CN108416974B - Automatic alarm device and method based on wireless channel state information - Google Patents

Abstract

The invention discloses an automatic alarm device and method based on wireless channel state information, the wireless channel state information is the channel state information of communication sub-carrier between the transmitting terminal and the receiving terminal in the wireless communication network, the automatic alarm device includes: the alarm module is used for acquiring the channel state information, identifying the human gestures or human actions in the coverage area of the wireless communication network according to the channel state information and generating alarm information according to the identification result; and the communication module is used for receiving the alarm information and giving an alarm according to the alarm information. The automatic alarm device can conveniently and effectively realize automatic alarm and has lower cost.

Description

Automatic alarm device and method based on wireless channel state information

technical field

The invention relates to the technical field of wireless communication, in particular to an automatic alarm device based on wireless channel state information and an automatic alarm method based on wireless channel state information.

Background technique

Home security has always been a concern of people. With the development of technology, people's requirements for security are getting higher and higher. The mainstream alarm systems are based on button-based active alarms and sensor-based passive alarms. There are certain limitations in the active alarm through the button type. When monitoring a large area, the passive alarm based on sensors requires a certain number of sensors to be arranged in the monitoring area, and as the number of sensors increases, the cost will increase, and the passive alarm system based on sensors also requires people to pass Press the button to dial the communication device to alarm. When personal freedom is restricted, there will be certain limitations.

SUMMARY OF THE INVENTION

The present invention aims to at least to a certain extent solve the technical problems such as high cost of the alarm system, inconvenient alarm, etc. Therefore, one object of the present invention is to propose an automatic alarm device based on wireless channel state information, which can conveniently and effectively realize automatic alarming. Alarm, and low cost.

The second object of the present invention is to propose an automatic alarm method based on wireless channel state information.

In order to achieve the above object, an automatic alarm device based on wireless channel state information proposed by the embodiment of the first aspect of the present invention, wherein the wireless channel state information is a channel of a communication subcarrier between a transmitting terminal and a receiving terminal in a wireless communication network status information, the automatic alarm device includes: an alarm module, the alarm module is configured to acquire the channel status information, and identify the gestures or movements of persons in the coverage area of the wireless communication network according to the channel status information , and generate alarm information according to the identification result; a communication module, the communication module is used for receiving the alarm information and giving an alarm according to the alarm information.

According to the automatic alarm device based on the wireless channel state information according to the embodiment of the present invention, the alarm module can obtain the channel state information, identify the gestures or movements of the people in the coverage area of the wireless communication network according to the channel state information, and generate a signal according to the identification result. Alarm information, the communication module can alarm according to the alarm information. Therefore, using the channel state information as the physical quantity for action recognition has the advantages of stability, reliability, high precision, and no identification blind spots. It makes reasonable use of the existing wireless communication equipment. Low cost, easy to popularize, and does not require the human body to carry any active equipment, further reducing the cost. Generally speaking, the automatic alarm device of the embodiment of the present invention can realize automatic alarm conveniently and effectively, and the cost is low.

In addition, the automatic alarm device based on wireless channel state information proposed according to the above embodiments of the present invention may also have the following additional technical features:

Further, the alarm module includes: a data processing unit for performing data preprocessing on the channel state information; a continuous action recognition unit for performing data preprocessing on the channel state information data obtained by the continuous action recognition algorithm Perform continuous action recognition to obtain valid action data segments; an action segmentation unit for segmenting the valid action data segments to obtain a plurality of individual action data; an identification unit for based on the individual action data and the deep learning algorithm The human gesture or the human action is recognized.

Specifically, the channel state information is obtained by sampling at the preset sampling frequency by the receiving terminal, and the channel state information data obtained by each sampling is represented as a matrix, and the data processing unit is used for: using Hampel filter, Butterworth The filter performs noise reduction processing on the channel state information data; and reconstructs the channel state information data after the noise reduction processing by a weighted moving average method.

Specifically, the continuous motion recognition unit is used for: intercepting the channel state information data obtained after data preprocessing through a sliding window with a preset window size and a preset step size, so as to obtain a plurality of segmentation matrices; Multiply the segmentation matrix by its own transpose to obtain the corresponding correlation matrix; calculate the eigenvalues and eigenvectors of each correlation matrix, and judge the changes of the eigenvalues and eigenvectors of multiple correlation matrices; according to the changes It is determined whether there is a specific action or substantial action to determine the valid action data segment.

