Channel data transmission methods, electronic devices and computer-readable storage media
By receiving uplink pilot signals and updating antenna phase deviation and angular power spectrum parameters in a frequency division duplex massive MIMO system, and reconstructing the downlink channel covariance matrix, the problem of low channel data transmission accuracy under non-ideal hardware environments is solved, and higher channel estimation accuracy and data transmission reliability are achieved.
Patent Information
- Application Number
- CN202611504959.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-24
- Publication Date
- 2026-11-17
AI Technical Summary
In frequency division duplex massive MIMO systems, existing technologies cannot effectively address phase distortion issues under non-ideal hardware environments, leading to decreased downlink channel estimation accuracy and decoding accuracy, and lower accuracy of channel data transmission.
By receiving the uplink pilot signal sent by the user equipment, the observed sample values of the uplink channel covariance matrix are determined, the initial antenna phase deviation vector and angular power spectrum parameters are updated, the target antenna phase deviation and angular power spectrum parameters are processed using the antenna array response matrix of the downlink band, the downlink channel covariance matrix is reconstructed, and the downlink precoding parameters are determined based on this for data transmission.
In a non-ideal hardware environment, accurate reconstruction of the downlink channel covariance matrix was achieved, improving the accuracy of downlink precoding parameters and enhancing the accuracy of channel data transmission and communication quality.
Smart Images

Figure CN123122598A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of communication technology, and more specifically, to a channel data transmission method, an electronic device, and a computer-readable storage medium. Background Technology
[0002] In the field of wireless communication, massive MIMO (Multiple-Input Multiple-Output) technology improves spectral efficiency and system capacity by deploying large-scale antenna arrays at base stations. However, related technologies have significant shortcomings in downlink channel estimation for Frequency Division Duplexing (FDD) MIMO systems. They cannot effectively address phase distortion issues in non-ideal hardware environments, leading to a severe decrease in estimation accuracy under massive antenna scenarios, which in turn reduces the accuracy of downlink channel decoding. Therefore, the accuracy of channel data transmission in related technologies is relatively low.
[0003] There is currently no effective solution to the above problems. Summary of the Invention
[0004] This invention provides a channel data transmission method, an electronic device, and a computer-readable storage medium to at least solve the technical problem of low accuracy in channel data transmission in related technologies.
[0005] According to one aspect of the present invention, a channel data transmission method is provided, comprising: receiving an uplink pilot signal transmitted by a user equipment; determining observation sample values of an uplink channel covariance matrix based on the uplink pilot signal, wherein the uplink channel covariance matrix is used to represent a statistical estimation matrix of the uplink channel space under the presence of antenna phase deviation and noise interference; updating an initial antenna phase deviation vector and an initial angular power spectrum parameter based on the observation sample values to obtain a target antenna phase deviation vector and a target angular power spectrum parameter; processing the target antenna phase deviation vector and the target angular power spectrum parameter using an antenna array response matrix of a downlink band to determine a downlink channel covariance matrix, wherein the downlink channel covariance matrix is used to represent a statistical estimation matrix of the reconstructed downlink channel space; determining downlink precoding parameters based on the downlink channel covariance matrix; and transmitting downlink data to the user equipment using the downlink precoding parameters.
[0006] Furthermore, the initial antenna phase deviation vector and initial angular power spectrum parameters are updated based on the observed sample values to obtain the target antenna phase deviation vector and target angular power spectrum parameters. This includes: in the current iteration, fixing the initial antenna phase deviation vector, using an iterative adaptive algorithm to update the initial angular power spectrum parameters based on the observed sample values to obtain updated angular power spectrum parameters; fixing the updated angular power spectrum parameters, using the coordinate descent method to update the initial antenna phase deviation vector based on the observed sample values to obtain updated antenna phase deviation vector; and determining the target antenna phase deviation vector and target angular power spectrum parameters based on the updated angular power spectrum parameters and updated antenna phase deviation vector.
[0007] Further, based on the updated angular power spectrum parameters and the updated antenna phase deviation vector, the target antenna phase deviation vector and the target angular power spectrum parameters are determined, including: in response to the updated angular power spectrum parameters and the updated antenna phase deviation vector not meeting the preset iteration conditions, the updated angular power spectrum parameters are determined as the initial angular power spectrum parameters, the updated antenna phase deviation vector is determined as the initial antenna phase deviation vector, and the execution steps of the current iteration round are repeated until the updated angular power spectrum parameters and the updated antenna phase deviation vector meet the preset iteration conditions; in response to the updated angular power spectrum parameters and the updated antenna phase deviation vector meeting the preset iteration conditions, the updated angular power spectrum parameters are determined as the target angular power spectrum parameters, and the updated antenna phase deviation vector is determined as the target antenna phase deviation vector.
[0008] Furthermore, the initial angular power spectrum parameters are updated using an iterative adaptive algorithm based on the observed sample values to obtain the updated angular power spectrum parameters. This includes: constructing an initial estimation model of the uplink channel covariance matrix based on the initial antenna phase deviation vector and the observed sample values; determining the signal decorrelation estimate of the discrete spatial angle points based on the initial angular power spectrum parameters and the initial estimation model; and updating the initial angular power spectrum parameters based on the signal decorrelation estimate to obtain the updated angular power spectrum parameters.
[0009] Furthermore, the initial antenna phase deviation vector is updated using the coordinate descent method based on the observed sample values to obtain the updated antenna phase deviation vector. This process includes: traversing multiple antennas in the base station array, using the initial antenna phase deviation vector corresponding to the currently traversed antenna as the current variable to be updated, and using the initial antenna phase deviation vectors corresponding to the other antennas as fixed variables; constructing a target cost function based on the current variable to be updated, the fixed variables, and the observed sample values; converting the target cost function into a univariate function of the current variable to be updated; determining the target phase deviation vector of the currently traversed antenna based on the extremum condition of the univariate function; and determining the updated antenna phase deviation vector based on the target phase deviation vectors of multiple antennas.
[0010] Furthermore, the target antenna phase deviation vector and target angular power spectrum parameters are processed using the antenna array response matrix of the downlink band to determine the downlink channel covariance matrix. This includes: using the target antenna phase deviation vector to perform phase compensation correction on the antenna array response matrix of the downlink band to obtain the corrected downlink array response; and determining the downlink channel covariance matrix based on the corrected downlink array response and the target angular power spectrum parameters.
[0011] Furthermore, the method also includes: constructing the antenna array response matrix of the downlink band based on the center frequency of the downlink band and the geometry of the base station antenna array.
