Identity identification method based on millimeter waves, electronic equipment and identification system

Abstract

The invention belongs to the technical field of identity recognition, and discloses a millimeter wave-based identity recognition method, which comprises the following steps of: S1, acquiring an echo signal received by a millimeter wave radar; s2, acquiring phase information of a reflection target object at a specific position in any antenna signal channel of the echo signal at a certain moment; s3, acquiring a plurality of continuous phase information of a reflection target object at a specific position in any antenna signal channel of the echo signal in a time period according to the operation in the step S2, wherein the plurality of continuous phase information form a phase fluctuation signal of the reflection target object at the specific position; s4, subtracting phase information of adjacent positions of the phase fluctuation signal to obtain a vibration signal of a reflection target object at a specific position; s5, extracting signal features of the vibration signals; and S6, comparing the signal features with a signal feature database or inputting the signal features into a pre-trained target object category identification model to obtain the category of the reflection target object. According to the invention, the frequency domain feature or / and the time domain feature of the vibration signal generated by the reflection target object is / are used as the core feature to realize target object identity recognition, and the recognition accuracy is high.

Description

Technical Field

[0001] This invention relates to the field of biometric identification technology, specifically to a millimeter-wave-based identification method, electronic device, and identification system. Background Technology

[0002] With the large-scale maturation of radio frequency devices, millimeter-wave devices in the 60-120 GHz frequency band have begun to enter the large-scale commercial field, appearing extensively in industries such as smart homes, elderly care, and healthcare. Millimeter-wave devices utilize the characteristics of electromagnetic wave reflection to not only use signal processing to represent reflection points in space using point clouds, thereby tracking the position of targets; they can also acquire vibration signals at a reflection location in a non-physical contact manner, which can be used to measure human respiration and heartbeat.

[0003] The accuracy and quantity of point clouds formed by millimeter-wave radar are far lower than those of visible light and lidar. For a human target within its coverage area, only a few or dozens of points can be formed for the entire body, making it impossible to identify the target based on facial features or limb length. This technical characteristic makes millimeter-wave radar less likely to infringe on personal privacy during application, expanding its application scenarios, such as allowing it to be installed in bedrooms and bathrooms.

[0004] This technical feature poses a significant challenge to identifying "who is who" in practical application scenarios, that is, given a detected target, who exactly is she, him, or it.

[0005] Patent CN 113591760 A calculates step distance, step frequency, and leg distance data for each lower limb by acquiring point cloud data sets and coordinates corresponding to human lower limb movement detected by millimeter-wave radar, and constructs a neural network machine learning model, using human lower limb movement characteristics as the basis for identity recognition. However, this method of identifying human identity through lower limb movement or specific actions has poor recognition rates for body postures and identities beyond specific movements, limiting its application in comprehensive scenarios, especially in home settings. Patent CN 114259225 A utilizes millimeter-wave radar to collect radar echo data from multiple different users; it preprocesses the collected radar echo data to decompose radar cardiac mechanical activity waveform data (RCG); it performs heart rate estimation and heartbeat localization on the RCG waveform data; based on the located heartbeat position, it performs heartbeat segmentation and alignment, ultimately generating a heartbeat template for the current user, which is then normalized and matched with templates in a template library for recognition. However, cardiac mechanical activity waveform data is based on time domain parameters such as wavelength and waveform, resulting in significant data fluctuations and low recognition accuracy.

[0006] Moreover, there are two existing millimeter-wave identity verification methods on the market. One is based on human body contours, the accuracy of which is often directly proportional to the number of RF front-end antenna elements. More antennas not only lead to higher costs and larger size but also cause problems with processing speed and power consumption. The other is based on scenario-specific recognition, such as requiring each person to walk back and forth a few meters along a prescribed route, which greatly limits its application scope. Therefore, there is an urgent need to provide a new millimeter-wave-based identity verification method. Summary of the Invention

[0007] In view of the problems and shortcomings of the existing technology, the purpose of this invention is to provide an identity recognition method, electronic device and recognition system based on millimeter wave.