其中，t_i ^s、t_i ^e分别为预估的第i个动作的起始点和结束点，且t_i ^s、t_i ^e构成一个数据对；设定安全间隔参数T^b，并通过所述安全间隔参数对所述预估集合中的每个数据对进行扩展，得到新的集合：

以根据所述新的集合得到多个单独动作数据。Specifically, the action segmentation unit is used for: estimating the start point and the end point of each action in the valid action data segment to obtain an estimated set:

Among them, t _i ^s and t _i ^e are the estimated starting point and end point of the ith action respectively, and t _i ^s and t _i ^e form a data pair; set the safety interval parameter T ^b , and pass the The safety interval parameter extends each data pair in the estimated set to obtain a new set:

to obtain a plurality of individual action data according to the new set.

Specifically, a sample data set is stored in the identifying unit, wherein the sample data representing each person's gestures and actions are learned in a deep learning algorithm in advance, and corresponding sample feature vectors are extracted, so as to obtain the sample data according to the sample feature vectors. The sample data set is established, and the identification unit is used for: using a deep learning algorithm to train each of the individual action data to obtain a feature vector of the corresponding action to be identified; The sample feature vectors in the sample data set are compared to identify the person gesture or the person action.

Further, the alarm module further includes: an active alarm unit, which is configured to recognize an alarm number according to the person gesture when the recognition unit recognizes the person gesture, and send the alarm number to the alarm number. Send to the communication module; passive alarm unit, the passive alarm unit is used to further judge the type of the personnel action when the identification unit recognizes the personnel action, and judges according to the type of the personnel action When danger occurs, corresponding alarm information is generated, and the alarm information is sent to the communication module.

Further, the communication module dials an alarm when receiving the alarm number, and sends an alarm to a corresponding alarm receiving agency when receiving the alarm information.

In order to achieve the above object, an embodiment of the second aspect of the present invention proposes an automatic alarm method based on wireless channel state information, wherein the wireless channel state information is a communication subcarrier between a transmitting terminal and a receiving terminal in a wireless communication network. The automatic alarm method includes: acquiring the channel state information; recognizing the hand gestures or human actions in the coverage area of the wireless communication network according to the channel state information, and generating alarm information according to the recognition result ; alarm according to the alarm information.

According to the automatic alarm method based on the wireless channel state information according to the embodiment of the present invention, the channel state information can be obtained, and according to the channel state information, the gestures or movements of the people in the coverage area of the wireless communication network can be recognized, and the alarm information can be generated according to the recognition result. , and can make an alarm according to the alarm information. Therefore, using the channel state information as the physical quantity for action recognition has the advantages of stability, reliability, high precision, and no identification blind area. It makes reasonable use of the existing wireless communication equipment, low cost and easy It is popular, and the human body does not need to carry any active equipment, which further reduces the cost. In general, the automatic alarm method of the embodiment of the present invention can realize automatic alarm conveniently and effectively.

In addition, the automatic alarm method based on wireless channel state information proposed according to the above-mentioned embodiments of the present invention may also have the following additional technical features:

Further, recognizing human gestures or human actions in the coverage area of the wireless communication network according to the channel state information specifically includes: performing data preprocessing on the channel state information; performing data preprocessing on the channel state information; The channel state information data obtained after processing is subjected to continuous action recognition to obtain valid action data segments; the valid action data segments are segmented to obtain a plurality of individual action data; the individual action data and the deep learning algorithm are identified according to the A person gesture or said person action.

Description of drawings

1 is a schematic structural diagram of a wireless communication network according to an embodiment of the present invention;

2 is a schematic block diagram of an automatic alarm device based on wireless channel state information according to an embodiment of the present invention;

3 is a schematic structural diagram of an automatic alarm device based on wireless channel state information according to an embodiment of the present invention;

FIG. 4 is a flowchart of an automatic alarm method based on wireless channel state information according to an embodiment of the present invention.

Detailed ways

The following describes in detail the embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals refer to the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary, and are intended to explain the present invention and should not be construed as limiting the present invention.

The automatic alarm device and method based on wireless channel state information according to the embodiments of the present invention will be described below with reference to the accompanying drawings.

The wireless channel state information in the embodiment of the present invention is the channel state information of the communication subcarriers between the transmitting terminal and the receiving terminal in the wireless communication network. The wireless communication network may be WIFI (Wireless Fidelity, a wireless local area network), the transmitting terminal may include a router, and the receiving terminal may include a terminal device configured with a wireless network card and connected to the router, such as a mobile phone, a tablet computer, and a notebook computer.

