A super-massive MIMO synchronization method, electronic equipment and storage medium

Abstract

This invention discloses a method, electronic device, and storage medium for ultra-large-scale MIMO synchronization, relating to the fields of wireless communication and channel measurement technology. The method includes: first, estimating the delay and direction parameters of at least one strong path as an anchor path based on reference subarray channel data; then, for each pair of adjacent subarrays, extracting the phase of the same path on the boundary elements based on the anchor path parameters, calculating the observed phase difference, and thus obtaining the observed value of the common phase drift between subarrays; next, fusing multiple observations to obtain a robust estimate of the common phase difference between adjacent subarrays; finally, accumulating the estimated values ​​of each segment with the first subarray as a reference to obtain the common phase drift of each subarray, performing phase compensation, and then stitching the data of each subarray according to coordinates into an equivalent ultra-large-scale array to achieve ultra-large-scale MIMO synchronization. Therefore, this invention can estimate and compensate for the common phase drift of subarrays without additional hardware time synchronization.

Description

Technical Field

[0001] This invention relates to the field of wireless communication and channel measurement technology, and in particular to a very large-scale MIMO synchronization method, electronic device, and storage medium. Background Technology

[0002] In new intermediate frequency, millimeter wave (mmWave), and terahertz (THz) communication systems, to obtain larger array apertures and higher angular resolution, a set of movable, small-scale subarrays are often virtually translated along a sliding rail, sequentially measured at multiple docking positions (subarray positions), and then the data from each subarray are stitched together to form an equivalent ultra-large-scale array. Since a scan within each subarray is completed in milliseconds, the phase within the subarray is essentially consistent; however, the measurement interval between adjacent subarrays is typically on the order of seconds to minutes. Even with a highly stable clock, this introduces a phase shift (common phase drift) between different subarrays. This shift is consistent across all ports / polarizations within the same subarray, but is unknown and different between different subarrays.

[0003] Directly stitching without correction will lead to failure of coherent superposition across subarrays, segmentation of the equivalent aperture, manifested as main lobe distortion and side lobe uplift, severely affecting beamforming, angle estimation, and imaging quality. Existing methods mostly rely on additional hardware for timing and external reference sources, making it difficult to implement robustly in practical measurement and communication systems. Summary of the Invention

[0004] The purpose of this invention is to provide a method, electronic device and storage medium for ultra-large-scale MIMO synchronization, which aims to solve or improve at least one of the above-mentioned technical problems.

[0005] To achieve the above objectives, the present invention provides the following solution: A method for MIMO synchronization of ultra-large scale includes: Acquire channel measurement data of the reference subarray and perform strong path estimation to obtain the delay and direction parameters of at least one strong path as anchor path parameters; For each pair of adjacent subarrays, the phase of the same path is extracted on the boundary elements on both sides according to the anchor path parameters to obtain the observation phase of the two subarrays. The observation phase difference across subarrays is calculated based on the observed phase, and the common phase drift observation value between adjacent subarrays is calculated based on the anchor path parameters. By fusing multiple common phase drift observations, an estimated common phase difference between adjacent subarrays is obtained. Using the first subarray as a reference, the estimated common phase difference between each adjacent subarray is accumulated to obtain the common phase drift of each subarray. The channel response of each subarray is phase compensated according to the common phase drift, and the compensated subarray data is spliced ​​together according to the array element coordinates to form an equivalent ultra-large scale array.

[0006] Optionally, the step of acquiring channel measurement data of the reference subarray and performing strong path estimation to obtain the delay and direction parameters of at least one strong path as anchor path parameters specifically includes: The channel measurement data of the reference subarray are spatially averaged, and the time delay domain response is obtained by inverse Fourier transform. The peak time delay is extracted as the path delay estimate. The path direction is estimated by beam scanning, subspace method or parameterized iterative algorithm to obtain the path direction estimate.

