A near-field channel parameter estimation method and system considering spatial non-stationary characteristics

Abstract

The application provides a near-field channel parameter estimation method considering spatial non-stationary characteristics, which comprises the following steps: obtaining a multipath parameter initial value by using a far-field SAGE algorithm; calculating propagation distances of array elements at a transmitting end and a receiving end to scattering points, constructing a near-field channel response, and performing refined estimation on angle and distance parameters, time delay parameters and complex gain parameters of the multipath. Correlation operation is performed on a current residual signal and the near-field channel response in a frequency domain to obtain an array element level correlation metric. Continuous spatial non-stationary coefficients with a value of 0 to 1 are updated based on the correlation metric, and weighting iteration is performed on residual construction and parameter updating until a stop criterion is met, and near-field geometric parameters, time delay parameters, complex gain parameters of each multipath and spatial non-stationary coefficients of both transmitting and receiving ends are output. The method solves the problem of inconsistent multipath visible domains under XL-MIMO arrays, can suppress the influence of invisible array elements on parameter estimation, and improves the estimation accuracy and stability of the SAGE algorithm under the condition of near-field spatial non-stationary channels.

Description

Technical Field

[0001] This invention belongs to the field of wireless communication channel measurement and parameter estimation technology, and relates to a near-field channel parameter estimation method and system that considers spatial non-stationary characteristics. Background Technology

[0002] As array apertures increase and operating frequencies expand to higher frequencies such as millimeter waves and terahertz, very large-scale arrays (VLS) can improve beamforming gain and spatial resolution. However, they also make electromagnetic propagation more prone to exhibiting near-field spherical wave characteristics, potentially introducing modeling biases under the far-field plane wave assumption. Under broadband conditions, the coupling between the spatial and frequency dimensions becomes more pronounced. Near-field broadband effects further increase the difficulty of jointly estimating parameters such as angular distance and time delay, and may lead to a decline in estimation performance. On the other hand, VLS may also exhibit spatial non-stationarity, manifesting as the same scatterer or multipath component contributing significant energy only in certain regions of the array, forming visible domain characteristics. If the entire array observation is treated with equal weighting during estimation, invisible or weakly visible elements may introduce additional interference, thus affecting the estimation stability of parameters such as angular distance, time delay, and gain.

[0003] In existing technologies, some methods employ simplified near-field models or ignore spatial non-stationary factors, making it difficult to simultaneously consider both broadband near-field effects and visible domain differences. Some methods use sparse representations or dictionary searches for near-field channel estimation, but under near-field conditions, angular domain sparsity may weaken, introducing new estimation challenges. Other works consider the problem of spatially non-stationary channel parameter estimation within an iterative estimation framework, but trade-offs still exist regarding parameter dimensionality, update strategies, and computational complexity.

[0004] Therefore, there is an urgent need to propose a broadband near-field channel parameter estimation method that considers the spatial non-stationary characteristics, so as to better characterize the spatial non-stationary near-field channel under ultra-large-scale arrays and improve the accuracy and stability of parameter estimation. Summary of the Invention

