Method for measuring wall parameters of ultra-wideband through-the-wall radar

Abstract

The invention discloses an ultra wide band through-the-wall radar wall parameter measurement method. The method is characterized by comprising the following steps: S1, data acquisition and input of a hardware part; s2, clutter and noise suppression; s3, performing data imaging; and S4, identifying and outputting parameters. According to the invention, a feature enhancement technology and a high-precision identification technology of aiming through-the-wall radar human body activity identification are optimized; parameters can be obtained more quickly, calculation is more accurate, and the method has good performance.

Description

Technical Field

[0001] This invention relates to the field of radar technology, and in particular to a method for measuring wall parameters in ultra-wideband through-wall radar. Background Technology

[0002] Through-wall radar is a device that detects targets behind walls or other cover using electromagnetic wave signals. It utilizes the strong penetrating power of ultra-wideband radar signals through non-metallic obstacles such as walls. After penetrating these obstacles, it generates radar echoes upon encountering human targets. Subtle movements such as human motion, breathing, heartbeat, body swaying, and limb movements create a Doppler effect on the echoes. Based on this, the received echoes are processed and analyzed to obtain information such as the location, number, and attitude of the targets behind the wall, which is then displayed on the control terminal client software interface.

[0003] A Chinese patent with publication number CN109696672B discloses a high-resolution through-wall radar imaging method based on spatial structure correlation. The method includes the following steps: acquiring the echo signal of the observed scene received by the nth array element in an N-dimensional uniform linear array; vectorizing the echo signals received by the N array elements and M frequency points; acquiring the echo signal received by the N-dimensional array under multipath propagation caused by electromagnetic wave reflection from the inner wall; downsampling the measurement signal to obtain a sparsely reconstructed observation vector; adding a set of sparse prior distributions to the direct wave and K-1 multipath reflection coefficient vectors to simulate the set sparse structure of the vector; iteratively solving for each parameter using parameter estimation formulas to obtain high-resolution through-wall radar imaging results. This method fully utilizes the set sparse structure of multipath propagation and the spatial distribution structure of target continuity during through-wall radar imaging to acquire high-resolution radar images.

[0004] The aforementioned patent has some advantages, but also some disadvantages: the parameter measurement is complex and the accuracy is low. Summary of the Invention

[0005] In view of the problems mentioned in the background art, the purpose of this invention is to provide a method for measuring wall parameters of ultra-wideband through-wall radar, so as to solve the problems mentioned in the background art.

[0006] The above-mentioned technical objective of the present invention is achieved through the following technical solution:

[0007] A method for measuring wall parameters using ultra-wideband through-wall radar includes the following steps:

[0008] S1, Data acquisition input for the hardware section;

[0009] S2, noise and clutter suppression;

[0010] S3, Data Imaging;

[0011] S4, Parameter Identification Output.

[0012] Preferably, step S1 includes the following steps:

[0013] S11. Determine the coordinate system: In the spherical coordinate system, the target azimuth is represented by (R, α, β), where α is the azimuth angle of the measurement point from the radar, β is the elevation angle of the measurement point from the radar, and R is the distance of the measurement point from the radar.

[0014] S12. Calculation of the distance from the measurement point to the radar: Calculate the distance R from the measurement point to the radar, where R is the distance between the radar and the target, C is the speed of light, and t... r The distance from the measurement point to the radar is expressed as the time delay between the transmitted and received pulses.

[0015] S13. Target Velocity Measurement: Velocity is measured using the Doppler effect, specifically the frequency of the radar's echo signal and the frequency of the main wave transmitted signal. When the target moves radially, there is a frequency difference, defined as: f d =f r -f t ;

[0016] Among them, f d f is the Doppler frequency. r To receive the echo signal frequency, f t For the transmitted signal frequency, S is the radio frequency signal. t (t) has:

[0017] S t (t)=A×cos(2πf c t+φ);

[0018] If we express this as mutual assistance and guidance, then we have:

[0019]

[0020] Radar typically uses linear frequency modulated signals, represented as:

[0021]

[0022] Then its frequency changes with time as f t , is represented as:

[0023] f(t) = f0 + μt.

[0024] Preferably, the angle range of α is 0-360°; the angle range of β is 0-180°.

[0025] Preferably, when S4 outputs the parameter identification, it uses a machine learning method to identify the parameters, and the machine learning algorithm includes SVM, CNN, and ViT.

[0026] Preferably, the process of S2 in suppressing clutter and noise includes: converting the received signal from the frequency domain to the time domain, local oscillator mixing, low-pass filtering, splicing along slow time, and displaying the moving target.

[0027] Preferably, the display of the animated target includes the following steps: First, SVD decomposition is performed: The input image is first split into three subspaces using SVD, which represent the wall subspace, the human motion feature subspace, and the noise subspace, respectively; Second, the image is sequentially input into the coordinate attention and LISTA networks. The coordinate attention embeds the position information into the channels, and the coordinate attention decomposes the channel attention into two one-dimensional feature encoding processes, which aggregate features along two spatial directions respectively.