As shown in Figure 1, the channel between the transmitting terminal and the receiving terminal is composed of multiple sub-carriers. When the human body is in the signal transmission space and makes an action, it can affect the propagation of the signal. In the orthogonal frequency division multiplexing system according to an embodiment of the present invention, the channel of each pair of transmitting terminal and receiving terminal is composed of 30 sub-carriers. The change of the phase value or the amplitude value of the channel state information CSI, and this change has a significant influence on the weak action. Therefore, in the embodiment of the present invention, the action identification can be performed according to the channel state information of the subcarriers.

通过内嵌的正交频分复用系统，将该信道状态信息数据保存，接下来将以该信道状态信息数据为基础进行动作识别和报警等。In an embodiment of the present invention, the channel state information CSI may be obtained by sampling at a preset sampling frequency by the receiving terminal, and the channel state information data obtained by each sampling may be represented as a matrix. Specifically, the receiving terminal may sample the channel state information data at a sampling frequency of 2000 packets/second, and obtain a matrix for each sampling

Through the built-in orthogonal frequency division multiplexing system, the channel state information data is saved, and then the action recognition and alarm will be performed based on the channel state information data.

FIG. 2 is a schematic block diagram of an automatic alarm device based on wireless channel state information according to an embodiment of the present invention.

As shown in FIG. 2 , the automatic alarm device based on wireless channel state information according to the embodiment of the present invention includes an alarm module 10 and a communication module 20 .

Wherein, the alarm module 10 is used to obtain the channel state information, and according to the channel state information, identify the gestures or movements of people in the coverage area of the wireless communication network, and generate alarm information according to the identification result; the communication module 20 is used to receive the alarm information, And according to the alarm information to alarm.

The alarm module 10 can be installed in the receiving terminal, or can be independently installed outside the receiving terminal. When independently installed outside the receiving terminal, the alarm module 10 can wirelessly communicate with the receiving terminal to receive the channel state information collected by it.

Further, as shown in FIG. 2 , the alarm module 10 may include a data processing unit 11 , a continuous action identification unit 12 , an action segmentation unit 13 and an identification unit 14 . Among them, the data processing unit 11 is used for data preprocessing on the channel state information; the continuous motion recognition unit 12 is used for performing continuous motion recognition on the channel state information data obtained after data preprocessing through the continuous motion recognition algorithm, so as to obtain effective motions data segment; the action segmentation unit 13 is used for segmenting the valid action data segment to obtain a plurality of individual action data; the recognition unit 14 is used for recognizing human gestures or actions according to the individual action data and the deep learning algorithm.

重构信道状态信息数据，从而使信道状态信息数据变化更平滑。In an embodiment of the present invention, the data processing unit 11 is specifically configured to perform noise reduction processing on the channel state information data through a Hampel filter and a Butterworth filter, and use a weighted moving average method to perform noise reduction processing on the channel state information. Data is reconstructed. Due to the influence of factors such as the change of the transmission power and the adaptation selection of the transmission rate during the signal conversion process, it is inevitable that there will be some sudden changes in the sampled channel state information data. According to the principle of Hampel filter, define μ as the median of the sampled data, define σ as the absolute deviation of the median of the sampled data, and define [μ-γ×σ, μ+γ×σ] as the sampling channel The normal range of the state information data, outside this range is the outlier. In addition, since the frequency of human action is usually very low, the high frequency point in the channel state information data is also an abnormality. According to the principle of the Butterworth filter, a cutoff frequency can be set, and the data can be further processed to eliminate the high frequency. High abnormal data. Through the above two data processing methods, although most of the abnormal data have been eliminated, there is still noise in the data that will affect the subsequent processing of the data, so the principle of the weighted moving average method can be used to pass

The channel state information data is reconstructed, so that the change of the channel state information data is smoother.

In an embodiment of the present invention, the continuous action recognition unit 12 is specifically configured to intercept the channel state information data obtained after data preprocessing through a sliding window of a preset window size with a preset step size, so as to obtain a plurality of segmentation matrices, Multiply each segmentation matrix with its own transpose to obtain the corresponding correlation matrix, and calculate the eigenvalues and eigenvectors of each correlation matrix, and judge the changes of the eigenvalues and eigenvectors of multiple correlation matrices, Then, it is determined whether there is a specific action or a large action according to the changing situation, so as to determine the valid action data segment. The preset window size may be 500 packets, and the preset step size may be 400 packets.

