Human body gesture recognition method and device based on double-array RFID

Abstract

The invention discloses a human body gesture recognition method and device based on double-array RFID, and the method comprises the steps: receiving first signal data corresponding to a first antenna and second signal data corresponding to a second antenna, which are analyzed by a reader; preprocessing the first signal data and the second signal data to obtain corresponding first time sequence data and second time sequence data; inputting the first time sequence data and the second time sequence data into a pre-trained gesture recognition model, so that the gesture recognition model outputs a corresponding human body gesture recognition result; therefore, gestures can be detected from multiple angles, the influence caused by the external environment can be reduced, in addition, a user does not need to wear an RFID tag or an antenna, and the experience of the user is greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of gesture recognition, and in particular to a human gesture recognition method based on dual-array RFID, a computer-readable storage medium, a computer device, and a human gesture recognition device based on dual-array RFID. Background Art

[0002] In the related technology, radio frequency identification (RFID) technology can realize human behavior recognition due to its contactless, low power consumption and high sensitivity. An RFID recognition system can be formed by using an antenna and a reader. In addition, the RFID tag needs to be attached to the human skin. As the human body gestures, the RFID tag moves and is detected by the antenna and the reader, so that some relatively simple gestures can be recognized, such as sliding up and down or left and right, writing letters in the air, etc. However, these gestures are basically completed in a plane in space, and there are very few gestures in all directions such as up, down, left, right, front and back at the same time. Using only one antenna to detect the user's posture is very limited. It can not recognize the above-mentioned complex gestures in multiple directions, and it is also easily interfered by the external environment. Attaching the tag to the human skin will also bring inconvenience to the user. Summary of the invention

[0003] The present invention aims to solve at least one of the technical problems in the above-mentioned technologies to a certain extent. To this end, one object of the present invention is to propose a human gesture recognition method based on dual-array RFID, which can detect gestures from multiple angles and reduce the influence of the external environment. In addition, the user does not need to wear an RFID tag or antenna, which greatly improves the user experience.

[0004] A second object of the present invention is to provide a computer-readable storage medium.

[0005] A third object of the present invention is to provide a computer device.

[0006] A fourth objective of the present invention is to provide a human gesture recognition device based on dual-array RFID.

[0007] To achieve the above-mentioned purpose, an embodiment of the first aspect of the present invention proposes a human gesture recognition method based on dual-array RFID, which is applied to a terminal device, wherein the terminal device is connected to a first antenna and a second antenna through a reader, wherein the first antenna is arranged relative to a first tag array, and the second antenna is arranged relative to a second tag array, and the first tag array and the second tag array are perpendicular to each other and face a gesture recognition environment, and the method comprises: receiving first signal data corresponding to the first antenna and second signal data corresponding to the second antenna analyzed by the reader, wherein the first signal data comprises a signal reception strength and a signal phase collected by each tag in the first tag array at each sampling moment, and the second signal data comprises a signal reception strength and a signal phase collected by each tag in the second tag array at each sampling moment; preprocessing the first signal data and the second signal data to obtain corresponding first time series data and second time series data; and inputting the first time series data and the second time series data into a pre-trained gesture recognition model so that the gesture recognition model outputs a corresponding human gesture recognition result.

[0008] According to the human gesture recognition method based on dual-array RFID in an embodiment of the present invention, two orthogonally placed tag arrays are firstly used to form a gesture recognition environment, so that when the human body makes various gestures in the gesture recognition environment, they are received by the antenna arranged opposite to the tag array, and then the received signal is analyzed and recognized; thereby, gestures can be detected from multiple angles, and the influence of the external environment can be reduced. In addition, the user does not need to carry an RFID tag or antenna with him / her, which greatly improves the user experience.

[0009] In addition, the human gesture recognition method based on dual-array RFID proposed in the above embodiment of the present invention may also have the following additional technical features:

[0010] Optionally, the gesture recognition model includes a first self-attention module, a second self-attention module, a third self-attention module and a multi-layer perceptron.

