detection module 1304 is further configured to: performing singular spectrum analysis on the CSI amplitude time sequence matrix to detect candidate respiratory frequencies of the static human body; determining that a stationary human body exists in the target environment in response to the candidate respiratory rate being within the respiratory rate range; alternatively, in response to the candidate respiratory rate not being within the respiratory rate range, it is determined that no person is within the target environment.

Further, the detection module 1304 is further configured to: transforming the CSI amplitude time sequence matrix to obtain a transformed target matrix; singular value decomposition is carried out on the target matrix to obtain a plurality of decomposition matrices; acquiring a feature triplet of the decomposition matrix, wherein the feature triplet comprises a feature value, a left feature vector and a right feature vector of the decomposition matrix; screening a target feature triplet from the feature triplet, and acquiring a spectrogram of one of a left feature vector and a right feature vector in the target feature triplet; candidate respiratory frequencies are extracted from the spectrogram.

Further, the detection module 1304 is further configured to: sorting the feature triples according to the feature values in the feature triples; sequentially selecting feature triples in sequence, and acquiring a first sum of feature values in the selected feature triples; and stopping backward continuous selection in response to the currently acquired first sum value and the characteristic value of the next characteristic triplet to be selected meeting a third setting condition, and taking the currently selected characteristic triplet as a target characteristic triplet.

Further, the detection module 1304 is further configured to: obtaining second sum values of the eigenvalues of all the decomposition matrixes; acquiring a first ratio of the first sum value to the second sum value and a second ratio of the characteristic value of the next characteristic triplet to be selected to the second sum value; and determining a third setting condition in response to the first ratio being greater than or equal to the second setting threshold and the second ratio being less than the third setting threshold, and stopping continuing to select backwards.

Further, the detection module 1304 is further configured to: and identifying the peak position of the spectrogram, and determining the frequency corresponding to the peak position as the candidate respiratory frequency.

Further, the control module 1305: receiving a detection result of a second household appliance accessed into the Wi-Fi network, wherein the detection result is used for indicating whether a person exists in a target environment; acquiring the total equipment number of home appliances accessed into a Wi-Fi network; acquiring the number of unmanned equipment in the detection result indication target environment in the second household electrical appliance; voting is carried out according to the number of the devices and the total device data, and a voting result is obtained; and controlling the first household appliance to be closed in response to the voting result indicating no person in the target environment.

Further, the control module 1305 is further configured to: and acquiring the duration of the unmanned aerial vehicle continuously detected in the target environment, and controlling the first household appliance to be closed in response to the duration reaching the preset duration.

Further, the detection module 1304 is further configured to: and sending the detection result of the first household appliance to a second household appliance in the Wi-Fi network.

Fig. 14 is a schematic diagram of a receiving device with human body detection according to the present application, and as shown in fig. 14, the receiving device 1400 with human body detection includes: a receiver 1401 and a controller 1402, wherein:

a receiver 1401 for receiving a plurality of Wi-Fi signals transmitted by Wi-Fi communication on each subcarrier;

the controller 1402 is configured to acquire channel state information CSI of subcarriers according to Wi-Fi signals, generate a CSI amplitude time series matrix based on the sampled CSI, wherein each row of CSI amplitude values of each subcarrier acquired at a same sampling time in the CSI amplitude time series matrix, acquire a mean variance of the CSI amplitude time series matrix, and perform human body detection on a target environment according to the mean variances of a plurality of CSI amplitude time series matrices that are continuous in time.

Fig. 15 is a schematic diagram of a human body detection system according to the present application, and as shown in fig. 15, the human body detection system 1500 includes: a receiving device 1501 and a transmitting device 1502 with human detection. A transmitting device 1502, configured to send a plurality of Wi-Fi signals through each subcarrier of the Wi-Fi network.

In order to implement the foregoing embodiments, an embodiment of the present application further provides an electrical home appliance 1600, as shown in fig. 16, where the electrical home appliance 1600 includes: the processor 1601 is communicatively connected to a memory 1602, and the memory 1602 stores instructions executable by at least one processor, and the instructions are executed by the at least one processor 1601 to implement the human detection method and the power-off method of the home device as described in the above embodiments.

In order to achieve the above embodiments, the embodiments of the present application also provide a non-transitory computer-readable storage medium storing computer instructions for causing a computer to implement the human body detection method and the shutdown method of the home appliance device as shown in the above embodiments.

