RFID-based human behavior detection method, device, medium and equipment

Abstract

The embodiments of the present application provide a method, apparatus, medium and equipment for detecting human behavior based on RFID. The method includes: receiving signal feature information within a predetermined time length transmitted by the antenna, the signal feature information including the signal receiving strength and signal phase of the signal received by the antenna within the time period corresponding to the predetermined time length; preprocessing the signal feature information to obtain a data set to be identified; inputting the data set to be identified into a pre-trained human behavior recognition model so that the human behavior recognition model outputs the corresponding human behavior recognition result, and the human behavior recognition model is trained based on a probabilistic sparse self-attention mechanism. The technical solution of the embodiment of the present application can reduce the difficulty of equipment deployment, reduce the computational complexity of the model and ensure the accuracy of human behavior recognition.

Description

Technical Field

[0001] The present application relates to the field of computer and communication technology, and more specifically, to a method, apparatus, medium, and device for detecting human behavior based on RFID. Background Art

[0002] Human behavior recognition has a wide range of applications and is a hot research topic in the field of artificial intelligence. It is a foundational technology for numerous applications, including intelligent surveillance, human-computer interaction, and robotics. Current technical solutions generally categorize human behavior recognition devices into two types: contact and non-contact. Contact devices typically include various accelerometers and gesture sensors, while non-contact devices include cameras, RFID, Wi-Fi, LiDAR, and more. RFID can perform both contact and non-contact human behavior recognition. Contact-based human behavior recognition often involves attaching detection tags to the subject's clothing, but this is highly device-dependent. Non-contact methods, on the other hand, place tags in fixed locations and analyze the multipath signals backscattered by the tags to achieve behavior recognition. This requires a large number of antennas and tags for comprehensive data collection to achieve basic recognition accuracy, but increases the difficulty of device deployment and the computational complexity of the model. Summary of the Invention

[0003] The embodiments of the present application provide an RFID-based human behavior detection method, apparatus, medium, and device, which can, at least to a certain extent, reduce the equipment required to be deployed, reduce the difficulty of equipment deployment, reduce the computational complexity of the model, and ensure the accuracy of human behavior recognition.

[0004] Other features and advantages of the present application will become apparent from the following detailed description, or may be learned in part by practice of the present application.

[0005] According to one aspect of an embodiment of the present application, a human behavior detection method based on RFID is provided, which is applied to a terminal device, wherein the terminal device is communicatively connected to a single antenna via a reader, and the antenna is used to receive signals transmitted by a tag array set in an environment;

[0006] The method includes:

[0007] Receiving signal characteristic information transmitted by the antenna within a predetermined time length, the signal characteristic information including a signal receiving strength and a signal phase of the signal received by the antenna within a time period corresponding to the predetermined time length;

[0008] Preprocessing the signal feature information to obtain a data set to be identified;

[0009] The data set to be identified is input into a pre-trained human behavior recognition model so that the human behavior recognition model outputs a corresponding human behavior recognition result. The human behavior recognition model is trained based on a probabilistic sparse self-attention mechanism.

[0010] According to one aspect of an embodiment of the present application, there is provided an RFID-based human behavior detection device, which is applied to a terminal device. The terminal device is communicatively connected to a single antenna via a reader, and the antenna is used to receive signals transmitted by a tag array set in an environment.

[0011] The device comprises:

[0012] a receiving module, configured to receive signal characteristic information transmitted by the antenna within a predetermined time length, the signal characteristic information including a signal receiving strength and a signal phase of the signal received by the antenna within a time period corresponding to the predetermined time length;

[0013] A preprocessing module, used to preprocess the signal feature information to obtain a data set to be identified;

[0014] A processing module is used to input the data set to be identified into a pre-trained human behavior recognition model so that the human behavior recognition model outputs a corresponding human behavior recognition result. The human behavior recognition model is trained based on a probabilistic sparse self-attention mechanism.

[0015] According to one aspect of an embodiment of the present application, a computer-readable medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the RFID-based human behavior detection method as described in the above embodiment is implemented.

[0016] According to one aspect of an embodiment of the present application, an electronic device is provided, comprising: one or more processors; a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the RFID-based human behavior detection method as described in the above embodiments.

[0017] According to one aspect of an embodiment of the present application, a computer program product or computer program is provided. The computer program product or computer program includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the RFID-based human behavior detection method provided in the above-described embodiment.

