Signal Processing Method and Device Based on Integrated Communication and Sensing

Abstract

This disclosure proposes a signal processing method and apparatus based on integrated communication and sensing. In response to acquiring an echo signal, the method denoises the echo signal using a pre-constructed sensing model to obtain a target-denoised echo signal, thus removing noise to a certain extent and solving the problem of poor echo signal quality. Further, the target-denoised echo signal is subjected to multilinear decomposition to obtain a subcarrier factor matrix, a symbol factor matrix, and an antenna factor matrix. Based on these matrices, the elevation angle, azimuth angle, range, and velocity of multiple targets included in the echo signal are estimated. By utilizing structured prior knowledge of the signal, stable and high-precision extraction of target parameters from low-quality echo signals is achieved.

Description

Technical Field

[0001] This disclosure relates to the field of signal processing technology, and specifically to a signal processing method and apparatus based on integrated communication and sensing. Background Technology

[0002] Integrated Sensing and Communication (ISAC) refers to the use of existing communication base stations to transmit detection signals to the surrounding environment and to achieve real-time perception of the surrounding environment based on the echo signals of the detection signals.

[0003] Since the bandwidth of a communication base station is primarily designed for data transmission rather than detection, the frame structure (time slots, symbols, and cyclic prefixes, etc.) of the base station is fixed, making it impossible to continuously transmit detection signals indefinitely specifically for environmental sensing. These limitations result in poor echo signal quality, leading to low accuracy in sensing the surrounding environment using traditional estimation algorithms (such as FFT-based spectrum analysis). Summary of the Invention

[0004] This disclosure proposes a signal processing method and apparatus based on integrated communication and sensing.

[0005] The first aspect of this disclosure proposes a signal processing method based on integrated communication and sensing, the method comprising: In response to the acquisition of the echo signal, the echo signal is denoised based on a pre-built sensing model to obtain the target denoised echo signal; the sensing model is constructed based on the system parameters of the communication base station and the channel theory of the antenna array; the system parameters include the number of transmit antennas, the number of receive antennas, the number of subcarriers, the number of coherent processing intervals, the subcarrier interval, and the symbol period; The target denoised echo signal is decomposed into multiple linear components to obtain the subcarrier factor matrix, symbol factor matrix and antenna factor matrix. Based on the subcarrier factor matrix, the symbol factor matrix, and the antenna factor matrix, the elevation angle, azimuth angle, range, and velocity of multiple targets included in the echo signal are estimated.

[0006] In this embodiment of the disclosure, the step of denoising the echo signal based on a pre-built sensing model to obtain a target denoised echo signal includes: The echo signal is converted into a frequency domain echo signal; The pre-acquired transmitted signal is removed from the frequency domain echo signal to obtain the initial denoised echo signal; The echo signal tensor is obtained by stacking the initial denoised echo signals. The echo signal tensor is denoised to obtain the target denoised echo signal.

[0007] In this embodiment of the disclosure, the step of denoising the echo signal tensor to obtain the target denoised echo signal includes: The echo signal tensor is decomposed to obtain the initial factor matrix and the initial core tensor; Update the initial factor matrix and the initial core tensor respectively to obtain the target factor matrix and the target core tensor; The denoised echo signal of the target is obtained using the target factor matrix and the target core tensor.

[0008] In this embodiment of the disclosure, the decomposition of the echo signal tensor to obtain the initial factor matrix and the initial core tensor includes: Expand the echo signal tensor in the antenna dimension to obtain a first matrix; expand the decomposition factor matrix in the symbol dimension to obtain a second matrix; expand the decomposition factor matrix in the subcarrier dimension to obtain a third matrix; The first matrix, the second matrix, and the third matrix are denoised and decomposed respectively to obtain an initial factor matrix; the initial factor matrix includes the denoised first matrix, the denoised second matrix, and the denoised third matrix. The initial factor matrix is ​​obtained based on the initial factor matrix and the echo signal tensor.

[0009] In this embodiment of the disclosure, the step of performing multilinear decomposition on the target denoised echo signal to obtain the subcarrier factor matrix, symbol factor matrix, and antenna factor matrix includes: In the target denoised echo signal, the subcarrier factor matrix and the symbol factor matrix are determined according to the characteristics of the subcarrier and the characteristics of the symbol, respectively; The target denoised echo signal is expanded along the antenna dimension to obtain the observation matrix; the antenna factor matrix is ​​determined based on the observation matrix and the subcarrier factor matrix.

[0010] In this embodiment of the disclosure, determining the subcarrier factor matrix and the symbol factor matrix in the target denoised echo signal based on the characteristics of the subcarrier and the characteristics of the symbol, respectively, includes: For the denoised echo signal of the target, data are extracted from the symbol dimension and subcarrier dimension to obtain a low-rank observation matrix; The low-rank observation matrix is ​​decomposed to obtain the signal subspace matrix; The subcarrier factor matrix is ​​determined in the signal subspace matrix based on the characteristics of the subcarriers, and the symbol factor matrix is ​​determined in the signal subspace matrix based on the characteristics of the symbols.

