Wi-fi signal-based sensing method, electronic device, and program product

Abstract

Embodiments of the present disclosure provide a Wi-Fi signal-based sensing method, an electronic device, and a program product. The Wi-Fi signal-based sensing method is applicable to an electronic device, and the electronic device is configured with an antenna and a laminated intelligent metasurface, the laminated intelligent metasurface comprising multiple layers of metasurfaces. The Wi-Fi signal-based sensing method comprises: controlling the laminated intelligent metasurface to reflect a Wi-Fi signal to a region of interest of a target object; acquiring first Wi-Fi CSI data of the region of interest by means of the antenna; and, on the basis of the first Wi-Fi CSI data, acquiring sensing information of the target object.

Description

Wi-Fi signal-based sensing method, electronic device, and program product

[0001] This application claims priority to Chinese patent application No. 202410343344.6, filed on March 22, 2024, the entire contents of which are incorporated herein by reference. Technical Field

[0002] The present disclosure relates to the field of communication technologies, and in particular to a Wi-Fi signal-based sensing method, electronic device, and program product. Background Art

[0003] Currently, a terminal can sense an object based on an acquired Wi-Fi (wireless fidelity) signal, thereby obtaining sensing information of the object (eg, a motion state of the object). Summary of the Invention

[0004] In one aspect, a Wi-Fi signal sensing method is provided, which is applied to an electronic device equipped with an antenna and a laminated smart metasurface, wherein the laminated smart metasurface comprises multiple layers. The Wi-Fi signal sensing method includes: controlling the laminated smart metasurface to reflect Wi-Fi signals to a region of interest (ROI) of a target object; collecting first Wi-Fi CSI data from the region of interest via the antenna; and obtaining sensing information about the target object based on the first Wi-Fi CSI data.

[0005] In another aspect, a Wi-Fi signal-based sensing device is provided for use in an electronic device comprising an antenna and a laminated smart metasurface comprising multiple layers of metasurfaces. The Wi-Fi signal-based sensing device comprises a control module, a collection module, and an acquisition module. The control module is configured to control the laminated smart metasurface to reflect Wi-Fi signals to an area of ​​interest of a target object. The collection module is configured to collect first Wi-Fi CSI data of the area of ​​interest via the antenna. The acquisition module is configured to acquire sensing information of the target object based on the first Wi-Fi CSI data.

[0006] In another aspect, an electronic device is provided, comprising: a memory and a processor. The memory is coupled to the processor; the memory is configured to store a computer program; and the processor implements the aforementioned Wi-Fi signal-based sensing method when executing the computer program.

[0007] On the other hand, a computer-readable storage medium is provided, on which computer program instructions are stored. When the computer program instructions are executed by a processor, the above-mentioned Wi-Fi signal-based perception method is implemented.

[0008] In yet another aspect, a computer program product is provided. The computer program product includes computer program instructions, and when the computer program instructions are executed by a processor, the computer program instructions implement the above-mentioned Wi-Fi signal-based perception method. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] To more clearly illustrate the technical solutions of the present disclosure, the following briefly introduces the drawings required for use in some embodiments of the present disclosure. Obviously, the drawings described below are only drawings of some embodiments of the present disclosure, and those skilled in the art can also derive other drawings based on these drawings.

[0010] FIG1 is a system architecture diagram of a Wi-Fi signal-based perception system according to some embodiments of the present disclosure.

[0011] FIG2 is a schematic structural diagram of an electronic device according to some embodiments of the present disclosure.

[0012] FIG3 is a schematic structural diagram of another electronic device according to some embodiments of the present disclosure.

[0013] FIG4 is a schematic structural diagram of another electronic device according to some embodiments of the present disclosure.

[0014] FIG5 is a schematic structural diagram of another electronic device according to some embodiments of the present disclosure.

[0015] FIG6 is a flowchart of a Wi-Fi signal-based sensing method according to some embodiments of the present disclosure.

[0016] FIG7 is a flowchart of another Wi-Fi signal-based sensing method according to some embodiments of the present disclosure.

[0017] FIG8 is a schematic diagram of channel distribution among an electronic device, an area of ​​interest, and a Wi-Fi device according to some embodiments of the present disclosure.

[0018] FIG9 is a flowchart of another Wi-Fi signal-based sensing method according to some embodiments of the present disclosure.

[0019] FIG10 is a flowchart of another Wi-Fi signal-based sensing method according to some embodiments of the present disclosure.

[0020] FIG11 is a schematic structural diagram of a first feature extraction model according to some embodiments of the present disclosure.

[0021] FIG12 is a flowchart of another Wi-Fi signal-based sensing method according to some embodiments of the present disclosure.

[0022] FIG13 is a flowchart of another Wi-Fi signal-based sensing method according to some embodiments of the present disclosure.

[0023] FIG14 is a flowchart of another Wi-Fi signal-based sensing method according to some embodiments of the present disclosure.

[0024] FIG15 is a schematic structural diagram of a perception information recognition model according to some embodiments of the present disclosure.

[0025] FIG16 is a flowchart of another Wi-Fi signal-based sensing method according to some embodiments of the present disclosure.

[0026] FIG17 is a schematic diagram of the structure of a long short-term memory network according to some embodiments of the present disclosure.

[0027] FIG18 is a schematic structural diagram of a Wi-Fi signal-based sensing device according to some embodiments of the present disclosure.

[0028] FIG19 is a schematic structural diagram of another electronic device according to some embodiments of the present disclosure. DETAILED DESCRIPTION

[0029] The following will clearly and completely describe the technical solutions of this disclosure in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of this disclosure, not all of them. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of this disclosure without inventive effort are within the scope of protection of this disclosure.

[0030] It should be noted that, in this disclosure, words such as "exemplary" or "for example" are used to describe examples, illustrations, or explanations. Any embodiment or design described in this disclosure using words such as "exemplary" or "for example" should not be interpreted as being more preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0031] In the following, the terms "first," "second," etc. are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the quantity of the technical features being described. Thus, a feature described by terms such as "first," "second," etc. may explicitly or implicitly include one or more of the features.

[0032] In this disclosure, unless otherwise specified, " / " means "or." For example, A / B can mean either A or B. "And / or" is used herein solely to describe an association between related objects, indicating that three possible relationships exist. For example, "A and / or B" can mean: only A, A and B, or only B. Furthermore, "at least one" means one or more, and "a plurality" means two or more.

[0033] Currently, terminals with Wi-Fi sensing capabilities can perceive objects in a spatial environment based on Wi-Fi signals. However, there's a problem with insufficient Wi-Fi signal strength in the spatial environment, preventing the terminal from accurately perceiving the objects. Specifically, in the aforementioned method, the terminal's ability to perceive objects based solely on Wi-Fi signals is limited, and it may not be able to accurately determine the object's perception information or more detailed perception information.

[0034] In addition, the surface area of ​​the terminal with Wi-Fi sensing function is small. In the solution of adding an intelligent metasurface to the terminal, the surface area of ​​the addable intelligent metasurface is small, which cannot effectively enhance the signal strength of the Wi-Fi signal scattered by objects in the spatial environment.

[0035] In addition, some terminals with Wi-Fi awareness requirements do not have Wi-Fi awareness capabilities. For example, Wi-Fi awareness cannot be performed because the processor's processing of the CSI (channel state information) data of the Wi-Fi signal is blocked, or the terminal lacks a Wi-Fi module.

[0036] Based on the above description, it can be seen that the following problems currently exist in the terminal's perception of objects in the spatial environment: the terminal does not have the perception function and cannot perceive; or the terminal has the perception function but the perception accuracy is low due to insufficient signal strength; or for terminals with small surface areas, the smart metasurface added to the terminal cannot effectively enhance the signal strength of the Wi-Fi signal scattered by the object.

[0037] To solve the above technical problems, the embodiments of the present disclosure provide a Wi-Fi signal-based perception method, which is applied to an electronic device. The electronic device is configured with an antenna and a laminated smart metasurface, and the laminated smart metasurface includes multiple layers of metasurfaces. The Wi-Fi signal-based perception method includes: controlling the laminated smart metasurface to reflect the Wi-Fi signal to the target object's area of ​​interest; collecting first Wi-Fi CSI data of the area of ​​interest through the antenna; and obtaining perception information of the target object based on the first Wi-Fi CSI data. Since the laminated smart metasurface includes multiple layers of metasurfaces, it can improve the beamforming effect of the Wi-Fi signal by reflecting the Wi-Fi signal through multiple layers when the surface area is small, and better enhance the signal strength of the Wi-Fi signal in the area of ​​interest. In this way, the electronic device controls the stacked smart metasurface to reflect the Wi-Fi signal to the target object's area of ​​interest, thereby enhancing the signal strength of the Wi-Fi signal in the area of ​​interest. The electronic device then collects first Wi-Fi CSI data of the area of ​​interest through the antenna. Because the signal strength of the Wi-Fi signal in the area of ​​interest is relatively strong, the first Wi-Fi CSI data contains more precise feature information of the target object in the area of ​​interest. Consequently, the electronic device can accurately and reliably obtain more precise perception information of the target object based on the first Wi-Fi CSI data, thereby ensuring the perception accuracy of the target object.