[0018] In the technical solutions provided in some embodiments of the present application, a terminal device is connected to a single antenna via a reader, and the antenna is used to receive signals transmitted by a tag array set in the environment. The terminal device can receive signal feature information within a predetermined time length transmitted by the antenna, which includes the signal reception strength and signal phase of the signal received by the antenna within a time period corresponding to the predetermined time length, and pre-process the signal feature information to obtain a corresponding data set to be identified. The data to be identified is then input into a pre-trained human behavior recognition model so that the human behavior recognition model can output a corresponding human behavior recognition result, wherein the human behavior recognition model is trained based on a probabilistic sparse self-attention mechanism. As a result, the human behavior recognition model obtained by training based on the probabilistic sparse self-attention mechanism can learn accurate behavior features from a small amount of data set, reducing the computational complexity of the model and ensuring the accuracy of the human behavior recognition results. In addition, only a single antenna and tag array need to be set up, which reduces the difficulty of device deployment.

[0019] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The accompanying drawings are incorporated into and constitute a part of the specification, illustrating embodiments consistent with the present application and, together with the specification, explaining the principles of the present application. Obviously, the drawings described below are only some embodiments of the present application, and those skilled in the art can derive other drawings based on these drawings without inventive effort. In the drawings:

[0021] Figure 1 A schematic diagram of a process of a human behavior detection method based on RFID according to an embodiment of the present application is shown;

[0022] Figure 2 A schematic diagram illustrating a process flow of a human behavior recognition model according to an embodiment of the present application is shown;

[0023] Figure 3 A block diagram of a human behavior detection device based on RFID according to an embodiment of the present application is shown;

[0024] Figure 4 A schematic diagram of the structure of a computer system suitable for implementing an electronic device according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0025] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this application will be thorough and complete and will fully convey the concepts of the example embodiments to those skilled in the art.

[0026] In addition, described feature, structure or characteristic can be combined in one or more embodiments in any suitable manner.In the following description, many specific details are provided so as to provide a full understanding of the embodiments of the present application. However, it will be appreciated by those skilled in the art that the technical scheme of the present application can be put into practice without one or more of the specific details, or other methods, components, devices, steps etc. can be adopted. In other cases, known methods, devices, implementations or operations are not shown or described in detail to avoid blurring the various aspects of the application.

[0027] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically separate entities. That is, these functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0028] The flowcharts shown in the accompanying drawings are for illustrative purposes only and do not necessarily include all contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may be decomposed, while others may be combined or partially combined. Therefore, the actual execution order may vary depending on the actual situation.

[0029] Figure 1 The flowchart of the RFID-based human behavior detection method according to an embodiment of the present application is shown. The RFID-based human behavior detection method is applied to a terminal device. It should be noted that the terminal device may include but is not limited to one or more of a smart phone, a tablet computer, a portable computer, and a desktop computer. The antenna can be connected to a reader, and the reader is connected to the terminal device, whereby the terminal device can communicate with a single antenna through the reader to receive data transmitted by the antenna. The antenna can be placed on one side of the tag array, and the subject stands outside the direct connection line between the antenna and the tag array, that is, the subject will not directly block the signal emitted by the antenna.

[0030] Please refer to Figure 1 The RFID-based human behavior detection method includes at least steps S110 to S130, which are described in detail as follows:

[0031] In step S110, signal characteristic information transmitted by the antenna within a predetermined time length is received, where the signal characteristic information includes a signal receiving strength and a signal phase of the signal received by the antenna within a time period corresponding to the predetermined time length.

[0032] Among them, the predetermined time length can be a time length pre-set by technical personnel in this field based on prior experience. The terminal device can divide the data transmitted by the antenna into several data segments with a time length equal to the predetermined time length for subsequent processing.

[0033] In this embodiment, the activities of the person being measured in the environment will cause changes in the signal reception strength and phase value of the signal transmission in the environment. The antenna can collect this part of the fluctuation information and return it to the reader, which is then returned to the terminal device by the reader. The terminal device can then present the fluctuation information in a numerical form and determine the corresponding signal characteristic information, which includes the signal reception strength and signal phase of the signal received by the antenna during the time period corresponding to the predetermined time length.

[0034] In step S120, the signal feature information is preprocessed to obtain a data set to be identified.

[0035] In this embodiment, after determining the signal feature information, the terminal device may pre-process the signal feature information to obtain a data set to be subsequently input into the human behavior recognition model.

[0036] In one example, the preprocessing includes at least one of data segmentation, data completion, phase correction, fast Fourier transform, and data filtering.