[0011] In this embodiment of the disclosure, the estimation of the elevation angle, azimuth angle, range, and velocity of multiple targets included in the echo signal based on the subcarrier factor matrix, the symbol factor matrix, and the antenna factor matrix includes: For the antenna factor matrix, the elevation angle corresponding to each angle is determined using a preset angle change rule; among the elevation angles corresponding to each angle change, the elevation angles and directions of multiple targets with the largest elevation angles are determined. For the symbol factor matrix, using a preset velocity change rule, the velocity change value corresponding to each velocity is determined; among the velocity change values, the velocity with the largest angular velocity change value is determined to be the velocity of multiple targets; For the subcarrier factor matrix, a preset distance change rule is used to determine the distance change value corresponding to each distance; among the distance change values, the distance with the largest distance change value is determined to be the distance of multiple targets.

[0012] A second aspect of this disclosure provides a signal processing apparatus based on integrated communication and sensing, the apparatus comprising: The denoising module is used to denoise the echo signal in response to the acquisition of the echo signal, based on a pre-built sensing model, to obtain the target denoised echo signal; the sensing model is constructed based on the system parameters of the communication base station and the channel theory of the antenna array; the system parameters include the number of transmit antennas, the number of receive antennas, the number of subcarriers, the number of coherent processing intervals, the subcarrier interval, and the symbol period; The decomposition module is used to perform multilinear decomposition on the target denoised echo signal to obtain the subcarrier factor matrix, symbol factor matrix and antenna factor matrix. The estimation module is used to estimate the elevation angle, azimuth angle, range, and velocity of multiple targets included in the echo signal based on the subcarrier factor matrix, the symbol factor matrix, and the antenna factor matrix.

[0013] An embodiment of the third aspect of this disclosure provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, the processor executing the computer program to implement the method described in the first aspect above.

[0014] An embodiment of the fourth aspect of this disclosure provides a computer-readable storage medium having a computer program stored thereon, the program being executed by a processor to implement the method described in the first aspect above.

[0015] The technical solutions provided in this disclosure have at least the following technical effects or advantages: In response to the acquired echo signal, the echo signal is denoised based on a pre-built sensing model to obtain a denoised target echo signal, which removes noise to a certain extent and solves the problem of poor echo signal quality. Further, the denoised target echo signal is subjected to multilinear decomposition to obtain a subcarrier factor matrix, a symbol factor matrix, and an antenna factor matrix. Based on the subcarrier factor matrix, the symbol factor matrix, and the antenna factor matrix, the elevation angle, azimuth angle, range, and velocity of multiple targets included in the echo signal are estimated. By utilizing structured prior knowledge of the signal, target parameters can be stably and accurately extracted from low-quality echo signals.

[0016] Additional aspects and advantages of this disclosure will be set forth in part in the description which follows, and in part will be obvious from the description or may be learned by practice of this disclosure. Attached Figure Description

[0017] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this disclosure. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 A schematic flowchart of a signal processing method based on integrated communication and sensing provided in an embodiment of this disclosure is shown. Figure 2A and Figure 2B A schematic diagram of a signal processing method based on integrated communication and sensing provided in an embodiment of this disclosure is shown. Figure 3A The power spectrum of the echo signal in the range-velocity dimension without implementing the Tucker decomposition-based noise suppression method; Figure 3B The power spectrum of the echo signal in the range-velocity dimension for implementing a Tucker decomposition-based noise suppression method; Figure 4A The power spectrum of the echo signal in the 2D angular dimension without performing the Tucker decomposition-based noise suppression method; Figure 4B The power spectrum of the echo signal in the 2D angular dimension for implementing a noise suppression method based on Tucker decomposition; Figure 5 The present disclosure provides an embodiment with a number of transmit and receive antennas. , the number of subcarriers Number of coherent processing intervals Target number The graph shows the relationship between the root mean square error of distance estimation and the signal-to-noise ratio under the given conditions. Figure 6 The present disclosure provides an embodiment with a number of transmit and receive antennas. , the number of subcarriers Number of coherent processing intervals Target number The graph showing the relationship between the root mean square error of speed estimation and the signal-to-noise ratio under the given conditions; Figure 7 The present disclosure provides an embodiment with a number of transmit and receive antennas. , the number of subcarriers Number of coherent processing intervals Target number The graph shows the relationship between the root mean square error of azimuth estimation and the signal-to-noise ratio under the given conditions. Figure 8 The present disclosure provides an embodiment with a number of transmit and receive antennas. , the number of subcarriers Number of coherent processing intervals Target number The graph shows the relationship between the root mean square error of pitch angle estimation and the signal-to-noise ratio under the given conditions. Figure 9 This diagram illustrates the structure of a signal processing device based on integrated communication and sensing according to an embodiment of the present disclosure. Figure 10 A schematic diagram of the structure of an electronic device provided in an embodiment of the present disclosure is shown; Figure 11 A schematic diagram of a storage medium provided according to an embodiment of the present disclosure is shown. Detailed Implementation

[0018] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0019] It should be noted that, unless otherwise stated, the technical or scientific terms used in this disclosure shall have the ordinary meaning as understood by one of ordinary skill in the art to which this disclosure pertains.

