Doppler shift-based behavior recognition method and device, electronic device, and storage medium

Abstract

The application provides a behavior recognition method and device based on Doppler frequency shift, an electronic device and a storage medium, comprising: for the frequency domain channel response of a Wi-Fi signal, cleaning the frequency domain channel response to obtain a cleaned power delay spectrum and a cleaned channel impulse response corresponding to the cleaned frequency domain channel response; determining a power sum corresponding to an NLOS channel based on the cleaned power delay spectrum; determining a correlation power value corresponding to the NLOS channel based on the cleaned channel impulse response; determining a multipath channel fading factor according to the power sum and the correlation power value; determining a corresponding Doppler frequency shift in a zero-order Bessel function based on the multipath channel fading factor; and determining a behavior recognition result of a passive target according to Doppler frequency shifts at multiple observation times. The application scheme realizes accurate behavior recognition of a passive target.

Description

Technical Field

[0001] This application relates to the field of wireless positioning technology, and in particular to a behavior recognition method and device, electronic device, and computer-readable storage medium based on Doppler frequency shift. Background Technology

[0002] A key development direction for next-generation wireless communication technology is the integration of communication sensing and radar. Future Wi-Fi signals will not only function as data transmission devices but will also possess the ability to sense passive targets. Utilizing wireless signal-based behavior recognition has broad application prospects in indoor scenarios. For example, it can play a role in the smart home field, helping users get rid of remote controls and obtain a better long-distance control experience; or it can play a role in health monitoring, detecting users' breathing, heart rate, and other states. Summary of the Invention

[0003] The purpose of this application is to provide a behavior recognition method, device, electronic device, and computer-readable storage medium based on Doppler frequency shift, for accurately recognizing the behavior of passive targets.

[0004] On the one hand, this application provides a behavior recognition method based on Doppler frequency shift, including:

[0005] For the frequency domain channel response of Wi-Fi signals, the frequency domain channel response is cleaned to obtain the cleaned power delay spectrum and the cleaned channel impulse response corresponding to the cleaned frequency domain channel response;

[0006] Based on the cleaned power delay spectrum, determine the total power corresponding to the NLOS channel;

[0007] Based on the channel impulse response after cleaning, the relevant power value corresponding to the NLOS channel is determined.

[0008] The multipath channel fading factor is determined based on the sum of the power and the relevant power values.

[0009] Based on the multipath channel fading factor, the corresponding Doppler frequency shift is determined in the zero-order Bessel function;

[0010] The behavior recognition results of passive targets are determined based on the Doppler frequency shift at multiple observation times.

[0011] In one embodiment, the frequency domain channel response for the Wi-Fi signal is cleaned to obtain a cleaned power delay spectrum and a cleaned channel impulse response, which includes:

[0012] The corresponding power delay spectrum is determined for the frequency domain channel response, and the sampling point with the highest power in the power delay spectrum is searched as the centroid sampling point;

[0013] Several sampling points before the centroid sampling point are truncated and spliced ​​to the end of the power delay spectrum to obtain the adjusted power delay spectrum;

[0014] Based on a preset first-length front window and a preset second-length rear window, target sampling points are selected from the adjusted power delay spectrum, and sampling points other than the target sampling points are used as noise sampling points; wherein, the front window is placed at the very beginning of the adjusted power delay spectrum, and the rear window is placed at the very end of the adjusted power delay spectrum;

[0015] The noise threshold power is determined based on the power of multiple noise sampling points;

[0016] Target sampling points whose power exceeds the noise threshold are selected from the front window and the rear window and designated as sampling points.

[0017] The front and back ends of the adjusted power delay spectrum are cyclically connected. It is determined whether the distance between the specified sampling point closest to the front end in the back window and the centroid sampling point at the junction exceeds the sampling point number threshold.

[0018] If so, the designated sampling point closest to the front end within the rear window is determined as the LOS sampling point, and the remaining designated sampling points are determined as NLOS sampling points;

[0019] Set the power of all sampling points in the power delay spectrum except for the NLOS sampling points to zero to obtain the cleaned power delay spectrum;

[0020] The time-domain channel estimate values ​​of the sampling points other than the NLSO sampling points in the channel impulse response corresponding to the frequency domain channel response are set to zero to obtain the cleaned channel impulse response.

[0021] In one embodiment, the method further includes:

[0022] If not, the centroid sampling point is determined to be a LOS sampling point, and the remaining designated sampling points are NLOS sampling points;

[0023] Return to the step of setting the power of all sampling points in the power delay spectrum except for the NLOS sampling points to zero to obtain the cleaned power delay spectrum.

[0024] In one embodiment, determining the relevant power value corresponding to the NLOS channel based on the cleaned channel impulse response includes:

[0025] The cleaned channel impulse responses of two symbols separated by a specified number of symbols within the same data packet in a Wi-Fi signal are multiplied by their conjugates to obtain the correlation power value corresponding to the NLOS channel.

[0026] In one embodiment, before determining the corresponding Doppler frequency shift in the zero-order Bessel function based on the multipath channel fading factor, the method further includes:

[0027] In the case of calculating multiple related power values ​​by using symbol pairs corresponding to multiple specified symbol numbers, the average number of several Doppler frequency shifts corresponding to each specified symbol number is calculated to obtain the average Doppler frequency shift corresponding to each specified symbol number.