其中，t_i ^s、t_i ^e分别为预估的第i个动作的起始点和结束点，且t_i ^s、t_i ^e构成一个数据对。同时设定安全间隔参数T^b，并通过安全间隔参数对预估集合中的每个数据对进行扩展，得到新的集合：

即确定每个动作的起始点和结束点，从而可根据该新的集合得到多个单独动作数据。In one embodiment of the present invention, the action segmentation unit 13 is specifically configured to estimate the start point and end point of each action in the valid action data segment, and obtain the estimated set:

Among them, t _i ^s and t _i ^e are the estimated start point and end point of the i-th action, respectively, and t _i ^s and t _i ^e constitute a data pair. At the same time, the safety interval parameter T ^b is set, and each data pair in the estimated set is expanded by the safety interval parameter to obtain a new set:

That is, the start point and the end point of each action are determined, so that a plurality of individual action data can be obtained according to the new set.

In an embodiment of the present invention, the identification unit 14 stores a sample data set, wherein the sample data representing each person's gestures and movements can be learned in a deep learning algorithm in advance, and the corresponding sample feature vectors can be extracted to obtain The sample data set is established according to the sample feature vector. The identification unit 14 is specifically configured to use a deep learning algorithm to train each individual action data to obtain a corresponding feature vector of the action to be identified, and compare the feature vector of the action to be identified with the sample feature vector in the sample data set to identify Human gestures or human actions.

In an embodiment of the present invention, as shown in FIG. 2 , the alarm module 10 may further include an active alarm unit 15 and a passive alarm unit 16 .

The active alarm unit 15 is configured to recognize the alarm number according to the gesture of the person when the recognition unit recognizes the gesture of the person, and send the alarm number to the communication module 20 . The communication module 20 can dial alarm when receiving the alarm number.

The passive alarm unit 16 is used for further judging the type of the personnel action when the identification unit recognizes the action of the personnel, and generating corresponding alarm information when judging the occurrence of danger according to the type of the personnel action, and sending the alarm information to the communication module 20 . The communication module 20 can send an alarm to the corresponding alarm receiving agency when receiving the alarm information.

In a specific embodiment of the present invention, an automatic alarm device can be applied in a home or public place. The transmitting terminal and the receiving terminal can be placed at two positions on the same horizontal line, respectively. As shown in Figure 3, the transmitting terminal can be an ordinary wireless router, and the receiving terminal can be a smart device equipped with an Intel 5300 wireless network card and a receiving antenna. composition.

As shown in Figure 3, when the personnel in the communication coverage area continuously and repeatedly make the specified digital gesture language, the smart device can identify the continuous number string according to the channel state information, and analyze the operator field, network point field and serial number of the number. field, so as to identify the alarm number, and then connect the communication module to realize the active alarm of gesture dialing.

When the smart device detects a large-scale physical attack in the communication coverage area, such as hitting, falling, running violently, etc., it can determine the case. At this time, the communication module can report the case to the local public security organ through the Internet to achieve Passive alarm. The intelligent device can also determine the nature of the case according to the identified physical attack actions, such as fire, first aid, robbery, etc. The communication module can report to the corresponding person in charge and the responsible department according to the nature of the case.

It should be noted that the receiving terminal and the automatic alarm device may be different or the same intelligent terminal equipment. For the case of the same smart terminal device, for example, when the receiving terminal is a mobile phone, the alarm module 10 can be set in the mobile phone, and the communication module 20 can also be set in the mobile phone, that is, the mobile phone can not only communicate with the wireless communication according to the channel state information Personnel gestures or personnel actions in the network coverage area are recognized, and alarm information is generated according to the recognition result, and an alarm can be issued according to the alarm information.

According to the automatic alarm device based on the wireless channel state information according to the embodiment of the present invention, the alarm module can obtain the channel state information, identify the gestures or movements of the people in the coverage area of the wireless communication network according to the channel state information, and generate a signal according to the identification result. Alarm information, the communication module can alarm according to the alarm information. Therefore, using the channel state information as the physical quantity for action recognition has the advantages of stability, reliability, high precision, and no identification blind spots. It makes reasonable use of the existing wireless communication equipment. Low cost, easy to popularize, and does not require the human body to carry any active equipment, further reducing the cost. Generally speaking, the automatic alarm device of the embodiment of the present invention can realize automatic alarm conveniently and effectively, and the cost is low.