[0005] In view of this, the purpose of this invention is to provide a near-field channel parameter estimation method and system that considers spatial non-stationary characteristics. Based on a broadband near-field spherical wave channel model, continuous spatial non-stationary coefficients at both the transmitting and receiving ends are introduced to characterize the visible domain differences of multipath propagation on the array. First, the far-field SAGE algorithm is used to obtain initial values ​​for parameters such as angle and time delay. Then, the near-field SAGE algorithm is used to refine the multipath geometric parameters and time delay parameters, thereby obtaining more reliable initialization results. Subsequently, joint iteration is performed under the log-likelihood criterion, updating the spatial non-stationary coefficients using array element-level normalized correlation metrics. Under the weighted condition involving these spatial non-stationary coefficients, the multipath angle parameters, distance parameters, time delay parameters, and complex gain parameters are iteratively updated until the stopping criterion is met, at which point the multipath parameters and the spatial non-stationary coefficients at both the transmitting and receiving ends are output. Through the above process, this invention can more accurately characterize spatially non-stationary channels and improve the stability and accuracy of parameter estimation.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A near-field channel parameter estimation method considering spatial non-stationary characteristics, the method specifically includes the following steps: S1. Obtain broadband frequency domain observation signals and the geometric coordinate information of the array elements of the transmitting and receiving arrays; S2. During the initialization phase, the far-field SAGE algorithm is executed based on the far-field plane wave channel model to obtain the initial values ​​of the multipath parameters. S3. Based on the near-field spherical wave channel model, calculate the propagation distance between the array elements at the transmitting and receiving ends and construct the near-field channel response of each multipath. At the same time, introduce the spatial non-stationary coefficients at both the transmitting and receiving ends into the model. S4. Under the condition that the spatial non-stationary coefficient is fixed at 1, the near-field SAGE algorithm is used to refine the estimation of the near-field geometric parameters, time delay parameters and complex gain parameters of each multipath, and the parameter refinement results are obtained. S5. Based on the refined parameter results, enter the joint iteration stage. Perform correlation calculation on the residual and the near-field channel response in the frequency domain to obtain the array element-level correlation metric. Update the continuous spatial non-stationary coefficient with a value of 0 to 1 based on the array element-level correlation metric. Update the multipath channel parameters under the weighted condition involving the spatial non-stationary coefficient. S6. Repeat step S5 until the stopping criterion is met, and output the near-field geometric parameters, time delay parameters, complex gain parameters, and spatial non-stationary coefficients of each multipath.

[0007] Furthermore, step S2 specifically includes: S21. Based on the far-field plane wave channel model, perform preliminary estimation of each multipath of the observed signal and obtain angle parameters. , Initial values ​​of delay parameters ; S22. Based on the initial values ​​of the angle parameter and time delay parameter, the initial values ​​of the complex gain parameter of each multipath are calculated according to the maximum likelihood criterion. .

[0008] Furthermore, the spatially non-stationary near-field spherical wave channel model described in step S3 satisfies: at the frequency point Place, No. l Multipath at the receiver m Each element and the transmitter n The frequency domain channel response between the array elements is (1) in and These are the spatial non-stationary coefficients for the receiver and transmitter, respectively. For complex gain, and For the responses of the receiving and transmitting antennas, the propagation delay satisfies (2) in c At the speed of light, For the first l The time delay of the stripe and the propagation distance at the array element level satisfy the following conditions: (3) (4) in and The receiving end is the first m Array element and transmitter n The coordinate vector of an array element relative to a reference array element. For the first l The distance from the last hop scatterer of the multipath to the reference element at the receiver. No. l The distance from the first hop scatterer of the multipath to the reference element at the transmitter; the unit vector of the element-level direction is... (5) (6) And the direction unit vector From pitch angle and azimuth Confirmed, the expression is (7).

[0009] Furthermore, in terms of frequency points First l The frequency domain channel matrix of a multipath is defined as follows: (8) The total frequency domain channel matrix satisfies: (9) Furthermore, regarding the first l Multipath at frequency points Definition without complex gain Template matrix: (10) And construct a spatially nonstationary matrix: (11) in Thus the first l The matrix form of a multipath can be written as: (12) Where ⊙ represents Hadamard element-wise multiplication.

[0010] Furthermore, step S4 specifically includes: S41. Fix the spatial nonstationary coefficients at both the transmitting and receiving ends to 1, that is, for any multipath and any array element index... m , n ,satisfy , and at frequency points The first structure l Multipath without complex gain Path response matrix ; S42, Construction of the l Residual observation matrix of multiple paths ,satisfy (13) in For frequency point The observed channel matrix at that location, for i The reconstruction matrix of the multipath; S43, Definition of the l The log-likelihood function of the multipath is: (14) in , For noise variance, For parameter set (15) S44, Near-field SAGE estimation, given , Under the conditions, and let Where D is the array aperture and λ is the carrier wavelength corresponding to the center frequency, the receiver parameters are updated. ,satisfy (16) S45, In the given , , Under the condition of updating the transmitter parameters ,satisfy (17) S46, In the given , , , Under the condition of updating the delay parameter ,,satisfy (18) S47. Given, satisfying , , , , Under the condition of complex gain The log-likelihood function right Find the partial derivative and set it to 0. The complex gain can be obtained. The maximum likelihood estimate is a closed-form solution: (19) in .