[0007] Optionally, for each pair of adjacent subarrays, extracting the phase of the same path on the boundary elements on both sides according to the anchor path parameters to obtain the observation phase of the two subarrays specifically includes: For each anchor path, the frequency domain channel response is demodulated using the estimated time delay on the boundary elements of the adjacent subarrays, and coherently accumulated in the frequency domain to obtain the complex values ​​of the path on the two subarrays. The phase of the complex value is taken as the observation phase of the path on the two subarrays.

[0008] Optionally, the step of calculating the observation phase difference across subarrays based on the observation phase, and calculating the common phase drift observation value between adjacent subarrays based on the anchor path parameters, specifically includes: For the same path, calculate the observation phase difference between the two subarrays based on the observed phase, and take the principal value; Calculate the geometric phase difference at the center frequency based on the path direction estimate and the boundary element coordinates; Subtract the observed phase difference from the geometric phase difference and take the principal value to obtain the common phase drift observation value corresponding to the path.

[0009] Optionally, fusing multiple common phase drift observations to obtain an estimate of the common phase difference between adjacent subarrays specifically includes: Outlier observations of multiple common phase drift values ​​of the same pair of subarrays are removed, the remaining observations are mapped onto the unit circle, and then weighted and summed according to preset weights; The phase angle of the weighted sum vector is taken as the estimated value of the common phase difference between the adjacent subarrays.

[0010] Optionally, the specific processing steps for phase compensation include: For all elements and all frequency points of each subarray, multiply them uniformly by the phase rotation factor corresponding to the common phase drift.

[0011] The present invention also provides an electronic device, including a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to perform a large-scale MIMO synchronization method according to the above description.

[0012] The present invention also provides a computer-readable storage medium, characterized in that it stores a computer program, which, when executed by a processor, implements a large-scale MIMO synchronization method as described above.

[0013] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects: This invention discloses a method, electronic device, and storage medium for ultra-large-scale MIMO synchronization. The method includes acquiring channel measurement data of a reference subarray and performing strong path estimation to obtain the delay and direction parameters of at least one strong path as anchor path parameters; for each pair of adjacent subarrays, extracting the phase of the same path on the boundary elements on both sides according to the anchor path parameters to obtain the observed phase of the two subarrays; calculating the observed phase difference across subarrays based on the observed phase, and calculating the common phase drift observation value between the adjacent subarrays based on the anchor path parameters; fusing multiple common phase drift observation values ​​to obtain an estimated common phase difference value between the adjacent subarrays; using a first subarray as a reference, accumulating the estimated common phase difference values ​​between each adjacent subarray to obtain the common phase drift amount of each subarray; performing phase compensation on the channel response of each subarray based on the common phase drift amount; and stitching the compensated subarray data into an equivalent ultra-large-scale array according to the element coordinates. This invention provides a method for estimating and compensating for common phase drift between subarrays without additional hardware time synchronization when phase drift exists. It is applicable to channel measurement, beamforming, and high-precision AoD extraction in new intermediate frequency / millimeter wave bands. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a schematic diagram of the framework of the ultra-large-scale MIMO synchronization method of the present invention; Figure 2 This is a diagram illustrating the effect of the ultra-large-scale MIMO synchronization method in this embodiment. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] The purpose of this invention is to provide a method, electronic device and storage medium for ultra-large-scale MIMO synchronization, which aims to solve or improve at least one of the above-mentioned technical problems.

[0018] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0019] As a first aspect, the present invention provides, as follows Figure 1 The illustrated method for ultra-large-scale MIMO synchronization includes the following steps: Step 1: Acquire channel measurement data of the reference subarray and perform strong path estimation to obtain the time delay and direction parameters of at least one strong path as anchor path parameters. Specifically, the channel measurement data of the reference subarray is spatially averaged, and the time delay domain response is obtained through inverse Fourier transform. The peak time delay is extracted as the path time delay estimate. The direction of the path is estimated by beam scanning, subspace method, or parameterized iterative algorithm to obtain the path direction estimate.