[0028] A new Doppler frequency shift is obtained by weighting the average Doppler frequency shifts corresponding to various specified symbol numbers.

[0029] In one embodiment, the step of weighting the average Doppler shifts corresponding to multiple specified symbol numbers to obtain a new Doppler shift includes:

[0030] Based on the Doppler frequency shift corresponding to various specified symbol numbers, a Doppler frequency shift function is obtained by fitting; wherein, the Doppler frequency shift function characterizes the mapping relationship between the specified symbol number and the Doppler frequency shift;

[0031] The function value corresponding to each specified number of symbols is selected from the Doppler frequency shift function, and a weighted average is performed to obtain the new Doppler frequency shift.

[0032] In one embodiment, the step of weighting the average Doppler shifts corresponding to multiple specified symbol numbers to obtain a new Doppler shift includes:

[0033] Based on the target signal-to-noise ratio corresponding to each specified number of symbols, a weighted average is performed on the average Doppler frequency shifts corresponding to multiple specified number of symbols to obtain a new Doppler frequency shift.

[0034] In one embodiment, determining the multipath channel fading factor based on the sum of power and the relevant power value includes:

[0035] Based on the total power and the noise threshold power, a specified signal-to-noise ratio is determined, and the relationship between the specified signal-to-noise ratio and the first and second signal-to-noise ratio thresholds is determined; wherein the second signal-to-noise ratio threshold is greater than the first signal-to-noise ratio threshold.

[0036] If the specified signal-to-noise ratio is less than the first signal-to-noise ratio threshold, the noise threshold power is subtracted from the sum of power to obtain the adjusted sum of power, and the relevant power value is divided by the adjusted sum of power to obtain the multipath channel fading factor;

[0037] If the specified signal-to-noise ratio is not less than the first signal-to-noise ratio threshold and is less than the second signal-to-noise ratio threshold, a power adjustment parameter is determined based on the noise threshold power, the specified signal-to-noise ratio and the preset adjustment factor. The power adjustment parameter is then subtracted from the relevant power value and divided by the sum of the powers to obtain the multipath channel fading factor.

[0038] If the specified signal-to-noise ratio is not less than the second signal-to-noise ratio threshold, the relevant power value is divided by the sum of the powers to obtain the multipath channel fading factor.

[0039] In one embodiment, determining the behavior recognition result of a passive target based on the Doppler frequency shift at multiple observation times includes:

[0040] Construct a Doppler frequency shift-time spectrum based on the Doppler frequency shift at multiple observation times;

[0041] The sub-spectral image generated within a specified time period is cropped from the Doppler frequency shift-time spectrum and used as the image to be processed;

[0042] The image to be processed is input into a trained behavior recognition model to obtain the behavior recognition result.

[0043] On the other hand, this application provides an electronic device, the electronic device comprising:

[0044] processor;

[0045] Memory used to store processor-executable instructions;

[0046] The processor is configured to execute the above-described behavior recognition method based on Doppler frequency shift.

[0047] In addition, this application also provides a computer-readable storage medium storing a computer program that can be executed by a processor to perform the above-described behavior recognition method based on Doppler frequency shift.

[0048] The proposed solution can generate a Doppler frequency shift representing the behavior state of a passive target based on the frequency shift channel response at multiple observation times, thereby performing behavior recognition based on the Doppler frequency shift at multiple observation times and obtaining the behavior recognition result. Attached Figure Description

[0049] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly described below.

[0050] Figure 1 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application;

[0051] Figure 2A flowchart illustrating a behavior recognition method based on Doppler frequency shift provided in an embodiment of this application;

[0052] Figure 3 A schematic diagram of a zero-order Bessel function provided in an embodiment of this application;

[0053] Figure 4 Provided for an embodiment of this application Figure 2 A detailed flowchart of step 210 is shown below;

[0054] Figure 5 A schematic diagram of the power delay spectrum provided in an embodiment of this application;

[0055] Figure 6 A schematic diagram of the adjusted power delay spectrum provided in an embodiment of this application;

[0056] Figure 7 A schematic diagram illustrating the sampling point selection process provided in one embodiment of this application;

[0057] Figure 8 Provided for an embodiment of this application Figure 2 A detailed flowchart of step 260;

[0058] Figure 9 A block diagram of a behavior recognition device based on Doppler frequency shift provided in an embodiment of this application. Detailed Implementation

[0059] The technical solutions in the embodiments of this application will now be described with reference to the accompanying drawings.

[0060] Similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0061] like Figure 1 As shown, this embodiment provides an electronic device 1, including: at least one processor 11 and a memory 12. Figure 1 Taking a processor 11 as an example, the processor 11 and the memory 12 are connected via a bus 10. The memory 12 stores instructions that can be executed by the processor 11. The instructions are executed by the processor 11 to enable the electronic device 1 to perform all or part of the process of the method in the following embodiments. In one embodiment, the electronic device 1 may be a router, a communication terminal device (e.g., a mobile phone, a tablet computer), etc., used to execute the behavior recognition method based on Doppler frequency shift. The following description uses the electronic device as the execution subject to describe the scheme.