[0012] Furthermore, receiving uplink pilot signals transmitted by user equipment includes: receiving uplink pilot signals transmitted by user equipment through a uniform linear array; transmitting downlink data to user equipment using downlink precoding parameters includes: transmitting downlink data to user equipment through a uniform linear array using downlink precoding parameters.
[0013] Further, determining downlink precoding parameters based on the downlink channel covariance matrix includes: extracting the downlink beamforming vector corresponding to the largest eigenvalue according to the characteristic distribution of the downlink channel covariance matrix, and determining the downlink precoding parameters based on the downlink beamforming vector; or, using the minimum mean square error to determine the downlink precoding matrix based on the downlink channel covariance matrix, noise power, and downlink transmission power, and determining the downlink precoding matrix as the downlink precoding parameters.
[0014] According to another aspect of the present invention, an electronic device is also provided, comprising: a memory storing an executable program; and a processor for running the program, wherein the program executes the methods of various embodiments of the present invention during runtime.
[0015] According to another aspect of the present invention, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored executable program, wherein, when the executable program is executed, it controls the device where the computer-readable storage medium is located to perform the methods of various embodiments of the present invention.
[0016] According to another aspect of the present invention, a computer program product is also provided, including a computer program that, when executed by a processor, implements the methods of various embodiments of the present invention.
[0017] According to another aspect of the present invention, a computer program product is also provided, including a non-volatile computer-readable storage medium storing a computer program that, when executed by a processor, implements the methods of various embodiments of the present invention.
[0018] According to another aspect of the present invention, a computer program is also provided, which, when executed by a processor, implements the methods of the various embodiments of the present invention.
[0019] In this embodiment of the invention, an uplink pilot signal transmitted by a user equipment is received. Based on the uplink pilot signal, observed sample values of the uplink channel covariance matrix are determined. The uplink channel covariance matrix represents the statistical estimation matrix of the uplink channel space under conditions of antenna phase deviation and noise interference. Based on the observed sample values, the initial antenna phase deviation vector and initial angular power spectrum parameters are updated to obtain the target antenna phase deviation vector and target angular power spectrum parameters. The target antenna phase deviation vector and target angular power spectrum parameters are processed using the antenna array response matrix of the downlink band to determine the downlink channel covariance matrix. The downlink channel covariance matrix represents the statistical estimation matrix of the reconstructed downlink channel space. Downlink precoding parameters are determined based on the downlink channel covariance matrix, and downlink data is transmitted to the user equipment using the downlink precoding parameters. This application utilizes the reciprocity of uplink and downlink angular power spectra to perform phase compensation correction on the antenna array response matrix of the downlink channel using the high-precision angular power spectrum parameters and antenna phase deviation vector obtained by joint estimation on the uplink. This accurately maps and reconstructs the statistical estimation of the uplink channel space into the statistical estimation matrix of the downlink channel space, achieving the goal of reconstructing the downlink channel covariance matrix under non-ideal hardware conditions. This improves the accuracy of downlink precoding parameters and solves the technical problem of low accuracy of channel data transmission in related technologies. Attached Figure Description
[0020] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:
[0021] Figure 1 This is a flowchart of a channel data transmission method according to an embodiment of the present invention;
[0022] Figure 2 This is a schematic diagram of an optional channel data transmission method according to an embodiment of the present invention;
[0023] Figure 3 This is a schematic diagram of an optional channel data transmission method according to an embodiment of the present invention;
[0024] Figure 4 This is a schematic diagram of an optional channel data transmission simulation result according to an embodiment of the present invention;
[0025] Figure 5This is a schematic diagram of an optional channel data transmission simulation result according to an embodiment of the present invention;
[0026] Figure 6 This is a schematic diagram of a channel data transmission device according to an embodiment of the present invention. Detailed Implementation
[0027] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. 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 should fall within the scope of protection of the present invention.
[0028] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.
[0029] According to an embodiment of the present invention, a method embodiment for channel data transmission is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.
[0030] Figure 1 This is a flowchart of a channel data transmission method according to an embodiment of the present invention, such as... Figure 1 As shown, the method includes the following steps:
[0031] Step S102: Receive the uplink pilot signal sent by the user equipment, and determine the observed sample values of the uplink channel covariance matrix based on the uplink pilot signal.
[0032] The uplink channel covariance matrix is used to represent the statistical estimation matrix of the uplink channel space under the condition of antenna phase deviation and noise interference.
[0033] The aforementioned user equipment refers to mobile terminals or IoT nodes located at the edge of a wireless communication network. The user equipment's role is to transmit uplink pilot signals according to base station scheduling instructions. The user equipment also acts as a transmitter of channel propagation characteristics within the communication link. The user equipment's behavior pattern determines the signal structure during the uplink training phase.
[0034] The aforementioned uplink pilot signal is a known reference signal sequence sent by the user equipment to the base station during the uplink training phase of the frequency division duplex system. The uplink pilot signal serves as a reference input for the base station receiver to estimate channel parameters. By providing a known waveform with specific statistical characteristics, it enables the base station to separate the channel response features from the noisy received signal, thereby providing raw data reference for subsequent channel covariance matrix calculation.
[0035] The observed sample values described above are approximate values of the uplink channel covariance matrix obtained through statistical averaging or matrix operations on the received uplink pilot signals. These observed sample values reflect the statistical correlation characteristics of signals between antenna ports in the uplink channel space under actual physical conditions with antenna phase deviation and noise interference. The observed sample values of the channel covariance matrix serve as input observation data for subsequent iterative improvement algorithms. They are used to approximate the true channel statistical characteristics.
[0036] The uplink channel covariance matrix described above is a mathematical model describing the second-order statistical characteristics of the received signal vectors from multiple antennas in the uplink. The uplink channel covariance matrix represents the statistical estimation matrix of the uplink channel space under conditions of antenna phase deviation and noise interference. Its function is to quantify the structural characteristics of the channel space, reflecting the degree of correlation between signal amplitude and phase between different antenna channels. The uplink channel covariance matrix can characterize the energy distribution and angular distribution characteristics of the channel.
[0037] The antenna phase deviation described above is the vector difference between the actual phase response and the ideal phase response of each antenna element in the antenna array, caused by manufacturing tolerances, temperature variations, or assembly errors. The antenna phase deviation serves to correct the ideal array response model, reflecting the impact of non-ideal hardware environments on the signal phase. Without compensation for the antenna phase deviation, channel estimation errors will accumulate. Accurately estimating and compensating for antenna phase deviation can improve the channel estimation accuracy of large-scale multiple-input multiple-output systems in frequency division duplex mode.