[0020] The specific steps for coarse estimation of strong paths in the reference subarray are as follows: From the actual channel measurements, strong paths that can also be observed in adjacent subarrays are selected, and their time delay and direction are roughly estimated, providing anchor points for subsequent phase alignment across subarrays, sharing the same path. First, a small number of strong paths that can also be stably observed in adjacent subarrays are identified using measurement data from the first subarray. The frequency domain response of all elements in this subarray is spatially averaged, and then transformed to the time delay domain using IFFT. Finding the peak yields several distinguishable time delays. Subsequently, for each path's angle (direction of arrival), a mature method can be selected to estimate it based on system resource and accuracy requirements. Matched correlation / beam scanning (GoB / CBF / MVDR): Calculates the matched output power on the directional grid and takes the peak direction as sl; simple, robust, and easy to use online.

[0021] MUSIC / ESPRIT (subspace class): High-resolution DOA estimation based on covariance decomposition; high resolution, clear requirements for array geometry, suitable for scenarios with high SNR / sufficient snapshots.

[0022] SAGE / Maximum Likelihood (Parameterized Iteration): Jointly estimates parameters such as delay, angle, amplitude, and phase, with strong resolution, suitable for scenarios with dense multipath propagation but where computational resources allow.

[0023] Without specifying a particular algorithm, using the actual measured data of this subarray as input, the output is a rough estimate of the time delay-angle for each strong path. Only a small number of paths (e.g., 2-8) that are high in power, separated from each other, and detectable in adjacent subarrays are retained as "anchor paths".

[0024] Step 2: For each pair of adjacent subarrays, extract the phase of the same path on the boundary elements on both sides according to the anchor path parameters to obtain the observed phase of the two subarrays. Specifically, for each anchor path, demodulate the frequency domain channel response on the boundary elements of the adjacent subarrays using the estimated time delay, and coherently accumulate it in the frequency domain to obtain the complex value of the path on the two subarrays; take the phase of the complex value as the observed phase of the path on the two subarrays.

[0025] The specific steps for extracting the same-path phase of adjacent subarray boundary elements are as follows: For any pair of subarrays Select one representative element from each side. For each anchor path in step 1... Repeat "by time delay" on both sides. Demodulation and cross-frequency coherence processing: In the formula, For the first The first in the formation The element corresponds to the first The complex coefficients of the path, K For the number of subcarriers, For the first The first in the formation Each element at discrete frequency points Frequency domain channel response at that location For the first The demodulation factor for delay phase compensation along the path. For discrete frequency points, Subarray number The array element number is the number of the array element in the first position. The first in each sub-array The three-dimensional coordinates of each array element , It is the field of real numbers.

[0026] Discrete frequency points , .in Center frequency,B For bandwidth, For frequency point index, This represents the frequency interval between adjacent discrete frequency points.

[0027] The observed phases of the two subarrays at adjacent array elements are obtained, which include the inherent propagation phase of each path, the geometric phase of the array, and the drift phase of the subarray.

[0028] Step 3: Calculate the observed phase difference across subarrays based on the observed phase, and calculate the common phase drift observation value between adjacent subarrays based on the anchor path parameters. Specifically, for the same path, calculate the observed phase difference between the two subarrays based on the observed phase and take the principal value; calculate the geometric phase difference at the center frequency based on the path direction estimate and the boundary element coordinates; subtract the observed phase difference from the geometric phase difference and take the principal value to obtain the common phase drift observation value corresponding to the path.

[0029] The specific steps for constructing the phase difference across subarrays and subtracting the geometric phase to obtain common drift observations are as follows: For the same path l First, calculate the phase difference across subarrays: In the formula, For the first Sub-arrays and the first The individual formations are in the... The observed phase difference of each path, The principal value operation for phase is used to normalize the phase to... interval, Phase angle operations for complex numbers.