[0011] Optionally, the first time series data and the second time series data are input into a pre-trained gesture recognition model so that the gesture recognition model outputs a corresponding human gesture recognition result, including: inputting the first time series data into the first self-attention module to obtain a first high-dimensional feature representation corresponding to a first direction; inputting the second time series data into the second self-attention module to obtain a second high-dimensional feature representation corresponding to a second direction perpendicular to the first direction; splicing the first high-dimensional feature representation and the second high-dimensional feature representation, and inputting them into the third self-attention module to extract the features of the overall gesture to obtain the overall high-dimensional features; inputting the overall high-dimensional features into the multi-layer perceptron for feature concentration to obtain the probability of the overall high-dimensional features corresponding to each gesture category in all gesture categories.

[0012] Optionally, the preprocessing includes deconvolution, upsampling, filtering and smoothing, normalization and construction processing.

[0013] To achieve the above-mentioned objectives, a second aspect of the present invention proposes a computer-readable storage medium on which a human gesture recognition program based on a dual-array RFID is stored. When the human gesture recognition program based on a dual-array RFID is executed by a processor, the human gesture recognition method based on a dual-array RFID as described above is implemented.

[0014] According to the computer-readable storage medium of the embodiment of the present invention, by storing the human gesture recognition program based on the dual-array RFID, the processor implements the human gesture recognition method based on the dual-array RFID as described above when executing the human gesture recognition program based on the dual-array RFID. As a result, gestures can be detected from multiple angles, and the influence of the external environment can be reduced. In addition, the user does not need to carry an RFID tag or antenna with him / her, which greatly improves the user experience.

[0015] To achieve the above objectives, the third aspect of the present invention proposes a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the human gesture recognition method based on dual-array RFID as described above is implemented.

[0016] According to the computer device of the embodiment of the present invention, the human gesture recognition program based on the dual-array RFID is stored in the memory, so that when the processor executes the human gesture recognition program based on the dual-array RFID, the human gesture recognition method based on the dual-array RFID as described above is implemented. As a result, gestures can be detected from multiple angles, and the influence of the external environment can be reduced. In addition, the user does not need to carry an RFID tag or antenna with him / her, which greatly improves the user experience.

[0017] To achieve the above-mentioned purpose, the fourth aspect of the present invention proposes a human gesture recognition device based on dual-array RFID, which is applied to a terminal device, wherein the terminal device is connected to a first antenna and a second antenna through a reader, wherein the first antenna is arranged relative to a first tag array, and the second antenna is arranged relative to a second tag array, and the first tag array and the second tag array are perpendicular to each other and face a gesture recognition environment, and the device includes: a receiving module, which is used to receive first signal data corresponding to the first antenna and second signal data corresponding to the second antenna analyzed by the reader, wherein the first signal data includes a signal receiving strength and a signal phase collected by each tag in the first tag array at each sampling time, and the second signal data includes a signal receiving strength and a signal phase collected by each tag in the second tag array at each sampling time; a preprocessing module, which is used to preprocess the first signal data and the second signal data to obtain corresponding first time series data and second time series data; and a gesture recognition module, which is used to input the first time series data and the second time series data into a pre-trained gesture recognition model so that the gesture recognition model outputs a corresponding human gesture recognition result.

[0018] According to an embodiment of the present invention, a human gesture recognition device based on a dual-array RFID is applied to a terminal device, wherein the terminal device is connected to a first antenna and a second antenna through a reader, wherein the first antenna is arranged relative to a first tag array, and the second antenna is arranged relative to a second tag array, wherein the first tag array and the second tag array are perpendicular to each other and face a gesture recognition environment, wherein first signal data corresponding to the first antenna and second signal data corresponding to the second antenna analyzed by the reader are received through a receiving module, and the first signal data and the second signal data are preprocessed by a preprocessing module to obtain corresponding first time series data and second time series data, and the first time series data and the second time series data are input into a pre-trained gesture recognition model through a gesture recognition module, so that the gesture recognition model outputs corresponding human gesture recognition results; thereby, gestures can be detected from multiple angles, and the influence of the external environment can be reduced, and in addition, users do not need to carry RFID tags or antennas with them, which greatly improves the user experience.