[0062] The memory 12 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable red-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0063] This application also provides a computer-readable storage medium storing a computer program that can be executed by a processor 11 to perform the behavior recognition method based on Doppler frequency shift provided in this application.

[0064] See Figure 2 This is a flowchart illustrating a behavior recognition method based on Doppler frequency shift provided in an embodiment of this application, as shown below. Figure 2 As shown, the method may include steps 210-260.

[0065] Step 210: For the frequency domain channel response of the Wi-Fi signal, clean the frequency domain channel response to obtain the cleaned power delay spectrum and the cleaned channel impulse response corresponding to the cleaned frequency domain channel response.

[0066] When applying this solution, the electronic device executing the solution is placed in a designated location within an indoor environment. For example, the electronic device could be a router placed in a corner of a room. During the transmission and reception of Wi-Fi signals, the electronic device can determine the channel frequency response (CFR) of the Wi-Fi signal. The CFR is the channel state information (CSI) of the Wi-Fi signal. Each data packet transmitted by the Wi-Fi signal contains multiple symbols, and a CFR can be determined based on each symbol. Various objects in the indoor environment (such as people, furniture, appliances, pets, etc.) reflect Wi-Fi signals; therefore, multiple wireless channels exist during the transmission and reception of Wi-Fi signals. When passive targets move within the indoor environment, they change the reflection of Wi-Fi signals. Since stationary objects remain stationary, the CFR can characterize the impact of the passive target's behavior on the wireless channels. In one embodiment, after obtaining the frequency domain channel response, the frequency domain channel response can be windowed based on a window function (including but not limited to a Hamming window) to make the spectral leakage in the frequency domain channel response more concentrated, which facilitates subsequent processing.

[0067] After obtaining the frequency domain channel response of the symbols transmitted by the Wi-Fi signal, the electronic device can clean the frequency domain channel response to remove data from the noisy channel and the LOS (Line of Sight) channel, thereby obtaining the cleaned power-delay profile (PDP) and the cleaned channel impulse response (CIR) corresponding to the cleaned frequency domain channel response.

[0068] Step 220: Based on the cleaned power delay spectrum, determine the total power corresponding to the NLOS channel.

[0069] The cleaned power delay spectrum includes the power corresponding to multiple sampling points, each sampling point corresponding to a wireless channel. The power of the sampling point corresponding to a noise channel is zero, and the power of the sampling point corresponding to a LOS channel is also zero. Electronic devices can sum the power of multiple sampling points in the cleaned power delay spectrum to obtain the total power corresponding to the NLOS channel.

[0070] Step 230: Based on the channel impulse response after cleaning, determine the relevant power value corresponding to the NLOS channel.

[0071] The relevant power value is the power correlation parameter of the NLOS channel calculated from the channel impulse response after cleaning.

[0072] In one embodiment, the electronic device can perform conjugate multiplication on the cleaned channel impulse responses of two symbols separated by a specified number of symbols within the same data packet in a Wi-Fi signal to obtain the relevant power value corresponding to the NLOS channel.

[0073] Here, the specified number of symbols can be configured as needed. For example, the specified number of symbols can be any combination of 1, 2, 3...10, etc. The electronic device can select two symbols with a specified number of symbols between them from the multiple symbols transmitted in the same data packet, clean the frequency domain channel response of these two symbols, and multiply the cleaned channel impulse response of these two symbols by their conjugate to obtain the relevant power value.

[0074] For example, the calculation process of the relevant power value can be represented by the following formula (1):

[0075]

[0076] Where x represents the specified number of signs; P carr (x) represents the relevant power value; h l q h represents the channel impulse response after cleaning the l-th symbol in the data packet; l+x q This represents the channel impulse response after cleaning the (l+x)th symbol in the data packet.

[0077] Step 240: Determine the multipath channel fading factor based on the total power and related power values.

[0078] At any given observation time, after obtaining the power value related to the sum of powers, the electronic device can divide the related power value by the sum of powers to obtain the multipath channel fading factor. For example, the calculation process of the multipath channel fading factor can be represented by the following formula (2):

[0079] σ fading (x)=P carr (x) / P clean (2)

[0080] Where, σ fading (x) is the multipath channel fading factor; x represents the specified number of symbols; P carr (x) represents the relevant power value; P clean This represents the total power.

[0081] If the passive target is stationary during the transmission time of the two symbols, the cleaned channel impulse responses of the two symbols are theoretically identical. In this case, the correlation power calculated based on the cleaned channel impulse responses of the two symbols is equal to the sum of the powers, and the multipath channel fading factor is 1. If the passive target is active during the transmission time of the two symbols, the cleaned channel impulse responses of the two symbols are different. In this case, the correlation power calculated based on the cleaned channel impulse responses of the two symbols is not equal to the sum of the powers, and the multipath channel fading factor is not 1.

[0082] Step 250: Determine the corresponding Doppler frequency shift in the zero-order Bessel function based on the multipath channel fading factor.

[0083] The mapping relationship between the multipath channel fading factor and the target parameter satisfies the zeroth-order Bessel function. Electronic devices can operate within the zeroth-order Bessel function J0(2πf). D The target parameter corresponding to the multipath channel fading factor is determined in *x), and then the Doppler frequency shift is calculated from the target parameter. The target parameter can be expressed as f. D *τ; Here, f D Represents the Doppler frequency shift; τ is x*t symbol x represents the specified number of signs, t symbol The transmission duration of a known symbol.