Corresponding to the above embodiments, the present invention also proposes an automatic alarm method based on wireless channel state information.

As shown in FIG. 4 , the automatic alarm method based on wireless channel state information according to an embodiment of the present invention includes the following steps:

S1, acquire channel state information.

通过内嵌的正交频分复用系统，将该信道状态信息数据保存，接下来将以该信道状态信息数据为基础进行动作识别和报警等。In an embodiment of the present invention, the channel state information CSI may be obtained by sampling at a preset sampling frequency by the receiving terminal, and the channel state information data obtained by each sampling may be represented as a matrix. Specifically, the receiving terminal may sample the channel state information data at a sampling frequency of 2000 packets/second, and obtain a matrix for each sampling

Through the built-in orthogonal frequency division multiplexing system, the channel state information data is saved, and then the action recognition and alarm will be performed based on the channel state information data.

S2: Identify the gestures or movements of the personnel in the coverage area of the wireless communication network according to the channel state information, and generate alarm information according to the identification result.

Specifically, data preprocessing can be performed on the channel state information, and continuous motion recognition can be performed on the channel state information data obtained after the data preprocessing through the continuous motion recognition algorithm, so as to obtain a valid action data segment, and the valid action data segment can be processed Segmentation to obtain multiple individual action data, and recognize human gestures or human actions based on the individual action data and deep learning algorithms.

重构信道状态信息数据，从而使信道状态信息数据变化更平滑。In an embodiment of the present invention, a Hampel filter and a Butterworth filter may be used to perform noise reduction processing on the channel state information data, and a weighted moving average method may be used to reconstruct the channel state information data after the noise reduction processing. Due to the influence of factors such as the change of the transmission power and the adaptation selection of the transmission rate during the signal conversion process, it is inevitable that there will be some sudden changes in the sampled channel state information data. According to the principle of Hampel filter, define μ as the median of the sampled data, define σ as the absolute deviation of the median of the sampled data, and define [μ-γ×σ, μ+γ×σ] as the sampling channel The normal range of the status information data, outside this range is the outlier. In addition, since the frequency of human action is usually very low, the high frequency point in the channel state information data is also an abnormality. According to the principle of the Butterworth filter, a cutoff frequency can be set, and the data can be further processed to eliminate the high frequency. High abnormal data. Through the above two data processing methods, although most of the abnormal data have been eliminated, there is still noise in the data that will affect the subsequent processing of the data, so the principle of the weighted moving average method can be used to pass

The channel state information data is reconstructed, so that the change of the channel state information data is smoother.

In an embodiment of the present invention, the continuous action recognition unit 12 is specifically configured to intercept the channel state information data obtained after data preprocessing through a sliding window of a preset window size with a preset step size, so as to obtain a plurality of segmentation matrices, Multiply each segmentation matrix with its own transpose to obtain the corresponding correlation matrix, and calculate the eigenvalues and eigenvectors of each correlation matrix, and judge the changes of the eigenvalues and eigenvectors of multiple correlation matrices, Then, it is determined whether there is a specific action or a large action according to the changing situation, so as to determine the valid action data segment. The preset window size may be 500 packets, and the preset step size may be 400 packets.

其中，t_i ^s、t_i ^e分别为预估的第i个动作的起始点和结束点，且t_i ^s、t_i ^e构成一个数据对。同时设定安全间隔参数T^b，并通过安全间隔参数对预估集合中的每个数据对进行扩展，得到新的集合：

即确定每个动作的起始点和结束点，从而可根据该新的集合得到多个单独动作数据。In one embodiment of the present invention, the starting point and ending point of each action in the valid action data segment can be estimated to obtain the estimated set:

Among them, t _i ^s and t _i ^e are the estimated start point and end point of the i-th action, respectively, and t _i ^s and t _i ^e constitute a data pair. At the same time, the safety interval parameter T ^b is set, and each data pair in the estimated set is expanded by the safety interval parameter to obtain a new set:

That is, the start point and the end point of each action are determined, so that a plurality of individual action data can be obtained according to the new set.

In one embodiment of the present invention, the sample data representing each person's gestures and movements can be learned in a deep learning algorithm in advance, and corresponding sample feature vectors can be extracted to establish a sample data set according to the sample feature vectors and store them. . In recognition, the deep learning algorithm can be used to train each individual action data to obtain the feature vector of the corresponding action to be recognized, and the feature vector of the action to be recognized is compared with the sample feature vector in the sample data set to recognize the human gesture. or human action.