[0019] In addition, the human gesture recognition device based on dual-array RFID according to the above embodiment of the present invention may also have the following additional technical features:

[0020] Optionally, the gesture recognition model includes a first self-attention module, a second self-attention module, a third self-attention module and a multi-layer perceptron.

[0021] Optionally, the gesture recognition module is also used to input the first time series data into the first self-attention module to obtain a first high-dimensional feature representation corresponding to a first direction; input the second time series data into the second self-attention module to obtain a second high-dimensional feature representation corresponding to a second direction perpendicular to the first direction; splice the first high-dimensional feature representation and the second high-dimensional feature representation, and input them into the third self-attention module to extract the features of the overall gesture to obtain the overall high-dimensional features; input the overall high-dimensional features into the multi-layer perceptron for feature concentration to obtain the probability of the overall high-dimensional features corresponding to each gesture category in all gesture categories.

[0022] Optionally, the preprocessing includes deconvolution, upsampling, filtering and smoothing, normalization and construction processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 A schematic diagram of a flow chart of a human gesture recognition method based on dual-array RFID according to an embodiment of the present invention;

[0024] Figure 2 is a structural schematic diagram of a gesture recognition system according to an embodiment of the present invention;

[0025] Figure 3 A schematic diagram of a flow chart of a recognition process of a gesture recognition model according to an embodiment of the present invention;

[0026] Figure 4 4 is a block diagram of a human gesture recognition device based on dual-array RFID according to an embodiment of the present invention. DETAILED DESCRIPTION

[0027] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0028] In order to better understand the above technical solution, exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present invention and to fully convey the scope of the present invention to those skilled in the art.

[0029] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0030] Figure 1 : is a flow chart of a human gesture recognition method based on dual-array RFID according to an embodiment of the present invention, the human gesture recognition method based on dual-array RFID is applied to a terminal device, the terminal device is connected to a first antenna and a second antenna through a reader, the first antenna is arranged relative to a first tag array, the second antenna is arranged relative to a second tag array, the first tag array and the second tag array are perpendicular to each other and face a gesture recognition environment; Figure 1 As shown, the method comprises the following steps:

[0031] S101, receiving first signal data corresponding to the first antenna and second signal data corresponding to the second antenna analyzed by the reader, wherein the first signal data includes the signal reception strength and signal phase collected by each tag in the first tag array at each sampling moment, and the second signal data includes the signal reception strength and signal phase collected by each tag in the second tag array at each sampling moment.

[0032] That is to say, since the user's activities in the gesture recognition environment will cause the signal reception strength and phase value of the signal transmission in the gesture recognition environment to change, the antenna can collect this part of the fluctuating RF signal over a period of time at a fixed frequency and return it to the reader. The reader then parses the RF signal, extracts the corresponding signal reception strength and phase value from it, and returns it to the terminal device.

[0033] It should be noted that in order to be able to recognize complex gestures, while not imposing significant restrictions on human body movements, it can also recognize the horizontal and vertical movements of the hands. The present application uses two orthogonal tag arrays, each of which is placed in front of the antenna. The human body makes various gestures in front of the tag array. By combining the two types of information read by the two antennas respectively over a period of time, a comprehensive judgment can be made for the various gestures made by the user during the time period. In this way, gestures can be detected from multiple angles and the impact of the external environment can be reduced. In addition, users do not need to wear RFID tags or antennas, which greatly improves the user experience.

[0034] As a specific example, Figure 2As shown, the gesture recognition system includes two tag arrays, and the two tag arrays have the same structure, which are composed of nine H47-U tags arranged in a 3×3 grid, two RFID antennas, an Impinj R420 RFID reader and a computer for collecting and calculating data; the tag arrays are posted on acrylic stand boards, placed perpendicular to the desktop, and the two acrylic stand boards where the two tag arrays are located are also placed perpendicular to each other. The user makes gestures in the space formed by the two tag arrays. The RFID antenna detects the corresponding signal and transmits it to the RFID reader. The RFID reader parses the signal and transmits it to the connected computer and sends it to the gesture recognition model.