[0011] Furthermore, step S5 specifically includes: S51. Initialize the joint iteration variables: Let the initial state at the start of the joint iteration be marked as round 1 (i.e., µ =1), the refined parameter result obtained in step S4 is directly used as the initial multipath parameter for the joint iteration, that is (20) Meanwhile, assuming the entire array energy is stationary in the initial state, the initial values ​​of the spatial non-stationarity coefficients at both the transmitting and receiving ends are set to... (twenty one) S52. Given the spatial non-stationary coefficients from the previous round, update the geometric and time delay parameters; in the... µ In rounds of iteration ( ), using the non-stationary matrix from the previous round Construct the equivalent weighted path response matrix: (twenty two) Under the condition of the equivalent weighted path response matrix, the transmitter angle and distance, receiver angle and distance, and time delay parameters of the multipath are sequentially searched and updated in the neighborhood to obtain the first... µ Updated values ​​of the wheel's geometry and time delay parameters; S53. Estimate the complex gain of the current round based on the non-stationary weights updated with delay; after obtaining the... µ After considering the geometric and time delay parameters of the wheel, and combining them with the spatial non-stationary matrix of the previous wheel... By taking the partial derivative of the log-likelihood function with respect to the complex gain and setting it to zero, we obtain the first... µ Cyclic gain Closed-form solution: (twenty three) S54. Based on the updated multipath parameters, estimate the spatial nonstationary factor of the current round: in the... µ After the multipath angle, distance, time delay, and complex gain of the wheel are updated, the clean signal of the current wheel is reconstructed using these parameters. The element-level correlation metrics at the receiver and transmitter are calculated separately using a method that "introduces coefficients from the other end when calculating one end's metric," satisfying the following conditions: (twenty four) (25) Determine the noise threshold based on the noise level. The effective signal threshold is obtained by adding 6dB to the noise threshold. ,satisfy (26) Construct the maximum normalized correlation metrics for the receiver and transmitter respectively. (27) The normalized correlation metric is normalized and truncated to the interval [0, 1] using the effective signal threshold as the lower bound, resulting in updated values ​​of non-stationary coefficients in continuous space, satisfying the following conditions: (28) (29) in To prevent constants with a denominator of 0; S55, Block-based alternating update and gating matrix construction; To ensure iterative stability, an alternating update strategy is adopted for the non-stationary coefficients at both ends of the transmission and reception: First, in a given... Update under the condition Then in the given Update under the condition ; and update ; S56. Update the multipath parameters and residuals under a given gating matrix: Under a given gating matrix... Under these conditions, the updated geometric parameters, time delay parameters, and complex gain parameters are combined to reconstruct the first... l Multipath at frequency points The contribution matrix at point satisfies (30) Based on this, the residual observation matrix is ​​updated according to the residual construction method in step 42. This is used for the next iteration.

[0012] Furthermore, the joint iteration in step S5 uses the gain of the log-likelihood function as the stopping criterion. Iteration stops when the gain of the log-likelihood function in two consecutive iterations is not greater than a preset threshold, satisfying... (31) in This is a preset threshold.

[0013] The present invention also provides a near-field channel parameter estimation system that takes into account spatial non-stationary characteristics.

[0014] The beneficial effects of this invention are as follows: 1) This invention addresses the parameter estimation requirements of broadband near-field ultra-large-scale array channels. Based on the near-field spherical wave parameterized channel model, it introduces continuous spatial non-stationary coefficients at both the transmitting and receiving ends to characterize the visible domain differences of multipaths on the array. This effectively suppresses the interference of weakly visible array elements on parameter estimation and improves the estimation robustness under spatial non-stationary conditions.