[0020] The following describes the technical scenarios involved in the embodiments of this disclosure.

[0021] With the rapid development of 5G / 6G wireless networks and the rise of the low-altitude economy, the applications of UAVs (Unmanned Aerial Vehicles) and other vehicles are becoming increasingly widespread. However, this also brings related problems, making the need for monitoring UAVs and other targets increasingly urgent to address related security and privacy issues. ISAC (Integrated Sensing and Communication), as an emerging technology, enables communication base stations to perform sensing functions by sharing radio resources and hardware, achieving real-time perception of the surrounding environment. It offers advantages such as low cost, wide coverage, and flexible deployment. This integrated characteristic facilitates large-scale coverage sensing, making MIMO (Multiple-Input Multiple-Output) and OFDM (Orthogonal Frequency Division Multiplexing) ISAC systems the mainstream choice for target detection. Therefore, achieving high-precision parameter estimation for multiple targets is crucial.

[0022] However, due to limitations in time-frequency spatial resources and OFDM waveforms, coupled with the large variety of target types and the small RCS (Radar Cross Section) characteristics of each target, the signal-to-noise ratio of the sensed echoes is low, posing a significant challenge to improving the accuracy of traditional estimation algorithms. Among related technologies, periodogram-based methods, such as 3D-FFT (3D-Fast Fourier Transform), struggle to address the resolution grid problem in multi-target scenarios. Specifically, when multiple targets are close together, their echo signals overlap in the time-frequency spatial domain, making effective differentiation difficult.

[0023] In view of this, the present disclosure proposes a signal processing method based on integrated communication and sensing. In the embodiments of the present disclosure, the signal processing method based on integrated communication and sensing proposed in the present disclosure is described below with reference to examples.

[0024] Figure 1 This disclosure illustrates a signal processing method based on integrated communication and sensing, which may include the following steps: In step S11, in response to the acquisition of the echo signal, the echo signal is denoised based on the pre-built perception model to obtain the target denoised echo signal.

[0025] The perception model is constructed based on the system parameters of the communication base station and the channel theory of the antenna array; the system parameters include the number of transmit antennas, the number of receive antennas, the number of subcarriers, the number of coherent processing intervals, the subcarrier interval, and the symbol period.

[0026] For example, the echo signal refers to the wireless signal received by the base station after being reflected by the target, which is a mixture of useful target signals and noise. The pre-built perception model refers to the established tensor-based unified signal model, which reveals that the echo signal can be decomposed into a superposition of multiple target components (each component consists of an angle, velocity, and range factor vector).

[0027] Before denoising the echo signal, a sensing model needs to be constructed. The following process describes how to construct a sensing model. Figure 2A and Figure 2B The image shows a schematic diagram of the signal corresponding to the perception model. Establish system parameters for the communication base station. The base station uses N... t =N tx ×N tz Uniform Planar Array (UPA) Transmit Antenna and N r =N rx ×N rz The UPA receiving antenna uses N subcarriers and K coherent processing intervals, each containing one OFDM symbol. Let it be denoted as the subcarrier spacing. Let this be denoted as the OFDM symbol period. For the k-th OFDM symbol, the ISAC dual-function baseband signal transmitted on the n-th subcarrier is represented as: ,satisfy , where n=0,1,2,…,N-1; k=0,1,2,…,K-1.

[0028] Initialize the sensing scene parameters. The communication base station transmits a shared radio signal and simultaneously senses U targets within the downlink beam. The echo signals reflected by the targets are collected by spatially separated transmit and receive antennas to avoid self-interference. For the u-th target, This is denoted as complex amplitude. This is recorded as the propagation delay. This is denoted as Doppler shift. and These are the guide vectors for the receiving and transmitting arrays, respectively. Furthermore, the propagation delay, Doppler shift, azimuth, and elevation angle of the u-th target are respectively... , , , , and For a single base station setup, the azimuth and elevation angles of arrival are equal to the azimuth and elevation angles of departure, respectively.

[0029] Define the UPA steering vector; the array steering vectors for the target and the base station can be expressed as follows: (1) (2) in , , , . It is the carrier wavelength. It is the antenna spacing between adjacent antenna elements, which is usually half the carrier wavelength.

[0030] Construct the target channel model. The channel response of the nth subcarrier within the kth symbol can be expressed as: (3) Furthermore, after obtaining the echo signal, the representation of the echo signal is generated according to formula (3), and the echo signal is further subjected to initial denoising.

[0031] In some embodiments, the echo signal is denoised based on a pre-built sensing model to obtain a target denoised echo signal, including: converting the echo signal into a frequency domain echo signal; removing the pre-acquired transmitted signal from the frequency domain echo signal to obtain an initial denoised echo signal; stacking the initial denoised echo signals to obtain an echo signal tensor; and denoising the echo signal tensor to obtain the target denoised echo signal.

[0032] For example, the representation of the echo signal generated by formula (3) is obtained by using N-point inverse DFT. Transform to the time domain. Finally, insert. A dotted cyclic prefix (CP) is used to avoid inter-symbol interference (ISI). Therefore, after removing the CP at the base station and performing a discrete Fourier transform, the echo signal is received. It can be represented as: (4) in It is sending the precoding matrix. It's noise.