[0038] The Wi-Fi signal-based sensing method provided in the embodiments of the present disclosure can be applied to a sensing system as shown in FIG1 . As shown in FIG1 , the sensing system includes: an electronic device 101 , a Wi-Fi device 102 , and a target object 103 .

[0039] Electronic device 101 includes a laminated smart metasurface 1011 and an antenna 1012. Laminated smart metasurface 1011 is configured to reflect Wi-Fi signals toward an area of ​​interest; antenna 1012 is configured to receive Wi-Fi signals from the area of ​​interest, thereby collecting first Wi-Fi CSI data for the area of ​​interest. Electronic device 101 can control laminated smart metasurface 1011 and antenna 1012 and, based on the first Wi-Fi CSI data, obtain sensory information of the target object.

[0040] The Wi-Fi device 102 is used to transmit a Wi-Fi signal so that the electronic device 101 can control the stacked smart metasurface 1011 to reflect the Wi-Fi signal toward the sensing area.

[0041] The target object 103 is located in the region of interest and can reflect the Wi-Fi signal sent to the region of interest. The target object 103 can be a person or an object.

[0042] The solid lines in FIG. 1 are used to represent Wi-Fi signals between the electronic device 101 , the Wi-Fi device 102 , and the target object 103 .

[0043] In some embodiments, the stacked smart metasurface 1011 includes multiple layers of smart metasurfaces, wherein the innermost layer of the multi-layer smart metasurface is a reflective smart metasurface, and the other (outer) layers of the smart metasurface except the innermost layer are omnidirectional smart metasurfaces (IOS, also referred to as omnidirectional single-unit smart metasurfaces). Omnidirectional smart metasurfaces can simultaneously transmit and / or reflect electromagnetic waves.

[0044] In an exemplary implementation, the omnidirectional smart metasurface may have other names, such as a transmissive smart metasurface.

[0045] In some embodiments, the omnidirectional intelligent metasurface includes a transparent substrate and two upper and lower transparent thin film functional layers. The transparent substrate can be made of one or more of the following transparent flexible materials: ordinary glass, quartz glass, organic glass, and polyethylene terephthalate. The symmetrical or reciprocal artificial electromagnetic structure between the upper and lower transparent thin film functional layers can be made of one or more of the following materials: indium tin oxide, aluminum zinc oxide, fluorine-doped tin dioxide, antimony tin oxide, graphene film, and metal nanowires. In addition, the artificial electromagnetic structure on the upper and lower transparent thin film functional layers can be formed based on one or more of the following methods: etching, photolithography, chemical corrosion, and electroplating.

[0046] In one implementation, as the stacked smart metasurface 1011 reflects Wi-Fi signals toward the region of interest, each outer layer of the smart metasurface sequentially transmits the Wi-Fi signal inward until the Wi-Fi signal reaches the innermost layer of the smart metasurface. Thereafter, the innermost layer and each outer layer of the smart metasurface simultaneously reflect the Wi-Fi signal toward the sensing area.

[0047] In some embodiments, the stacked smart metasurface operates in various modes, including energy splitting (ES), mode switching (MS), and time switching (TS). In the energy splitting mode, each metaatom in the stacked smart metasurface reflects and transmits simultaneously; in the mode switching mode, some metaatoms in the stacked smart metasurface reflect while others transmit; and in the time switching mode, metaatoms in the stacked smart metasurface reflect and transmit at different times based on a time sequence.

[0048] The superimposable coupling paths between metaatoms within the stacked smart metasurface allow for modulation and control, such as the amplitude superposition of Wi-Fi signals. Furthermore, the stacked smart metasurface is an ultra-thin smart metasurface based on the Huygens principle (where each point on the electromagnetic wave's wavefront is treated as a new secondary wave source, and the superposition of secondary waves emitted by each secondary wave source is used as a new wave source after a period of electromagnetic wave propagation).

[0049] It should be understood that compared to related metal diode technologies, stacked smart metasurfaces are not based on the Huygens principle. Therefore, each layer of the smart metasurface is relatively thick and cannot be deployed on portable mobile devices. However, the thickness of each layer of the stacked smart metasurface based on the Huygens principle can be controlled to be a few tenths of the wavelength of the reflected electromagnetic wave. At the same time, the spacing between each two layers of the smart metasurface can be set to a few tenths of the wavelength of the reflected electromagnetic wave.

[0050] In some embodiments, the laminated smart metasurface 1011 can be transparent or non-transparent. The transparent laminated smart metasurface is used to cover the display screen surface of the mobile terminal, while the non-transparent laminated smart metasurface is used to cover the housing surface (non-display screen area) of the mobile terminal.

[0051] In some embodiments, each layer of the stacked smart metasurface 1011 includes multiple reflective units (or meta-atoms, metamaterials, etc.), each reflective unit has multiple working states, and each working state corresponds to a phase. For example, the working state of a reflective unit can be represented by a bit, 1 bit represents two working states, 2 bits represent 4 working states, and 3 bits represent 8 working states. The more working states of the reflective units, the more configurable reflection schemes of the stacked smart metasurface and the more flexible they are. In order to enable the stacked smart metasurface to be applied to more complex beamforming scenarios, the working state of each reflective unit in the stacked smart metasurface can be 2 bits or 3 bits.

[0052] In one implementation, the size of each reflective unit in the stacked smart metasurface is made to be one-quarter or one-eighth the wavelength of a Wi-Fi signal.

[0053] In an exemplary implementation, the size of each reflective unit in the stacked smart metasurface can be 15mm×15mm, and has four working states (including 00, 01, 10, and 11, four bit values, which can correspond to four phases of 0, π / 2, π, and 3π / 2, and one bit value can correspond to any phase). A layer of smart metasurface can deploy 8×4 reflective units (or metaatoms) on the display screen of a 5.5-inch mobile terminal. The stacked smart metasurface can include two or more layers of smart metasurfaces. In some embodiments, each layer of the smart metasurface can also be divided into two subcells, and one subcell includes 4×4 reflective units.

[0054] In one implementation, as shown in FIG2 , the electronic device provided by an embodiment of the present disclosure may include a stacked smart metasurface 201 , an antenna 202 , a controller 203 , and a power management module 204 .

[0055] The laminated smart metasurface 201 is used to reflect Wi-Fi signals toward an area of ​​interest.

[0056] Antenna 202 is used to collect Wi-Fi CSI data from an area of ​​interest.

[0057] In one implementation, the antenna 202 may include a first antenna 2021 and a second antenna 2022. The electronic device may perform noise cancellation on the Wi-Fi signal based on the Wi-Fi signal received by the first antenna 2021 and the second antenna 2022.

[0058] In some embodiments, the first antenna is used to collect Wi-Fi signals containing characteristic information of the target object, and the second antenna is used to collect existing interfering Wi-Fi signals. The electronic device performs noise cancellation processing on the Wi-Fi signals collected by the first antenna based on the Wi-Fi signals collected by the second antenna.

[0059] The controller 203 is used to control the stacked smart metasurface 201 and the antenna 202, and to process the Wi-Fi CSI data from the area of ​​interest to obtain perception information of the target object in the area of ​​interest.

[0060] The controller 203 may further include a radio frequency front-end module 2031 , a control processing module 2032 , and a memory 2033 .

[0061] The RF front-end module 2031 is used to pre-process Wi-Fi signals received by the antenna 202, so that the control processing module 2032 can collect Wi-Fi CSI data and parse it to obtain sensing information. The RF front-end module 2031 may include an RF transceiver, an RF front-end circuit, and a baseband processor to implement Wi-Fi signal reception and pre-processing.

[0062] The control processing module 2032 can be a high-performance processor, such as a field programmable gate array (FPGA) or a graphics processing unit (GPU). The control processing module 2032 can also send data via a high-speed signal channel to a processor on the mobile terminal for processing to achieve data analysis or control the stacked intelligent metasurface 201 and antenna 202.

[0063] The memory 2033 is used to store data to provide support for the control processing module. The memory 2033 can be an external high-speed access memory or a built-in high-speed memory.

[0064] In an exemplary implementation, the controller 203 may further include an interface circuit 2034 to connect the controller 203 to the laminated smart metasurface 201, to communicate with a mobile terminal, or to power the controller 203 from an external power source. The interface circuit may include one or more of the following: a universal serial bus (USB) interface, a smart metasurface interface, and a power supply interface.

[0065] In one implementation, if the mobile terminal does not have a short-range communication module, the controller 203 may further include a short-range communication module 2035, so that the electronic device provided by the embodiment of the present disclosure can communicate and interact with the mobile terminal or other communication devices (such as Wi-Fi devices, routers, etc.), and can send to other remote devices (such as servers, etc.). In an exemplary implementation, the mobile terminal can complete the communication and interaction with the electronic device provided by the embodiment of the present disclosure and control the electronic device provided by the embodiment of the present disclosure based on the application installed on the mobile terminal.