[0037] Specifically, with respect to data segmentation, it is understandable that since the collected signal feature information is a long time series data, the signal feature information can be segmented using a sliding window method so that each sample after segmentation has the same length. These samples can be further divided into n (for example, n is 10) non-overlapping time steps X, with a time interval width of τ, for example, τ is 0.5s, X = {X1,...,Xt,....,Xn}, Xt represents the data collected at time step t for k labels in the label array, which can be further divided into Xt = {xt1,xt2,...,xtk}. Thus, the collected long time series data can be segmented into multiple training samples through data segmentation.

[0038] For data completion, due to the transmission characteristics of radio frequency signals, the signal reception strength and signal phase of each tag in the scene cannot be read identically. Linear interpolation can be used to sample the signal feature information to the same dimension for subsequent reading.

[0039] For phase correction, the antenna can use a frequency hopping mechanism for data collection, obtaining richer multipath information. However, due to the introduction of frequency hopping technology, the phase will fluctuate significantly even in an inactive environment. To this end, the signal phase can be unwrapped based on the Unwrap algorithm, and phase correction can be performed based on the following formula:

[0040] θ(t)=θ j (t)-θ j +θ d

[0041] Among them, θ j (t) represents the time t with frequency f j The measured phase, θ j and θ d Respectively represent the frequency f within the latest predetermined time interval j and the default frequency f d The phase median value measured under the above phase correction method can greatly reduce the problem of large phase fluctuations and make the phase data smoother.

[0042] Regarding fast Fourier transform, the original RFID signal is a mixture of various objects with high noise, which cannot be directly understood and used. The fast Fourier transform can convert time domain data into frequency domain data, which is beneficial for the subsequent distinction between the activity characteristics of the object and multipath noise.

[0043] For data filtering, in order to further eliminate the noise of the collected original RFID signal, a filter can be used to filter it, such as Gauss filtering, Kalman filtering, mean filtering, median filtering, Hampel filtering, etc. Preferably, Kalman filtering can be used to reduce the noise of the original RFID signal.

[0044] In step S130, the data set to be identified is input into a pre-trained human behavior recognition model so that the human behavior recognition model outputs a corresponding human behavior recognition result. The human behavior recognition model is trained based on a probabilistic sparse self-attention mechanism.

[0045] In this embodiment, the terminal device can input the data set to be identified after preprocessing into the human behavior recognition model, and the human behavior recognition model can be pre-trained by a person skilled in the art. Thus, based on the data set to be identified, the human behavior recognition model can output the corresponding human behavior recognition result. It is worth noting that the human behavior recognition model is trained based on the probabilistic sparse self-attention mechanism. Therefore, even in the case of a small data set, the human behavior recognition model can learn accurate behavioral features, reducing the computational complexity of the model while also ensuring the accuracy of the human behavior recognition results.

[0046] In one embodiment of the present application, the human behavior recognition model includes a separate convolution layer, a data flattening and merging layer, a variant layer, and a linear variation output layer.

[0047] Specifically, a separate convolution layer can perform alternating dilated convolution and traditional convolution operations on the k data-preprocessed labels at timestamp t to extract and fuse activity feature information on different frequency bands in the frequency domain. The data flattening and merging layer can flatten the spatial features of the k labels (i.e., the aforementioned activity feature information) into one-dimensional data, and then further merge them, and then extract the correlation in their spatial features through convolution operations. The variant layer is based on the encoder module of the transformer, and the encoder module adopts a probabilistic sparse self-attention mechanism for learning. Compared with other attention mechanisms, the probabilistic sparse self-attention mechanism can reduce the computational complexity of the native transformer, focus the attention on the top-n timestamps selected by the model, and finally complete the probabilistic output of the behavior prediction. The output of the variant layer is O = {o1, o2, ..., ot}, and the linear change output layer y^c = OA T +b can map O to y^c (where C is the set of all classes), and then y^c is normalized by softmax to obtain the final behavior prediction.

[0048] Please refer to Figure 2 In actual use, the data set to be processed after being processed by the data preprocessing module is arranged in time sequence and input into the human behavior recognition model. It should be noted that since the operations performed in each time step are the same, only the operations in time step t will be described later. Figure 2 As shown in , the pre-processed data collected by k labels in time step t is first input into a single convolution layer for feature extraction operation, where the single convolution layer consists of a hole convolution layer and two two-dimensional convolution layers. The hole convolution is first used to fuse the virtual and real features of the data after FFT change in the frequency domain, and then the fused data is further input into two two-dimensional convolution layers for high-dimensional feature extraction.