[0030] Then use the direction obtained in step 1 to roughly estimate Calculate the geometric phase difference at the center frequency point using the coordinates of the two array elements: In the formula, For discrete frequency points Below, the array-guided phase of the plane wave model Indicates the first l The direction of arrival of the path (e.g., spherical coordinates can be mapped from azimuth / elevation angles); The speed of light; To quantity Transpose operation.

[0031] Subtract the two and unify to : In the formula, For the first Sub-arrays and the first The individual formations are in the... Common phase drift observations obtained from the path.

[0032] Obtain the common phase difference of this pair of subarrays The first observation; the geometric terms can be directly calculated from coordinates and direction, and are therefore accurately subtracted. As a specific implementation, the common phase drift is the first... The entire constant phase of each subarray (All ports and frequencies within the same subarray are shared), with the first subarray as a reference. .

[0033] Step 4: Fuse multiple common phase drift observations to obtain an estimate of the common phase difference between adjacent subarrays. Specifically, outlier observations of multiple common phase drift observations of the same pair of subarrays are removed, the remaining observations are mapped onto the unit circle, and a weighted sum is performed according to a preset weight; the phase angle of the weighted sum vector is taken as the estimate of the common phase difference between adjacent subarrays.

[0034] The specific steps for obtaining the inter-subarray drift through multi-observation fusion are as follows: The same pair of subarrays typically has observation phase differences from multiple paths and multiple pairs of representative array elements. First, perform outlier removal (e.g., based on the absolute deviation of the median), then perform weighted summation on the unit circle.

[0035] In the formula, This is a weighted sum vector of all observed phases on the unit circle. As weight, For the first The complex exponent of each observed phase on the unit circle It represents the difference between adjacent pairs of subarrays.

[0036] This yields a more robust estimate of the common phase difference between the two subarrays. Weights Weighted averaging based on multipath power, using an exponential form, avoids phase bias. The problem of surrounding.

[0037] Step 5: Using the first subarray as a reference, accumulate the estimated common phase difference between each adjacent subarray to obtain the common phase drift of each subarray. Perform phase compensation on the channel response of each subarray based on the common phase drift, and then stitch the compensated subarray data together according to the element coordinates to form an equivalent ultra-large scale array. Specifically, the phase compensation process is as follows: multiply all elements and all frequency points of each subarray by a phase rotation factor corresponding to the common phase drift.

[0038] The specific steps for common phase reconstruction, integer array despinning, and splicing are as follows: The difference between each pair of adjacent subarrays By accumulating the values ​​sequentially along the subarrays, the common phase of each subarray relative to the reference subarray is obtained. (The meaning of unwrap (phase "unwrapping") is: to eliminate the ±π jumps in the phase sequence caused by taking the principal value of (-π,π] (or [0,2π)) by automatically adding / subtracting 2π, so that the phase curve is continuous and smooth.): Subsequently, regarding the first p Multiply all elements and all frequencies of a subarray by a spin factor: in, For the first The first in the formation Each element at discrete frequency points The frequency domain channel response after common phase compensation. To offset the first Phase rotation factor of common phase drift of each subarray For the first The first in the formation Each element at discrete frequency points Frequency domain channel response before compensation.

[0039] The data from all subarrays after compensation are stitched together according to their respective three-dimensional coordinates to form an equivalent large-aperture array. At this point, the signals across subarrays return to the same phase scale, the main lobe is recovered, the side lobes are reduced, and the beamforming and angle / imaging quality are significantly improved.

[0040] Based on the above technical solution, the following embodiments are provided.