[0066] The short-range communication module can be prepared using one or more of the following technical means: Star Flash technology, Bluetooth technology or Wi-Fi technology.

[0067] The power management module 204 is used to supply power to components (such as controllers, etc.) in the electronic device that require power. In an exemplary implementation, the power module in the electronic device may be a built-in power supply.

[0068] The application scenarios of the embodiments of the present disclosure can be smart home scenarios, Internet of Things scenarios, etc. For example, the electronic device provided by the embodiments of the present disclosure can be a mobile terminal or a household appliance. The electronic device can sense people or objects in the home environment to obtain biometric information (respiration, heartbeat, etc.) or motion characteristic information (movement, expression, etc.) of the people in the home environment.

[0069] In some embodiments, the mobile terminal may include one or more of the following: a mobile phone, a tablet computer, a laptop computer, and customer premise equipment (CPE, such as a router, a modem, an access point gateway, etc.).

[0070] In some embodiments, household appliances may include sweeping robots, refrigerators, televisions, smart screens, smart homes, and other devices with Wi-Fi capabilities.

[0071] In one implementation, as shown in FIG3 , when the electronic device provided by an embodiment of the present disclosure is a mobile terminal, an embodiment of the present disclosure provides a structural schematic diagram of an electronic device including a stacked intelligent metasurface, a control circuit board, and an antenna covering the surface of the mobile terminal.

[0072] The laminated smart metasurface can be a thin layer of smart metasurface, and covers the AA area (active area) of the display screen of the mobile terminal. The control circuit board and the antenna can be installed on a flexible circuit board, which is integrally connected to the smart metasurface, and the flexible circuit board is located in the non-AA area of ​​the mobile terminal. The antenna includes a main antenna and an auxiliary antenna, which are respectively used to receive Wi-Fi signals scattered by the target object and reflected by the smart metasurface and receive ambient scattered signals. The main antenna and the auxiliary antenna are also placed on the non-AA area surface of the display screen.

[0073] In another implementation, as shown in FIG4 , in an embodiment of the present disclosure, a laminated smart metasurface may be added to the back of the mobile terminal, and the main antenna is compatible with the mobile terminal's built-in Wi-Fi antenna (i.e., the mobile terminal's built-in Wi-Fi antenna functions as the main antenna).

[0074] As shown in FIG5 , in an embodiment of the present disclosure, the electronic device is a sweeping robot, and the laminated intelligent metasurface is located on the side of the sweeping robot.

[0075] In an exemplary implementation, the stacked intelligent metasurface can obtain the contour and grayscale distribution characteristics of the target object based on the low-frequency coefficients of the two-dimensional discrete Fourier transform (2D-DFT); and obtain the shape details of the target object based on the high-frequency coefficients.

[0076] In some embodiments, a sweeping robot including a laminated intelligent metasurface is used to detect whether a target object falls.

[0077] The following will describe in detail the Wi-Fi signal sensing method provided by the embodiments of the present disclosure with reference to the accompanying drawings.

[0078] The Wi-Fi signal sensing method provided in the embodiments of the present disclosure can be applied to the electronic device 101 in the sensing system shown in Figure 1. Figure 6 shows a schematic flow chart of a Wi-Fi signal sensing method. As shown in Figure 6, the Wi-Fi signal sensing method includes the following steps S601 to S603.

[0079] In S601 , the stacked smart metasurface is controlled to reflect Wi-Fi signals to an area of ​​interest of a target object.

[0080] It should be understood that the stacked smart metasurface includes multiple layers of smart metasurface, which can reflect Wi-Fi signals with stronger signal strength to the area of ​​interest of the target object compared to a single-layer smart metasurface.

[0081] In some embodiments, the target object may be a person, a pet, an object, etc.; it may also be a part of the human body, such as the chest, heart, hand, foot, etc.

[0082] In one implementation, the spatial area that can be sensed by the electronic device may be a living or activity area of ​​people and / or pets, or an area where items are stored.

[0083] In an exemplary implementation, the number and distribution of reflective units in each layer of the multi-layer smart metasurface in the stacked smart metasurface are the same. The Wi-Fi signal reflected by the stacked smart metasurface can be expressed based on the following formula (1):

[0084] Among them, r is the observation point in the observation area of ​​the stacked intelligent metasurface, F tot (r) represents the radiation field corresponding to the Wi-Fi signal reflected by the stacked smart metasurface, F in (r) represents the Wi-Fi signal received by the stacked smart metasurface at observation point r, Indicates the number (n) on a smart super surface in a stacked smart super surface. x ,n y ) positions of the reflection units, n x and n yrepresents the number of reflective units in the x-direction and the y-direction of a layer of smart metasurface, L represents the number of layers of the stacked smart metasurface, N represents the number of reflective units in the x-direction, and M represents the number of reflective units in the y-direction. represents the transmission matrix, Indicates the (n x ,n y ) Wi-Fi signals received by the reflection units, Indicates the (n x ,n y ) current of each reflective unit in the on or off state.

[0085] When the stacked smart metasurface operates in an energy sharing mode and the reflectance and transmittance in the energy sharing are 1:1, the Wi-Fi signal reflected by the stacked smart metasurface based on formula (1) should be modified to formula (2):

[0086] That is, the innermost smart metasurface can reflect the full proportion of Wi-Fi signals, and the smart metasurfaces other than the innermost smart metasurface reflect only half of the Wi-Fi signals. nx,ny ) represents the transmission matrix of the innermost smart metasurface.

[0087] When the stacked smart metasurface operates in a time-switching mode and the ratio between the reflection time slot and the transmission time slot is 1:1, the Wi-Fi signal reflected by the stacked smart metasurface during the time period consisting of the reflection time slot and the transmission time slot also satisfies formula (2). However, the Wi-Fi signal reflected by the stacked smart metasurface during the reflection time slot and the Wi-Fi signal reflected by the stacked smart metasurface during the transmission time slot satisfy formulas (3) and (4), respectively:

[0088] Here, q represents the transmission time slot (at this time, only the innermost layer of the smart metasurface reflects the Wi-Fi signal), and q+1 represents the reflection time slot (at this time, each layer of the smart metasurface reflects the Wi-Fi signal).

[0089] In S602 , first Wi-Fi CSI data of an area of ​​interest is collected through an antenna.

[0090] It should be understood that the first Wi-Fi CSI data includes characteristic information of a target object in the region of interest, and the characteristic information of the target object is used to represent information of a distance dimension and an angle dimension of the target object.

[0091] It can be understood that the stronger the signal strength of the Wi-Fi signal reflected by the laminated smart metasurface to the target object's area of ​​interest, the higher the accuracy of the feature information of the target object in the area of ​​interest contained in the first Wi-Fi CSI data.

[0092] In S603 , perception information of the target object is acquired based on the first Wi-Fi CSI data.

[0093] In the disclosed embodiment, since the signal strength of the Wi-Fi signal reflected by the stacked smart metasurface to the target object's area of ​​interest is stronger than the signal strength of the Wi-Fi signal reflected by the single-layer smart metasurface, the feature information of the target object in the area of ​​interest contained in the first Wi-Fi CSI data is more accurate, and higher-precision perception information of the target object can be reliably obtained.

[0094] In combination with FIG6 , as shown in FIG7 , in the above S601 , controlling the laminated smart metasurface to reflect the Wi-Fi signal to the region of interest of the target object includes S701 and S702 .

[0095] In S701, position information of a region of interest is determined, and configuration information of the stacked smart metasurface is determined based on the position information of the region of interest.

[0096] The position information of the region of interest is used to characterize the position of the region of interest relative to the laminated smart metasurface. The configuration information of the laminated smart metasurface is used to configure the operating parameters of the reflective unit on the laminated smart metasurface.

[0097] It should be understood that the operating parameters of a reflective unit are used to indicate the operating state of the reflective unit, and each operating state of the reflective unit corresponds to a phase. Based on different operating parameters, the reflective units on the laminated smart metasurface can reflect Wi-Fi signals of different phases. The superposition of Wi-Fi signals of different phases reflected by multiple reflective units on the laminated smart metasurface results in a Wi-Fi signal with stronger signal strength in a certain direction.

[0098] In some embodiments, the channel of the Wi-Fi signal reflected by the stacked smart metasurface is estimated based on the location information of the region of interest, and the codebook information (i.e., configuration information) of the stacked smart metasurface is determined based on the channel estimation result.