[0049] Then the features of the processed k labels are Input data flattening and merging layer, flatten these k label features into k one-dimensional data Then merge them, that is, arrange these k one-dimensional data in rows according to the position of the corresponding label in the label array (from left to right, from top to bottom), and arrange them into k rows as follows Figure 2 Medium V t Arranged in V tThe form can facilitate the subsequent extraction of spatial characteristics between k labels through convolution, and the use of hole convolution in a single convolution layer has already extracted the information of different frequency bands of each individual label in the environment in the frequency domain, so V t The convolution of column data actually fuses the spatiotemporal features collected by different labels in the environment. Then, multiple two-dimensional convolution layers are used to establish the spatial feature connections among k labels to obtain v′ t , the output {v′1,v′2,…,v′ n} Arrange the input in time sequence and input it into the variant layer based on the encoder module of the transformer.

[0050] The variant layer retains the basic framework of the transformer's encoder module and employs sparse ProbSparse self-attention, enabling the model to learn accurate behavioral features from a small dataset and predict input time series data. The final results are then fed into the linear variant output layer for dimensionality compression. A softmax function is then used to output the maximum predicted value, resulting in the final behavior prediction result, or human behavior recognition result.

[0051] In addition, cross-entropy can be used as an estimate of the loss function, the Adam algorithm can be used as the optimizer in the model iteration, and the weighted F1 score can be used as the evaluation indicator of the model.

[0052] The following describes an embodiment of the device of the present application, which can be used to implement the RFID-based human behavior detection method in the above-mentioned embodiment of the present application. For details not disclosed in the embodiment of the device of the present application, please refer to the embodiment of the RFID-based human behavior detection method in the above-mentioned embodiment of the present application.

[0053] Figure 3 A block diagram of an RFID-based human behavior detection device according to an embodiment of the present application is shown.

[0054] Reference Figure 3 As shown, according to an embodiment of the present application, an RFID-based human behavior detection device is applied to a terminal device, wherein the terminal device is communicatively connected to a single antenna via a reader, and the antenna is used to receive signals transmitted by a tag array set in an environment;

[0055] The device includes:

[0056] a receiving module 310 configured to receive signal characteristic information transmitted by the antenna within a predetermined time length, the signal characteristic information including a signal receiving strength and a signal phase of the signal received by the antenna within a time period corresponding to the predetermined time length;

[0057] A preprocessing module 320 is used to preprocess the signal feature information to obtain a data set to be identified;

[0058] The processing module 330 is used to input the data set to be identified into a pre-trained human behavior recognition model so that the human behavior recognition model outputs a corresponding human behavior recognition result. The human behavior recognition model is trained based on a probabilistic sparse self-attention mechanism.

[0059] In one embodiment of the present application, the human behavior recognition model includes a separate convolutional layer, a data flattening and merging layer, a variant layer, and a linear change output layer, wherein the variant layer is based on the encoder module of the transformer, and the encoder module adopts a probabilistic sparse self-attention mechanism for learning.

[0060] In one embodiment of the present application, the data flattening and merging layer flattens the multiple label features obtained by the separate convolutional layer into one-dimensional data, and merges the multiple label features according to the positions of the labels corresponding to each label feature in the label array.

[0061] In one embodiment of the present application, the preprocessing includes at least one of data segmentation, data completion, phase correction, fast Fourier transform, and data filtering.

[0062] In one embodiment of the present application, performing phase correction includes:

[0063] The signal phase is unwrapped based on the Unwrap algorithm, and phase correction is performed based on the following formula:

[0064] θ(t)=θ j (t)-θ j +θ d

[0065] Among them, θ j (t) represents the time t with frequency f j The measured phase, θ j and θ d Respectively represent the frequency f within the latest predetermined time interval j and the default frequency f d The median phase value measured below.

[0066] Figure 4 A schematic diagram of the structure of a computer system suitable for implementing an electronic device according to an embodiment of the present application is shown.

[0067] It should be noted that Figure 4The computer system of the electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0068] like Figure 4 As shown, the computer system includes a central processing unit (CPU) 401, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 402 or the program loaded from the storage part 408 into the random access memory (RAM) 403, such as executing the method described in the above embodiment. Various programs and data required for system operation are also stored in the RAM 403. The CPU 401, ROM 402 and RAM 403 are connected to each other via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0069] The following components are connected to the I / O interface 405: an input section 406 including a keyboard, a mouse, and the like; an output section 407 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and a speaker; a storage section 408 including a hard disk and the like; and a communication section 409 including a network interface card such as a LAN (Local Area Network) card or a modem. The communication section 409 performs communication processing via a network such as the Internet. A drive 410 is also connected to the I / O interface 405 as needed. Removable media 411, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 410 as needed, so that computer programs read therefrom can be installed into the storage section 408 as needed.