[0041] As a specific implementation method, the simulation parameters in this embodiment are configured as follows: Carrier / Bandwidth: ( ;Number of subcarriers K=1001, Each subarray has 16 elements (U=16), the number of subarrays is P=4, and the total number of equivalent elements is N=64; the spacing between elements is... , Subarray translation step size Array element coordinates Set the truth values ​​for the two path parameters to: , , Subarray common phase: The remaining subarrays drift to random values. The signal-to-noise ratio (SNR) is 25 dB. The specific simulation process and parameters are as follows: Before synchronization, data synthesis is required to generate the broadband signal to be measured. First, perform ideal path superposition for each frequency point. Sub-array p Array element The frequency domain responses of the two paths are constructed and geometric phase is applied according to the array element coordinates; then, subarray common phase injection is performed to inject the entire phase drift factor. Ride to all of the subarrays Finally, noise is added, and the target SNR is calculated for each... Injecting complex white Gaussian noise, the observed channel response is obtained. .

[0042] Execute the specific synchronization process: Step 1: Coarse estimation of strong paths in the reference subarray.

[0043] 1. Frequency domain preprocessing (only in the reference subarray) Above): Spatial averaging of multiple array elements within a subarray. Multiply by a Hann window and perform energy-preserving normalization (only once here), then IFFT to the time-delay domain. .

[0044] 2. Time Delay Peak Extraction: In The peak finding function is used to find the two maximum peaks, resulting in... .

[0045] 3. Angle estimation: Perform matched correlation / beam scan: On the angle grid, construct a steering vector using the coordinates of the reference subarray elements, match it with the observed path vector, and take the peak direction as the multipath angle direction.

[0046] 4. Output: Anchor path parameter set .

[0047] Step 2: Extract the phase of adjacent subarray boundary elements with the same path.

[0048] For each pair of subarrays : 1. Selection of representative array elements: Use boundary array elements .

[0049] 2. Frequency domain demodulation + cross-frequency coherence: For each anchor path On both sides respectively Demodulate and cross-frequency coherently to obtain , Take its principal phase value.

[0050] 3. Output: The phase of the same path in the two subarrays.

[0051] Step 3: Construct the phase difference across subarrays and subtract the geometric phase to obtain common drift observations.

[0052] 1. Observation phase difference: .

[0053] 2. Geometric phase difference: using Calculate the geometric phase at the center frequency using the coordinates of the two array elements. .

[0054] 3. Common phase drift: .

[0055] 4. Output: Common phase drift of each pair of subarrays (multipath × multiple elements) .

[0056] Step 4: Multi-observation fusion yields the inter-array drift.

[0057] For a fixed pair of subarrays Perform weighted fusion of multiple sample points in a circular domain: perform weighted summation on the unit circle. The weights are the power weights of the two paths.

[0058] Output: Estimated common phase difference for each pair of subarrays .

[0059] Step 5: Common phase reconstruction, despinning and splicing of the whole array.

[0060] Reconstruction: For reference, add sequentially. .

[0061] Subarray rotation: For the first All elements of the sub-array take ,get .

[0062] Splicing and Beamforming: H is spliced ​​into an equivalent large array according to the element coordinates. After frequency focusing at the sub-band frequency points, the multiple elements are coherently combined to output the power angle spectrum. For example... Figure 2 As shown, the estimated phase drift of each subarray is consistent with the original set true value. The beamforming results of the calibrated and uncalibrated results are compared as follows: Without calibration, signals across subarrays cannot be coherently superimposed, the equivalent aperture is segmented, the main lobe is distorted, ringing near the peak is significant, the peak sidelobe level is raised, the peak value in the target direction is lower, and the angular spectrum peak shape diverges. After calibration, a unified phase scale is restored across subarrays, signals achieve coherent synthesis, the main lobe shape returns to the theoretical aperture, the peak value is significantly raised (coherent gain recovery), the sidelobes decrease overall, and ringing in the peak neighborhood is weakened. Overall, the beamforming and angle estimation quality after calibration is significantly better than that after uncalibration, meeting the performance requirements of coherent processing of equivalent ultra-large aperture.