[0038] The aforementioned noise interference refers to random thermal noise and other interfering signals superimposed on the received signal. Noise interference is typically modeled as zero-mean complex Gaussian white noise. Its role is to simulate the unavoidable background noise in real-world communication environments. Noise interference causes observed sample values to deviate from the true channel covariance matrix. During channel estimation, noise interference reduces the signal-to-noise ratio and increases the estimation variance. Channel estimation algorithms need to be robust to accurately extract channel statistical features and antenna hardware error parameters even in the presence of noise.
[0039] In one optional embodiment, the user equipment transmits an orthogonal pilot sequence, i.e., an uplink pilot signal, according to a preset time-frequency resource grid. The base station receives the composite signal containing multi-user pilots through a uniform linear array. The receiving module preprocesses the received RF signal, for example, through down-conversion, analog-to-digital conversion, and pilot removal, to obtain an equivalent baseband received signal vector. The signal processing unit uses the equivalent baseband received signal vectors from multiple time snapshots and calculates the observed sample values of the uplink channel covariance matrix using the sample covariance matrix calculation formula, which is to sum the outer product of the received signal vector of each time snapshot and its conjugate transpose and divide by the total number of samples. The observed sample values include off-diagonal phase distortion information caused by antenna phase deviation.
[0040] In another alternative embodiment, the user equipment transmits random sequence pilots on non-orthogonal pilot resources. The base station performs sparse reconstruction of the received signal using compressed sensing principles. The base station first constructs an overcomplete angular power spectrum dictionary matrix. The base station solves for the sparse coefficient vector based on the received signal vector. The base station then reconstructs the observed sample values of the uplink channel covariance matrix using the sparse coefficient vector and the array response matrix. The determination of the observed sample values of the uplink channel covariance matrix considers not only the statistical characteristics of the directly received signal but also sparse prior knowledge of the signal.
[0041] This application's embodiments, by accurately determining the observed sample values of the uplink channel covariance matrix, can truly reflect the channel statistical characteristics under non-ideal hardware environments. This application's embodiments provide a reliable basis for subsequent joint estimation of antenna phase deviation and angular power spectrum.
[0042] Step S104: Update the initial antenna phase deviation vector and the initial angle power spectrum parameters based on the observed sample values to obtain the target antenna phase deviation vector and the target angle power spectrum parameters.
[0043] The initial antenna phase deviation vector mentioned above is the estimated phase error of each antenna element in the antenna array at the initial moment or a preset zero vector. The initial antenna phase deviation vector is usually expressed as a diagonal matrix or vector. The role of the initial antenna phase deviation vector is to serve as the starting variable for the alternating iterative improvement algorithm, providing the initial search direction for the coordinate descent method. By continuously iterating and updating the initial antenna phase deviation vector, it gradually approximates the actual antenna hardware phase error, thereby correcting the ideal array response model, reflecting the influence of the non-ideal hardware environment on the signal phase, and ensuring the accuracy of the final estimation result.
[0044] The aforementioned initial angular power spectrum parameter describes the initial estimate or uniform distribution assumption of the signal energy distribution in the spatial angular domain. The initial angular power spectrum parameter is used to characterize the initial state of the received or transmitted power per unit angle in a given physical direction. The initial angular power spectrum parameter serves as another starting variable for the alternating iterative improvement algorithm, providing an initial power spectrum estimate for the iterative adaptive method. By fixing the initial antenna phase deviation vector and updating the initial angular power spectrum parameter, the spatial angular distribution estimate is gradually improved, thereby providing angular domain support for constructing an accurate channel covariance matrix.
[0045] The target antenna phase deviation vector mentioned above is the final antenna phase error estimate obtained after convergence of the iteratively improved algorithm. The target antenna phase deviation vector represents the final difference vector between the actual phase response and the ideal phase response of each antenna element in the antenna array due to manufacturing tolerances, temperature changes, or assembly errors. The target antenna phase deviation vector is used to compensate for the distortion of the signal phase caused by the non-ideal hardware environment. By correcting the downlink array response through the accurate antenna phase deviation vector, the base station can accurately generate downlink beams, thereby improving the channel estimation accuracy and communication quality of the large-scale multiple-input multiple-output system in frequency division duplex mode.
[0046] The target angle power spectrum parameters mentioned above are the final spatial angle energy distribution estimates obtained after convergence of the iteratively improved algorithm. These parameters represent the true power distribution of the signal across various discrete angle grids under optimal phase compensation conditions. The target angle power spectrum parameters accurately reflect the spatial propagation characteristics of the uplink channel, utilizing the reciprocity of the uplink and downlink angle power spectra. These parameters can be directly used to reconstruct the downlink channel covariance matrix, thereby overcoming the channel non-reciprocity problem caused by different uplink and downlink frequency bands in frequency division duplex systems, and providing crucial information support for downlink precoding.
[0047] In one optional embodiment, the signal processing unit employs an alternating iterative framework. First, it fixes the initial antenna phase deviation vector and uses an iterative adaptive algorithm to update the initial angular power spectrum parameters based on observed sample values, obtaining updated angular power spectrum parameters. Then, the signal processing unit fixes the updated angular power spectrum parameters and uses a coordinate descent method to update the initial antenna phase deviation vector based on observed sample values, obtaining updated antenna phase deviation vector. This process is repeated until a preset iteration condition is met, and the final updated angular power spectrum parameters and target antenna phase deviation vector are determined as the target angular power spectrum parameters and target antenna phase deviation vector, respectively.
[0048] In another optional embodiment, the signal processing unit introduces regularization constraints to enhance the robustness of the estimation. In each iteration, the initial antenna phase deviation vector is fixed, and the initial angular power spectrum parameters are updated based on the observed sample values using an iterative adaptive algorithm with sparsity constraints, resulting in sparsified updated angular power spectrum parameters. Then, the sparsified updated angular power spectrum parameters are fixed, and the initial antenna phase deviation vector is updated based on the observed sample values using a coordinate descent method with modulus constraints, resulting in an updated antenna phase deviation vector that satisfies the unit modulus constraint. By alternately improving the initial antenna phase deviation vector and the initial angular power spectrum parameters, the covariance matrix mismatch objective function is gradually reduced until a high-precision target antenna phase deviation vector and target angular power spectrum parameters are obtained.