[0070] In particular, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 409, and / or installed from a removable medium 411. When the computer program is executed by the central processing unit (CPU) 401, the various functions defined in the system of the present application are executed.

[0071] It should be noted that the computer-readable medium shown in the embodiments of the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or component, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries a computer-readable computer program. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. A computer program embodied on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, or any suitable combination thereof.

[0072] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. Among them, each box in the flowchart or block diagram can represent a module, program segment, or part of the code, and the above-mentioned module, program segment, or part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0073] The units involved in the embodiments described in this application may be implemented by software or hardware, and the units described may also be set in a processor. In some cases, the names of these units do not constitute limitations on the units themselves.

[0074] As another aspect, the present application further provides a computer-readable medium, which may be included in the electronic device described in the above embodiments, or may exist independently without being incorporated into the electronic device. The computer-readable medium carries one or more programs, and when the one or more programs are executed by the electronic device, the electronic device implements the method described in the above embodiments.

[0075] It should be noted that, although several modules or units of the device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiment of the application, the features and functions of two or more modules or units described above can be concretized in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into multiple modules or units to be concretized.

[0076] Through the description of the above embodiments, it is easy for those skilled in the art to understand that the example embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solution according to the embodiments of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, a touch terminal, or a network device, etc.) to execute the method according to the embodiments of the present application.

[0077] Those skilled in the art will readily conceive of other embodiments of the present application after considering the specification and practicing the embodiments disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present application that follow the general principles of this application and include common knowledge or customary techniques in the art that are not disclosed herein.

[0078] It should be understood that the present application is not limited to the exact structures described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A human behavior detection method based on RFID, characterized in that: Applied to a terminal device, the terminal device is connected to a single antenna via a reader, and the antenna is used to receive signals transmitted by an array of tags set in the environment; The method comprises: Receiving signal characteristic information transmitted by the antenna within a predetermined time length, the signal characteristic information including a signal receiving strength and a signal phase of the signal received by the antenna within a time period corresponding to the predetermined time length; Preprocessing the signal feature information to obtain a data set to be identified; Inputting the to-be-recognized data set into a pre-trained human behavior recognition model so that the human behavior recognition model outputs a corresponding human behavior recognition result, wherein the human behavior recognition model is trained based on a probabilistic sparse self-attention mechanism; The human behavior recognition model includes a single convolutional layer, a data flattening and merging layer, a variant layer, and a linear variation output layer. The variant layer is based on the encoder module of the transformer, and the encoder module adopts a probabilistic sparse self-attention mechanism for learning. The data flattening and merging layer flattens the label features obtained by the individual convolutional layers into one-dimensional data, and merges the label features according to the positions of the labels corresponding to the respective label features in the label array; Wherein, the preprocessing includes at least one of data segmentation, data completion, phase correction, fast Fourier transform and data filtering; Among them, the data set to be identified is arranged in time sequence and input into the human behavior recognition model. Since the operations performed in each time step are the same, only the operations in time step t will be described later; the preprocessed data collected by k labels in time step t are first input into a separate convolution layer for feature extraction operation, where the separate convolution layer consists of a void convolution layer and two two-dimensional convolution layers. The void convolution is first used to fuse the virtual and real features of the data after FFT change in the frequency domain, and then the fused data is further input into two two-dimensional convolution layers for high-dimensional feature extraction. Then the features of the processed k labels are Input data flattening and merging layer, flatten these k label features into k one-dimensional data Then merge them, that is, arrange these k one-dimensional data into rows according to the position of the corresponding label in the label array, and arrange them into k rows of V t ; Arranged in a V t The form can facilitate the subsequent extraction of spatial characteristics between k labels through convolution, and the use of hole convolution in a single convolution layer has already extracted the information of different frequency bands of each individual label in the environment in the frequency domain, so V t The convolution of column data actually fuses the spatiotemporal features collected by different labels in the environment; then, multiple two-dimensional convolution layers are used to establish the spatial feature connections among k labels to obtain v′ t , the output obtained in the remaining time steps {v′1,v′2,…,v′ n The data are arranged in time sequence and input into the variant layer based on the encoder module of the transformer. In the variant layer, the basic framework of the encoder module of the transformer is retained, and sparse ProbSparse self-attention is used to enable the model to learn accurate behavioral features on a small amount of data set and predict the input time series data. The final result is then input into the linear variation output layer for dimensionality compression, and the maximum prediction value is output through the softmax function to obtain the final behavior prediction result, that is, the human behavior recognition result.