[0084] See Figure 4 This is a schematic diagram of a zero-order Bessel function provided in an embodiment of this application, as shown below. Figure 4 As shown, after obtaining the multipath fading factor, it can be found on the vertical axis to determine the horizontal axis coordinate corresponding to the point on the found curve. This horizontal axis coordinate is the target parameter. Since the specified number of symbols and the transmission duration of one symbol are known, dividing the target parameter by the specified number of symbols and the transmission duration yields the Doppler frequency shift corresponding to the multipath fading factor.

[0085] Step 260: Determine the behavior recognition result of the passive target based on the Doppler frequency shift at multiple observation times.

[0086] Electronic devices can determine the Doppler frequency shift over multiple consecutive observation times, and then classify the behavior of passive targets based on the Doppler frequency shifts over these multiple observation times to obtain behavior recognition results. Here, the observation times can be configured as needed. For example, each data packet can correspond to one observation time, and each observation time generates one Doppler frequency shift. During Wi-Fi signal transmission, Doppler frequency shifts over multiple consecutive observation times can be generated. Depending on the application scenario, recognition of various behavior categories can be achieved. For example, it can recognize gestures such as waving, pushing, and clapping of passive targets (users), or behavioral states such as breathing and heartbeat, or behavioral states such as falling, gait, and running.

[0087] Through the above measures, Doppler frequency shifts representing the behavior state of passive targets can be generated based on the frequency shift channel response at multiple observation times. Thus, behavior recognition can be performed based on the Doppler frequency shifts at multiple observation times to obtain behavior recognition results.

[0088] In one embodiment, see Figure 3 This is provided as an embodiment of the present application. Figure 2 A detailed flowchart of step 210 is shown below. Figure 3 As shown, when performing step 210, steps 211 to 219 can be specifically performed.

[0089] Step 211: Determine the corresponding power delay spectrum for the frequency domain channel response, and search for the sampling point with the largest power in the power delay spectrum as the centroid sampling point.

[0090] For any symbol to be processed, the electronic device can perform an inverse fast fourier transform (IFFT) on the frequency domain channel response to obtain the channel impulse response.

[0091] After obtaining the channel impulse response, the electronic device can calculate the power delay spectrum based on the channel impulse response. In one embodiment, the transmission of the Wi-Fi signal may be accomplished through multiple receiving antennas and multiple transmitting antennas. There are multiple frequency domain channel responses for the same symbol, and multiple power delay spectra can be calculated. In this case, the average of the multiple power delay spectra can be calculated as the power delay spectrum of the symbol in subsequent processing.

[0092] Electronic devices can find the sampling point with the highest power in the power delay spectrum and use the found sampling point as the centroid sampling point.

[0093] Step 212: Take several sampling points before the centroid sampling point and splice them to the end of the power delay spectrum to obtain the adjusted power delay spectrum.

[0094] After determining the centroid sampling point, the electronic device can extract several sampling points before the centroid sampling point and splice these extracted sampling points to the end of the power delay spectrum to obtain the adjusted power delay spectrum.

[0095] See Figure 5 This is a schematic diagram of the power delay spectrum provided in an embodiment of this application, as shown below. Figure 5 As shown, the power delay spectrum includes 256 sampling points, and the area within the rectangle in the lower left corner contains the sampling points before the centroid sampling point.

[0096] See Figure 6 This is a schematic diagram of the adjusted power delay spectrum provided in an embodiment of this application, as shown below. Figure 6 As shown, Figure 5 After the sampling points before the centroid sampling point are truncated, they are spliced ​​to the end of the power delay spectrum to obtain the adjusted power delay spectrum.

[0097] Step 213: Based on the preset first length front window and the preset second length rear window, select the target sampling point in the adjusted power delay spectrum, and take the sampling points other than the target sampling points as noise sampling points; wherein, the front window is placed at the front end of the adjusted power delay spectrum, and the rear window is placed at the back end of the adjusted power delay spectrum.

[0098] The first length can be obtained by multiplying the length of the cyclic prefix (CP) by a first coefficient, which is greater than one; for example, the first coefficient is 2. The second length can be obtained by multiplying the length of the cyclic prefix by a second coefficient, which is less than one; for example, the second coefficient is 0.5.

[0099] Electronic devices can place the front window at the very beginning of the adjusted power delay spectrum and the rear window at the very end, using the sampling points within both windows as target sampling points. Sampling points other than the target sampling points can be identified as noise sampling points.

[0100] See Figure 7 This is a schematic diagram of sampling point screening provided in an embodiment of this application, as shown below. Figure 7 As shown, multiple sampling points within the front window are target sampling points, multiple sampling points within the rear window are target sampling points, and multiple sampling points between the front and rear windows are noise sampling points.

[0101] Step 214: Determine the noise threshold power based on the power of multiple noise sampling points.

[0102] Electronic devices can calculate the average power of multiple noise sampling points and use this average noise power as the noise threshold power.

[0103] In one embodiment, during the determination of the noise threshold power, the noise threshold power corresponding to multiple consecutive symbols can be obtained, and the multiple noise threshold powers can be smoothed to obtain a new noise threshold power. Here, the number of symbols used for smoothing can be configured as needed.