[0099] As shown in Figure 8, an embodiment of the present disclosure provides a schematic diagram of the channel distribution between an electronic device, an area of ​​interest, and a Wi-Fi device. The channel between the stacked smart metasurface in the electronic device and the Wi-Fi device is h, the channel between the stacked smart metasurface in the electronic device and the area of ​​interest is g, the channel between the antenna in the electronic device and the area of ​​interest is r, the number of layers of the stacked smart metasurface in the electronic device is i, the number of reflective units in each layer of the stacked smart metasurface is Q = M × N, the number of transmitting antenna units of the Wi-Fi device is K, and K, M, and N are positive integers. Suppose the vector of the signal after the orthogonal multi-carrier signal s transmitted by the Wi-Fi device is modulated by the MIMO (multiple-input multiple-output) antenna is x, and the power limit of x is P max Indicates the predetermined power. The phase shift matrix of each layer of omnidirectional intelligent metasurface i∈N≡{1,2,...,I},I∈Z in the stacked intelligent metasurface satisfies formula (5) Θ i =diag(θ i )=diag([θ i,1 ,θ i,n ,...,θ i,I ] T ) (5)

[0100] Among them, θ i,n represents the phase shift coefficient of the nth reflective unit of the i-th layer smart metasurface, The Wi-Fi signal received by the stacked smart metasurface satisfies formula (6):

[0101] Where h is the channel between the stacked smart metasurface and the Wi-Fi device, h H is the transposed conjugate of h, k is the loss coefficient of penetrating the stacked smart metasurface, f1 is the channel coefficient of the outermost smart metasurface, and f i is the channel coefficient of the innermost intelligent metasurface, f1∈C N×M , f i ∈C N×M , Θ1 is the phase shift coefficient of the first layer of smart metasurface, Θ i represents the phase shift coefficient of the i-th layer of the smart metasurface, and n is Gaussian white noise.

[0102] The Wi-Fi signal reflected by the stacked smart metasurface satisfies formula (7):

[0103] Where u represents the Wi-Fi signal reflected by the stacked smart metasurface.

[0104] The antenna receives the Wi-Fi signal from the area of ​​interest, satisfying formula (8):

[0105] Where v represents the Wi-Fi signal received by the antenna from the area of ​​interest.

[0106] Based on the above formulas (5) to (8), we can obtain the problem of maximizing the antenna detection rate under the conditions of maximum transmission power and phase shift, γ = log2(1 + SNR), which can be equivalent to maximizing the signal-to-noise ratio (SNR). The maximization of the signal-to-noise ratio problem satisfies formula (9):

[0107] in, For formula (9), the alternating optimization method can be used to solve it, that is, for r, g, h, Θ1, ..., Θ I In an exemplary implementation, a suboptimal solution is obtained based on a singular value decomposition (SVD) method and a relaxation method for solving the maximum eigenvector of a semi-positive definite matrix with a rank of 1 when necessary.

[0108] In some embodiments, the electronic device includes two antennas, and for a multi-carrier signal, only some of the carriers are selected as signals to be processed, and the other signals are treated as interference signals. In this case, the problem of maximizing the signal-to-noise ratio satisfies formula (10):

[0109] Among them, r l represents the channel between antenna l and the region of interest, r l' In this way, the signal-to-noise ratio corresponding to the carrier signal to be processed can be determined.

[0110] In one implementation, the electronic device determines the configuration information of the stacked smart metasurface based on the location information of the region of interest and the expected signal strength (the signal strength of the Wi-Fi signal expected in the region of interest). Based on the location information of the region of interest and the expected signal strength, the spatial distribution of the signal strength of the expected Wi-Fi signal in the surrounding environment is conditionally constrained to satisfy formula (11):

[0111] Among them, E o (x, y) is used to characterize the spatial distribution of the signal strength of the desired Wi-Fi signal in the surrounding environment, R focus represents the region of interest, and R is the area in the surrounding environment except the region of interest.

[0112] The electronic device obtains the switching state of each reflective unit in the stacked intelligent metasurface and combines formula (1) and formula (11) to obtain the optimal switching state S that satisfies the constraints expressed by formula (11). In an exemplary implementation, the optimal switching state S that satisfies the constraints expressed by formula (11) can be solved based on a genetic algorithm.

[0113] In S702 , based on the configuration information of the laminated smart metasurface, the laminated smart metasurface is controlled to reflect the Wi-Fi signal to the area of ​​interest of the target object.

[0114] It is understood that by configuring the operating state of the reflective units on the laminated smart metasurface, the signal strength of the superimposed Wi-Fi signals reflected by the laminated smart metasurface can be made stronger in the direction of the region of interest, or the signal strength of the superimposed Wi-Fi signals reflected by the laminated smart metasurface can be made stronger when reaching the region of interest. In this way, the laminated smart metasurface can reliably control the reflected signal to the region of interest of the target object, enhance the strength of the Wi-Fi signal in the region of interest, and thus ensure the accuracy of obtaining the perception information of the target object.

[0115] In an exemplary implementation, the electronic device performs coding design based on the Gerchberg-saxton algorithm (phase recovery through amplitude) and the configuration information of the stacked intelligent metasurface, and configures the optimal switching state S that satisfies the constraints expressed by formula (11) into the stacked intelligent metasurface.

[0116] 7 , as shown in FIG9 , in the above S701 , determining the location information of the region of interest includes S901 to S904 .

[0117] In S901 , second Wi-Fi CSI data of the surrounding environment is collected through an antenna.

[0118] It should be understood that the surrounding environment is a spatial area within a preset range centered on the antenna (or electronic device), and the surrounding environment includes an area of ​​interest.

[0119] In some embodiments, the surrounding environment may be a room where the electronic device is located, or a collection of multiple rooms.

[0120] It can be understood that the second Wi-Fi CSI data is CSI data of Wi-Fi signals from all directions of the surrounding environment.

[0121] In one implementation, before collecting the second Wi-Fi CSI data of the surrounding environment through the antenna, the electronic device can control the stacked smart metasurface to reflect the Wi-Fi signal into the surrounding environment of the electronic device based on the initial configuration information of the stacked smart metasurface.

[0122] The initial configuration parameters indicate the operating parameters of each reflective unit in the stacked smart metasurface. These parameters include the reflection coefficient (or transmission coefficient) and the operating status. The reflection coefficient of a reflective unit indicates the ratio of the Wi-Fi signal strength reflected by the unit to the signal strength transmitted by it.

[0123] It can be understood that the diffraction neural network model can make the Wi-Fi signal reflected by the stacked smart metasurface have a more efficient energy focusing mode, which can better improve the signal strength in the surrounding environment of the electronic device, and thus reliably collect the second Wi-Fi CSI data of the surrounding environment through the antenna.

[0124] In some embodiments, the electronic device obtains initial configuration parameters based on the diffraction neural network model. Each reflection unit of the laminated intelligent metasurface corresponds to a model parameter of a diffraction neural network model, and the reflection coefficient (or transmission coefficient) of each reflection unit can be used as a learning parameter of the model. The incident field and the outgoing field of the laminated intelligent metasurface serve as the input and output of the diffraction neural network model. The diffraction neural network model includes multiple layers, each layer is connected based on the Rayleigh-Sommer non-formula, and the diffraction neural network model is trained and iterated based on the back propagation algorithm. In this way, the diffraction neural network model can predict the initial configuration parameters corresponding to the laminated intelligent metasurface reflecting the Wi-Fi signal to the surrounding environment of the electronic device, so that the surrounding environment of the electronic device has a strong signal strength, and the antenna can reliably collect the second Wi-Fi CSI data of the surrounding environment.

[0125] In one implementation, the electronic device writes preset configuration parameters into the controller, or fixes the focusing mode of the laminated smart metasurface based on 3D printing technology.

[0126] In S902, second Doppler shift characteristic information is obtained based on the second Wi-Fi CSI data.

[0127] The second Doppler frequency shift characteristic information is used to characterize the Doppler frequency shift characteristics of the environment surrounding the electronic device.

[0128] It should be understood that the Doppler frequency shift characteristic of the environment surrounding the electronic device is used to characterize the frequency change of the Wi-Fi signal caused by the scattering of the Wi-Fi signal by obstacles in the surrounding environment.

[0129] In S903 , the target object is detected based on the second Doppler frequency shift characteristic information to obtain characteristic information of the target object.

[0130] Feature information includes the location information and shape information of the target object.

[0131] It should be understood that the Doppler shift characteristics of the electronic device's surroundings, as represented by the second Doppler shift characteristic information, include the Doppler characteristics of the target object. The Doppler characteristics of the target object differ from the Doppler characteristics of other obstacles in the surrounding environment. The electronic device can determine characteristic information of the target object based on the Doppler characteristics of the target object.

[0132] In S904 , the location information of the region of interest is determined based on the feature information of the target object.

[0133] It should be understood that since the characteristic information of the target object includes its location and shape information, and the region of interest needs to ensure that the Wi-Fi signal reflected from the region of interest can completely cover the target object, the electronic device determines the location information of the region of interest based on the location and shape information of the target object. Based on the location information of the region of interest, the electronic device can determine the configuration information of the laminated smart metasurface corresponding to the Wi-Fi signal that can cover the target object, thereby ensuring the reliability and effectiveness of the target object's perception information.