[0008] To achieve the objectives of this invention, the technical solution adopted is as follows: The first aspect of this invention provides a millimeter-wave-based identity recognition method, comprising the following steps: S1: Acquire the echo signal received by the millimeter-wave radar; S2: Obtain the position information p of the target object reflecting the echo signal at a specific location in any antenna signal channel at a certain moment, which can reflect the distance from the radar. S3: Following the operation of step S2, obtain multiple consecutive position information p of the target object reflecting at a specific position in any antenna signal channel of the echo signal within a time period. The multiple consecutive position information p constitute the position fluctuation signal P of the target object reflecting at a specific position. S4: Expand the position fluctuation signal P obtained in step S3, and subtract the position information of adjacent positions to obtain the vibration signal V of the target object reflected at a specific position; S5: Extract features from the position fluctuation signal P obtained in step S3 or the vibration signal V obtained in step S4 to obtain signal features F. Signal features F are frequency domain features or mixed frequency domain features that include the time domain. S6: Compare the signal feature F with the signal feature database or input the signal feature F into a pre-trained target category recognition model to obtain the category of the reflective target.

[0009] Furthermore, the millimeter-wave-based identification method of this invention extracts frequency features of the change in the relative position of the reflecting target with respect to the millimeter-wave radar, and then performs identification; therefore, information in the echo signal that reflects the change in position between the reflecting target and the millimeter-wave radar at a specific location can be used as position information p. More specifically, the position information p is phase information and / or signal strength information.

[0010] Furthermore, the frequency domain signal features include wavelet transform coefficients. Even further, the wavelet transform coefficients are coefficients obtained by performing wavelet transform processing on the frequency domain signal features.

[0011] Furthermore, the location can be a one-dimensional spatial location, a two-dimensional spatial location, or a three-dimensional spatial location.

[0012] Furthermore, the antenna signal is a physical antenna signal or a virtual antenna signal formed by a physical antenna array.

[0013] Furthermore, the specific operation for obtaining the phase information p of the reflected target at a specific location in any antenna signal channel of the echo signal at a certain moment is as follows: S21: Select any one antenna signal channel in the echo signal and obtain the time domain signal A(n) of the antenna signal channel; S22: Perform an M-point Fourier transform on the time-domain signal A(n) obtained in step S21, and divide the maximum reachable distance of the time-domain signal A(n) into M / 2 distance units; S23: Select any distance unit and obtain the real and imaginary parts of the Fourier transform corresponding to the distance unit; based on the real and imaginary parts, obtain the position information p that can reflect the target object at that moment. Wherein, the position information p is the phase information composed of the real and imaginary parts, or the signal strength information of the distance unit reflected by the real part alone, or the signal strength information of the distance unit reflected by the imaginary part alone, or the strength information constructed by all real parts, or the strength information constructed by all imaginary parts.

[0014] Furthermore, in step S2, before acquiring the position information p of the reflecting target at a specific location in any signal channel of the echo signal at a certain moment, the echo signal needs to be preprocessed. Even further, the preprocessing methods include filtering and AD sampling.

[0015] Furthermore, the echo signal is a signal formed by the reflection of the millimeter-wave radar's transmitted signal onto the target object.

[0016] A second aspect of the present invention provides an electronic device, the electronic device comprising a memory and a processor, wherein the memory stores a program executable on the processor, characterized in that the processor, when executing the program, implements the aforementioned electronic device, characterized in that it comprises a memory and a processor, the memory storing a program executable on the processor, characterized in that the processor, when executing the program, implements the identity recognition method according to any one of claims 1 to 8.

[0017] A third aspect of the present invention provides an identity recognition system, the identity recognition system comprising a millimeter-wave radar and the electronic device described in the second aspect above, the millimeter-wave radar being communicatively connected to the electronic device, the millimeter-wave radar being used to transmit and receive electromagnetic waves into the space to be measured, and to send the received electromagnetic wave signals to the electronic device; the electronic device being used to process the received electromagnetic wave signals to obtain the category of the reflecting target object.