[0049] The embodiments of this application can solve the problems of antenna hardware error and channel spatial distribution estimation by jointly updating the initial antenna phase deviation vector and the initial angular power spectrum parameters, thereby improving the channel estimation accuracy and system robustness.
[0050] Step S106: The target antenna phase deviation vector and target angle power spectrum parameters are processed using the antenna array response matrix of the downlink frequency band to determine the downlink channel covariance matrix.
[0051] The downlink channel covariance matrix is used to represent the statistical estimation matrix of the reconstructed downlink channel space.
[0052] The aforementioned downlink antenna array response matrix is a mathematical model describing the response characteristics of each element of the antenna array to a signal at a specific angle under the downlink carrier frequency. The downlink antenna array response matrix contains complex elements, reflecting the influence of antenna geometry, spacing, and downlink carrier frequency on the signal phase and amplitude. The function of the downlink antenna array response matrix is to establish the mapping relationship between spatial angles and antenna port signals, serving as the basic framework for reconstructing the downlink channel covariance matrix. Combined with angular power spectrum parameters, the spatial structure information of the downlink channel can be generated.
[0053] The downlink channel covariance matrix described above is a mathematical model that describes the second-order statistical characteristics of the transmitted signal vectors from multiple antennas in the downlink. The downlink channel covariance matrix represents the statistical estimation matrix of the reconstructed downlink channel space. Its function is to quantify the structural characteristics of the downlink channel space, reflecting the correlation between signal amplitude and phase between different antenna channels. Through the downlink channel covariance matrix, downlink precoding parameters can be determined, guiding the base station in beamforming, thereby enabling efficient downlink data transmission using uplink estimation information in a frequency division duplex system.
[0054] In one optional embodiment, the signal processing unit first performs phase compensation correction on the antenna array response matrix of the downlink band using the target antenna phase deviation vector. Specifically, the target antenna phase deviation vector is constructed as a diagonal matrix, and a Hadamard product is performed with the antenna array response matrix of the downlink band to obtain the corrected downlink array response. Subsequently, based on the corrected downlink array response and the target angular power spectrum parameters, the signal processing unit calculates the weighted sum of the array response outer product and power weight corresponding to each angular component by using a discrete angular grid summation method, thereby determining the downlink channel covariance matrix and achieving high-precision downlink channel spatial statistical estimation.
[0055] In another optional embodiment, the signal processing unit directly constructs a diagonal power matrix using the target angular power spectrum parameters. It then preprocesses the phase factor matrix corresponding to the target antenna phase deviation vector and the antenna array response matrix of the downlink band to obtain the phase-calibrated downlink array manifold matrix. Next, the signal processing unit performs matrix multiplication on the phase-calibrated downlink array manifold matrix and the diagonal power matrix—that is, multiplying the downlink array manifold matrix by the diagonal power matrix and then by the conjugate transpose of the phase-calibrated downlink array manifold matrix—to directly determine the downlink channel covariance matrix. This process fully utilizes the reciprocity of the uplink and downlink angular power spectra, simplifies the calculation steps, and improves the efficiency of determining the downlink channel covariance matrix.
[0056] This application embodiment utilizes the antenna array response matrix of the downlink band to process the target antenna phase deviation vector and target angular power spectrum parameters, which can accurately reconstruct the downlink channel covariance matrix, overcome the channel non-reciprocity of the frequency division duplex system, and improve downlink precoding performance.
[0057] Step S108: Determine downlink precoding parameters based on the downlink channel covariance matrix, and use the downlink precoding parameters to send downlink data to the user equipment.
[0058] The aforementioned downlink precoding parameters are matrix or vector coefficients used by the base station to perform linear processing on downlink transmitted data. The function of downlink precoding parameters is to map user data onto the transmitted signals of multiple antenna ports, thereby improving the signal distribution in space. Downlink precoding parameters can suppress multi-user interference and enhance the signal strength of the target user. The determination of downlink precoding parameters depends on the downlink channel covariance matrix and aims to maximize link spectral efficiency or the signal-to-interference-to-noise ratio, ensuring the reliability and high speed of downlink data transmission. It is a technical means to achieve spatial multiplexing and beamforming in wireless communication systems.
[0059] The downlink data mentioned above refers to the data or control information that the base station needs to transmit to the user equipment. Downlink data contains a user-specific bit stream. The function of downlink data is to achieve one-way information transmission through the wireless channel. After precoding, the downlink data forms a multi-antenna transmission signal. The transmission quality of downlink data is directly affected by the downlink precoding parameters, which are determined based on the downlink channel covariance matrix. By utilizing these parameters, the base station can accurately control the transmission direction and power allocation of the downlink data, thereby improving the signal-to-noise ratio (SNR) of the user equipment and ensuring the integrity and real-time performance of data transmission.
[0060] In one optional embodiment, the signal processing unit performs eigenvalue decomposition on the downlink channel covariance matrix based on its characteristic distribution, extracts the eigenvector corresponding to the largest eigenvalue as the downlink beamforming vector, and determines the downlink precoding parameters using the downlink beamforming vector. Subsequently, the signal processing unit performs matrix multiplication on the downlink data and the downlink precoding parameters to generate a multi-antenna transmit signal, and transmits the downlink data to the user equipment using a uniform linear array. This process achieves single-user beamforming, effectively focusing energy and improving link gain.
[0061] In another optional embodiment, the signal processing unit uses the minimum mean square error criterion, combined with the downlink channel covariance matrix, noise power, and downlink transmit power, to calculate the downlink precoding matrix, which is then determined as the downlink precoding parameters. Subsequently, the signal processing unit preprocesses the downlink data and the downlink precoding parameters to generate a multi-antenna precoded signal, and transmits the downlink data to the user equipment through the radio frequency front-end. This process considers noise and power constraints, aiming to minimize the mean square error at the receiver, making it suitable for interference management in multi-user scenarios and improving the overall system capacity.
[0062] This application embodiment determines downlink precoding parameters based on the downlink channel covariance matrix and transmits downlink data, achieving high-precision downlink transmission under non-ideal hardware conditions, thereby improving spectral efficiency and communication reliability.