[0035] S102, preprocessing the first signal data and the second signal data to obtain corresponding first time series data and second time series data.

[0036] That is to say, the signal receiving strength and signal phase of all tags collected at each sampling moment (sampling point) during the gesture process are preprocessed.

[0037] As an embodiment, the preprocessing includes deconvolution, upsampling, filtering and smoothing, normalization and construction processing.

[0038] Specifically, the phase data is first deconvolved to restore the phase change, then the data is upsampled to increase the number of data samples, and then the Savitzky-Golay filter is used to reduce noise and smooth the data. Finally, the data is normalized and constructed into a suitable data structure for use in subsequent processes.

[0039] S103: input the first time series data and the second time series data into a pre-trained gesture recognition model, so that the gesture recognition model outputs a corresponding human gesture recognition result.

[0040] As an embodiment, the gesture recognition model includes a first self-attention module, a second self-attention module, a third self-attention module and a multi-layer perceptron.

[0041] As an embodiment, the first time series data and the second time series data are input into a pre-trained gesture recognition model so that the gesture recognition model outputs a corresponding human gesture recognition result, including: inputting the first time series data into a first self-attention module to obtain a first high-dimensional feature representation corresponding to a first direction; inputting the second time series data into a second self-attention module to obtain a second high-dimensional feature representation corresponding to a second direction perpendicular to the first direction; splicing the first high-dimensional feature representation and the second high-dimensional feature representation, and inputting them into a third self-attention module to extract features of the overall gesture to obtain an overall high-dimensional feature; inputting the overall high-dimensional feature into a multi-layer perceptron for feature concentration to obtain the probability of the overall high-dimensional feature corresponding to each gesture category in all gesture categories.

[0042] That is to say, after the data is preprocessed, the data corresponding to the two labels are sent to the self-attention module with position labels to extract the high-dimensional feature representation of gestures in different directions. Then, the two high-dimensional features are spliced ​​together and sent to the next self-attention module to achieve the feature extraction of the overall gesture. After that, the feature output of each moment in the data is concentrated using a multi-layer perceptron, and finally a sequence including the action features of all moments is formed, which is sent to another multi-layer perceptron for gesture classification to obtain the final result.

[0043] Specifically, Figure 3 As shown, the sampled data C after preprocessing has the following structure:

[0044] C=[c1,c2]

[0045] Among them, c i , i=1,2 specifically:

[0046] c i =[c i1 ,…,c ij ]

[0047] c ij =[rssi ij1 ,rssi ij2 ,…,rssi ij9 ,phase ij1 ,phase ij2 ,…,phase ij9 ]

[0048] Where 1≤j≤m, m is the number of samples after sampling, c1 and c2 are the time series data corresponding to the first label array and the second label array, which store the signal receiving strength and signal phase data of the 9 labels at each sample point after preprocessing. iIt is actually a time series, in which the data corresponding to each moment is an 18-dimensional vector, which represents the gesture information observed by each tag at that moment.

[0049] After the data is preprocessed, the data corresponding to the two label arrays are sent to the positional self-attention module with positional labels to extract high-dimensional feature representations of gestures in different directions. The two label arrays record the horizontal and vertical motion features of 3D gestures, as shown below:

[0050] Taking c1 as an example, after feature extraction, we get the high-level feature representation f1=F f1 (c1), f1=[f 11 ,…,f 1j ], where 1≤j≤m,f ij =[f i,j,1 ,f i,j,2 ,…,f i,j,18 ]

[0051]

[0052] Among them, QKV are pre-trained model parameters, d k is the vector dimension.

[0053] After the same process as above, f2 can be obtained from c2. Then, the two high-dimensional features are spliced ​​together to obtain the overall spatial motion high-level feature representation of the 3D gesture:

[0054] f=[[f 1,1,1 ,f 1,1,2 ,…,f 1,1,18 ,f 2,1,1 ,f 2,1,2 ,…,f 2,1,18 ],…,[f 1,j,1 ,f 1,j,2 ,…,f 1,i,18 ,f 2,j,1 ,f 2,h,2 ,…,f 2,j,18 ]], where 1≤j≤m. The whole process can be expressed as f=F f (c1,c2).