In order to implement the above embodiments, the embodiments of the present application further provide a computer program product, including a computer program, which when executed by a processor implements the human body detection method and the power-off method of the home appliance device as shown in the above embodiments.

Furthermore, the terms "first," "second," and the like, are used for descriptive purposes only and are not to be construed as indicating or implying a relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include one or more such feature. In the description of the present application, the meaning of "a plurality" is two or more, unless explicitly defined otherwise.

In the description of the present specification, a description referring to terms "one embodiment," "some embodiments," "examples," "specific examples," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present application. In this specification, schematic representations of the above terms are not necessarily directed to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, the different embodiments or examples described in this specification and the features of the different embodiments or examples may be combined and combined by those skilled in the art without contradiction.

While embodiments of the present application have been shown and described above, it will be understood that the above embodiments are illustrative and not to be construed as limiting the application, and that variations, modifications, alternatives and variations may be made to the above embodiments by one of ordinary skill in the art within the scope of the application.

Claims

1. A human body detection method, comprising:

Sampling Channel State Information (CSI) of each subcarrier of Wi-Fi communication in an environment to be detected for multiple times;

generating a CSI amplitude time sequence matrix based on the sampled CSI, wherein each line of the CSI amplitude time sequence matrix is a CSI amplitude value of each subcarrier acquired at the same sampling moment;

acquiring the mean variance of the CSI amplitude time sequence matrix;

generating a total mean variance vector based on a plurality of the mean variances that are temporally continuous;

acquiring a first number of first mean variances from the total mean variance vector in a time sequence from late to early;

generating a first mean variance vector based on the first mean variance;

determining that an active human body exists in the environment to be detected in response to the first mean variance vector meeting a first set condition;

responding to the first mean variance vector not meeting a first set condition, and acquiring a third number of second mean square values from the total mean variance vector according to a time sequence from late to early;

generating a second mean variance vector based on the second mean squared value;

determining that no active human body exists in the environment to be detected in response to the second mean variance vector meeting a second set condition;

And responding to the second mean variance vector not meeting the second set condition, and carrying out static personnel detection on the environment to be detected based on the CSI amplitude time sequence matrix.

2. The method according to claim 1, wherein before human body detection is performed on the environment to be detected according to a mean variance of a plurality of CSI amplitude time series matrices that are continuous in time, further comprising:

updating the CSI amplitude time sequence matrix of the last sampling time based on the CSI acquired at the current sampling time, and generating the CSI amplitude time sequence matrix of the current sampling time;

and acquiring the CSI amplitude time sequence matrixes of a plurality of sampling moments which are continuous in time, and acquiring the mean variance of each CSI amplitude time sequence matrix to acquire a plurality of mean variances which are continuous in time.

3. The method according to claim 2, wherein updating the CSI amplitude time series matrix at a previous sampling time based on the CSI acquired at a current sampling time, to generate the CSI amplitude time series matrix at a current sampling time, comprises:

shifting out a first row of matrix elements in the CSI amplitude time sequence matrix corresponding to the last sampling moment, and sequentially shifting forward each remaining row in the CSI amplitude time sequence matrix by one row;

Based on the CSI acquired at the current sampling moment, generating a last row of matrix elements of the CSI amplitude time sequence matrix, writing the last row of matrix elements into the CSI time sequence matrix, and generating the CSI amplitude time sequence matrix at the current sampling moment.

4. A method according to any of claims 1-3, wherein said obtaining the mean variance of the CSI amplitude time series matrix comprises:

acquiring a first variance of each column based on the CSI amplitude value on each column of the CSI amplitude time sequence matrix to form a first variance vector corresponding to the CSI amplitude time sequence matrix;

and carrying out weighted average on the first variance in the first variance vector to obtain the mean variance of the CSI amplitude time sequence matrix.

5. The method according to claim 1, wherein the method further comprises:

counting a second number of the first mean variance vectors, wherein the second number of the first mean variance vectors is larger than a first threshold value;

and in response to the second number exceeding a first number threshold and the first mean variance of the first mean variance vector with the latest sampling time being greater than a second threshold, determining that the first mean variance vector meets a first set condition, and determining that a movable human body exists in the environment to be detected.

6. The method according to claim 1, wherein the method further comprises:

acquiring a second variance of the second mean variance vector;

and responding to the fact that the second mean variance of the second mean variance vector with the latest sampling time is smaller than a third threshold value, wherein the second variance is smaller than a variance threshold value, determining that the second mean variance vector meets a second set condition, and determining that no person exists in the environment to be detected.