[0033] Matching the echo signal. By autocorrelation with the transmitted signal at the receiver, the echo signal can be further represented as: (5) in .

[0034] Multi-subcarrier shared transmission precoding is used, meaning that the transmission precoding for multiple OFDM symbols is the same. It can be further expressed as unified. Therefore, by superimposing K OFDM echo symbols in the antenna dimension, the following transformation matrix can be obtained: (6) in This represents the multi-antenna and multi-symbol echo on the nth subcarrier.

[0035] It is sending the precoding matrix.

[0036] It is a diagonal matrix generated by the symbolic Doppler extension. Furthermore, the noise term is reconstructed as... .

[0037] The echo signal tensor can be obtained by converting the echo signal matrix. The third-order echo signal tensor obtained by superposition along the subcarrier dimension It can be represented as: (7) in It's a noise tensor. Factor vector. , and Including azimuth and elevation angle of arrival, Doppler, and time delay (estimated later), it can be expressed as: (8) (9) (10) Formulas (8), (9) and (10) above are expressions for the obtained echo signal tensor.

[0038] The sensing model represents the echo signal as expressed in equations (8), (9), and (10). Based on this, the echo signal is further denoised.

[0039] In some embodiments, denoising the echo signal tensor to obtain a target denoised echo signal includes: decomposing the echo signal tensor to obtain an initial factor matrix and an initial core tensor; updating the initial factor matrix and the initial core tensor respectively to obtain a target factor matrix and a target core tensor; and using the target factor matrix and the target core tensor to obtain the target denoised echo signal.

[0040] Specifically, the echo signal tensor is expanded in the antenna dimension to obtain the first matrix; the factorization matrix is ​​expanded in the symbol dimension to obtain the second matrix; the factorization matrix is ​​expanded in the subcarrier dimension to obtain the third matrix; denoising decomposition is performed on the first, second, and third matrices respectively to obtain the initial factor matrix; the initial factor matrix includes the denoised first, second, and third matrices; based on the initial factor matrix and the echo signal tensor, the initial factor matrix is ​​obtained. For example, the process of decomposing the echo signal tensor refers to constructing a Tucker decomposition model of the echo signal tensor. For instance, the echo signal tensor can be decomposed into the form of formula (11).

[0041] (11) in It is the initial core tensor that embodies the fundamental interactions. , , It is the initial factor matrix corresponding to the antenna, symbol, and subcarrier modes (which are orthogonal). , and Represents the pre-specified Tucker rank for each pattern.

[0042] The specific representation of each core tensor and factor matrix in formula (11) can be determined as follows. For example, the factor matrix is ​​initialized using higher-order singular value decomposition on the echo tensor. This is achieved by expanding along each mode. This allows us to obtain the intermediate variable matrix: (12) Among them, X1 is obtained by expanding from the antenna dimension, X2 is obtained by expanding from the symbol dimension, and X3 is obtained by expanding from the subcarrier number dimension. Let the pattern k be the expansion operator. Then, compute the truncated SVD for each expansion matrix and initialize the factor matrix as follows: (13) in The core tensor is initialized using orthographic projection as follows: (14) The initial factor matrix is ​​the one corresponding to formula (13). , where k is 1, 2, 3. The initial core tensor is shown in formula (14).

[0043] The process of updating the initial factor matrix and the initial core tensor is as follows: Refinement is performed using higher-order orthogonal iterations. The factor matrix is ​​iteratively updated to minimize the reconstruction error. In each iteration, the factor matrix can be calculated as follows: (denoising) (15) in It is an intermediate variable tensor. This represents the SVD and left singular matrix extraction operators. Then, the core tensor can be computed. .

[0044] The denoised tensor is recovered using the optimized core tensor and factor matrix. The denoised tensor is as follows: (16) In step S12, the target denoised echo signal is decomposed into multiple linear components to obtain the subcarrier factor matrix, symbol factor matrix, and antenna factor matrix.

[0045] For example, sliding windows are defined on both the symbol dimension and the subcarrier dimension of the denoised echo signal of the target. Then, by sliding these windows, numerous overlapping sub-data blocks are extracted from the original data. These sub-data blocks are then concatenated into a new, larger observation matrix. This observation matrix effectively increases the sample size of the data, aiding in discorrelation—that is, distinguishing targets with very similar parameters that result in high correlation of the echo signals. This provides the data foundation for the next step of accurately extracting the signal subspace, thereby significantly improving the algorithm's resolution in dense target scenes.

[0046] The constructed observation matrix is ​​then subjected to singular value decomposition (SVD). SVD decomposes the matrix into the product of a left singular vector, a singular value matrix, and a right singular vector. By retaining the left singular vectors corresponding to the top U (where U is the number of targets) largest singular values, the signal subspace matrix is ​​obtained. The signal subspace consists of the principal components with the most concentrated energy in the data, further filtering out residual noise.