[0015] 2) This invention adopts a phased estimation process that combines far-field SAGE initialization and near-field SAGE refinement. Before entering the spatial non-stationary joint iteration, relatively reliable initial parameter values ​​are obtained first, which reduces the instability caused by direct joint estimation of high-dimensional coupled parameters and improves the stability of iterative convergence.

[0016] 3) In the joint iteration stage, the present invention constructs a normalized correlation metric at the array element level to update the spatial non-stationary coefficients, and adopts a noise threshold plus a 6 dB effective signal threshold constraint weight mapping to make the visibility update insensitive to antenna gain differences and amplitude scale changes, avoiding misjudging gain differences as visibility domain differences, thereby further improving the estimation accuracy and consistency of multipath angle parameters, distance parameters, time delay parameters and complex gain parameters.

[0017] Other advantages, objectives, and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination, or may be learned from practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description

[0018] To make the objectives, technical solutions, and advantages of the present invention clearer, the preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings, wherein: Figure 1 This is a flowchart illustrating a broadband near-field channel parameter estimation method considering spatial non-stationary characteristics according to an embodiment of the present invention. Figure 2 This is a schematic diagram illustrating the specific execution flow of the near-field spatial non-stationary SAGE algorithm provided in an embodiment of the present invention. Detailed Implementation

[0019] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0020] The accompanying drawings are for illustrative purposes only and are schematic diagrams, not actual pictures. They should not be construed as limiting the invention. To better illustrate the embodiments of the invention, some parts in the drawings may be omitted, enlarged, or reduced, and do not represent the actual product dimensions. It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0021] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components. In the description of the present invention, it should be understood that if terms such as "upper," "lower," "left," "right," "front," and "rear" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting the present invention. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.

[0022] Please see Figure 1 and Figure 2 This invention provides a broadband near-field channel parameter estimation method that considers spatial non-stationary characteristics.

[0023] This embodiment first introduces a broadband near-field channel parameter estimation method that considers spatial non-stationary characteristics, such as... Figure 1As shown, the method includes the following steps: Step 1: Obtain the broadband frequency domain observation signal and the geometric coordinate information of the array elements of the transmitting and receiving arrays; Step 2: In the initialization phase, the far-field SAGE algorithm is executed based on the far-field plane wave channel model to obtain the initial values ​​of the multipath parameters; Step 3: Calculate the propagation distance between the array elements at the transmitting and receiving ends based on the near-field spherical wave channel model and construct the near-field channel response for each multipath. At the same time, introduce the spatial non-stationary coefficients at both the transmitting and receiving ends into the model. Step 4: Under the condition that the spatial non-stationary coefficient is fixed at 1, the near-field SAGE algorithm is used to refine the estimation of the near-field geometric parameters, time delay parameters and complex gain parameters of each multipath, and the parameter refinement results are obtained. Step 5: Based on the refined parameter results, proceed to the joint iteration stage. Perform correlation calculation on the residual and the near-field channel response in the frequency domain to obtain the array element-level correlation metric. Update the continuous spatial non-stationary coefficients with values ​​from 0 to 1 based on the array element-level correlation metric, and update the multipath channel parameters under the weighted condition involving the spatial non-stationary coefficients. Step 6: Repeat step S5 until the stopping criterion is met, and output the near-field geometric parameters, time delay parameters, complex gain parameters, and spatial non-stationary coefficients of each multipath.

[0024] In this embodiment, step 1 specifically includes the following detailed steps: In step 1, the broadband frequency domain observation signal and the element geometric information of the transceiver array are acquired. In this embodiment, the center frequency is taken as... f c =15 GHz, bandwidth taken B =1 GHz, the number of frequency domain sampling points is taken as K =511 to match A PN sequence structure of length, with frequency point spacing of... and define the frequency point set. (1) At frequency f k The observation channel matrix at point is denoted as In this embodiment, the number of array elements at the receiving end is taken. M =256, number of transmitting array elements N =128. Obtain the three-dimensional element coordinate information of the transceiver array and establish a three-dimensional coordinate system. The receiver... m The coordinate vectors of the array elements are denoted as follows: The first transmitter n The coordinate vectors of the array elements are denoted as follows: The geometric coordinates are used to calculate the propagation distance and direction parameters of the array elements in the subsequent near-field spherical wave model.