2. The method according to claim 1, characterized in that Performing phase correction includes: The signal phase is unwrapped based on the Unwrap algorithm, and phase correction is performed based on the following formula: θ(t)=θ j (t)-θ j +θ d Among them, θ j (t) represents the time t with frequency f j The measured phase, θ j and θ d Respectively represent the frequency f within the latest predetermined time interval j and the default frequency f d The median phase value measured below.

3. A human behavior detection device based on RFID, characterized in that: Applied to a terminal device, the terminal device is connected to a single antenna via a reader, and the antenna is used to receive signals transmitted by an array of tags set in the environment; The device comprises: a receiving module, configured to receive signal characteristic information transmitted by the antenna within a predetermined time length, the signal characteristic information including a signal receiving strength and a signal phase of the signal received by the antenna within a time period corresponding to the predetermined time length; A preprocessing module, used to preprocess the signal feature information to obtain a data set to be identified; A processing module is used to input the data set to be recognized into a pre-trained human behavior recognition model so that the human behavior recognition model outputs a corresponding human behavior recognition result, and the human behavior recognition model is trained based on a probabilistic sparse self-attention mechanism; The human behavior recognition model includes a single convolutional layer, a data flattening and merging layer, a variant layer, and a linear variation output layer. The variant layer is based on the encoder module of the transformer, and the encoder module adopts a probabilistic sparse self-attention mechanism for learning. The data flattening and merging layer flattens the label features obtained by the individual convolutional layers into one-dimensional data, and merges the label features according to the positions of the labels corresponding to the respective label features in the label array; Wherein, the preprocessing includes at least one of data segmentation, data completion, phase correction, fast Fourier transform and data filtering; Among them, the data set to be identified is arranged in time sequence and input into the human behavior recognition model. Since the operations performed in each time step are the same, only the operations in time step t will be described later; the preprocessed data collected by k labels in time step t are first input into a separate convolution layer for feature extraction operation, where the separate convolution layer consists of a void convolution layer and two two-dimensional convolution layers. The void convolution is first used to fuse the virtual and real features of the data after FFT change in the frequency domain, and then the fused data is further input into two two-dimensional convolution layers for high-dimensional feature extraction. Then the features of the processed k labels are Input data flattening and merging layer, flatten these k label features into k one-dimensional data Then merge them, that is, arrange these k one-dimensional data into rows according to the position of the corresponding label in the label array, and arrange them into k rows of V t ; Arranged in a V t The form can facilitate the subsequent extraction of spatial characteristics between k labels through convolution, and the use of hole convolution in a single convolution layer has already extracted the information of different frequency bands of each individual label in the environment in the frequency domain, so V t The convolution of column data actually fuses the spatiotemporal features collected by different labels in the environment; then, multiple two-dimensional convolution layers are used to establish the spatial feature connections among k labels to obtain v′ t , the output obtained in the remaining time steps {v′1,v′2,…,v′ n The data are arranged in time sequence and input into the variant layer based on the encoder module of the transformer. In the variant layer, the basic framework of the encoder module of the transformer is retained, and sparse ProbSparse self-attention is used to enable the model to learn accurate behavioral features on a small amount of data set and predict the input time series data. The final result is then input into the linear variation output layer for dimensionality compression, and the maximum prediction value is output through the softmax function to obtain the final behavior prediction result, that is, the human behavior recognition result.

4. The device according to claim 3, characterized in that The human behavior recognition model includes a single convolutional layer, a data flattening and merging layer, a variant layer, and a linear variation output layer, wherein the variant layer is based on the encoder module of the transformer, and the encoder module adopts a probabilistic sparse self-attention mechanism for learning.

5. The device according to claim 4, characterized in that The data flattening and merging layer flattens the multiple label features obtained by the individual convolutional layers into one-dimensional data, and merges the multiple label features according to the positions of the labels corresponding to the respective label features in the label array.

6. A computer-readable medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the RFID-based human behavior detection method according to any one of claims 1 to 2 is implemented.

7. An electronic device, characterized in that: include: one or more processors; A storage device for storing one or more programs, which, when executed by the one or more processors, enables the one or more processors to implement the RFID-based human behavior detection method as described in any one of claims 1 to 2.

Images