[0047] After rearranging the signal subspace matrix, two submatrices U with a fixed phase difference (i.e., "rotation") can be constructed. prev and U next This phase difference is caused by the target's range (causing inter-carrier phase changes) and velocity (causing inter-symbol phase changes). This is achieved by solving U... prev + U next The eigenvalues ​​can be used to directly estimate the phase factor representing the distance, and then reconstruct the subcarrier factor matrix with a Vandermonde structure. Sign factor matrix A similar method can be used for estimation. This method does not require iteration; it directly obtains the parameters by solving an eigenvalue problem, resulting in high computational efficiency and avoiding the risk of convergence in iterative algorithms.

[0048] Construct a new observation matrix Y along the subcarrier dimension Sel To focus on spatial information. Then, calculate its signal subspace. The estimated subcarrier factor matrix is ​​then used. Introducing this known information, the antenna factor matrix is ​​finally calculated using a linear transformation formula combined with new signal subspace components. By incorporating the estimated distance information, the coupling between angle and distance is effectively decoupled, making angle estimation more accurate.

[0049] In some embodiments, the target denoised echo signal is subjected to multilinear decomposition to obtain a subcarrier factor matrix, a symbol factor matrix, and an antenna factor matrix, including: In the target denoised echo signal, the subcarrier factor matrix and symbol factor matrix are determined according to the characteristics of the subcarrier and the symbol, respectively; the target denoised echo signal is expanded in the antenna dimension to obtain the observation matrix; the antenna factor matrix is ​​determined according to the observation matrix and the subcarrier factor matrix.

[0050] Specifically, in the target denoised echo signal, the subcarrier factor matrix and symbol factor matrix are determined according to the characteristics of the subcarrier and the symbol, respectively. This includes: extracting data from the symbol dimension and subcarrier dimension of the target denoised echo signal to obtain a low-rank observation matrix; decomposing the low-rank observation matrix to obtain a signal subspace matrix; determining the subcarrier factor matrix in the signal subspace matrix according to the characteristics of the subcarrier, and determining the symbol factor matrix in the signal subspace matrix according to the characteristics of the symbol.

[0051] For example, a two-dimensional spatial smoothing is used to construct a low-rank observation matrix. The sliding window sizes along the symbol and subcarrier dimensions are defined as follows: and The number of effective sliding steps along the symbol and subcarrier dimensions can be calculated as follows: (17) For by Each sliding window of the index, where Used for symbolic dimension, For the subcarrier dimension, the binary selection matrix is ​​constructed as follows to extract the windowed tensor: (18) Sliding window via Kronecker product The combination choice matrix can be represented as: (19) Then, construct a 3D matrix. The third dimension indexes all Each window. For each selection matrix , , through from Extract windowed symbol-subcarrier data, then concatenate these windows to form a low-rank observation matrix. Calculated as: (20) Signal subspace matrix extraction. For Singular value decomposition is performed to separate the signal subspace from the noise subspace. The first U left singular vectors are retained to obtain the signal subspace matrix. .

[0052] Subcarrier and symbol dimension factor matrix estimation. Let... and It is a factor matrix with a Vandermonde structure in both symbolic and subcarrier dimensions, with its columns being factor vectors. and .

[0053] remember Then, the two sliding window matrices are calculated as follows: (twenty one) (twenty two) The eigenvalue decomposition results are then calculated as follows: (twenty three) The diagonal matrix of eigenvalues ​​is derived. And calculate the normalized generators as follows: (twenty four) Factor matrix The u-th column can be reconstructed as: (25) Factor matrix From In this case, a similar method is used to estimate.

[0054] Antenna dimension factor matrix estimation. Let... This is a factor matrix of a Vandermonde structure with antenna dimensions, and the columns are factor vectors. Define a new selection matrix along the subcarrier dimension. ,in And construct a three-dimensional matrix based on this selection matrix. Then construct a new observation matrix. as follows: (26) Subsequently, calculation SVD, extracting subspace , , and Factor matrix It can be refactored as: (27) in , .

[0055] In step S13, the elevation angle, azimuth angle, range, and velocity of multiple targets included in the echo signal are estimated based on the subcarrier factor matrix, symbol factor matrix, and antenna factor matrix.

[0056] For example, for each column of the antenna factor matrix (representing a spatial eigenvector of a target), its corresponding azimuth and elevation angles need to be estimated. This can be done by searching a pre-defined two-dimensional angle grid. Specifically, a dense grid is generated within the possible azimuth and elevation angle ranges. For each angle pair in the grid, its corresponding ideal steering vector is calculated. Then, the normalized correlation between this ideal steering vector and the column vectors to be estimated in the factor matrix is ​​calculated. The angle corresponding to the grid point with the highest correlation is determined as the estimated elevation and azimuth angles of the target.

[0057] For each column of the subcarrier factor matrix (distance) and the symbol factor matrix (velocity), a more efficient interval-reduction iterative search algorithm is used to estimate the time delay (corresponding to distance) and Doppler shift (corresponding to velocity). This algorithm consists of two stages: The first stage is a coarse search, which searches across the entire possible range of parameters (such as time delay) with a large step size. The normalized correlation ρ(τ) between the factor vector and the ideal vector generated by different parameter values ​​is calculated. The few candidate parameter values ​​with the highest correlation (e.g., two) are retained.