[0025] Step 2 includes initial time delay estimation, initial angle estimation, initial complex gain calculation, and multipath line determination. The result obtained in step S1... For input, for the th l First, estimate the common delay parameters of the multipath. The time delay search range is taken as ,in In this embodiment and K =511, therefore The latency estimation resolution is determined by the bandwidth and is taken as 1 / B. After obtaining... Afterwards, in the given Estimate the departure angle parameter under the condition With angle of arrival parameter When the direction is expressed as an azimuth angle θ With pitch angle When indicating, the search range is taken as and The initial angle value is obtained by scanning using a preset angle grid. Given... Given the initial angle value, calculate the initial value of the complex gain based on the maximum likelihood criterion. Multipath extraction employs a power threshold stopping criterion. Extraction stops when the estimated power of the multipath to be extracted falls below a preset effective signal threshold, thus determining the number of multipaths. L The effective signal threshold is determined by a noise threshold and a preset margin, preferably a noise threshold plus 6 dB. The final result is... As initial input for the near-field SAGE refinement stage.

[0026] The spatially non-stationary near-field spherical wave channel model described in step 3 satisfies: at frequency points Place, No. l Multipath at the receiver m Each element and the transmitter n The frequency domain channel response between the array elements is (2) in and These are the spatial non-stationary coefficients for the receiver and transmitter, respectively. For complex gain, and For the responses of the receiving and transmitting antennas, the propagation delay satisfies (3) in c At the speed of light, For the first l The time delay of the stripe and the propagation distance at the array element level satisfy the following conditions: (4) (5) in and The receiving end is the first m Array element and transmitter n The coordinate vector of an array element relative to a reference array element. For the first l The distance from the last hop scatterer of the multipath to the reference element at the receiver. No. l The distance from the first hop scatterer of the multipath to the reference element at the transmitter; the unit vector of the element-level direction is... (6) (7) And the direction unit vector From pitch angle and azimuth Confirmed, the expression is (8) Furthermore, at frequency points First l The frequency domain channel matrix of a multipath is defined as follows: (9) The total frequency domain channel matrix satisfies: (10) Furthermore, regarding the first l Multipath at frequency points Definition without complex gain Template matrix: (11) And construct a spatially nonstationary matrix: (12) in Thus the first l The matrix form of a multipath can be written as: (13) Where ⊙ represents Hadamard element-wise multiplication.

[0027] Combination Figure 2 In this embodiment, the specific algorithm execution flow, step 4 includes the following detailed steps: Step 41: Fix the spatial nonstationarity coefficients at both the transmitting and receiving ends to 1, that is, for any multipath and any array element index... m , n ,satisfy , and at frequency points The first structure l Multipath without complex gain Path response matrix ; Step 42: Construct the first l Residual observation matrix of multiple paths ,satisfy (14) in For frequency point The observed channel matrix at that location, for i The reconstruction matrix of the multipath; Step 43: Define the first l The log-likelihood function of the multipath is: (15) in , For noise variance, For parameter set (16) In step 4, to improve the stability and computational efficiency of the near-field refinement estimation, this embodiment limits the updates of the angle parameters, reference distance parameters, and time delay parameters to a preset range near the initial values ​​obtained in step 2. The departure angle parameters and arrival angle parameters are searched within the neighborhood of their far-field initial values; preferably, the azimuth search range is set to [range to be specified in the original text]. The pitch angle search range is Common delay parameters in Neighborhood search, preferred settings and the feasible range of delay Intersections are used to ensure physical feasibility. Reference distance parameters are searched within a preset feasible distance range, with the minimum distance preferably set. Maximum distance of the scene and order and The value does not exceed Under the above range constraints, the departure angle and transmitter reference distance parameters, the arrival angle and receiver reference distance parameters, and the common delay parameters are updated sequentially. Given the updated geometric parameters and delay parameters, the closed-form solution of the complex gain is obtained by the maximum likelihood criterion to complete the parameter refinement.