[0055] Next, it is sent to the next self-attention module to realize the feature extraction of the overall gesture. The overall feature extraction is consistent with the feature extraction process of a single label array, and the overall high-dimensional feature f is obtained. t =[[f 1,1 ,f 1,2 ,…,f 1,36 ],...[f j,1 ,f j,2,…,f j,36 ]], where 1≤j≤m.

[0056] After that, a multi-layer perceptron is used to condense the feature output at each moment in the data.

[0057] f T =MLP(f t )

[0058] Get f T =[p1,p2,…p n ], n is the number of gesture types, p n represents the probability that the current gesture is gesture n. So far, the predicted probability of each gesture is obtained.

[0059] It should be noted that although the currently widely used numerical modeling methods can realize gesture recognition, this modeling method is very dependent on the experience and attempts of the modeler, and the established model is highly specialized and can only cope with simple and less changing scenes; the gesture recognition model of this application is different from the traditional numerical analysis model, but uses a deep learning model combined with a self-attention mechanism; although many methods currently use the Transformer model for related research, the Transformer model has a large parameter scale, complex training, and slow processing speed. For this reason, this application extracts the core self-attention mechanism in the Transformer, which can not only reduce the defect of large parameter scale, but also make full use of its advantages.

[0060] In summary, according to the human gesture recognition method based on dual-array RFID in an embodiment of the present invention, two orthogonally placed tag arrays are first used to form a gesture recognition environment, so that when the human body makes various gestures in the gesture recognition environment, they are received by the antenna arranged opposite to the tag array, and then the received signal is analyzed and recognized; thereby, gestures can be detected from multiple angles, and the influence of the external environment can be reduced. In addition, the user does not need to wear an RFID tag or antenna, which greatly improves the user experience.

[0061] In order to implement the above embodiment, the embodiment of the present invention further proposes a computer-readable storage medium, on which a human gesture recognition program based on a dual-array RFID is stored. When the human gesture recognition program based on a dual-array RFID is executed by a processor, the human gesture recognition method based on a dual-array RFID as described above is implemented.

[0062] According to the computer-readable storage medium of the embodiment of the present invention, by storing the human gesture recognition program based on the dual-array RFID, the processor implements the human gesture recognition method based on the dual-array RFID as described above when executing the human gesture recognition program based on the dual-array RFID. As a result, gestures can be detected from multiple angles, and the influence of the external environment can be reduced. In addition, the user does not need to carry an RFID tag or antenna with him / her, which greatly improves the user experience.

[0063] In order to implement the above embodiment, an embodiment of the present invention proposes a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the human gesture recognition method based on dual-array RFID as described above is implemented.

[0064] According to the computer device of the embodiment of the present invention, the human gesture recognition program based on the dual-array RFID is stored in the memory, so that when the processor executes the human gesture recognition program based on the dual-array RFID, the human gesture recognition method based on the dual-array RFID as described above is implemented. As a result, gestures can be detected from multiple angles, and the influence of the external environment can be reduced. In addition, the user does not need to carry an RFID tag or antenna with him / her, which greatly improves the user experience.

[0065] In order to implement the above embodiment, the embodiment of the present invention proposes a human gesture recognition device based on dual-array RFID, which is applied to a terminal device. The terminal device is connected to a first antenna and a second antenna through a reader. The first antenna is arranged relative to the first tag array, and the second antenna is arranged relative to the second tag array. The first tag array and the second tag array are perpendicular to each other and face the gesture recognition environment, such as Figure 4 As shown, the human gesture recognition device based on dual-array RFID includes: a receiving module 10 , a preprocessing module 20 and a gesture recognition module 30 .