[0018] Compared with the prior art, the technical effects achieved by the present invention are as follows: (1) Due to the limitations of available electromagnetic wave bandwidth and the number of antennas, millimeter wave devices can only distinguish targets with a certain precision (e.g., targets at a distance of 8 cm or an angle of 12 degrees). This invention reflects subtle displacements based on the changes in the phase signal of the reflected target at a specific location. Depending on the millimeter wave band, this displacement can reach a resolution of 0.001 mm, which is sufficient to reflect the positional changes of the chest cavity caused by heartbeat and respiration, i.e., the so-called vibration signal. Therefore, the identification method of this invention uses the frequency domain features and / or time domain features of the vibration signal generated by the reflected target as the core features to achieve target identification. It can distinguish individuals more precisely, and the target identification accuracy is high, overcoming the problem that the BPM feature space is too small to accurately identify individual targets.

[0019] (2) The identity recognition method of the present invention can accurately identify the target as her or him, or even it, without increasing the number of antennas or requiring the target person to perform a specific action. It has high recognition accuracy and high precision. Attached Figure Description

[0020] Figure 1 This refers to the positional fluctuation signal of a reflecting target at different locations within the same antenna signal channel over a period of time; where Target represents the positional fluctuation signal of the reflecting target (human body), Wall represents the positional fluctuation signal of the wall, Noisy Before represents the positional fluctuation signal of the no-object (before the reflecting target (human body) position), and Noisy After represents the positional fluctuation signal of the no-object (after the reflecting target (human body) position). Figure 2 To reflect the vibration signals of the target object, the human body. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the embodiments and accompanying drawings. Here, the illustrative embodiments and descriptions of this invention are used to explain the invention, but are not intended to limit the invention.

[0022] It should also be noted that, in order to avoid obscuring the invention with unnecessary details, only the structures and / or processing steps closely related to the solution according to the invention are shown in the accompanying drawings, while other details that are not closely related to the invention are omitted.

[0023] It should be emphasized that the term "including / comprises" as used herein refers to the presence of a feature, element, step, or component, but does not exclude the presence or addition of one or more other features, elements, steps, or components.

[0024] It should also be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other.

[0025] A millimeter-wave-based identity recognition method includes the following steps: S1: Acquire the echo signal received by the millimeter-wave radar. The echo signal is the signal reflected off the target after the millimeter-wave radar transmits its signal. Millimeter-wave radar equipment can transmit signals into the air according to a certain modulation method (such as FMCW), frequency (such as 62.1GHz), and bandwidth (1GHz). After acquiring the echo signal, it can be preprocessed. Preprocessing methods include filtering and AD sampling.

[0026] S2: Obtain the position information p of the reflective target at a specific location in any antenna signal channel of the echo signal, which reflects its distance from the radar at a certain moment. In this invention, the antenna signal is a physical antenna signal or a virtual antenna signal formed by a physical antenna array. Depending on the number and placement of the antennas in the millimeter-wave radar equipment, and the different echo signal processing algorithms, the position in this invention refers to the one-dimensional, two-dimensional, or three-dimensional spatial position of the reflecting target relative to the millimeter-wave radar equipment.

[0027] The millimeter-wave-based identification method of this invention extracts frequency features of the relative position change of a reflecting target with respect to a millimeter-wave radar, and then performs identification. Therefore, information in the echo signal that reflects the position change of the reflecting target relative to the millimeter-wave radar at a specific location can be used as position information p. Furthermore, the position information p can be phase information and / or signal strength information.

[0028] The specific steps of step S2 are as follows: S21: Select any antenna signal channel n in the echo signal and obtain the time domain signal (i.e., digital signal) A(n) of the antenna signal channel n; S22: Perform an M-point Fourier transform on the time-domain signal A(n) obtained in step S21, and divide the maximum reachable distance of the time-domain signal A(n) into M / 2 distance units; S23: Select any distance unit m (m≤M / 2), and obtain the real and imaginary parts of the Fourier transform corresponding to distance unit m; based on the real and imaginary parts, obtain the position information p that can reflect the target object at that moment. Wherein, the position information p is the phase information composed of the real and imaginary parts, or the signal strength information of the distance unit m reflected by the real part alone, or the signal strength information of the distance unit m reflected by the imaginary part alone, or the strength information constructed by all real parts, or the strength information constructed by all imaginary parts. In this embodiment, the position information p is the phase information composed of the real and imaginary parts.