[0028] Preferably, step S3, during data imaging, includes the following steps:

[0029] Input: The LISTA network receives a corrupted or compressed signal as input;

[0030] Sparse representation initialization: The network first performs a linear transformation on the input signal, mapping it to a sparse representation space;

[0031] Iterative process: The LISTA network improves the quality of sparse representation through multiple iterations. Each iteration includes two steps: sparsification and update.

[0032] Output: Use this sparse representation for signal reconstruction or other subsequent tasks.

[0033] Preferably, in the S3 step, during data imaging, the images of the three subspaces are fused together by summation to determine the weights applied to the images, and the network used is constructed from LeNet.

[0034] In summary, the present invention has the following main beneficial effects:

[0035] This invention optimizes the feature enhancement technology and high-precision recognition technology for human activity recognition in through-wall radar; the parameters are acquired more quickly and the calculation is more accurate, resulting in superior performance. Attached Figure Description

[0036] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0037] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0038] Example 1

[0039] refer to Figure 1 A method for measuring wall parameters using ultra-wideband through-wall radar includes the following steps:

[0040] S1, Data acquisition input for the hardware section;

[0041] S2, noise and clutter suppression;

[0042] S3, Data Imaging;

[0043] S4, Parameter Identification Output.

[0044] Preferably, step S1 includes the following steps:

[0045] S11. Determine the coordinate system: In the spherical coordinate system, the target azimuth is represented by (R, α, β), where α is the azimuth angle of the measurement point from the radar, β is the elevation angle of the measurement point from the radar, and R is the distance of the measurement point from the radar.

[0046] S12. Calculation of the distance from the measurement point to the radar: Calculate the distance R from the measurement point to the radar, where R is the distance between the radar and the target, C is the speed of light, and t... r The distance from the measurement point to the radar is expressed as the time delay between the transmitted and received pulses.

[0047] S13. Target Velocity Measurement: Velocity is measured using the Doppler effect, specifically the frequency of the radar's echo signal and the frequency of the main wave transmitted signal. When the target moves radially, there is a frequency difference, defined as: f d =f r -f t ;

[0048] Among them, f d f is the Doppler frequency. r To receive the echo signal frequency, f t For the transmitted signal frequency St(t), we have:

[0049] S t (t)=A×cos(2πf c t+φ);

[0050] If we express frequency as the derivative of phase, then we have:

[0051]

[0052] Radar typically uses linear frequency modulated signals, represented as:

[0053]

[0054] Then its frequency changes with time as f t , is represented as:

[0055] f(t) = f0 + μt

[0056] Wherein, the angle range of α is 0-360°; the angle range of β is 0-180°.

[0057] Specifically, S4 uses machine learning methods to identify parameters during parameter identification output. The machine learning algorithms include SVM, CNN, and ViT.

[0058] The process of S2 in suppressing clutter and noise includes: converting the received signal from the frequency domain to the time domain, local oscillator mixing, low-pass filtering, splicing along slow time, and displaying the moving target.

[0059] The process of displaying the animated target includes the following steps: First, SVD decomposition is performed: The input image is first split into three subspaces using SVD, which represent the wall subspace, the human motion feature subspace, and the noise subspace, respectively; Second, the image is sequentially input into the coordinate attention and LISTA networks. The coordinate attention embeds the position information into the channels, and the coordinate attention decomposes the channel attention into two one-dimensional feature encoding processes, which aggregate features along two spatial directions respectively.

[0060] Specifically, S3, during data imaging, includes the following steps:

[0061] Input: The LISTA network receives a corrupted or compressed signal as input;

[0062] Sparse representation initialization: The network first performs a linear transformation on the input signal, mapping it to a sparse representation space;

[0063] Iterative process: The LISTA network improves the quality of sparse representation through multiple iterations. Each iteration includes two steps: sparsification and update.

[0064] Output: Use this sparse representation for signal reconstruction or other subsequent tasks.

[0065] In the S3 method, during data imaging, the images of the three subspaces are fused together by summation to determine the weights applied to the images. The network used is constructed from LeNet.

[0066] To obtain the spatial location of targets behind walls and achieve more accurate imaging, through-wall radar typically operates in ultra-wideband frequencies, enabling it to acquire high-resolution signal data. Furthermore, through-wall radar can employ MIMO arrays to create virtual apertures, achieving higher resolution within a limited array aperture. The detection of moving targets by through-wall radar is primarily based on the Doppler effect, where the wavelength of radiation emitted by an object changes with the relative motion of the source and the observer. When the receiver moves towards the transmitting radar, the electromagnetic waves are compressed, resulting in a shorter wavelength and higher frequency received by the receiver. Conversely, when the receiver moves away from the transmitting radar, the electromagnetic waves are stretched, resulting in a longer wavelength and lower frequency received by the receiver. The higher the relative speed, the stronger the Doppler effect, allowing the relative direction of motion to be calculated from the frequency difference. Because the movement of targets behind walls is unpredictable, in order to avoid the influence of the radar's own movement on the detection of target movement, according to the core idea of ​​the controlled variable method, through-wall radar is usually fixed in a certain position to conduct long-term stable observation of the target area. This working mode of through-wall radar is usually called staring mode.