[0134] In one exemplary implementation, the second Wi-Fi CSI data of the surrounding environment corresponds to multiple Wi-Fi signal scattering paths. The multiple scattering paths include a target scattering path and other scattering paths. The target scattering path is the path of the Wi-Fi signal scattered by the target object. The other scattering paths are the paths of interference signals, which may be Wi-Fi signals scattered by other obstacles (such as walls).

[0135] In some embodiments, the Wi-Fi signal is a multi-carrier signal with an orthogonal frequency division multiplexing (OFDM) waveform. Multiple data can be transmitted on the channel corresponding to each subcarrier, and the multi-carrier Wi-Fi signal is based on QAM modulation (quadrature amplitude modulation). The Wi-Fi signal based on QAM modulation is composed of two orthogonal subcarriers. The Wi-Fi signal with K subcarriers based on QAM modulation can be expressed based on formula (12):

[0136] Where s(t) represents the Wi-Fi signal, s krepresents the kth subcarrier, s k is plural (s k1 ,s k2 ), f k is the frequency of the kth subcarrier, t represents the arrival time, and j is a complex unit.

[0137] The channel frequency response of the Wi-Fi signal satisfies the formula (13): H=(H(f1),H(f2)…H(f k )…H(f K )) (13)

[0138] in, H(f k ) represents the frequency channel response of the kth subcarrier, including phase information and amplitude information, and K is the total number of subcarriers.

[0139] In one implementation, the electronic device determines the variance of the frequency channel response of each of the multiple subcarriers based on formula (13), and processes the subcarrier with the smallest variance to obtain Wi-Fi CSI data.

[0140] The channel frequency response of a Wi-Fi signal with multiple transmission paths satisfies formula (14):

[0141] Among them, f is the frequency of the subcarrier, t is the arrival time, a m (f, t) is the signal attenuation on the mth path, v m is the speed of change of the length of the mth path (i.e., the speed of change of the path length caused by the movement of the obstacle), c is the speed of light, j is a complex unit, M is the total number of transmission paths, 1≤m≤M, and m and M are positive integers.

[0142] When the transmission path includes the target transmission path (the transmission path passing through the target object) and other transmission paths (the path of the environmental interference signal, the direct path, etc.), combining formula (13) and formula (14), formula (15) representing the channel frequency response of the Wi-Fi signal can be obtained:

[0143] Among them, H s (f, t) represents other transmission paths, H d (f, t) represents the target transmission path, For the mth d Signal attenuation on the target transmission path, M d is the total number of target transmission paths, v m is the speed of change of the length of the mth path, c is the speed of light, j is a complex unit, m d 、Md is a positive integer, 1≤m d ≤M d In addition, other transmission paths also include the path of the environmental interference signal and the direct path. The path of the environmental interference signal can include the path of the stacked intelligent metasurface beamforming signal reflected by the environment and the path of the non-stacked intelligent metasurface beamforming signal reflected by the environment. Based on formula (15), formulas (16) and (17) can be further obtained: H s (f,t)=H s,Los (f,t)+H s,NLos (f,t)=H s,Los (f,t)+H s,RISCas (f,t)+H s,NRIScas (f,t) (16)

[0144] Among them, H s,Los (f, t) is the channel frequency response of the direct path, H s,NLos (f, t) is the channel frequency response of the path of the environmental interference signal, The channel frequency response of the path of the stacked smart metasurface beamforming signal reflected by the environment, H s,NRIScas (f, t) is the channel frequency response of the path of the non-stacked smart metasurface beamforming signal reflected by the environment; H d,RISCas (f, t) represents the channel frequency response of the path of the stacked smart metasurface beamforming signal reflected by the target object, H d,NRISCas (f, t) represents the channel frequency response of the path of the non-stacked smart metasurface beamforming signal reflected by the target object, For the mth d1 The signal attenuation corresponding to the path of the stacked intelligent metasurface beamforming signal reflected by the target object, For the mth d2 Signal attenuation corresponding to the path of the non-laminated smart metasurface beamforming signal reflected by the target object, m d1 、m d2 is a positive integer, m d1 、m d2 ≤M d The channel frequency response of the path of the stacked smart metasurface beamforming signal reflected by the target object in formula (17) can be used as a reference item for the electronic device to determine the target object; the channel frequency response of the path of the non-stacked smart metasurface beamforming signal reflected by the target object in formula (17) can be used as noise filtering, or it can be based on filter matching to fill the phase between the channel frequency response of the path of the stacked smart metasurface beamforming signal reflected by the target object, which can enhance the signal strength of the signal scattered by the target object.

[0145] 9 , as shown in FIG10 , in the above S902 , obtaining the second Doppler shift characteristic information based on the second Wi-Fi CSI data includes S1001 and S1002 .

[0146] In S1001, the second Wi-Fi CSI data is preprocessed to obtain preprocessed second Wi-Fi CSI data.

[0147] It should be understood that the second Wi-Fi CSI data includes noise and useful signals carrying characteristic information of the target object. The electronic device pre-processes the second Wi-Fi CSI data to decompose the original signal containing the noise signal into signals of different levels, thereby filtering out the noise signal.

[0148] In one exemplary implementation, since the wavelet coefficients of the useful signal are strongly correlated with the scale, while the wavelet coefficients of the noise signal are weakly correlated with the scale (i.e., since the useful signal is the same in terms of scale, the trend of change of the wavelet coefficients of the useful signal at different scales is consistent with the change of scale; while the trend of change of the wavelet coefficients of the noise signal at different scales is obviously more random and independent with the change of scale), the amplitude of the useful signal is significantly greater than the amplitude of the noise signal in the result of multiplying the signals of different wavelet coefficients at adjacent scales, so the noise signal and the useful signal can be distinguished. Processing the noise signal with a large amplitude based on the short-term Fourier transform (STFT) or discrete wavelet transform (DWT) can eliminate the noise therein.

[0149] In one implementation, the antenna may include a first antenna and a second antenna, and the preprocessing may include noise cancellation. The noise cancellation may be implemented by performing conjugate multiplication or division on the second Wi-Fi CSI data collected by the first antenna and the second Wi-Fi CSI data collected by the second antenna to obtain the second Wi-Fi CSI data after noise cancellation.

[0150] In an exemplary implementation, a result of conjugate multiplication of the second Wi-Fi CSI data collected by the first antenna and the second Wi-Fi CSI data collected by the second antenna satisfies formula (18):

[0151] Wherein, H1(f, t) is the channel frequency response of the second Wi-Fi CSI data collected by the first antenna, H2(f, t) is the channel frequency response of the second Wi-Fi CSI data collected by the second antenna, and Hs,1 (f) is the channel frequency response of other transmission paths received by the first antenna, H s,2 (f) is the channel frequency response of other transmission paths received by the second antenna, a m (f, t) is the signal attenuation of the mth path in the target transmission path received by the first antenna, a n (f, t) is the signal attenuation of the nth path in the target transmission path received by the second antenna, v m is the speed of change of the length of the mth path in the target transmission path received by the first antenna, v n M is the speed of change of the length of the nth path in the target transmission path received by the second antenna, d,1 is the number of target transmission paths received by the first antenna, M d,2 is the number of target transmission paths received by the second antenna, M d,1 、M d,2 , m, n are positive integers, n≤M d,2 ,m≤M d,1 Among the four terms after the second equal sign in formula (18), the last three terms contain feature information related to the target object, and the last two terms contain more feature information.

[0152] In some embodiments, the first antenna can be a primary antenna and the second antenna can be a secondary antenna. In this case, the product of the channel frequency response corresponding to the target transmission path received by the first antenna and the channel frequency response of the other transmission paths received by the second antenna contains the most characteristic information of the target object. In this case, the product term of the channel frequency response corresponding to the target transmission path received by the first antenna and the channel frequency response of the other transmission paths received by the second antenna in formula (18) satisfies formula (19):

[0153] Among them, H d,RISCas (f, t) represents the channel frequency response of the path of the stacked smart metasurface beamforming signal reflected by the target object, H d,NRISCas (f, t) represents the channel frequency response of the path of the non-stacked smart metasurface beamforming signal reflected by the target object, For the mth d1 The signal attenuation corresponding to the path of the stacked intelligent metasurface beamforming signal reflected by the target object, For the mth d2 Signal attenuation corresponding to the path of the non-laminated smart metasurface beamforming signal reflected by the target object, m d1 、m d2 、M d,1 is a positive integer, m d1 ≤M d,1 , m d2 ≤Md,1 .

[0154] It should be understood that since the signal amplitude corresponding to the product term of the channel frequency response of the target transmission path received by the first antenna and the target transmission path received by the second antenna is small, useful characteristic information cannot be obtained. The product term of the channel frequency response of other transmission paths received by the first antenna and the other transmission paths received by the second antenna itself has no useful characteristic information, and the first antenna is the main antenna and the second antenna is the secondary antenna. The characteristic information contained in the product term of the channel frequency response of other transmission paths received by the first antenna and the target transmission path received by the second antenna is not true characteristic information.