[0028] Step 44: In the first iteration, given , Under the conditions, and let Taking the far-field boundary order, preferably... Where D is the array aperture and λ is the carrier wavelength corresponding to the center frequency, the receiver parameters are updated. ,satisfy (17) Step 45: In the given , , Under the condition of updating the transmitter parameters ,satisfy (18) Step 46: In the given , , , Under the condition of updating the delay parameter ,,satisfy (19) Step 47: Given, satisfying , , , , Under the condition of complex gain The log-likelihood function right Find the partial derivative and set it to 0. The complex gain can be obtained. The maximum likelihood estimate is a closed-form solution: (20) in ; Furthermore, step 5 specifically includes the following sub-steps: In the joint iteration process of step 5, the search center of each multipath parameter is taken as the parameter estimate after the previous iteration, and is updated within the preset neighborhood range of the parameter estimate.

[0029] Step 51: Initialize the joint iteration variables: Let the initial state at the start of the joint iteration be marked as round 1 (i.e., µ =1), the refined parameter result obtained in step 4 is directly used as the initial multipath parameter for the joint iteration, that is (twenty one) Meanwhile, assuming the entire array energy is stationary in the initial state, the initial values ​​of the spatial non-stationarity coefficients at both the transmitting and receiving ends are set to... (twenty two) Step 52: Given the spatial non-stationary coefficients from the previous round, update the geometric and temporal delay parameters. In the... µ In rounds of iteration ( ), using the non-stationary matrix from the previous round Construct the equivalent weighted path response matrix: (twenty three) Under the condition of the equivalent weighted path response matrix, the transmitter angle and distance, receiver angle and distance, and time delay parameters of the multipath are sequentially searched and updated in the neighborhood to obtain the first... µ Updated values ​​for the wheel's geometry and time delay parameters.

[0030] Step 53: Estimate the complex gain for the current round based on the delayed-updated non-stationary weights. After obtaining the... µ After considering the geometric and time delay parameters of the wheel, and combining them with the spatial non-stationary matrix of the previous wheel... By taking the partial derivative of the log-likelihood function with respect to the complex gain and setting it to zero, we obtain the first... µ Cyclic gain Closed-form solution: (twenty four) Step 54: Based on the updated multipath parameters, estimate the spatial nonstationary factor of the current round: in the... µ After the multipath angle, distance, time delay, and complex gain of the wheel are updated, the clean signal of the current wheel is reconstructed using these parameters. The element-level correlation metrics at the receiver and transmitter are calculated separately using a method that "introduces coefficients from the other end when calculating one end's metric," satisfying the following conditions: (25) (26) Determine the noise threshold based on the noise level. The effective signal threshold is obtained by adding 6dB to the noise threshold. ,satisfy (27) Construct the maximum normalized correlation metrics for the receiver and transmitter respectively. (28) The normalized correlation metric is normalized and truncated to the interval [0, 1] using the effective signal threshold as the lower bound, resulting in updated values ​​of non-stationary coefficients in continuous space, satisfying the following conditions: (29) (30) in To prevent constants with a denominator of 0; Step 55: Block-based alternating update and gating matrix construction. To ensure iterative stability, an alternating update strategy is adopted for the non-stationary coefficients at both the transmitting and receiving ends: first, in a given... Update under the condition Then in the given Update under the condition ; and update ; Step 56: Update the multipath parameters and residuals under a given gating matrix: Under these conditions, the updated geometric parameters, time delay parameters, and complex gain parameters are combined to reconstruct the first... l Multipath at frequency points The contribution matrix at point satisfies (31) Based on this, the residual observation matrix is ​​updated according to the residual construction method in step 42. This is used for the next iteration.

[0031] Furthermore, in step 6, the joint iteration of step 5 is executed cyclically, and the gain of the log-likelihood function is used as the stopping criterion. The iteration stops when the gain of the log-likelihood function in two consecutive iterations is not greater than a preset threshold, satisfying the condition... (32) in For the preset threshold, the preferred value is... .