S3, alarm according to the alarm information.

In one embodiment of the present invention, an active alarm can be realized by recognizing a person's gesture, and a passive alarm can also be realized by recognizing a person's action. Specifically, when the recognition unit recognizes the gesture of the person, the alarm number can be recognized according to the gesture of the person, and then the alarm number can be dialed according to the alarm number. When the identification unit recognizes the action of the person, the type of the action of the person is further judged, and when the danger is judged according to the type of the action of the person, an alarm is sent to the corresponding alarm receiving agency.

According to the automatic alarm method based on the wireless channel state information according to the embodiment of the present invention, the channel state information can be obtained, and according to the channel state information, the gestures or movements of the people in the coverage area of the wireless communication network can be recognized, and the alarm information can be generated according to the recognition result. , and can make an alarm according to the alarm information. Therefore, using the channel state information as the physical quantity for action recognition has the advantages of stability, reliability, high precision, and no identification blind area. It makes reasonable use of the existing wireless communication equipment, low cost and easy It is popular, and the human body does not need to carry any active equipment, which further reduces the cost. In general, the automatic alarm method of the embodiment of the present invention can realize automatic alarm conveniently and effectively.

In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", " Back, Left, Right, Vertical, Horizontal, Top, Bottom, Inner, Outer, Clockwise, Counterclockwise, Axial , "radial", "circumferential" and other indicated orientations or positional relationships are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying the indicated device or Elements must have a particular orientation, be constructed and operate in a particular orientation and are therefore not to be construed as limitations of the invention.

In addition, the terms "first" and "second" are only used for descriptive purposes, and should not be construed as indicating or implying relative importance or implying the number of indicated technical features. Thus, a feature defined as "first" or "second" may expressly or implicitly include one or more of that feature. In the description of the present invention, "plurality" means two or more, unless otherwise expressly and specifically defined.

In the present invention, unless otherwise expressly specified and limited, the terms "installed", "connected", "connected", "fixed" and other terms should be understood in a broad sense, for example, it may be a fixed connection or a detachable connection , or integrated; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, and it can be the internal connection of the two elements or the interaction relationship between the two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

In the present invention, unless otherwise expressly specified and limited, a first feature "on" or "under" a second feature may be in direct contact between the first and second features, or the first and second features indirectly through an intermediary touch. Also, the first feature being "above", "over" and "above" the second feature may mean that the first feature is directly above or obliquely above the second feature, or simply means that the first feature is level higher than the second feature. The first feature being "below", "below" and "below" the second feature may mean that the first feature is directly below or obliquely below the second feature, or simply means that the first feature has a lower level than the second feature.

In the description of this specification, description with reference to the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples", etc., mean specific features described in connection with the embodiment or example , structure, material or feature is included in at least one embodiment or example of the present invention. 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, those skilled in the art may combine and combine the different embodiments or examples described in this specification, as well as the features of the different embodiments or examples, without conflicting each other.

Although the embodiments of the present invention have been shown and described above, it should be understood that the above-mentioned embodiments are exemplary and should not be construed as limiting the present invention. Embodiments are subject to variations, modifications, substitutions and variations.

Claims (5)

1. An automatic alarm device based on wireless channel state information, wherein the wireless channel state information is channel state information of communication subcarriers between a transmitting terminal and a receiving terminal in a wireless communication network, the automatic alarm device comprising:

the alarm module is used for acquiring the channel state information, identifying the human gestures or human actions in the coverage area of the wireless communication network according to the channel state information, and generating alarm information according to the identification result;

a communication module for receiving the alarm information and giving an alarm according to the alarm information,

the alarm module includes: the data processing unit is used for carrying out data preprocessing on the channel state information; the continuous action identification unit is used for carrying out continuous action identification on the channel state information data obtained after data preprocessing by using a continuous action identification algorithm so as to obtain an effective action data section; an action division unit, which is used for dividing the effective action data segment to obtain a plurality of single action data; a recognition unit for recognizing the person gesture or the person action according to the individual action data and a deep learning algorithm,

the channel state information is obtained by sampling the channel state information by the receiving terminal at a preset sampling frequency, the channel state information data obtained by each sampling is represented as a matrix, and the data processing unit is specifically configured to: carrying out noise reduction processing on the channel state information data through a Hampel filter and a Butterworth filter; reconstructing the channel state information data after the noise reduction processing by a weighted moving average method,