[0104] For example, smoothing can be performed using the following formula (3):

[0105] m l =α noise *n1+(1-α noise )*n l-1 (3)

[0106] Where, n l α represents the noise threshold power corresponding to the l-th symbol in the data packet; noise This represents the noise smoothing factor, which is greater than zero and less than one; n l-1 This represents the noise threshold power corresponding to the (l-1)th symbol in the data packet; m l This represents the noise threshold power after smoothing.

[0107] When smoothing is performed using the noise threshold power corresponding to two consecutive symbols, it can be done directly using formula (3). If smoothing is performed using the noise threshold power corresponding to at least three consecutive symbols, it can be done first based on the noise threshold power corresponding to the first two symbols, and then the smoothing result can be smoothed with the noise threshold power corresponding to the next symbol, and so on, until the noise threshold power of all symbols is smoothed.

[0108] A new noise threshold power can be obtained by smoothing the noise threshold power corresponding to multiple symbols, which can yield a more accurate noise threshold power.

[0109] Step 215: Select target sampling points in the front and rear windows whose power exceeds the noise threshold, and use them as designated sampling points.

[0110] Electronic devices can check whether the power corresponding to each target sampling point exceeds the noise threshold power, thereby filtering out target sampling points whose power exceeds the noise threshold power. These filtered target sampling points are called designated sampling points.

[0111] Step 216: Connect the front and back ends of the adjusted power delay spectrum in a loop, and determine whether the distance between the specified sampling point closest to the front end in the back window and the centroid sampling point exceeds the sampling point number threshold.

[0112] Step 217: If yes, determine the specified sampling point closest to the front end in the back window as the LOS sampling point, and the remaining specified sampling points as NLOS sampling points.

[0113] The electronic device can cyclically connect the front and back ends of the adjusted power delay spectrum. In this case, the last sampling point of the adjusted power delay spectrum is connected to the first sampling point. The electronic device can determine whether the distance between the specified sampling point closest to the front end within the back window and the centroid sampling point exceeds a sampling point number threshold. Here, the sampling point number threshold can be configured as needed; an exemplary sampling point number threshold is 2.

[0114] by Figure 7 For example, the designated sampling point closest to the front end in the back window is the first designated sampling point from left to right. The distance between this designated sampling point and the centroid sampling point after passing the 256th sampling point from left to right is checked to see if it exceeds the sampling point number threshold.

[0115] On the one hand, if so, it means that the distance between the designated sampling point closest to the front end in the back window and the centroid sampling point is large enough that the power is not affected by the power of the centroid sampling point. In this case, it can be determined that the designated sampling point closest to the front end in the back window is the LOS sampling point, while the other designated sampling points are NLOS sampling points.

[0116] On the other hand, if not, it indicates that the distance between the designated sampling point closest to the front end within the rear window and the centroid sampling point is small, and the power may be affected by the power of the centroid sampling point. In this case, if there are other designated sampling points within the rear window that are even closer to the centroid sampling point, they will also be affected by the centroid sampling point. Therefore, the centroid sampling point can be directly determined as the LOS sampling point, and the remaining designated sampling points can be determined as NLOS sampling points.

[0117] Furthermore, if there are no designated sampling points in the back window, in other words, if the power of all target sampling points in the back window does not exceed the noise power threshold, the centroid sampling point can be directly determined as the LOS sampling point, and the remaining designated sampling points can be determined as NLOS sampling points.

[0118] Step 218: Set the power of all sampling points in the power delay spectrum except for the NLOS sampling points to zero to obtain the cleaned power delay spectrum.

[0119] After determining multiple NLOS sampling points, the electronic device can find the NLOS sampling points in the original power delay spectrum based on the index of the NLOS sampling points. It can also set the power of sampling points other than these NLOS sampling points to zero in the original power delay spectrum, retain the power corresponding to the NLOS sampling points, and obtain the cleaned power delay spectrum.

[0120] Step 219: Set the time-domain channel estimate values ​​of the sampling points other than the NLOS sampling points in the channel impulse response corresponding to the frequency domain channel response to zero to obtain the cleaned channel impulse response.

[0121] In the channel impulse response corresponding to the frequency domain channel response, the electronic device can find the NLOS sampling point in the original channel impulse response based on the index of the NLOS sampling point, set the time domain channel estimate value of the sampling points other than the NLOS sampling point to zero, and retain the time domain channel estimate value corresponding to the NLOS sampling point to obtain the cleaned channel impulse response.

[0122] The above measures can be used to clean the frequency domain channel response to obtain the cleaned power delay spectrum and the cleaned channel impulse response.

[0123] In one embodiment, after executing step 240 and before executing step 250, when multiple related power values ​​are calculated using symbol pairs corresponding to multiple specified symbol numbers, the average number of several Doppler frequency shifts corresponding to each specified symbol number can be calculated to obtain the average Doppler frequency shift corresponding to each specified symbol number.

[0124] Each symbol pair consists of two symbols spaced a specified number of symbols apart. A specified number of symbols can correspond to multiple symbol pairs. In this case, a Doppler shift can be calculated for each symbol pair. The specific calculation process can be found in the previous description and will not be repeated here. The electronic device can calculate the average of the multiple Doppler shifts corresponding to the specified number of symbols to obtain the average Doppler shift corresponding to that specified number of symbols.