[0058] The second stage is fine-tuning, where the search step size is significantly reduced within a small interval centered on the candidate values ​​obtained from the coarse search, and a fine-tuned grid search is performed again. This process can be iterated multiple times, each time narrowing the interval and increasing the resolution near the previously found best value, until a preset accuracy threshold is reached.

[0059] In some embodiments, estimation is performed based on the subcarrier factor matrix, symbol factor matrix, and antenna factor matrix to obtain the elevation angle, azimuth angle, range, and velocity of multiple targets included in the echo signal. This includes: for the antenna factor matrix, determining the elevation angle corresponding to each angle using a preset angle variation rule; determining the elevation angle and direction of the multiple targets with the largest elevation angle among the elevation angles corresponding to each angle variation; for the symbol factor matrix, determining the velocity variation value corresponding to each velocity using a preset velocity variation rule; determining the velocity of the multiple targets with the largest angular velocity variation value among the velocity variation values; for the subcarrier factor matrix, determining the distance variation value corresponding to each distance using a preset distance variation rule; and determining the distance of the multiple targets with the largest distance variation value among the distance variation values.

[0060] For example, azimuth and pitch angle of arrival estimation. This is achieved through a uniform grid. Find the guiding vector that produces the greatest correlation from the spatial factor matrix. Each column of estimated angle parameters The specific calculations are as follows: (28) in It is a uniform grid. The guiding vector in.

[0061] Set coarse grid resolution ,in and These represent the maximum and minimum radial range values ​​of the sensing area, respectively. The time delay grid used for distance estimation is calculated as follows: (29) in This is the number of coarse grid points. For each The factor vector is constructed as follows: (30) Then, for each signal component u, the normalized correlation is calculated: (31) Among them, the highest The two delays were kept as coarse candidates. Then, the fine-grid resolution calculation is repeated until the interval resolution decreases by a pre-designed threshold to obtain the final refined candidate estimated delay, from which the estimated distance can be seen. Therefore, by using a similar method, distance and velocity... The following estimates are made: (32) To verify the effectiveness of the embodiments of this disclosure, a simulation was conducted using a three-target sensing scenario as an example. The simulation was performed with a subcarrier spacing of 60 kHz, a carrier frequency of 4.9 GHz, 64 subcarriers, and a CPI(K) count of 16. It was configured to select one symbol for sensing within every five standard time slots. Each standard time slot consists of 14 symbols. The CP duration was set to 1 / 4 of the OFDM symbol duration. Both the transceiver and transmit antenna arrays were... The distances, velocities, pitch angles, and azimuth angles of targets 1, 2, and 3 are randomly distributed in [50m, 500m], [2m / s, 10m / s], [5°, 35°], and [5°, 35°]. The Tucker rank is set to 3 in advance, and the number of higher-order orthogonal iterations is 5.

[0062] The three-target scene simulation was tested on test datasets with signal-to-noise ratios of -15dB, -10dB, -5dB, 0dB, 5dB, 10dB, and 15dB.

[0063] Algorithms that can be compared with the method described in this invention include Vandermonde-based CP decomposition (VCD), traditional CP-ALS, Tucker-ALS based on Tucker decomposition, and improved 3D-FFT.

[0064] As shown in Figures 3 and 4, the noise severely obscures the useful signal, causing targets 1, 2, and 3 to appear as blurry, disordered energy clusters intertwined with random noise fluctuations. By employing the noise suppression method based on Tucker decomposition, the range-velocity and 2D angular contour features become clearer, with concentrated energy clusters and well-defined boundaries between targets. Experimental results demonstrate that the embodiments of this disclosure (the method shown in the figures) effectively improve the signal-to-noise ratio and sharpen target features in both the range-velocity and 2D angular domains, laying the foundation for accurate estimation of multi-target parameters.

[0065] Depend on Figure 5 , Figure 6 , Figure 7 as well as Figure 8As can be seen, with the increase of signal-to-noise ratio (SNR), the root mean square error (RMSE) of the 4D parameter estimation of the method and the comparison algorithm of the present invention gradually decreases. However, the embodiments of the present disclosure consistently exhibit the lowest RMSE across the entire SNR range, indicating higher accuracy in estimating all four parameters. Taking the case with an SNR of 0 dB as an example, the distance estimation RMSE of the method is approximately 1.49 m, and the velocity estimation RMSE is 0.026 m / s, significantly lower than the velocity estimation RMSE of the VCD method (0.071 m / s). The azimuth and elevation angle arrival estimation RMSEs of the method are 0.22 degrees and 0.25 degrees, respectively, also significantly better than the 0.58 degrees and 0.35 degrees from the best comparison method.

[0066] As can be seen, by implementing the above-described signal processing method based on integrated communication and sensing, in response to the acquisition of echo signals, the echo signals are denoised using a pre-built sensing model to obtain denoised target echo signals, which removes noise to a certain extent and solves the problem of poor echo signal quality. Furthermore, the denoised target echo signals are subjected to multilinear decomposition to obtain subcarrier factor matrices, symbol factor matrices, and antenna factor matrices. Based on these matrices, the elevation angle, azimuth angle, range, and velocity of multiple targets included in the echo signals are estimated. By utilizing structured prior knowledge of the signal, target parameters can be stably and accurately extracted from low-quality echo signals.