[0066] Among them, the receiving module 10 is used to receive the first signal data corresponding to the first antenna and the second signal data corresponding to the second antenna analyzed by the reader, the first signal data includes the signal reception strength and signal phase collected by each tag in the first tag array at each sampling time, and the second signal data includes the signal reception strength and signal phase collected by each tag in the second tag array at each sampling time; the preprocessing module 20 is used to preprocess the first signal data and the second signal data to obtain corresponding first time series data and second time series data; the gesture recognition module 30 is used to input the first time series data and the second time series data into a pre-trained gesture recognition model so that the gesture recognition model outputs the corresponding human gesture recognition result.

[0067] Optionally, the gesture recognition model includes a first self-attention module, a second self-attention module, a third self-attention module and a multi-layer perceptron.

[0068] Optionally, the gesture recognition module 30 is also used to input the first time series data into the first self-attention module to obtain a first high-dimensional feature representation corresponding to the first direction; input the second time series data into the second self-attention module to obtain a second high-dimensional feature representation corresponding to a second direction perpendicular to the first direction; concatenate the first high-dimensional feature representation and the second high-dimensional feature representation, and input them into the third self-attention module to extract the features of the overall gesture to obtain the overall high-dimensional features; input the overall high-dimensional features into the multi-layer perceptron for feature concentration to obtain the probability of the overall high-dimensional features corresponding to each gesture category in all gesture categories.

[0069] Optionally, preprocessing includes deconvolution, upsampling, filtering and smoothing, normalization and construction processing.

[0070] It should be noted that the above Figure 1 The description of the human gesture recognition method based on dual-array RFID is also applicable to the human gesture recognition device based on dual-array RFID, and will not be repeated here.

[0071] In summary, according to the human gesture recognition device based on dual-array RFID according to an embodiment of the present invention, it is applied to a terminal device, the terminal device is connected to the first antenna and the second antenna through a reader, the first antenna is arranged relative to the first tag array, the second antenna is arranged relative to the second tag array, the first tag array and the second tag array are perpendicular to each other and face the gesture recognition environment, the first signal data corresponding to the first antenna and the second signal data corresponding to the second antenna analyzed by the reader are received by the receiving module, the first signal data and the second signal data are preprocessed by the preprocessing module to obtain the corresponding first time series data and the second time series data, and the first time series data and the second time series data are input into the pre-trained gesture recognition model through the gesture recognition module, so that the gesture recognition model outputs the corresponding human gesture recognition result; thereby, gestures can be detected from multiple angles, and the influence of the external environment can be reduced. In addition, the user does not need to wear an RFID tag or antenna, which greatly improves the user experience.

[0072] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0073] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0074] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0075] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0076] It should be noted that in the claims, any reference signs placed between brackets shall not be construed as limiting the claims. The word "comprising" does not exclude the presence of components or steps not listed in the claim. The word "a" or "an" preceding a component does not exclude the presence of a plurality of such components. The invention may be implemented by means of hardware comprising several different components and by means of a suitably programmed computer. In a unit claim enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third etc. does not indicate any order. These words may be interpreted as names.

[0077] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0078] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

[0079] In the description of the present invention, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0080] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0081] In the present invention, unless otherwise clearly specified and limited, a first feature being "above" or "below" a second feature may mean that the first and second features are in direct contact, or the first and second features are in indirect contact through an intermediate medium. Moreover, a first feature being "above", "above" or "above" a 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 higher in level than the second feature. A first feature being "below", "below" or "below" a second feature may mean that the first feature is directly below or obliquely below the second feature, or simply means that the first feature is lower in level than the second feature.

[0082] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms should not be understood as necessarily referring to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, unless they are contradictory.

[0083] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. A human gesture recognition method based on dual array RFID, characterized in that: Applied to a terminal device, the terminal device is connected to a first antenna and a second antenna through a reader, the first antenna is arranged relative to a first tag array, the second antenna is arranged relative to a second tag array, the first tag array and the second tag array are perpendicular to each other and face a gesture recognition environment, the method includes: Receive first signal data corresponding to the first antenna and second signal data corresponding to the second antenna parsed by the reader, wherein the first signal data includes the signal reception strength and signal phase collected by each tag in the first tag array at each sampling time, and the second signal data includes the signal reception strength and signal phase collected by each tag in the second tag array at each sampling time; Preprocessing the first signal data and the second signal data to obtain corresponding first time series data and second time series data; The first time series data and the second time series data are input into a pre-trained gesture recognition model so that the gesture recognition model outputs a corresponding human gesture recognition result.