[0063] In this embodiment of the invention, an uplink pilot signal transmitted by a user equipment is received. Based on the uplink pilot signal, observed sample values of the uplink channel covariance matrix are determined. The uplink channel covariance matrix represents the statistical estimation matrix of the uplink channel space under conditions of antenna phase deviation and noise interference. Based on the observed sample values, the initial antenna phase deviation vector and initial angular power spectrum parameters are updated to obtain the target antenna phase deviation vector and target angular power spectrum parameters. The target antenna phase deviation vector and target angular power spectrum parameters are processed using the antenna array response matrix of the downlink band to determine the downlink channel covariance matrix. The downlink channel covariance matrix represents the statistical estimation matrix of the reconstructed downlink channel space. Downlink precoding parameters are determined based on the downlink channel covariance matrix, and downlink data is transmitted to the user equipment using the downlink precoding parameters. This application utilizes the reciprocity of uplink and downlink angular power spectra to perform phase compensation correction on the antenna array response matrix of the downlink channel using the high-precision angular power spectrum and antenna phase deviation vector obtained by joint estimation on the uplink. This accurately maps and reconstructs the statistical estimation of the uplink channel space into the statistical estimation matrix of the downlink channel space, achieving the goal of reconstructing the downlink channel covariance matrix under non-ideal hardware conditions. This improves the accuracy of downlink precoding parameters and solves the technical problem of low accuracy of channel data transmission in related technologies.
[0064] Optionally, the initial antenna phase deviation vector and the initial angular power spectrum parameter are updated based on the observed sample values to obtain the target antenna phase deviation vector and the target angular power spectrum parameter. This includes: in the current iteration, fixing the initial antenna phase deviation vector, using an iterative adaptive algorithm to update the initial angular power spectrum parameter based on the observed sample values to obtain the updated angular power spectrum parameter; fixing the updated angular power spectrum parameter, using the coordinate descent method to update the initial antenna phase deviation vector based on the observed sample values to obtain the updated antenna phase deviation vector; and determining the target antenna phase deviation vector and the target angular power spectrum parameter based on the updated angular power spectrum parameter and the updated antenna phase deviation vector.
[0065] The current iteration round mentioned above refers to a specific computational cycle within the alternating iterative improvement process. The current iteration round identifies the state to which the algorithm has reached. Its role is to decompose the complex non-convex joint improvement problem into two relatively simple subproblems, and to gradually approximate the globally optimal solution by alternately updating the variables. The current iteration round ensures that the algorithm utilizes the latest estimate from the previous round to improve the estimation of the current variables in each loop, thereby guaranteeing the stability and orderliness of the convergence process.
[0066] The iterative adaptive algorithm described above is a signal processing technique for discrete power spectrum estimation. It updates the angular power spectrum parameters under a fixed antenna phase offset vector. The algorithm works by adaptively adjusting the power estimate at each angle grid point using the principle of iterative weighted least squares to minimize the difference between the observation covariance matrix and the model covariance matrix. The iterative adaptive algorithm offers advantages such as high resolution and the ability to accurately estimate angular power distributions in complex multipath environments without requiring pre-defined signal sparsity.
[0067] The updated angular power spectrum parameters described above are intermediate results obtained in the current iteration after improving the initial angular power spectrum parameters using an iterative adaptive algorithm. These updated parameters reflect the optimal angular power distribution estimate under the current fixed antenna phase offset vector constraint. Their purpose is to provide more accurate spatial structure information for the next stage of phase offset improvement. By minimizing the covariance matrix mismatch objective function, the updated angular power spectrum parameters improve the accuracy of angle estimation.
[0068] The coordinate descent method described above is a numerical improvement algorithm for solving multivariate improvement problems. It improves the initial antenna phase deviation vector under the condition of fixed update angle power spectrum parameters. The coordinate descent method decomposes the multidimensional improvement problem into a series of single-variable improvement problems, updating the phase deviation factor of each antenna element one by one, making the objective function monotonically decrease in each single-variable improvement. The coordinate descent method has low computational complexity and is easy to implement, making it particularly suitable for solving non-convex improvement problems with separable variable structures, and can efficiently obtain approximate optimal solutions to antenna phase deviations.
[0069] The updated antenna phase deviation vector described above is an intermediate result obtained in the current iteration after improving the initial antenna phase deviation vector using the coordinate descent method. The updated antenna phase deviation vector reflects the optimal antenna phase error estimate under the current fixed angle power spectrum parameter constraints. Its purpose is to compensate for phase distortion introduced by non-ideal hardware environments. By improving each antenna element individually, the updated antenna phase deviation vector gradually approximates the true antenna phase deviation, providing a more accurate phase correction model for the next angle power spectrum estimation, thereby improving the robustness of the overall channel estimation.
[0070] In one optional embodiment, the signal processing unit is configured to store an alternating iterative improvement program. First, it initializes the initial antenna phase deviation vector and the initial angular power spectrum parameters, entering the current iteration round. The signal processing unit fixes the initial antenna phase deviation vector and calls the iterative adaptive algorithm module to calculate the updated angular power spectrum parameters based on the observed sample values. The signal processing unit then fixes the updated angular power spectrum parameters and calls the coordinate descent method module to traverse each antenna element, calculating the updated antenna phase deviation vector based on the univariate improvement conditions. The signal processing unit determines whether the updated angular power spectrum parameters and the updated antenna phase deviation vector satisfy the preset iteration conditions. If the updated angular power spectrum parameters and the updated antenna phase deviation vector do not satisfy the preset iteration conditions, the updated parameters are used as the initial values for the next round, and the process is repeated. If the updated angular power spectrum parameters and the updated antenna phase deviation vector satisfy the preset iteration conditions, the iteration terminates, and the final updated angular power spectrum parameters and the target antenna phase deviation vector are determined as the target angular power spectrum parameters and the target antenna phase deviation vector, respectively, completing the joint estimation process.
[0071] The embodiments of this application improve the joint estimation of antenna phase deviation and angular power spectrum through alternating iterations, thereby reducing algorithm complexity and improving channel estimation accuracy and system robustness in non-ideal hardware environments.
[0072] Optionally, determining the target antenna phase deviation vector and the target angle power spectrum parameter based on the updated angle power spectrum parameter and the updated antenna phase deviation vector includes: in response to the updated angle power spectrum parameter and the updated antenna phase deviation vector not meeting the preset iteration conditions, determining the updated angle power spectrum parameter as the initial angle power spectrum parameter, determining the updated antenna phase deviation vector as the initial antenna phase deviation vector, and repeating the execution steps of the current iteration round until the updated angle power spectrum parameter and the updated antenna phase deviation vector meet the preset iteration conditions; in response to the updated angle power spectrum parameter and the updated antenna phase deviation vector meeting the preset iteration conditions, determining the updated angle power spectrum parameter as the target angle power spectrum parameter, and determining the updated antenna phase deviation vector as the target antenna phase deviation vector.