[0067] Corresponding to the above implementation of the signal processing method based on integrated communication and sensing, this disclosure also provides a signal processing device based on integrated communication and sensing, which is used to perform the above-described signal processing method. Figure 1 The illustrated embodiment is a signal processing method based on integrated communication and sensing. For example... Figure 9 As shown, the signal processing device based on integrated communication and sensing includes: The denoising module 901 is used to denoise the echo signal based on a pre-built perception model in response to the acquisition of the echo signal, so as to obtain the target denoised echo signal. The decomposition module 902 is used to perform multilinear decomposition on the target denoised echo signal to obtain the subcarrier factor matrix, symbol factor matrix and antenna factor matrix. The estimation module 903 is used to estimate the elevation angle, azimuth angle, range and velocity of multiple targets included in the echo signal based on the subcarrier factor matrix, the symbol factor matrix and the antenna factor matrix.

[0068] As can be seen, by executing the above-mentioned signal processing device based on integrated communication and sensing, in response to acquiring the pathological image to be processed, the mask region of the pathological image to be processed is extracted. Since the mask region represents the region of the patient's identity characteristics corresponding to the pathological image to be processed, by using the pre-trained adjustment model to adjust the mask region corresponding to the pathological image to be processed, the generated target privacy-preserving image only adjusts the region of identity characteristics and does not adjust other regions, thus removing the patient's identity information while retaining the relevant pathological features in the pathological image.

[0069] The signal processing device based on integrated communication and sensing provided in the above embodiments of this disclosure and the signal processing method based on integrated communication and sensing provided in the embodiments of this disclosure are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the applications stored therein.

[0070] This disclosure also provides an electronic device for performing the above-described signal processing method based on integrated communication and sensing. Please refer to... Figure 10 This illustrates a schematic diagram of an electronic device provided by some embodiments of the present disclosure. For example... Figure 10 As shown, the electronic device includes: a processor 1000, a memory 1001, a bus 1002, and a communication interface 1003. The processor 1000, the communication interface 1003, and the memory 1001 are connected via the bus 1002. The memory 1001 stores a computer program that can run on the processor 1000. When the processor 1000 runs the computer program, it executes the aforementioned provisions of this disclosure. Figure 1 The illustrated embodiment provides a signal processing method based on communication and sensing integration.

[0071] The memory 1001 may include high-speed random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Communication between this system network element and at least one other network element is achieved through at least one communication interface 1003 (which can be wired or wireless), such as the Internet, wide area network, local area network, or metropolitan area network.

[0072] Bus 1002 can be an ISA bus, PCI bus, or EISA bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. Memory 1001 is used to store programs, and the processor 1000 executes the programs after receiving execution instructions. Figure 1 The illustrated embodiment reveals that the signal processing method based on communication and sensing integration can be applied to or implemented by the processor 1000.

[0073] The processor 1000 may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by the integrated logic circuitry in the hardware of the processor 1000 or by instructions in software form. The processor 1000 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), an off-the-shelf programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this disclosure. The general-purpose processor may be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this disclosure can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules may reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. The storage medium is located in memory 1001. Processor 1000 reads the information in memory 1001 and, in conjunction with its hardware, completes the steps of the above method.

[0074] The electronic device provided in this disclosure and the signal processing method based on integrated communication and sensing provided in this disclosure are based on the same inventive concept and have the same beneficial effects as the methods they employ, operate, or implement.

[0075] This disclosure also provides a computer-readable storage medium corresponding to the signal processing method based on integrated communication and sensing provided in the foregoing embodiments. Please refer to... Figure 11 The computer-readable storage medium shown is an optical disc 30, on which a computer program (i.e., a program product) is stored. When the computer program is run by a processor, it executes the signal processing method based on communication and sensing integration provided in any of the foregoing embodiments.

[0076] It should be noted that examples of the computer-readable storage medium may also include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other optical and magnetic storage media, which will not be elaborated here.

[0077] The computer-readable storage medium provided in the above embodiments of this disclosure and the signal processing method based on communication sensing integration provided in the embodiments of this disclosure are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the applications stored therein.

[0078] It should be noted that: Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of this disclosure may be practiced without these specific details. In some instances, well-known structures and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0079] Similarly, it should be understood that, in order to simplify this disclosure and aid in understanding one or more of the various inventive aspects, in the above description of exemplary embodiments of this disclosure, various features of this disclosure are sometimes grouped together in a single embodiment, figure, or description thereof. However, this approach to disclosure should not be construed as reflecting a schematic diagram in which the claimed disclosure requires more features than are expressly recited in each claim. Rather, as reflected in the following claims, inventive aspects lie in fewer than all features of a single foregoing disclosed embodiment. Therefore, the claims following the detailed description are hereby expressly incorporated into that detailed description, wherein each claim itself is a separate embodiment of this disclosure.

[0080] Furthermore, those skilled in the art will understand that although some embodiments described herein include certain features included in other embodiments but not others, combinations of features from different embodiments are intended to be within the scope of this disclosure and form different embodiments. For example, in the following claims, any of the claimed embodiments can be used in any combination.