[0125] For example, the selected number of symbols during the calculation is 1, 3, or 5, with each number of symbols having 4 symbol pairs. For symbol number 1, the average of the four Doppler shifts calculated from the 4 symbol pairs is used as the average Doppler shift corresponding to symbol number 1. For symbol number 3, the average of the four Doppler shifts calculated from the 4 symbol pairs is used as the average Doppler shift corresponding to symbol number 3. For symbol number 5, the average of the four Doppler shifts calculated from the 4 symbol pairs is used as the average Doppler shift corresponding to symbol number 5.

[0126] For example, the calculation process of the average Doppler frequency shift can be represented by the following formula (4):

[0127]

[0128] Where x represents the specified number of signs; k x f represents the number of symbol pairs; D (x) represents the Doppler frequency shift corresponding to the specified number of symbols x; It is the average Doppler frequency shift corresponding to the specified number of symbols x.

[0129] Electronic devices can perform weighted averaging of the average Doppler frequency shift corresponding to multiple specified symbol numbers to obtain a new Doppler frequency shift.

[0130] For example, the calculation process of the weighted average can be represented by the following formula (5):

[0131]

[0132] Where x represents the number of symbols; It is the average Doppler frequency shift corresponding to the specified number of symbols x; This represents the new Doppler shift obtained after the final weighted average.

[0133] In one embodiment, during the process of weighted summation of the average Doppler frequency shifts corresponding to multiple specified symbol numbers, the electronic device can perform a weighted average of the average Doppler frequency shifts corresponding to multiple specified symbol numbers based on the target signal-to-noise ratio corresponding to each specified symbol number, to obtain a new Doppler frequency shift.

[0134] The target signal-to-noise ratio (SNR) can be obtained by dividing the correlation power value corresponding to the NLOS channel by the noise threshold power. Different correlation power values ​​can be calculated for different specified symbol numbers; therefore, each specified symbol number corresponds to a different target SNR.

[0135] For example, the calculation process of the weighted average can be represented by the following formula (6):

[0136]

[0137] Where x represents the number of symbols; It is the average Doppler frequency shift corresponding to the specified number of symbols x; This represents the new Doppler frequency shift obtained after the final weighted average; SNR represents the target signal-to-noise ratio.

[0138] In one embodiment, during the weighted summation of the average Doppler frequency shifts corresponding to multiple specified symbol numbers, the electronic device can fit a Doppler frequency shift function based on the Doppler frequency shifts corresponding to multiple specified symbol numbers. The Doppler frequency shift function characterizes the mapping relationship between the specified symbol number and the Doppler frequency shift. For example, the Doppler frequency shift function can be a least-squares fitting equation, which can be represented by the following formula (7):

[0139]

[0140] Where x represents the number of symbols; is the average Doppler frequency shift corresponding to the specified number of symbols x; a and b are constants obtained from the fitting.

[0141] After obtaining the Doppler frequency shift function, the electronic device can select the function value corresponding to each specified number of symbols within the Doppler frequency shift function. This function value is the Doppler frequency shift determined by the Doppler frequency shift function. A weighted average of multiple Doppler frequency shifts is then performed to obtain the new Doppler frequency shift. The calculation process is described in the aforementioned formula (5), and will not be repeated here.

[0142] In one embodiment, when performing step 250, the electronic device can determine a specified signal-to-noise ratio based on the total power and the noise threshold power. For example, the specified signal-to-noise ratio can be calculated using the following formula (8):

[0143] SNR=10*lg (P clean / n l (8)

[0144] Where SNR is the specified signal-to-noise ratio; P clean The sum of power; n l This is the noise threshold power.

[0145] An electronic device can compare a specified signal-to-noise ratio (SNR) with a first SNR threshold and a second SNR threshold to determine the magnitude relationship between the specified SNR and the first and second SNR thresholds. The second SNR threshold is greater than the first SNR threshold. For example, the first SNR threshold is 0, and the second SNR threshold is 20.

[0146] In one scenario, if the specified signal-to-noise ratio (SNR) is less than a first SNR threshold, the electronic device can subtract the noise threshold power from the total power to obtain an adjusted total power. The electronic device can then divide the relevant power value by the adjusted total power to obtain the multipath channel fading factor.

[0147] For example, this calculation process can be represented by the following formula (9):

[0148] σ fading (x)=P carr (x) / (P clean -n l (9)

[0149] Where, σ fading (x) is the multipath channel fading factor; x represents the specified number of symbols; P carr (x) represents the relevant power value; P clean n represents the total power; l This represents the noise threshold power corresponding to the l-th symbol in the data packet.

[0150] In another scenario, if the specified signal-to-noise ratio (SNR) is not less than a first SNR threshold and is less than a second SNR threshold, the electronic device can determine the power adjustment parameters based on the noise threshold power, the specified SNR, and a preset adjustment factor. For example, the power adjustment parameters can be expressed by the following formula (10):

[0151]

[0152] Where w is the power adjustment parameter; γ is the adjustment factor, which can be set between 0 and 1 as needed; n l The signal-to-noise ratio (SNR) represents the noise threshold power corresponding to the l-th symbol in the data packet.