[0063] As a second aspect, the present invention also provides an electronic device, including a memory and a processor, the memory being used to store a computer program, and the processor running the computer program to cause the electronic device to perform a super-large-scale MIMO synchronization method according to the above description.

[0064] As a third aspect, the present invention also provides a computer-readable storage medium, characterized in that it stores a computer program, which, when executed by a processor, implements a large-scale MIMO synchronization method as described above.

[0065] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for synchronizing ultra-large-scale MIMO, characterized in that, include: Acquire channel measurement data of the reference subarray and perform strong path estimation to obtain the delay and direction parameters of at least one strong path as anchor path parameters; For each pair of adjacent subarrays, the phase of the same path is extracted on the boundary elements on both sides according to the anchor path parameters to obtain the observation phase of the two subarrays. The observation phase difference across subarrays is calculated based on the observed phase, and the common phase drift observation value between adjacent subarrays is calculated based on the anchor path parameters. By fusing multiple common phase drift observations, an estimated common phase difference between adjacent subarrays is obtained. Using the first subarray as a reference, the estimated common phase difference between each adjacent subarray is accumulated to obtain the common phase drift of each subarray. The channel response of each subarray is phase compensated according to the common phase drift, and the compensated subarray data is spliced ​​together according to the array element coordinates to form an equivalent ultra-large scale array.

2. The ultra-large-scale MIMO synchronization method according to claim 1, characterized in that, The process of acquiring channel measurement data of the reference subarray and performing strong path estimation to obtain the delay and direction parameters of at least one strong path as anchor path parameters specifically includes: The channel measurement data of the reference subarray are spatially averaged, and the time delay domain response is obtained by inverse Fourier transform. The peak time delay is extracted as the path delay estimate. The path direction is estimated by beam scanning, subspace method or parameterized iterative algorithm to obtain the path direction estimate.

3. The ultra-large-scale MIMO synchronization method according to claim 1, characterized in that, For each pair of adjacent subarrays, the phase along the same path is extracted from the boundary elements on both sides based on the anchor path parameters to obtain the observed phase of the two subarrays. Specifically, this includes: For each anchor path, the frequency domain channel response is demodulated using the estimated time delay on the boundary elements of the adjacent subarrays, and coherently accumulated in the frequency domain to obtain the complex values ​​of the path on the two subarrays. The phase of the complex value is taken as the observation phase of the path on the two subarrays.

4. The ultra-large-scale MIMO synchronization method according to claim 1, characterized in that, The step of calculating the observed phase difference across subarrays based on the observed phase, and calculating the common phase drift observation value between adjacent subarrays based on the anchor path parameters, specifically includes: For the same path, calculate the observation phase difference between the two subarrays based on the observed phase, and take the principal value; Calculate the geometric phase difference at the center frequency based on the path direction estimate and the boundary element coordinates; Subtract the observed phase difference from the geometric phase difference and take the principal value to obtain the common phase drift observation value corresponding to the path.

5. The ultra-large-scale MIMO synchronization method according to claim 1, characterized in that, The process of fusing multiple common phase drift observations to obtain an estimate of the common phase difference between adjacent subarrays specifically includes: Outlier observations of multiple common phase drift values ​​of the same pair of subarrays are removed, the remaining observations are mapped onto the unit circle, and then weighted and summed according to preset weights; The phase angle of the weighted sum vector is taken as the estimated value of the common phase difference between the adjacent subarrays.

6. The ultra-large-scale MIMO synchronization method according to claim 1, characterized in that, The specific processing steps for phase compensation include: For all elements and all frequency points of each subarray, multiply them uniformly by the phase rotation factor corresponding to the common phase drift.

7. An electronic device, characterized in that, The device includes a memory and a processor, the memory being used to store a computer program, and the processor running the computer program to cause the electronic device to perform a large-scale MIMO synchronization method according to any one of claims 1-6.

8. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed by a processor, implements a large-scale MIMO synchronization method as described in any one of claims 1-6.

Images