7. The method according to claim 1, wherein the performing static person detection on the environment to be detected based on the CSI amplitude time series matrix comprises:

performing singular spectrum analysis on the CSI amplitude time sequence matrix to detect candidate respiratory frequencies of the static human body;

determining that a stationary human body exists in the environment to be detected in response to the candidate respiratory frequency being within a respiratory frequency range; or,

and determining that no person exists in the environment to be detected in response to the candidate respiratory rate not being in the respiratory rate range.

8. The method of claim 7, wherein the performing singular spectrum analysis on the CSI amplitude time series matrix to detect candidate respiratory frequencies of a stationary human body comprises:

Transforming the CSI amplitude time sequence matrix to obtain a transformed target matrix;

singular value decomposition is carried out on the target matrix to obtain a plurality of decomposition matrices;

acquiring a feature triplet of the decomposition matrix, wherein the feature triplet comprises a feature value, a left feature vector and a right feature vector of the decomposition matrix;

screening a target feature triplet from the feature triplet, and acquiring a spectrogram of one of a left feature vector and a right feature vector in the target feature triplet;

and extracting the candidate respiratory frequency from the spectrogram.

9. The method of claim 8, wherein the screening the target feature triples from the feature triples comprises:

sorting the feature triples according to the feature values in the feature triples;

sequentially selecting the characteristic triples in sequence, and acquiring a first sum of characteristic values in the selected characteristic triples;

and stopping backward continuous selection in response to the currently acquired first sum value and the characteristic value of the next characteristic triplet to be selected meeting a third setting condition, and taking the currently selected characteristic triplet as the target characteristic triplet.

10. The method according to claim 9, wherein the method further comprises:

obtaining second sum values of the eigenvalues of all the decomposition matrixes;

acquiring a first ratio of the first sum value to the second sum value and a second ratio of the feature value of the next feature triplet to be selected to the second sum value;

and determining the third setting condition in response to the first ratio being greater than or equal to a second setting threshold and the second ratio being less than a third setting threshold, and stopping continuing to select backwards.

11. The method of claim 8, wherein the extracting the candidate respiratory rate from the spectrogram comprises:

and identifying the peak position of the spectrogram, and determining the frequency corresponding to the peak position as the candidate respiratory frequency.

12. A method of powering off a home device, adapted for a first home device, the method comprising:

sampling Channel State Information (CSI) of each subcarrier of Wi-Fi communication in a target environment where the first household appliance is located for multiple times;

Acquiring the mean variance of the CSI amplitude time sequence matrix;

generating a first mean variance vector based on the first mean variance;

determining that an active human body exists in the target environment in response to the first mean variance vector meeting a first set condition;

determining that no active human body exists in the target environment in response to the second mean variance vector meeting a second set condition;

responding to the second mean variance vector not meeting the second set condition, and performing static personnel detection on the target environment based on the CSI amplitude time sequence matrix;

and controlling the first household appliance to be closed in response to detecting that no person exists in the target environment.

13. The method of claim 12, wherein prior to human detection of the target environment based on the mean variance of the plurality of CSI amplitude time series matrices that are consecutive in time, further comprising:

14. The method of claim 13, wherein updating the CSI amplitude time series matrix for the last sampling time based on the CSI acquired at the current sampling time, to generate the CSI amplitude time series matrix for the current sampling time, comprises:

15. The method according to any of claims 12-14, wherein the obtaining the mean variance of the CSI amplitude time series matrix comprises:

16. The method according to claim 12, wherein the method further comprises:

and in response to the second number exceeding a first number threshold and the first mean variance of the first mean variance vector with the latest sampling time being greater than a second threshold, determining that the first mean variance vector meets a first set condition, and determining that an active human body exists in the target environment.

17. The method according to claim 12, wherein the method further comprises:

acquiring a second variance of the second mean variance vector;

And responding to the fact that the second mean variance of the second mean variance vector with the latest sampling time is smaller than a third threshold value, wherein the second variance is smaller than a variance threshold value, determining that the second mean variance vector meets a second set condition, and determining that no person exists in the target environment.

18. The method of claim 12, wherein the performing static person detection on the target environment based on the CSI amplitude time series matrix comprises:

determining that a stationary human body is present within the target environment in response to the candidate respiratory frequency being within a respiratory frequency range; or,

in response to the candidate respiratory rate not being within the respiratory rate range, determining that no person is within the target environment.