[0155] At this time, in order to extract the characteristic information of the target object in the path of the stacked intelligent metasurface beamforming signal reflected by the target object, the product term of the channel frequency response of other transmission paths received by the first antenna and the other transmission paths received by the second antenna in formula (18) and the product term of the channel frequency response of the target transmission path received by the first antenna and the target transmission path received by the second antenna can be eliminated through noise cancellation processing.

[0156] For the product term of the channel frequency response of other transmission paths received by the first antenna and the target transmission path received by the second antenna, different proportional coefficients can be set for the first antenna and the second antenna, so that the amplitude of the signal received by the first antenna is small and the variance is large, which is conducive to obtaining the response of the target path, and the amplitude of the signal received by the second antenna is large and the contrast is small, which is conducive to the response of other paths. Based on the proportional coefficient, the product term of the channel frequency response of other transmission paths received by the first antenna and the target transmission path received by the second antenna is eliminated, and finally the product term of the channel frequency response of the target transmission path received by the first antenna and the other transmission path received by the second antenna is obtained, thereby obtaining the preprocessed channel state information data (for example, Doppler frequency shift spectrum data corresponding to the channel state information).

[0157] In S1002 , the pre-processed second Wi-Fi CSI data is input into a first feature extraction model to obtain second Doppler frequency shift feature information.

[0158] The first feature extraction model can be constructed based on convolutional neural networks (CNN).

[0159] In some embodiments, the pre-processed second Wi-Fi CSI data may be Doppler frequency shift spectrum data corresponding to the second Wi-Fi CSI data.

[0160] In an exemplary implementation, the first feature extraction model may be composed of an encoder and a decoder. The encoder converts the Doppler frequency shift spectrum data corresponding to the input second Wi-Fi CSI data into a feature space Λ, and the decoder obtains the reconstructed data corresponding to the Doppler frequency shift spectrum data based on the feature space Λ. The existence of the feature space Λ can reduce the data dimension of the Doppler shift spectrum, thereby enabling better feature extraction. In an exemplary implementation, the difference between the input and output of the first feature extraction model is minimized, and the training iteration process of the first feature extraction model can be controlled based on the loss function shown in the following formula (20):

[0161] in, Represents input data U and output data The loss value between n is the batch size, and i is the current batch.

[0162] In some embodiments, Gaussian noise may be added before the encoder to improve the generalization performance of the first feature extraction model and avoid overfitting.

[0163] Figure 11 is a schematic diagram of a first feature extraction model based on a convolutional neural network according to an embodiment of the present disclosure. The Doppler shift spectrum data is sequentially processed through multiple convolutional layers, multiple pooling layers, and upsampling layers, undergoing multiple feature extractions and reducing the feature space size. This results in multiple channels, i.e., multiple types of features, accurately extracting the Doppler shift feature information corresponding to the Doppler shift spectrum data.

[0164] In combination with FIG9 , as shown in FIG12 , in the above S702 , controlling the stacked smart metasurface to reflect the Wi-Fi signal to the region of interest of the target object includes S1201 and S1202 .

[0165] In S1201 , the accuracy of the feature information of the target object is determined.

[0166] It should be understood that in S903, the electronic device detects the target object based on the second Doppler frequency shift characteristic information to obtain characteristic information of the target object. The electronic device can determine the accuracy of the characteristic information of the target object based on the characteristic information of the target object.

[0167] In some embodiments, the electronic device determines the type of the target object based on the feature information of the target object, and determines the accuracy of the feature information of the target object based on the type of the target object.

[0168] In an exemplary implementation, the target object type includes a biometric feature part (e.g., a chest or heart part, to obtain the target object's heart rate or respiratory rate) and a motion feature part (e.g., a hand, foot, or body part). When the target object type is a biometric feature part, the electronic device determines the target object's feature information with a first precision, which is the information precision for obtaining the feature information of the biometric feature part. When the target object type is a motion feature part, the electronic device determines the target object's feature information with a second precision, which is the information precision for obtaining the feature information of the motion feature part.

[0169] In S1202 , when the accuracy of the characteristic information of the target object does not meet the preset accuracy requirement, the laminated smart metasurface is controlled to reflect the Wi-Fi signal to the area of ​​interest of the target object.

[0170] It should be understood that if the accuracy of the target object's characteristic information does not meet the preset accuracy requirement, it means that the Wi-Fi CSI data corresponding to the current signal strength within the region of interest cannot obtain the perception information corresponding to the accuracy of the target object's characteristic information. In this case, the electronic device controls the laminated intelligent metasurface to reflect the signal to the region of interest of the target object, thereby enhancing the signal strength within the region of interest and improving the accuracy of the characteristic information corresponding to the Wi-Fi CSI data, thereby obtaining high-precision perception information.

[0171] In combination with FIG7 , as shown in FIG13 , the Wi-Fi signal-based sensing method provided by the embodiment of the present disclosure further includes S1301 to S1303 .

[0172] In S1301 , the location information of the area of ​​interest is updated based on the first Wi-Fi CSI data to obtain the updated location information of the area of ​​interest.

[0173] In some embodiments, the position information of the region of interest may be obtained by obtaining a picture of the target object based on a camera.

[0174] It should be understood that the first Wi-Fi CSI data includes the location information of the target object. Therefore, the electronic device can obtain more accurate location information of the region of interest based on the first Wi-Fi CSI data and update the location information of the region of interest.

[0175] In S1302 , based on the updated position information of the region of interest, the configuration information of the laminated smart metasurface is updated to obtain the updated configuration information of the laminated smart metasurface.

[0176] It should be understood that the updated position information of the region of interest has higher accuracy, and the electronic device can determine the configuration information of the stacked smart metasurface corresponding to the updated position information based on the higher accuracy position information.

[0177] In S1303, based on the updated configuration information of the laminated smart metasurface, the laminated smart metasurface is controlled to reflect the Wi-Fi signal to the area of ​​interest of the target object.

[0178] It can be understood that the updated configuration information is determined based on the location information of the area of ​​interest with higher precision. Therefore, the stacked intelligent metasurface can be controlled based on the updated configuration information of the stacked intelligent metasurface, which can reflect signals that are more concentrated on the target object. The antenna can collect CSI data containing more precise and richer feature information, so that electronic devices can obtain more accurate and reliable perception information.

[0179] 6 , as shown in FIG14 , in the above S603 , acquiring the perception information of the target object based on the first Wi-Fi CSI data includes S1401 to S1403 .

[0180] In S1401, first Doppler frequency shift characteristic information is obtained based on first Wi-Fi CSI data.

[0181] In S1402, a scale of the perception information is determined based on the first Doppler frequency shift feature information.

[0182] The first Doppler frequency shift characteristic information is used to characterize the Doppler frequency shift characteristic of the target object corresponding to each moment in multiple moments.

[0183] It should be understood that the Doppler shift characteristic is a characteristic of a moving object changing the frequency of a reflected signal when reflecting the signal, and different motion characteristics correspond to different frequency changes.

[0184] It is understandable that the electronic device may determine the scale of the perception information related to the motion type of the target object based on the motion feature information included in the first Doppler frequency shift feature information.

[0185] In S1403, if the perception information is large-scale perception information, the first Doppler frequency shift feature information is input into a perception information recognition model based on meta-learning technology to obtain large-scale perception information output by the perception information recognition model. Alternatively, if the perception information is small-scale perception information, time-frequency analysis is performed on the first Wi-Fi CSI data to obtain small-scale perception information.

[0186] The perceptual information recognition model is suitable for identifying perceptual information in multiple different spatial scenarios.

[0187] It should be understood that meta-learning technology can enable the perceptual information recognition model to adapt to multi-region recognition scenarios, such as an environment with multiple rooms. The perceptual information recognition model trained by meta-learning technology can better recognize perceptual information from different rooms.

[0188] In the disclosed embodiments, the electronic device can determine the scale of the target object's perception information based on the first Doppler shift characteristic information. Different perception information determination processes can be used for different scales. Thus, by identifying the scale of the perception information, an appropriate method can be selected to accurately and effectively determine the target object's perception information.

[0189] In some embodiments, meta-learning includes zero-shot (there are no samples of this category in the training samples) learning / one-shot (there is only one or very few training samples) learning, model-independent meta-learning, and meta-reinforcement learning, etc. Meta-learning may include two stages: stage one is model training based on the training task, and stage two is model training based on the test task. In stage one, the electronic device trains the model parameters corresponding to each sub-training task based on the support set of each sub-training task in the N sub-training tasks; and tests the performance based on the query set of each sub-training task to obtain the loss value between the predicted value and the true label. The loss functions of the N sub-training tasks can be integrated into one loss function, satisfying the following formula (21): Loss(φ)=l1+l n +…+l N (twenty one)

[0190] Among them, l n is the loss function for the nth sub-training task, where n and N are positive integers. Furthermore, the electronic device can update model parameters based on a gradient descent method and the loss function to obtain a perceptual information recognition model. Alternatively, the electronic device can obtain the perceptual information recognition model by solving for optimal model parameters based on reinforcement learning or an evolutionary algorithm.