[0032] After the iteration is completed, the multipath parameters are output and saved for channel characteristic analysis and modeling research.

[0033] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A near-field channel parameter estimation method considering spatial non-stationary characteristics, characterized in that, The method specifically includes the following steps: S1. Obtain broadband frequency domain observation signals and the geometric coordinate information of the array elements of the transmitting and receiving arrays; S2. During the initialization phase, the far-field SAGE algorithm is executed based on the far-field plane wave channel model to obtain the initial values ​​of the multipath parameters. S3. Based on the near-field spherical wave channel model, calculate the propagation distance between the array elements at the transmitting and receiving ends and construct the near-field channel response of each multipath. At the same time, introduce the spatial non-stationary coefficients at both the transmitting and receiving ends into the model. S4. Under the condition that the spatial non-stationary coefficient is fixed at 1, the near-field SAGE algorithm is used to refine the estimation of the near-field geometric parameters, time delay parameters and complex gain parameters of each multipath, and the parameter refinement results are obtained. S5. Based on the refined parameter results, enter the joint iteration stage. Perform correlation calculation on the residual and the near-field channel response in the frequency domain to obtain the array element-level correlation metric. Update the continuous spatial non-stationary coefficient with a value of 0 to 1 based on the array element-level correlation metric. Update the multipath channel parameters under the weighted condition involving the spatial non-stationary coefficient. S6. Repeat step S5 until the stopping criterion is met, and output the near-field geometric parameters, time delay parameters, complex gain parameters, and spatial non-stationary coefficients of each multipath.

2. The method of claim 1, wherein the spatial non-stationary characteristics are considered. Step S2 specifically includes: S21, based on the far-field plane wave channel model, the observation signal is preliminarily estimated on each multipath, and the angle parameter is obtained , and the initial value of the time delay parameter ; S22、based on the angle parameter and the initial value of the time delay parameter, calculate the initial value of the complex gain parameter of each multipath according to the maximum likelihood criterion .

3. The near-field channel parameter estimation method considering spatial non-stationary characteristics according to claim 2, characterized in that, The spatial non-stationary near-field spherical wave channel model in step S3 satisfies: at a frequency point , the first l th multipath is m between the first n th array element at the receiving end and the first array element at the transmitting end. (1) in and These are the spatial non-stationary coefficients for the receiver and transmitter, respectively. For complex gain, and For the responses of the receiving and transmitting antennas, the propagation delay satisfies (2) in c At the speed of light, For the first l The time delay of the stripe and the propagation distance at the array element level satisfy the following conditions: (3) (4) in and The receiving end is the first m Array element and transmitter n The coordinate vector of an array element relative to a reference array element. For the first l The distance from the last hop scatterer of the multipath to the reference element at the receiver. No. l The distance from the first hop scatterer of the multipath to the reference element at the transmitter; the unit vector of the element-level direction is... (5) (6) And the direction unit vector From pitch angle and azimuth Confirmed, the expression is (7)。 4. The near-field channel parameter estimation method considering spatial non-stationary characteristics according to claim 3, characterized in that, At frequency First l The frequency domain channel matrix of a multipath is defined as follows: (8) The total frequency domain channel matrix satisfies: (9) Furthermore, regarding the first l Multipath at frequency points Definition without complex gain Template matrix: (10) And construct a spatially nonstationary matrix: (11) in Thus the first l The matrix form of a multipath can be written as: (12) Where ⊙ represents Hadamard element-wise multiplication.