19. The method of claim 18, wherein the performing singular spectrum analysis on the CSI amplitude time series matrix to detect candidate respiratory frequencies of a stationary human body comprises:

and extracting the candidate respiratory frequency from the spectrogram.

20. The method of claim 19, wherein the screening the target feature triples from the feature triples comprises:

21. The method of claim 20, wherein the method further comprises:

22. The method of claim 19, wherein the extracting the candidate respiratory rate from the spectrogram comprises:

23. The method of claim 12, wherein the controlling the first home device to close in response to detecting the absence of a person within the target environment comprises:

receiving a detection result of a second household appliance accessed into the Wi-Fi network, wherein the detection result is used for indicating whether a person exists in the target environment;

acquiring the total equipment number of household appliances connected into the Wi-Fi network;

acquiring the detection result in the second household appliance to indicate the number of unmanned appliances in the target environment;

Voting according to the number of the devices and the total device data, and obtaining a voting result;

and controlling the first household appliance to be closed in response to the voting result indicating no human being in the target environment.

24. The method of claim 12 or 23, wherein the controlling the first home device to turn off further comprises:

and acquiring the duration of continuous detection of the unmanned in the target environment, and controlling the first household appliance to be closed in response to the duration reaching a preset duration.

25. The method of claim 12, wherein after the human body detection of the target environment, further comprising:

and sending the detection result of the first household appliance to a second household appliance which is accessed into the Wi-Fi network.

26. A human body detection apparatus, comprising:

the sampling module is used for sampling Channel State Information (CSI) of each subcarrier of Wi-Fi communication in the environment to be detected for a plurality of times;

the generation module is used for generating a CSI amplitude time sequence matrix based on the sampled CSI, wherein each line of the CSI amplitude time sequence matrix is the CSI amplitude value of each subcarrier acquired at the same sampling time;

The acquisition module is used for acquiring the mean variance of the CSI amplitude time sequence matrix;

the detection module is used for detecting the human body of the environment to be detected according to the mean variance of the time series matrixes of the CSI amplitude;

the device is particularly used for:

generating a first mean variance vector based on the first mean variance;

27. The apparatus of claim 26, wherein the generating module is further configured to:

28. A power-off apparatus for a home appliance, comprising:

the sampling module is used for sampling Channel State Information (CSI) of each subcarrier of Wi-Fi communication in a target environment where the first household appliance is located for a plurality of times;

the generation module is used for generating an amplitude time sequence matrix based on the sampled CSI, wherein each line of the CSI amplitude time sequence matrix is the CSI amplitude value of each subcarrier acquired at the same sampling time;

the detection module is used for detecting the human body of the target environment according to the mean variance of the time series matrixes of the CSI amplitude in time;

The control module is used for controlling the first household appliance to be closed in response to detecting that no person exists in the target environment;

the apparatus is specifically configured to, prior to controlling the first home device to close in response to detecting that no person is within the target environment:

generating a first mean variance vector based on the first mean variance;

and responding to the second mean variance vector not meeting the second set condition, and performing static personnel detection on the target environment based on the CSI amplitude time sequence matrix.

29. The apparatus of claim 28, wherein the detection module is further configured to:

30. A receiving device with human detection, comprising:

the receiver is used for receiving Wi-Fi signals sent by Wi-Fi communication on all subcarriers;

the controller is used for acquiring Channel State Information (CSI) of the subcarriers according to the Wi-Fi signals, generating a CSI amplitude time sequence matrix based on the sampled CSI, wherein each of the CSI amplitude time sequence matrix is a CSI amplitude value of each subcarrier acquired at the same sampling moment, acquiring a mean variance of the CSI amplitude time sequence matrix, and detecting a human body of a target environment according to the mean variances of a plurality of continuous time CSI amplitude time sequence matrices;

The receiving device is specifically configured to:

generating a first mean variance vector based on the first mean variance;

responding to the first mean variance vector to meet a first set condition, and determining that an active human body exists in the environment to be detected;

31. A human body detection system, comprising:

the receiving device with human detection of claim 30;

And the transmitting device is used for transmitting Wi-Fi signals through each subcarrier of Wi-Fi communication.

32. An electrical home appliance, comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-11 or claims 12-25.

33. A non-transitory computer readable storage medium storing computer instructions for causing the computer to perform the method of any one of claims 1-11 or claims 12-25.