[0191] In one implementation, when the perception information is small-scale perception information, time-frequency analysis is performed on the first Wi-Fi CSI data to obtain the small-scale perception information, including: preprocessing the first Wi-Fi CSI data to obtain a relatively smooth CSI sample curve; and the electronic device extracting amplitude-frequency / phase-frequency characteristics of the signal samples based on a fast Fourier transform (FFT).

[0192] In an exemplary implementation, the relationship between the phase change of respiration (or heartbeat) and the displacement of the heart (or chest cavity) satisfies the following formula (22):

[0193] Where Δχ is the phase change of respiration (or heartbeat), and ΔZ is the displacement of the heart (or chest cavity). The electronic device performs an FFT on the Wi-Fi signal reflected from the chest to obtain the phase change of respiration (or heartbeat), and then uses bandpass filtering to separate the respiration (or heartbeat) signal.

[0194] In some embodiments, the sampling frequency of the first Wi-Fi CSI data is 100 Hz or above, which can better capture the human respiratory frequency range (0.2-0.5 Hz, 12-30 times per minute). The electronic device can collect the respiratory frequency based on the signal spectrum resolution of 512-point fast Fourier transform.

[0195] In one implementation, when the perception information is small-scale perception information, the electronic device determines whether the small-scale perception information is measurable, and performs time-frequency analysis if the small-scale perception information is measurable.

[0196] In one implementation, when the perception information is large-scale perception information, the electronic device performs model training based on meta-learning to obtain the basic framework, task generation, parameter update and domain adaptation of the perception information recognition model.

[0197] In some embodiments, Wi-Fi signals are multi-carrier signals. Analyzing the Doppler characteristics of each subcarrier signal can complicate the underlying framework of the perception information recognition model. This disclosed embodiment uses principal component analysis (PCA) to reduce the complexity of the underlying framework.

[0198] In an exemplary implementation, as shown in FIG15 , the basic framework of the perceptual information recognition model can use two principal components to input Doppler frequency shift features, connect the features, and extract features through multiple residual blocks. The residual structure of multiple residual blocks includes multiple convolutional layer paths and a small number of convolutional paths (short connection paths). Based on the multiple convolutional layer paths and a small number of convolutional paths, irrelevant feature information can be eliminated. The features extracted are pooled, and through a smoothing layer, a fully connected layer, and a classifier, multiple labels with different probabilities can be obtained. The label with the highest probability is the predicted large-scale perceptual information.

[0199] Furthermore, the information extracted by the basic framework of the perceptual information recognition model includes cross-domain scenarios, that is, when there are multiple different regions, the perceptual information recognition model accurately predicts one region, but may have poor prediction results for cross-regional data. In order to enable the perceptual information recognition model to have a better prediction effect on cross-domain data, it is necessary to train the perceptual information recognition model based on a training set of tasks with multiple situations, so that the basic framework can learn knowledge from new domain scenarios based on a small number of samples. The electronic device evaluates the effect of the perceptual information recognition model and iteratively updates it based on the query set of tasks with multiple situations. At this time, the loss function of the perceptual information recognition model can be cross entropy and satisfies the following formula (23):

[0200] in, is the task-related parameter in the basic framework of the perceptual information recognition model, J is the number of samples in the task, k is the model parameter, and o is the regularization term used to avoid overfitting. Performing a gradient operation on the parameter θ can satisfy formula (24):

[0201] Where ξ is a hyperparameter that controls the learning speed of the model. The meta-objective function is optimized based on stochastic gradient descent (SGD). At this time, the update of the parameter θ satisfies formula (25):

[0202] In this way, based on formulas (23), (24), and (25), the parameter θ can be updated based on a variety of task training, so that the perceptual information recognition model has better recognition effects for a variety of tasks and cross-domain tasks.

[0203] 14 , as shown in FIG. 16 , in the above S1401 , obtaining first Doppler shift characteristic information based on the first Wi-Fi CSI data includes S1601 and S1602 .

[0204] In S1601, the first Wi-Fi CSI data is preprocessed to obtain preprocessed first Wi-Fi CSI data.

[0205] In one implementation, the antenna may include a first antenna and a second antenna, and the preprocessing may include noise reduction. The noise reduction may be implemented by performing conjugate multiplication or division on the first Wi-Fi CSI data collected by the first antenna and the first Wi-Fi CSI data collected by the second antenna to obtain the first Wi-Fi CSI data after noise reduction.

[0206] In S1602, the pre-processed first Wi-Fi CSI data is input into a second feature extraction model to obtain first Doppler frequency shift feature information.

[0207] The second feature extraction model is built based on convolutional neural networks and long short-term memory (LSTM) networks.

[0208] It should be understood that since the first Doppler frequency shift characteristic information is used to characterize the Doppler frequency shift characteristics corresponding to the target object at each of multiple moments, the Doppler frequency shift characteristics at different moments can be linked based on the long short-term memory network to obtain the first Doppler frequency shift characteristic information including the time dimension.

[0209] In one exemplary implementation, each of the multiple Doppler shift characteristics at each moment corresponds to a motion characteristic of the target object at that moment, and the first Doppler shift characteristic information corresponding to the multiple Doppler shift characteristics at each moment is the motion trend of the target object within the time interval corresponding to the multiple moments. For example, the first Doppler shift characteristic information corresponding to the multiple hand motion characteristics at each moment is the direction of the hand motion.

[0210] In some embodiments, the second feature extraction model is based on the angle mean square error and the distance mean square error as the loss function, satisfying the following formula (26):

[0211] Among them, MSE θ is the mean square error of angle, MSE d is the mean square error of the distance, λ1 and λ2 are weight parameters used to adjust the accuracy of angle estimation and distance estimation.

[0212] Figure 17 shows a schematic diagram of feature extraction based on a long short-term memory network. Different long short-term memory network layers (represented by LSTM in the figure) input Wi-Fi CSI data at different times (for example, the first, second, and third times), and then the fully connected layer (linear fully connected layer) outputs the predicted angle and distance.

[0213] It is understandable that, in order to realize the above functions, the sensing device based on Wi-Fi signals includes hardware structures and / or software modules corresponding to the execution of each function. It should be easy for those skilled in the art to realize that, in combination with the algorithm steps of each example described in the embodiments of the present disclosure, the present disclosure can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this disclosure.

[0214] The embodiment of the present disclosure can divide the functional modules of the Wi-Fi signal-based sensing device according to the above-mentioned method embodiment. For example, each functional module can be divided corresponding to each function, or two or more functions can be integrated into one functional module. The above-mentioned integrated module can be implemented in the form of hardware or software. It should be noted that the division of modules in the embodiment of the present disclosure is schematic and is only a logical function division. There may be other division methods in actual implementation. The following is an example of dividing each functional module corresponding to each function.

[0215] Figure 18 is a schematic diagram of the structure of a Wi-Fi signal-based sensing device according to an embodiment of the present disclosure. This Wi-Fi signal-based sensing device can implement the Wi-Fi signal-based sensing method provided in the above method embodiment. As shown in Figure 18, sensing device 180 includes a control module 1801, a collection module 1802, and an acquisition module 1803.

[0216] The control module 1801 is used to control the laminated smart metasurface to reflect Wi-Fi signals to the area of ​​interest of the target object.

[0217] The collecting module 1802 is configured to collect first Wi-Fi CSI data of the area of ​​interest through the antenna.

[0218] The acquisition module 1803 is configured to acquire the perception information of the target object based on the first Wi-Fi CSI data.

[0219] In the case of implementing the functions of the above-mentioned integrated modules in hardware, the embodiments of the present disclosure provide another structure of the electronic device involved in the above-mentioned embodiments. As shown in Figure 19, the electronic device 190 includes: a processor 1902 and a bus 1904. In an exemplary implementation, the electronic device may also include a memory 1901. In some embodiments, the electronic device may also include a communication interface 1903.

[0220] Processor 1902 may implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the embodiments of this disclosure. Processor 1902 may be a central processing unit, a general-purpose processor, a digital signal processor, an application-specific integrated circuit, a field programmable gate array, or other programmable logic device, a transistor logic device, a hardware component, or any combination thereof, and may implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the embodiments of this disclosure. Processor 1902 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP (digital signal processor) and a microprocessor, and the like.

[0221] The communication interface 1903 is used to connect to other devices via a communication network, such as Ethernet, wireless access network, or wireless local area network (WLAN).

[0222] The memory 1901 may be a read-only memory (ROM) or other type of static storage device that can store static information and instructions, a random access memory (RAM) or other type of dynamic storage device that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto.