[0067] When through-wall radar propagates through space and walls, the electromagnetic wave signals emitted by the radar undergo reflection, refraction, and attenuation. From the detection and imaging principle of through-wall radar, it can be seen that during the transmission and reception process, the signal undergoes several refractions and reflections between the air and walls. Simultaneously, other objects in the complex surrounding environment also reflect the signal. These refracted and reflected signals, after propagating through the air or walls, are received by the receiver. In other words, the signal received by the receiver contains multiple clutter signals. These clutter signals cause phenomena such as virtual images and ghosting when reconstructing the target; this is the multipath effect of through-wall radar.

[0068] This invention optimizes the feature enhancement technology and high-precision recognition technology for human activity recognition in through-wall radar; the parameters are acquired more quickly and the calculations are more accurate, resulting in superior performance.

[0069] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for measuring wall parameters using ultra-wideband through-wall radar, characterized in that: Includes the following steps: S1, Data acquisition input for the hardware section; S2, noise and clutter suppression; S3, Data Imaging; S4, Parameter Identification Output.

2. The method for measuring wall parameters of ultra-wideband through-wall radar according to claim 1, characterized in that: S1 includes the following steps: S11. Determine the coordinate system: In the spherical coordinate system, the target azimuth is represented by (R, α, β), where α is the azimuth angle of the measurement point from the radar, β is the elevation angle of the measurement point from the radar, and R is the distance of the measurement point from the radar. S12. Calculation of the distance from the measurement point to the radar: Calculate the distance R from the measurement point to the radar, where R is the distance between the radar and the target, C is the speed of light, and t... r The distance from the measurement point to the radar is expressed as the time delay between the transmitted and received pulses. S13. Target Velocity Measurement: Velocity is measured using the Doppler effect, specifically the frequency of the radar's echo signal and the frequency of the main wave transmitted signal. When the target moves radially, there is a frequency difference, defined as: f d =f r -f t ; Among them, f d f is the Doppler frequency. r To receive the echo signal frequency, f t For the transmitted signal frequency, S is the radio frequency signal. t (t) has: S t (t)=A×cos(2πf c t+φ); If we express frequency as the derivative of phase, then we have: Radar typically uses linear frequency modulated signals, represented as: Then its frequency changes with time as f t , is represented as: f(t) = f0 + μt.

3. The method for measuring wall parameters of an ultra-wideband through-wall radar according to claim 1, characterized in that: The angle range of α is 0-360°; the angle range of β is 0-180°.

4. The method for measuring wall parameters of ultra-wideband through-wall radar according to claim 1, characterized in that: When S4 outputs the parameter identification, it uses machine learning methods to identify the parameters. The machine learning algorithms include SVM, CNN, and ViT.

5. The method for measuring wall parameters of an ultra-wideband through-wall radar according to claim 1, characterized in that: The process of S2 in suppressing clutter and noise includes: converting the received signal from the frequency domain to the time domain, local oscillator mixing, low-pass filtering, splicing along slow time, and displaying the moving target.

6. The method for measuring wall parameters of an ultra-wideband through-wall radar according to claim 5, characterized in that: When displaying the animated target, the process includes the following steps: First, SVD decomposition is performed: The input image is first split into three subspaces using SVD, which represent the wall subspace, the human motion feature subspace, and the noise subspace, respectively; Second, the image is sequentially input into the coordinate attention and LISTA networks. The coordinate attention embeds the position information into the channels, and the coordinate attention decomposes the channel attention into two one-dimensional feature encoding processes, which aggregate features along two spatial directions respectively.

7. The method for measuring wall parameters of an ultra-wideband through-wall radar according to claim 1, characterized in that: S3 during data imaging Includes the following steps: Input: The LISTA network receives a corrupted or compressed signal as input; Sparse representation initialization: The network first performs a linear transformation on the input signal, mapping it to a sparse representation space; Iterative process: The LISTA network improves the quality of sparse representation through multiple iterations. Each iteration includes two steps: sparsification and update. Output: Use this sparse representation for signal reconstruction or other subsequent tasks.

8. The method for measuring wall parameters of an ultra-wideband through-wall radar according to claim 6, characterized in that: In the S3 process, the images from the three subspaces are fused together by summation during data imaging to determine the weights applied to the images. The network used is constructed from LeNet.

Images