[0153] After obtaining the power adjustment parameters, the electronic device can subtract the power adjustment parameters from the relevant power value and then divide by the sum of the powers to obtain the multipath channel fading factor.

[0154] For example, this calculation process can be represented by the following formula (11):

[0155] σ fading (x)=(P carr (x)-w) / P clean (11)

[0156] Where, σ fading (x) is the multipath channel fading factor; x represents the specified number of symbols; P carr (x) represents the relevant power value; P clean This represents the total power; w is the power adjustment parameter.

[0157] In another case, if the specified signal-to-noise ratio is not less than the second signal-to-noise ratio threshold, the electronic device can divide the relevant power value by the sum of the powers to obtain the multipath channel fading factor. The specific calculation process is as described in the aforementioned formula (2), and will not be repeated here.

[0158] By taking the above measures, the calculation method of the multipath channel fading factor can be adaptively adjusted according to the specified signal-to-noise ratio, reducing the impact of noise and thus obtaining a more accurate multipath channel fading factor.

[0159] In one embodiment, see Figure 8 This is provided as an embodiment of the present application. Figure 2 A detailed flowchart of step 260 is shown below. Figure 8 As shown, when executing step 260, steps 261 to 263 can be executed.

[0160] Step 261: Construct the Doppler frequency shift-time spectrum based on the Doppler frequency shift at multiple observation times.

[0161] Electronic devices can construct a Doppler shift-time spectrum from the Doppler frequency shifts of multiple consecutive observation times. The horizontal axis of the Doppler shift-time spectrum represents time, and the vertical axis represents the Doppler frequency shift.

[0162] Step 262: Cut out the sub-spectral image generated within the specified time period from the Doppler frequency shift-time spectrum as the image to be processed.

[0163] The specified time period can be configured as needed. For example, in an application scenario where all actions to be identified can be completed within 2 seconds, the specified time period would be the most recent 2 seconds.

[0164] The electronic device performs longitudinal cropping of the Doppler frequency shift-time spectrum based on a specified time period to obtain a sub-spectral image, and uses this sub-spectral image as the image to be processed.

[0165] Step 263: Input the image to be processed into the trained behavior recognition model to obtain the behavior recognition result.

[0166] The behavior recognition model can be trained from any network model used for classification, such as CNN (Convolutional Neural Networks), TCN (Temporal Convolutional Network), or a hybrid model of CNN and RNN (Recurrent Neural Networks).

[0167] After obtaining the image to be processed, the image is input into the behavior recognition model, which processes it to obtain the behavior recognition result.

[0168] By employing the above measures, a neural network model can be used to process the Doppler frequency shift at multiple observation times, thereby obtaining behavior recognition results.

[0169] Figure 9 This is a block diagram of a behavior recognition device based on Doppler frequency shift according to an embodiment of the present invention, such as... Figure 9 As shown, the device may include:

[0170] The cleaning module 910 is used to clean the frequency domain channel response of the Wi-Fi signal to obtain the cleaned power delay spectrum and the cleaned channel impulse response corresponding to the cleaned frequency domain channel response.

[0171] The first determining module 920 is used to determine the total power corresponding to the NLOS channel based on the cleaned power delay spectrum;

[0172] The second determining module 930 is used to determine the relevant power value corresponding to the NLOS channel based on the channel impulse response after cleaning.

[0173] The third determining module 940 is used to determine the multipath channel fading factor based on the total power and the relevant power value;

[0174] The fourth determining module 950 is used to determine the corresponding Doppler frequency shift in the zero-order Bessel function based on the multipath channel fading factor;

[0175] The identification module 960 is used to determine the behavior identification result of passive targets based on the Doppler frequency shift at multiple observation times.

[0176] The specific implementation process of the functions and roles of each module in the above-mentioned device can be found in the implementation process of the corresponding steps in the above-mentioned behavior recognition method based on Doppler frequency shift, and will not be repeated here.

[0177] The apparatuses and methods disclosed in the several embodiments provided in this application can also be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of apparatuses, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0178] In addition, the functional modules in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0179] If a function is implemented as a software module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

Claims

1. A behavior recognition method based on Doppler frequency shift, characterized in that, include: For the frequency domain channel response of Wi-Fi signals, the frequency domain channel response is cleaned to obtain the cleaned power delay spectrum and the cleaned channel impulse response corresponding to the cleaned frequency domain channel response; Based on the cleaned power delay spectrum, determine the total power corresponding to the NLOS channel; Based on the channel impulse response after cleaning, the relevant power value corresponding to the NLOS channel is determined. The multipath channel fading factor is determined based on the sum of the power and the relevant power values. Based on the multipath channel fading factor, the corresponding Doppler frequency shift is determined in the zero-order Bessel function; The behavior recognition results of passive targets are determined based on the Doppler frequency shift at multiple observation times; The step of determining the multipath channel fading factor based on the sum of power and the relevant power value includes: Based on the total power and the noise threshold power, a specified signal-to-noise ratio is determined, and the relationship between the specified signal-to-noise ratio and the first and second signal-to-noise ratio thresholds is determined; wherein the second signal-to-noise ratio threshold is greater than the first signal-to-noise ratio threshold. If the specified signal-to-noise ratio is less than the first signal-to-noise ratio threshold, the noise threshold power is subtracted from the sum of power to obtain the adjusted sum of power, and the relevant power value is divided by the adjusted sum of power to obtain the multipath channel fading factor; If the specified signal-to-noise ratio is not less than the first signal-to-noise ratio threshold and is less than the second signal-to-noise ratio threshold, a power adjustment parameter is determined based on the noise threshold power, the specified signal-to-noise ratio and the preset adjustment factor. The power adjustment parameter is then subtracted from the relevant power value and divided by the sum of the powers to obtain the multipath channel fading factor. If the specified signal-to-noise ratio is not less than the second signal-to-noise ratio threshold, the relevant power value is divided by the sum of the powers to obtain the multipath channel fading factor.