[0223] As an implementation, memory 1901 may exist independently of processor 1902. Memory 1901 may be connected to processor 1902 via bus 1904 to store instructions or program code. When processor 1902 calls and executes the instructions or program code stored in memory 1901, the perception method provided in the embodiments of the present disclosure can be implemented.

[0224] In another implementation, the memory 1901 may also be integrated with the processor 1902 .

[0225] Bus 1904 can be an Extended Industry Standard Architecture (EISA) bus, etc. Bus 1904 can be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, FIG19 shows only one thick line, but this does not mean that there is only one bus or only one type of bus.

[0226] Some embodiments of the present disclosure provide a computer-readable storage medium (e.g., a non-transitory computer-readable storage medium), which stores computer program instructions. When the computer program instructions are executed on a computer, the computer executes the perception method described in any of the above embodiments.

[0227] Exemplarily, the above-mentioned computer-readable storage media may include, but are not limited to: magnetic storage devices (e.g., hard disks, floppy disks, or magnetic tapes), optical disks (e.g., compact disks (CDs), digital versatile disks (DVDs), etc.), smart cards, and flash memory devices (e.g., erasable programmable read-only memories (EPROMs), cards, sticks, or key drives, etc.). The various computer-readable storage media described in the present disclosure may represent one or more devices and / or other machine-readable storage media for storing information. The term "machine-readable storage medium" may include, but is not limited to, wireless channels and various other media capable of storing, containing, and / or carrying instructions and / or data.

[0228] An embodiment of the present disclosure provides a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute the perception method described in any one of the above embodiments.

[0229] In the embodiment of the present disclosure, the electronic device controls the stacked smart metasurface to reflect the Wi-Fi signal to the area of ​​interest of the target object. Since the stacked smart metasurface has multiple layers of smart metasurfaces, each layer of the smart metasurface can reflect the Wi-Fi signal to the area of ​​interest, which can better enhance the signal strength of the Wi-Fi signal in the area of ​​interest compared to a single-layer smart metasurface. In addition, the first Wi-Fi CSI data of the area of ​​interest is collected through the antenna. Since the signal strength of the Wi-Fi signal in the area of ​​interest is strong, the first Wi-Fi CSI data contains more accurate feature information of the target object in the area of ​​interest. Thus, the electronic device can accurately and reliably obtain more accurate perception information of the target object based on the first Wi-Fi CSI data, thereby ensuring the perception accuracy of the target object.

[0230] The above is only a specific embodiment of the present disclosure, but the scope of protection of the present disclosure is not limited thereto. Any changes or replacements within the technical scope disclosed in the present disclosure should be included in the scope of protection of the present disclosure. Therefore, the scope of protection of the present disclosure should be based on the scope of protection of the claims.

Claims

1. A Wi-Fi signal sensing method, applied to electronic devices, wherein: The electronic device is configured with an antenna and a laminated smart metasurface, wherein the laminated smart metasurface includes multiple layers of metasurfaces, and the method includes: controlling the laminated intelligent metasurface to reflect Wi-Fi signals to an area of ​​interest of a target object; collecting, by the antenna, first Wi-Fi channel state information (CSI) data of the area of ​​interest; Acquire perception information of the target object based on the first Wi-Fi CSI data.

2. The method according to claim 1, wherein The innermost metasurface in the multi-layer metasurface is a reflective metasurface, and the other metasurfaces in the multi-layer metasurface except the innermost metasurface are omnidirectional metasurfaces.

3. The method according to claim 2, wherein: Each layer of the multi-layer metasurface of the stacked intelligent metasurface includes multiple reflection units, wherein each of the multiple reflection units has multiple working states, and each of the multiple working states corresponds to a phase.

4. The method according to any one of claims 1 to 3, wherein Before controlling the stacked smart metasurface to reflect the Wi-Fi signal to the area of ​​interest of the target object, the method further includes: Based on the initial configuration information of the laminated smart metasurface, the laminated smart metasurface is controlled to reflect the Wi-Fi signal to the surrounding environment of the electronic device.

5. The method according to any one of claims 1 to 4, wherein The controlling the laminated smart metasurface to reflect the Wi-Fi signal to the area of ​​interest of the target object includes: Determining location information of the region of interest, and determining configuration information of the laminated smart metasurface based on the location information of the region of interest; wherein the configuration information of the laminated smart metasurface is used to configure operating parameters of a reflective unit on the laminated smart metasurface; Based on the configuration information of the laminated smart metasurface, the laminated smart metasurface is controlled to reflect the Wi-Fi signal to the area of ​​interest of the target object.

6. The method according to claim 5, wherein: The determining the location information of the area of ​​interest includes: collecting second Wi-Fi CSI data of the surrounding environment of the electronic device through the antenna; obtaining second Doppler frequency shift characteristic information based on the second Wi-Fi CSI data, where the second Doppler frequency shift characteristic information is used to characterize a Doppler frequency shift characteristic of an environment surrounding the electronic device; Detecting the target object based on the second Doppler frequency shift characteristic information to obtain characteristic information of the target object, where the characteristic information of the target object includes position information and shape information of the target object; The position information of the region of interest is determined based on the feature information of the target object.

7. The method according to claim 6, wherein: The obtaining, based on the second Wi-Fi CSI data, the second Doppler frequency shift characteristic information includes: Preprocessing the second Wi-Fi CSI data to obtain preprocessed second Wi-Fi CSI data; The preprocessed second Wi-Fi CSI data is input into a first feature extraction model to obtain the second Doppler frequency shift feature information, where the first feature extraction model is constructed based on a convolutional neural network.

8. The method according to claim 6, wherein: The controlling the laminated smart metasurface to reflect the Wi-Fi signal to the area of ​​interest of the target object includes: Determining the accuracy of the characteristic information of the target object; When the accuracy of the characteristic information of the target object does not meet the preset accuracy requirement, the laminated intelligent metasurface is controlled to reflect the Wi-Fi signal to the area of ​​interest of the target object.

9. The method according to claim 5, further comprising: updating the location information of the area of ​​interest based on the first Wi-Fi CSI data to obtain updated location information of the area of ​​interest; Based on the updated position information of the region of interest, updating the configuration information of the laminated smart metasurface to obtain the updated configuration information of the laminated smart metasurface; Based on the updated configuration information of the laminated intelligent metasurface, the laminated intelligent metasurface is controlled to reflect the Wi-Fi signal to the area of ​​interest of the target object.

10. The method according to any one of claims 1 to 9, wherein The acquiring, based on the first Wi-Fi CSI data, the perception information of the target object includes: obtaining, based on the first Wi-Fi CSI data, first Doppler frequency shift characteristic information, where the first Doppler frequency shift characteristic information is used to characterize a Doppler frequency shift characteristic of the target object corresponding to each of a plurality of time instants; determining a scale of the perception information based on the first Doppler frequency shift characteristic information; If the perception information is large-scale perception information, the first Doppler frequency shift feature information is input into a perception information recognition model to obtain large-scale perception information output by the perception information recognition model, where the perception information recognition model is applicable to identifying perception information in multiple different spatial scenarios. Alternatively, if the perception information is small-scale perception information, time-frequency analysis is performed on the first Wi-Fi CSI data to obtain the small-scale perception information.

11. The method according to claim 10, wherein: The obtaining, based on the first Wi-Fi CSI data, the first Doppler frequency shift characteristic information includes: Preprocessing the first Wi-Fi CSI data to obtain preprocessed first Wi-Fi CSI data; The preprocessed first Wi-Fi CSI data is input into a second feature extraction model to obtain first Doppler frequency shift feature information, where the second feature extraction model is constructed based on a convolutional neural network and a long short-term memory network.

12. The method according to claim 7 or 11, wherein: The antenna includes a first antenna and a second antenna; the preprocessing includes noise cancellation processing, and the noise cancellation processing includes the following steps: Conjugate multiplication or division is performed on the first Wi-Fi CSI data collected by the first antenna and the first Wi-Fi CSI data collected by the second antenna to obtain the first Wi-Fi CSI data after the noise reduction process.

13. An electronic device comprising: A memory and a processor; wherein the memory is coupled to the processor; the memory is used to store instructions executable by the processor; and when the processor executes the instructions, the method according to any one of claims 1 to 12 is performed.

14. The electronic device according to claim 13, wherein: The electronic device is a mobile terminal, and the laminated intelligent metasurface of the mobile terminal includes multiple layers of metasurfaces. The laminated intelligent metasurface is located on the screen surface of the mobile terminal and / or the back panel surface of the mobile terminal; the portion of the laminated metasurface located on the screen surface of the mobile terminal is transparent.

15. The electronic device according to claim 13, wherein The electronic device is a sweeping robot, which includes a laminated intelligent metasurface and an antenna. The laminated intelligent metasurface includes multiple layers of metasurfaces, and the laminated intelligent metasurface is located on the outer surface of the side wall of the shell of the sweeping robot.

16. A computer program product, wherein The computer program product comprises computer program instructions, which, when executed by a processor, implement the method according to any one of claims 1 to 12.

Images