[0073] The aforementioned preset iteration conditions are a set of thresholds or criteria used to determine whether the alternating iterative improvement algorithm has converged. These conditions measure the difference between the updated parameters obtained in the current iteration and those of the previous iteration, or the stability of the objective function value. The preset iteration conditions determine whether the algorithm continues to the next iteration or terminates the computation. They ensure that the algorithm stops when sufficient accuracy is achieved, avoiding wasted computational resources, and guarantee that the final output objective parameters have an acceptable error range. This is a key mechanism for controlling the algorithm's convergence and computational efficiency.
[0074] In one optional embodiment, the signal processing unit is configured with an iteration control module. After the current iteration round ends, the signal processing unit calculates the difference norm between the updated angular power spectrum parameters and the updated antenna phase deviation vector and the corresponding parameters from the previous round. The signal processing unit compares the difference norm with a convergence threshold in the preset iteration conditions. If the difference norm is greater than the convergence threshold, it indicates that the preset iteration conditions are not met. The signal processing unit copies the updated angular power spectrum parameters and assigns them to the initial angular power spectrum parameter register, and copies the updated antenna phase deviation vector and assigns it to the initial antenna phase deviation vector register. Subsequently, the signal processing unit returns to the starting step of the current iteration round and re-executes the angular power spectrum update and phase deviation update operations. If the difference norm is less than or equal to the convergence threshold, it indicates that the preset iteration conditions are met. The signal processing unit terminates the iteration loop, locks the current updated angular power spectrum parameters as the target angular power spectrum parameters, locks the current updated antenna phase deviation vector as the target antenna phase deviation vector, and outputs the final estimation result for subsequent channel reconstruction.
[0075] This application embodiment dynamically determines the convergence state by preset iteration conditions, realizes the automatic termination of the alternating iteration process, ensures the estimation accuracy of the target parameters, avoids the waste of computing resources caused by excessive iteration, and improves the execution efficiency of the algorithm.
[0076] Optionally, an iterative adaptive algorithm is used to update the initial angle power spectrum parameters based on the observed sample values to obtain the updated angle power spectrum parameters. This includes: constructing an initial estimation model of the uplink channel covariance matrix based on the initial antenna phase deviation vector and the observed sample values; determining the signal decorrelation estimate of the discrete spatial angle points based on the initial angle power spectrum parameters and the initial estimation model; and updating the initial angle power spectrum parameters based on the signal decorrelation estimate to obtain the updated angle power spectrum parameters.
[0077] The aforementioned discrete spatial angle point signal decorrelation estimates are complex amplitude estimates calculated for each discrete angle grid point during the iterative adaptive algorithm process. These estimates are used to eliminate mutual interference between multipath components and the influence of noise. Their function is to separate independent signal components in a specific direction from the mixed received signal through adaptive weighted filtering. The discrete spatial angle point signal decorrelation estimates reflect the effective power contribution of the signal in that angular direction and are direct input data for updating the angular power spectrum parameters. Their accuracy determines the resolution and accuracy of the final angular power spectrum estimation.
[0078] In one optional embodiment, the signal processing unit first calculates the observed sample covariance matrix of the uplink channel covariance matrix based on the initial antenna phase deviation vector and the observed sample values. It then constructs the phase-calibrated array response matrix using the initial antenna phase deviation vector, thereby building an initial estimation model for the uplink channel covariance matrix. The signal processing unit initializes the signal decorrelation estimates for discrete spatial angle points and enters an iterative loop. In each iteration, the signal processing unit uses the inverse of the current covariance matrix to weight the observed samples, calculating the signal decorrelation estimate for each discrete angle point. Subsequently, the signal processing unit calculates and updates the initial angle power spectrum parameters based on the magnitude, square, and average of the signal decorrelation estimates until the convergence condition is met, obtaining the final updated angle power spectrum parameters.
[0079] The embodiments of this application achieve high-resolution spectral estimation by constructing an initial estimation model and updating the angular power spectrum using the signal decorrelation estimate, effectively suppressing interference and improving the accuracy and robustness of angular power spectrum estimation.
[0080] Optionally, the initial antenna phase deviation vector is updated using the coordinate descent method based on the observed sample values to obtain the updated antenna phase deviation vector. This includes: traversing multiple antennas in the base station array, using the initial antenna phase deviation vector corresponding to the currently traversed antenna as the current variable to be updated, and using the initial antenna phase deviation vectors corresponding to the other antennas as fixed variables; constructing a target cost function based on the current variable to be updated, the fixed variables, and the observed sample values; converting the target cost function into a univariate function of the current variable to be updated; determining the target phase deviation vector of the currently traversed antenna based on the extremum condition of the univariate function; and determining the updated antenna phase deviation vector based on the target phase deviation vectors of multiple antennas.
[0081] The aforementioned base station array antenna consists of multiple antennas physically deployed on the base station side for transmitting and receiving wireless signals. These antennas form the spatial degrees of freedom of a large-scale multiple-input multiple-output (MIMO) system. Their function is to enhance signal transmission gain through spatial diversity and beamforming. Each antenna element in the base station array corresponds to an independent phase deviation factor. These antenna elements collectively constitute the physical hardware foundation for channel estimation and precoding processing; their spatial distribution and quantity directly affect the system's performance.
[0082] The initial antenna phase deviation vector corresponding to the currently traversed antenna is the phase error estimate of a single antenna element that the coordinate descent method focuses on improving in the current iteration step. The initial antenna phase deviation vector corresponding to the currently traversed antenna is used to represent the phase offset introduced by the non-ideal characteristics of the hardware of a specific antenna channel. The role of the initial antenna phase deviation vector corresponding to the currently traversed antenna is to serve as the core variable of the single-variable improvement problem. By adjusting this variable individually, the global objective function is minimized. The update of the initial antenna phase deviation vector corresponding to the currently traversed antenna aims to eliminate the phase distortion of the specific antenna channel, thereby improving the accuracy of the overall array response.
[0083] The initial antenna phase deviation vectors for the other antennas mentioned above are the set of antenna element phase error estimates that remain unchanged in the current improvement step. These initial antenna phase deviation vectors represent the phase deviation states of all antenna channels except the antenna currently being improved. The purpose of these vectors is to serve as fixed parameters in the construction of the cost function for the current antenna, ensuring the integrity of the model structure when improving individual variables. Fixing the initial antenna phase deviation vectors for the other antennas decomposes the complex non-convex improvement problem into a series of simpler, single-variable improvement subproblems, reducing computational complexity.