2. The method according to claim 1, characterized in that, The frequency domain channel response for the Wi-Fi signal is cleaned to obtain the cleaned power delay spectrum and cleaned channel impulse response, including: The corresponding power delay spectrum is determined for the frequency domain channel response, and the sampling point with the highest power in the power delay spectrum is searched as the centroid sampling point; Several sampling points before the centroid sampling point are truncated and spliced ​​to the end of the power delay spectrum to obtain the adjusted power delay spectrum; Based on a preset first-length front window and a preset second-length rear window, target sampling points are selected from the adjusted power delay spectrum, and sampling points other than the target sampling points are used as noise sampling points; wherein, the front window is placed at the very beginning of the adjusted power delay spectrum, and the rear window is placed at the very end of the adjusted power delay spectrum; The noise threshold power is determined based on the power of multiple noise sampling points; Target sampling points whose power exceeds the noise threshold are selected from the front window and the rear window and designated as sampling points. The front and back ends of the adjusted power delay spectrum are cyclically connected. It is determined whether the distance between the specified sampling point closest to the front end in the back window and the centroid sampling point at the junction exceeds the sampling point number threshold. If so, the designated sampling point closest to the front end within the rear window is determined as the LOS sampling point, and the remaining designated sampling points are determined as NLOS sampling points; Set the power of all sampling points in the power delay spectrum except for the NLOS sampling points to zero to obtain the cleaned power delay spectrum; The time-domain channel estimate values ​​of the sampling points other than the NLSO sampling points in the channel impulse response corresponding to the frequency domain channel response are set to zero to obtain the cleaned channel impulse response.

3. The method according to claim 2, characterized in that, The method further includes: If not, the centroid sampling point is determined to be a LOS sampling point, and the remaining designated sampling points are NLOS sampling points; Return to the step of setting the power of all sampling points in the power delay spectrum except for the NLOS sampling points to zero to obtain the cleaned power delay spectrum.

4. The method according to claim 1, characterized in that, The determination of the relevant power value corresponding to the NLOS channel based on the cleaned channel impulse response includes: The cleaned channel impulse responses of two symbols separated by a specified number of symbols within the same data packet in a Wi-Fi signal are multiplied by their conjugates to obtain the correlation power value corresponding to the NLOS channel.

5. The method according to claim 4, characterized in that, Before determining the corresponding Doppler frequency shift in the zero-order Bessel function based on the multipath channel fading factor, the method further includes: In the case of calculating multiple related power values ​​by using symbol pairs corresponding to multiple specified symbol numbers, the average number of several Doppler frequency shifts corresponding to each specified symbol number is calculated to obtain the average Doppler frequency shift corresponding to each specified symbol number. A new Doppler frequency shift is obtained by weighting the average Doppler frequency shifts corresponding to various specified symbol numbers.

6. The method according to claim 5, characterized in that, The step of weighted averaging of the average Doppler frequency shifts corresponding to multiple specified symbol numbers to obtain a new Doppler frequency shift includes: Based on the Doppler frequency shift corresponding to various specified symbol numbers, a Doppler frequency shift function is obtained by fitting; wherein, the Doppler frequency shift function characterizes the mapping relationship between the specified symbol number and the Doppler frequency shift; The function value corresponding to each specified number of symbols is selected from the Doppler frequency shift function, and a weighted average is performed to obtain the new Doppler frequency shift.

7. The method according to claim 5, characterized in that, The step of weighted averaging of the average Doppler frequency shifts corresponding to multiple specified symbol numbers to obtain a new Doppler frequency shift includes: Based on the target signal-to-noise ratio corresponding to each specified number of symbols, a weighted average is performed on the average Doppler frequency shifts corresponding to multiple specified number of symbols to obtain a new Doppler frequency shift.

8. The method according to claim 1, characterized in that, The determination of the behavior recognition result of the passive target based on the Doppler frequency shift at multiple observation times includes: Construct a Doppler frequency shift-time spectrum based on the Doppler frequency shift at multiple observation times; The sub-spectral image generated within a specified time period is cropped from the Doppler frequency shift-time spectrum and used as the image to be processed; The image to be processed is input into a trained behavior recognition model to obtain the behavior recognition result.

9. An electronic device, characterized in that, The electronic device includes: processor; Memory used to store processor-executable instructions; The processor is configured to execute the behavior recognition method based on Doppler frequency shift as described in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that, The storage medium stores a computer program that can be executed by a processor to perform the behavior recognition method based on Doppler frequency shift as described in any one of claims 1-8.

Images