the continuous motion recognition unit is specifically configured to: intercepting channel state information data obtained after data preprocessing through a sliding window with a preset window size by a preset step length to obtain a plurality of partition matrixes; multiplying each partition matrix by the transpose of the partition matrix to obtain a corresponding correlation matrix; calculating the eigenvalue and the eigenvector of each correlation matrix, and judging the change conditions of the eigenvalue and the eigenvector of the multiple correlation matrices; determining whether there is a specific action or a large action according to the change condition to determine the effective action data segment,

the action segmentation unit is specifically configured to: estimating the starting point and the end point of each action in the effective action data segment to obtain an estimated set:

wherein, t_i ^s、t_i ^eRespectively estimated starting point and ending point of the ith action, and t_i ^s、t_i ^eForming a data pair; setting a safety interval parameter T^bAnd expanding each data pair in the pre-estimated set through the safety interval parameters to obtain a new set:

to derive a plurality of individual action data from the new set.

2. The wireless channel status information-based automatic alarm device according to claim 1, wherein the identification unit stores a sample data set, wherein sample data representing gestures and actions of people are learned in advance in a deep learning algorithm, and corresponding sample feature vectors are extracted to establish the sample data set according to the sample feature vectors, and the identification unit is specifically configured to:

training each individual action data by utilizing a deep learning algorithm to obtain a feature vector of a corresponding action to be recognized;

comparing the feature vector of the action to be recognized with the sample feature vectors in the sample data set to recognize the human gesture or the human action.

3. The wireless channel state information-based automatic warning device according to claim 1 or 2, wherein the warning module further includes:

the active alarm unit is used for identifying an alarm number according to the personnel gesture when the identification unit identifies the personnel gesture, and sending the alarm number to the communication module;

and the passive alarm unit is used for further judging the type of the personnel action when the identification unit identifies the personnel action, generating corresponding alarm information when a danger is judged according to the type of the personnel action, and sending the alarm information to the communication module.

4. The automatic alarm device based on wireless channel status information as claimed in claim 3, wherein the communication module dials an alarm upon receiving the alarm number and alarms to a corresponding alarm receiving mechanism upon receiving the alarm information.

5. An automatic alarm method based on wireless channel state information, wherein the wireless channel state information is channel state information of communication subcarriers between a transmitting terminal and a receiving terminal in a wireless communication network, the automatic alarm method comprising:

acquiring the channel state information;

identifying the personnel gestures or personnel actions in the coverage area of the wireless communication network according to the channel state information, and generating alarm information according to an identification result;

an alarm is given according to the alarm information,

identifying the human gestures or human actions in the coverage area of the wireless communication network according to the channel state information, which specifically comprises the following steps:

performing data preprocessing on the channel state information;

carrying out continuous action recognition on the channel state information data obtained after data preprocessing by using a continuous action recognition algorithm to obtain an effective action data section;

dividing the effective action data segment to obtain a plurality of single action data;

identifying the person gesture or the person action from the individual action data and a deep learning algorithm,

the channel state information is obtained by sampling the channel state information by the receiving terminal at a preset sampling frequency, channel state information data obtained by each sampling is represented as a matrix, and the data preprocessing of the channel state information specifically comprises the following steps:

carrying out noise reduction processing on the channel state information data through a Hampel filter and a Butterworth filter;

reconstructing the channel state information data after the noise reduction processing by a weighted moving average method,

intercepting channel state information data obtained after data preprocessing through a sliding window with the size of a preset window by a preset step length to obtain a plurality of partition matrixes; multiplying each partition matrix by the transpose of the partition matrix to obtain a corresponding correlation matrix; calculating the eigenvalue and the eigenvector of each correlation matrix, and judging the change conditions of the eigenvalue and the eigenvector of the multiple correlation matrices; determining whether there is a specific action or a large action according to the change condition to determine the effective action data segment,

estimating the starting point and the end point of each action in the effective action data segment to obtain an estimated set:

wherein, t_i ^s、t_i ^eRespectively estimated starting point and ending point of the ith action, and t_i ^s、t_i ^eForming a data pair; setting a safety interval parameter T^bAnd expanding each data pair in the pre-estimated set through the safety interval parameters to obtain a new set:

to derive a plurality of individual action data from the new set.

Images