[0084] The objective cost function described above is a mathematical function that measures the difference between the model covariance matrix and the observed sample covariance matrix. The objective cost function quantifies the estimation error under the current antenna phase deviation vector assumption. Its role is to provide a convergence criterion for improving the algorithm; by minimizing the objective cost function, the algorithm can find the antenna phase deviation parameters that best fit the observed data to the model. The objective cost function reflects the magnitude of the model residuals and is the basis for univariate improvement using the coordinate descent method, determining the accuracy of the final estimation result.
[0085] The aforementioned univariate function is a functional expression of the multidimensional objective cost function, with other variables fixed, that only relates to the phase deviation factor of the antenna to be updated. The univariate function simplifies the improvement process, transforming multivariate joint improvement into a single-variable extremum problem. Its role is to eliminate the influence of cross terms through mathematical derivation, allowing the algorithm to directly find a better solution through analytical methods or simple search. The univariate function ensures the stability and convergence of the improvement process, reducing the computational burden of real-time signal processing.
[0086] The extremum conditions of a single-variable function, as described above, refer to the mathematical conditions that derive the minimum or maximum value of the function. These conditions are used to determine the optimal phase deviation factor for the currently traversed antenna. The extremum conditions are typically obtained by taking the derivative and setting it to zero, or by applying the principle of maximizing the complex modulus. The purpose of these conditions is to provide a formula for calculating the analytical solution, allowing the algorithm to directly calculate the optimal antenna phase deviation without requiring time-consuming numerical searches. Furthermore, these extremum conditions ensure that each single-variable update reduces the objective function value, thus guaranteeing the monotonically convergent nature of the overall algorithm.
[0087] The target phase deviation vector of the currently traversed antennas is a better phase error estimate for a specific antenna element obtained after univariate improvement. The target phase deviation vector of the currently traversed antennas represents the best phase compensation value that minimizes the cost function in the current iteration step. The target phase deviation vector of the currently traversed antennas replaces the old initial value and updates the global vector. The iterative improvement of the phase deviation vector of the entire antenna array is achieved through the successive updates of the target phase deviation vector of the currently traversed antennas. By accumulating the better updates of all antennas, a high-precision antenna phase deviation estimate is finally obtained.
[0088] In one optional embodiment, the signal processing unit is configured with a coordinate descent improvement module. The signal processing unit initializes the antenna index to the first antenna and enters a traversal loop. The signal processing unit uses the initial phase deviation factor of the first antenna as the current variable to be updated, and the phase deviation factors of all other antennas as fixed variables. Based on the current variable to be updated, the fixed variables, and the observed sample values, the signal processing unit constructs a target cost function. Using matrix trace operations and the Hadamard product property, the signal processing unit simplifies the target cost function into a single-variable function about the current variable to be updated. The signal processing unit differentiates the single-variable function or derives the extremum condition using complex number properties to calculate the target phase deviation vector of the first antenna. The signal processing unit updates the target phase deviation vector of the first antenna into the vector and increments the antenna index by one, repeating the above steps until all antennas are traversed, finally outputting the updated antenna phase deviation vector.
[0089] The embodiments of this application improve the decomposition of multidimensional phase deviation into a single-variable problem by using the coordinate descent method, thereby reducing computational complexity and achieving efficient and high-precision antenna phase deviation estimation.
[0090] Optionally, the downlink channel covariance matrix is determined by processing the target antenna phase deviation vector and the target angular power spectrum parameters using the antenna array response matrix of the downlink frequency band. This includes: performing phase compensation correction on the antenna array response matrix of the downlink frequency band using the target antenna phase deviation vector to obtain the corrected downlink array response; and determining the downlink channel covariance matrix based on the corrected downlink array response and the target angular power spectrum parameters.
[0091] The corrected downlink array response described above is the downlink antenna array response matrix after compensation for the target antenna phase offset vector. The corrected downlink array response represents the true downlink array response characteristics after eliminating hardware phase errors. The corrected downlink array response serves as a correct mapping matrix connecting the spatial angle domain and the antenna port domain. By combining the target antenna phase offset vector with the downlink frequency band antenna array response matrix through Hadamard product operations, the corrected downlink array response ensures the accuracy of subsequent channel covariance matrix reconstruction based on angular power spectrum parameters, avoiding estimation errors caused by phase mismatch.
[0092] In one optional embodiment, the signal processing unit first constructs a diagonal matrix from the target antenna phase deviation vector, where the diagonal elements of the diagonal matrix are the phase factors in the target antenna phase deviation vector. The signal processing unit then performs a Hadamard product operation between the antenna array response matrix of the downlink band and the diagonal matrix to obtain the corrected downlink array response. Subsequently, the signal processing unit constructs a diagonal power matrix based on the target angular power spectrum parameters. The signal processing unit then performs matrix multiplication between the corrected downlink array response and the diagonal power matrix, i.e., multiplying the corrected downlink array response by the diagonal power matrix and then by the conjugate transpose of the corrected downlink array response, directly determining the downlink channel covariance matrix. This process fully utilizes the reciprocity of the uplink and downlink angular power spectra, simplifies the calculation steps, and improves the efficiency of determining the downlink channel covariance matrix.
[0093] This application embodiment utilizes the antenna array response matrix of the downlink band to process the target antenna phase deviation vector and target angular power spectrum parameters, which can accurately reconstruct the downlink channel covariance matrix, overcome the channel non-reciprocity of the frequency division duplex system, and improve downlink precoding performance.
[0094] Optionally, the method further includes: constructing an antenna array response matrix for the downlink band based on the center frequency of the downlink band and the geometry of the base station antenna array.
[0095] The center frequency of the downlink band mentioned above is the geometric center frequency value of the carrier band in the downlink communication link. The center frequency of the downlink band is used to define the main frequency points for downlink signal transmission. The center frequency of the downlink band determines the wavelength parameters of electromagnetic waves propagating in space. The center frequency of the downlink band determines the angular resolution and sidelobe level of beam scanning. The accuracy of the center frequency of the downlink band directly affects the accuracy of downlink channel covariance matrix reconstruction and is a fundamental parameter for frequency band planning and signal processing in wireless communication systems.