[0029] S3: Following step S2, acquire multiple consecutive position information p of a target object reflecting a specific location within a time period in any antenna signal channel of the echo signal. These multiple consecutive position information p constitute the position fluctuation signal P of the target object reflecting a specific location. In this embodiment, the position fluctuation signal P of the reflecting target object (such as a human body) within a time period is acquired. Figure 1 As shown in the figure, the position fluctuation signal P of objects at different locations in the same space as the reflecting target was also acquired within a time period.

[0030] S4: Expand the position fluctuation signal P of the reflected target obtained in step S3, and subtract the position information p of adjacent positions to obtain the vibration signal V of the reflected target at a specific position. Figure 2 for Figure 1 The vibration signal V is obtained by unfolding the positional fluctuation signal P of the reflected target object, which is the human body.

[0031] S5: Extract features from the position fluctuation signal P obtained in step S3 or the vibration signal V obtained in step S4 to obtain the frequency domain signal features F.

[0032] S6: Compare the signal feature F with a pre-built signal feature database, or input the signal feature F into a pre-trained target category recognition model (e.g., KVM or deep network model) for comparison to obtain the category of the reflecting target. The features used in the signal feature database or target category recognition model in this invention are all constructed or trained using the scheme of this invention.

[0033] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A millimeter wave-based identification method, characterized by, The method comprises the following steps: S1: obtaining echo signals received by a millimeter wave radar; S2: obtaining position information p of a specific position reflecting target object in a signal channel of the echo signals at a certain time; S3: obtaining continuous position information p of the specific position reflecting target object in a signal channel of the echo signals within a time period according to the operation of step S2, the continuous position information p forming a position fluctuation signal P of the specific position reflecting target object; S4: unfolding the position fluctuation signal P obtained in step S3, and subtracting position information of adjacent positions to obtain a vibration signal V of the specific position reflecting target object; S5: performing feature extraction on the position fluctuation signal P obtained in step S3 or the vibration signal V obtained in step S4 to obtain a signal feature F, the signal feature F being a frequency domain feature or a mixed frequency domain feature containing a time domain; S6: comparing the signal feature F with a signal feature database or inputting the signal feature F into a pre-trained target object category recognition model to obtain a category of the reflecting target object. 2.The millimeter-wave-based identification method of claim 1, wherein, The position information p is phase information or / and signal strength information. 3.The millimeter-wave-based identification method of claim 2, wherein, The frequency domain signal feature comprises a coefficient after wavelet transform. 4.The millimeter-wave-based identification method of claim 2, wherein, The position is a one-dimensional spatial position, a two-dimensional spatial position, or a three-dimensional spatial position. 5.The millimeter-wave-based identification method of claim 4, wherein, The antenna signal is a physical antenna signal or a virtual antenna signal formed by a physical antenna array. 6.The millimeter-wave-based identification method of claim 4, wherein, The specific operation of obtaining the position information p of the specific position reflecting target object in a signal channel of the echo signals at a certain time comprises the following steps: S21: selecting an arbitrary signal channel of the echo signals, and obtaining a time domain signal A(n) of the signal channel; S22: performing M-point Fourier transform on the time domain signal A(n) obtained in step S21, and dividing a maximum reachable distance of the time domain signal A(n) into M / 2 distance units; S23: selecting an arbitrary distance unit, obtaining real and imaginary parts of Fourier transform corresponding to the distance unit, and obtaining the position information p of the reflecting target object at the certain time according to the real and imaginary parts. 7.The millimeter-wave-based identification method of claim 6, wherein, Before step S2, the echo signals need to be preprocessed, and the preprocessing mode comprises filtering and AD sampling. 8.The millimeter-wave-based identification method of claim 7, wherein, The echo signals are signals reflected on the target object after the millimeter wave radar transmits the emission signals.

9. An electronic device, comprising: The electronic device comprises a memory and a processor, and the memory stores a program capable of running on the processor, wherein the processor implements the identity recognition method in any one of claims 1 to 8 when executing the program.

10. An identity recognition system characterized by, The electronic device comprises a millimeter wave radar and the electronic device in claim 9, the millimeter wave radar is in communication connection with the electronic device, the millimeter wave radar is used for transmitting and receiving electromagnetic waves to a space to be measured, and transmitting the received electromagnetic wave signals to the electronic device; and the electronic device is used for processing the received electromagnetic wave signals to obtain the category of the reflecting target object.

Images