2. The human gesture recognition method based on dual array RFID as claimed in claim 1, characterized in that: The gesture recognition model includes a first self-attention module, a second self-attention module, a third self-attention module and a multi-layer perceptron.

3. The human gesture recognition method based on dual array RFID as claimed in claim 2, characterized in that: Inputting the first time series data and the second time series data into a pre-trained gesture recognition model so that the gesture recognition model outputs a corresponding human gesture recognition result, including: Inputting the first time series data into the first self-attention module to obtain a first high-dimensional feature representation corresponding to a first direction; Inputting the second time series data into the second self-attention module to obtain a second high-dimensional feature representation corresponding to a second direction perpendicular to the first direction; The first high-dimensional feature representation and the second high-dimensional feature representation are concatenated and input into the third self-attention module to extract features of the overall gesture to obtain an overall high-dimensional feature; The overall high-dimensional feature is input into the multi-layer perceptron for feature concentration to obtain the probability of the overall high-dimensional feature corresponding to each gesture category in all gesture categories.

4. The method for human gesture recognition based on dual array RFID as claimed in claim 1, characterized in that: The preprocessing includes deconvolution, upsampling, filtering and smoothing, normalization and construction processing.

5. A computer-readable storage medium, characterized in that: A human gesture recognition program based on dual-array RFID is stored thereon, and when the human gesture recognition program based on dual-array RFID is executed by a processor, a human gesture recognition method based on dual-array RFID as described in any one of claims 1-4 is implemented.

6. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the human gesture recognition method based on dual-array RFID as described in any one of claims 1 to 4 is implemented.

7. A human gesture recognition device based on dual array RFID, characterized in that: Applied to a terminal device, the terminal device is connected to a first antenna and a second antenna through a reader, the first antenna is arranged relative to a first tag array, the second antenna is arranged relative to a second tag array, the first tag array and the second tag array are perpendicular to each other and face a gesture recognition environment, the device comprises: A receiving module, configured to receive first signal data corresponding to the first antenna and second signal data corresponding to the second antenna parsed by the reader, wherein the first signal data includes a signal receiving strength and a signal phase collected by each tag in the first tag array at each sampling time, and the second signal data includes a signal receiving strength and a signal phase collected by each tag in the second tag array at each sampling time; A preprocessing module, used for preprocessing the first signal data and the second signal data to obtain corresponding first time series data and second time series data; The gesture recognition module is used to input the first time series data and the second time series data into a pre-trained gesture recognition model so that the gesture recognition model outputs a corresponding human gesture recognition result.

8. The human gesture recognition device based on dual array RFID as claimed in claim 7, characterized in that: The gesture recognition model includes a first self-attention module, a second self-attention module, a third self-attention module and a multi-layer perceptron.

9. The human gesture recognition device based on dual array RFID as claimed in claim 8, characterized in that: The gesture recognition module is also used to: Inputting the first time series data into the first self-attention module to obtain a first high-dimensional feature representation corresponding to a first direction; Inputting the second time series data into the second self-attention module to obtain a second high-dimensional feature representation corresponding to a second direction perpendicular to the first direction; The first high-dimensional feature representation and the second high-dimensional feature representation are concatenated and input into the third self-attention module to extract features of the overall gesture to obtain an overall high-dimensional feature; The overall high-dimensional feature is input into the multi-layer perceptron for feature concentration to obtain the probability of the overall high-dimensional feature corresponding to each gesture category in all gesture categories.

10. The human gesture recognition device based on dual array RFID according to claim 7, characterized in that: The preprocessing includes deconvolution, upsampling, filtering and smoothing, normalization and construction processing.

Images