5. The near-field channel parameter estimation method considering spatial non-stationary characteristics according to claim 4, characterized in that, Step S4 specifically includes: S41. Fix the spatial nonstationary coefficients at both the transmitting and receiving ends to 1, that is, for any multipath and any array element index m , n ,satisfy , and at frequency points The first structure l Multipath without complex gain Path response matrix ; S42, Construction of the l Residual observation matrix of multiple paths ,satisfy (13) in For frequency point The observed channel matrix at that location, for i The reconstruction matrix of the multipath; S43, Definition of the l The log-likelihood function of the multipath is: (14) in , For noise variance, For parameter set (15) S44, Near-field SAGE estimation, given , Under the conditions, and let Where D is the array aperture and λ is the carrier wavelength corresponding to the center frequency, the receiver parameters are updated. ,satisfy (16) S45, In the given , , Under the condition of updating the transmitter parameters ,satisfy (17) S46, In the given , , , Under the condition of updating the delay parameter ,,satisfy (18) S47. Given, satisfying , , , , Under the condition of complex gain The log-likelihood function right Find the partial derivative and set it to 0. The complex gain can be obtained. The maximum likelihood estimate is a closed-form solution: (19) in .

6. The near-field channel parameter estimation method considering spatial non-stationary characteristics according to claim 5, characterized in that, Step S5 specifically includes: S51. Initialize joint iteration variables: Let the initial state at the start of the joint iteration be marked as round 1. Use the refined parameter results obtained in step S4 directly as the initial multipath parameters for the joint iteration, i.e. (20) Meanwhile, assuming the entire array energy is stationary in the initial state, the initial values ​​of the spatial non-stationarity coefficients at both the transmitting and receiving ends are set to... (21) S52. Given the spatial non-stationary coefficients from the previous round, update the geometric and time delay parameters; in the... µ In each iteration, the non-stationary matrix from the previous round is used. Construct the equivalent weighted path response matrix: (22) Under the condition of the equivalent weighted path response matrix, the transmitter angle and distance, receiver angle and distance, and time delay parameters of the multipath are sequentially searched and updated in the neighborhood to obtain the first... µ Updated values ​​of the wheel's geometry and time delay parameters; S53. Estimate the complex gain of the current round based on the non-stationary weights updated with delay; after obtaining the... µ After considering the geometric and time delay parameters of the wheel, and combining them with the spatial non-stationary matrix of the previous wheel... By taking the partial derivative of the log-likelihood function with respect to the complex gain and setting it to zero, we obtain the first... µ Cyclic gain Closed-form solution: (23) S54. Based on the updated multipath parameters, estimate the spatial nonstationary factor of the current round: in the... µ After the multipath angle, distance, time delay, and complex gain of the wheel are updated, the clean signal of the current wheel is reconstructed using these parameters. The element-level correlation metrics at the receiver and transmitter are calculated separately using a method that "introduces coefficients from the other end when calculating one end's metric," satisfying the following conditions: (24) (25) Determine the noise threshold based on the noise level. The effective signal threshold is obtained by adding 6dB to the noise threshold. ,satisfy (26) Construct the maximum normalized correlation metrics for the receiver and transmitter respectively. (27) The normalized correlation metric is normalized and truncated to the interval [0,1] using the effective signal threshold as the lower bound to obtain the updated values ​​of the non-stationary coefficients in the continuous space, satisfying the following conditions: (28) (29) in To prevent constants with a denominator of 0; S55, Block-based alternating update and gating matrix construction; To ensure iterative stability, an alternating update strategy is adopted for the non-stationary coefficients at both ends of the transmission and reception: First, in a given... Update under the condition Then in the given Update under the condition ; and update ; S56. Update the multipath parameters and residuals under a given gating matrix: Under a given gating matrix... Under these conditions, the updated geometric parameters, time delay parameters, and complex gain parameters are combined to reconstruct the first... l Multipath at frequency points The contribution matrix at point satisfies (30) Based on this, the residual observation matrix is ​​updated according to the residual construction method in step 42. This is used for the next iteration.

7. The near-field channel parameter estimation method considering spatial non-stationary characteristics according to claim 6, characterized in that, The joint iteration in step S5 uses the gain of the log-likelihood function as the stopping criterion. Iteration stops when the gain of the log-likelihood function in two consecutive iterations is not greater than a preset threshold, satisfying the condition... (31) in This is a preset threshold.

8. A near-field channel parameter estimation system considering spatial non-stationary characteristics, characterized in that, The system employs the method as described in any one of claims 1 to 7.

Images