[0081] The above description is merely a preferred embodiment of this disclosure, but the scope of protection of this disclosure is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this disclosure should be included within the scope of protection of this disclosure. Therefore, the scope of protection of this disclosure should be determined by the scope of the claims.

Claims

1. A signal processing method based on integrated communication and sensing, characterized in that, The method includes: In response to the acquisition of the echo signal, the echo signal is denoised based on a pre-built sensing model to obtain the target denoised echo signal; The target denoised echo signal is decomposed into multiple linear components to obtain the subcarrier factor matrix, symbol factor matrix and antenna factor matrix. Based on the subcarrier factor matrix, the symbol factor matrix, and the antenna factor matrix, the elevation angle, azimuth angle, range, and velocity of multiple targets included in the echo signal are estimated.

2. The method according to claim 1, characterized in that, The step of denoising the echo signal based on a pre-built perception model to obtain the target denoised echo signal includes: The echo signal is converted into a frequency domain echo signal; The pre-acquired transmitted signal is removed from the frequency domain echo signal to obtain the initial denoised echo signal; The echo signal tensor is obtained by stacking the initial denoised echo signals. The echo signal tensor is denoised to obtain the target denoised echo signal.

3. The method according to claim 2, characterized in that, The step of denoising the echo signal tensor to obtain the target denoised echo signal includes: The echo signal tensor is decomposed to obtain the initial factor matrix and the initial core tensor; Update the initial factor matrix and the initial core tensor respectively to obtain the target factor matrix and the target core tensor; The denoised echo signal of the target is obtained using the target factor matrix and the target core tensor.

4. The method according to claim 3, characterized in that, The process of decomposing the echo signal tensor to obtain the initial factor matrix and the initial core tensor includes: Expand the echo signal tensor in the antenna dimension to obtain a first matrix; expand the decomposition factor matrix in the symbol dimension to obtain a second matrix; expand the decomposition factor matrix in the subcarrier dimension to obtain a third matrix; The first matrix, the second matrix, and the third matrix are denoised and decomposed respectively to obtain an initial factor matrix; the initial factor matrix includes the denoised first matrix, the denoised second matrix, and the denoised third matrix. The initial factor matrix is ​​obtained based on the initial factor matrix and the echo signal tensor.

5. The method according to claim 1, characterized in that, The multilinear decomposition of the denoised echo signal of the target to obtain the subcarrier factor matrix, symbol factor matrix, and antenna factor matrix includes: In the target denoised echo signal, the subcarrier factor matrix and the symbol factor matrix are determined according to the characteristics of the subcarrier and the characteristics of the symbol, respectively; The target denoised echo signal is expanded along the antenna dimension to obtain the observation matrix; the antenna factor matrix is ​​determined based on the observation matrix and the subcarrier factor matrix.

6. The method according to claim 5, characterized in that, The step of determining the subcarrier factor matrix and the symbol factor matrix in the target denoised echo signal based on the characteristics of the subcarrier and the symbol, respectively, includes: For the denoised echo signal of the target, data are extracted from the symbol dimension and subcarrier dimension to obtain a low-rank observation matrix; The low-rank observation matrix is ​​decomposed to obtain the signal subspace matrix; The subcarrier factor matrix is ​​determined in the signal subspace matrix based on the characteristics of the subcarriers, and the symbol factor matrix is ​​determined in the signal subspace matrix based on the characteristics of the symbols.

7. The method according to claim 1, characterized in that, The estimation based on the subcarrier factor matrix, the symbol factor matrix, and the antenna factor matrix to obtain the elevation angle, azimuth angle, range, and velocity of multiple targets included in the echo signal includes: For the antenna factor matrix, the elevation angle corresponding to each angle is determined using a preset angle change rule; among the elevation angles corresponding to each angle change, the elevation angles and directions of multiple targets with the largest elevation angles are determined. For the symbol factor matrix, using a preset velocity change rule, the velocity change value corresponding to each velocity is determined; among the velocity change values, the velocity with the largest angular velocity change value is determined to be the velocity of multiple targets; For the subcarrier factor matrix, a preset distance change rule is used to determine the distance change value corresponding to each distance; among the distance change values, the distance with the largest distance change value is determined to be the distance of multiple targets.

8. The method according to claim 1, characterized in that, The perception model is constructed based on the system parameters of the communication base station and the channel theory of the antenna array; the system parameters include the number of transmit antennas, the number of receive antennas, the number of subcarriers, the number of coherent processing intervals, the subcarrier interval, and the symbol period.

9. A signal processing device based on integrated communication and sensing, characterized in that, The device includes: The denoising module is used to denoise the echo signal based on a pre-built perception model in response to the acquisition of the echo signal, so as to obtain the target denoised echo signal. The decomposition module is used to perform multilinear decomposition on the target denoised echo signal to obtain the subcarrier factor matrix, symbol factor matrix and antenna factor matrix. The estimation module is used to estimate the elevation angle, azimuth angle, range, and velocity of multiple targets included in the echo signal based on the subcarrier factor matrix, the symbol factor matrix, and the antenna factor matrix.

Images