[0096] The geometry of the base station antenna array described above is a topological shape describing the physical spatial arrangement of the antenna elements on the base station side. The geometry of the base station antenna array includes the number of antenna elements, their spacing, arrangement direction, and the position of the phase center. The geometry of the base station antenna array determines the radiation pattern characteristics and spatial beamforming capabilities of the antenna array. The geometry of the base station antenna array is the physical basis for constructing the antenna array response matrix for the downlink frequency band. By influencing the phase delay of the received signal in each antenna element, the geometry of the base station antenna array determines the phase relationship between the elements in the array response vector. The design of the base station antenna array's geometry directly affects the system's spatial diversity gain and interference suppression performance.
[0097] In one optional embodiment, the signal processing unit first obtains the center frequency parameter of the downlink band. The center frequency parameter is provided by the frequency band configuration module of the wireless communication system. The signal processing unit then reads the geometric parameters of the base station antenna array, including the total number of antenna elements, the physical spacing between adjacent antenna elements, and the array arrangement type. Based on the center frequency of the downlink band, the signal processing unit calculates the wavelength of the electromagnetic wave in vacuum. Combining the geometry of the base station antenna array, the signal processing unit uses the array response vector calculation formula to generate array response vectors corresponding to each discrete angle grid point. The signal processing unit concatenates the array response vectors corresponding to the discrete angle grid points column-wise to construct the antenna array response matrix of the downlink band. The signal processing unit stores the constructed antenna array response matrix of the downlink band in memory for subsequent use by the phase compensation correction and downlink channel covariance matrix reconstruction modules.
[0098] This application embodiment constructs the antenna array response matrix of the downlink frequency band based on the center frequency of the downlink frequency band and the geometry of the base station antenna array, providing an accurate physical model basis for downlink channel reconstruction and ensuring the accuracy of channel estimation and precoding.
[0099] Optionally, receiving the uplink pilot signal transmitted by the user equipment includes: receiving the uplink pilot signal transmitted by the user equipment through a uniform linear array; transmitting downlink data to the user equipment using downlink precoding parameters includes: transmitting downlink data to the user equipment through a uniform linear array using downlink precoding parameters.
[0100] The aforementioned uniform linear array is an antenna array structure composed of multiple omnidirectional antenna elements arranged at equal intervals along a straight line. Uniform linear arrays are used for signal reception and transmission at the base station side. Their function is to form a directional beam using a fixed phase difference between antenna indices. The mathematical model of a uniform linear array is simple and easy to process. Uniform linear arrays are the fundamental hardware platform for constructing the antenna array response matrix for the downlink frequency band. By adjusting the excitation phase and amplitude between each element, uniform linear arrays can achieve beamforming and interference suppression of spatial signals.
[0101] In one optional embodiment, the signal processing unit first configures the uniform linear array to operate in a receiving state. The user equipment (UE) sends uplink pilot signals to the uniform linear array according to scheduling instructions. Each antenna element of the uniform linear array synchronously captures the radio frequency (RF) signals and converts them into baseband digital signals. The signal processing unit performs depiloting and normalization processing on the baseband signals to obtain the observed sample values of the uplink channel covariance matrix. Subsequently, the signal processing unit executes a channel estimation algorithm to determine the downlink precoding parameters after the downlink data generation is complete. The signal processing unit applies the downlink precoding parameters to the downlink data to generate a multi-antenna transmit signal. The signal processing unit controls the uniform linear array to switch to transmit mode, using the uniform linear array to send the precoded downlink data to the target UE, completing one bidirectional communication interaction.
[0102] This application embodiment achieves efficient spatial beamforming by transmitting and receiving signals using a uniform linear array and combining it with downlink precoding parameters, thereby improving the system's transmission performance and anti-interference capability under non-ideal hardware conditions.
[0103] Optionally, determining downlink precoding parameters based on the downlink channel covariance matrix includes: extracting the downlink beamforming vector corresponding to the largest eigenvalue according to the characteristic distribution of the downlink channel covariance matrix, and determining the downlink precoding parameters based on the downlink beamforming vector; or, using the minimum mean square error to determine the downlink precoding matrix based on the downlink channel covariance matrix, noise power, and downlink transmission power, and determining the downlink precoding matrix as the downlink precoding parameters.
[0104] The characteristic distribution of the downlink channel covariance matrix described above is a set of statistical characteristics describing the magnitude of each eigenvalue and the direction of its corresponding eigenvector in the downlink channel covariance matrix. The characteristic distribution of the downlink channel covariance matrix reflects the main concentration direction and relative intensity of spatial energy in the downlink channel. The largest eigenvalue in the characteristic distribution of the downlink channel covariance matrix corresponds to the strongest signal propagation path. The characteristic distribution of the downlink channel covariance matrix determines the potential capacity of the system and the upper limit of beamforming gain. The characteristic distribution of the downlink channel covariance matrix is obtained through eigenvalue decomposition, and its accuracy directly affects the improvement effect of precoding parameters.
[0105] The downlink beamforming vector corresponding to the maximum eigenvalue mentioned above is the eigenvector corresponding to the maximum eigenvalue in the eigenvalue decomposition of the downlink channel covariance matrix. This vector represents the main propagation direction with the strongest energy in the downlink channel space. Its function is to focus the transmitted signal energy onto the main path to maximize the signal strength at the receiver. This vector is also the precoding vector for single-user beamforming. It overcomes the effects of channel fading, simplifies precoding calculations, and exhibits superior spectral efficiency in low signal-to-noise ratio or single-user scenarios.
[0106] The downlink beamforming vector described above is a complex vector used for spatial weighting of the downlink transmitted signal. The downlink beamforming vector controls the transmit phase and amplitude of each antenna element in a uniform linear array. By forming a specific spatial beam pattern, the downlink beamforming vector concentrates energy towards the target user. The design of the downlink beamforming vector is based on the characteristic distribution of the downlink channel covariance matrix. The selection of the downlink beamforming vector aims to maximize link gain and effectively suppress sidelobe interference. The accuracy of the downlink beamforming vector determines the transmission quality of the downlink data.
[0107] The minimum mean square error (MME) mentioned above measures the statistical average of the squared errors between the precoded output signal and the desired signal. MME is used to evaluate precoding performance. Its role is to provide the objective function for improving the precoding matrix. MME comprehensively considers the effects of signal power, interference, and noise, ensuring the reliability of the received signal and making it suitable for scenarios with noise and multi-user interference. The calculation of MME depends on the downlink channel covariance matrix, noise power, and downlink transmit power, aiming to balance signal enhancement and interference suppression. MME is the basis for determining the downlink precoding matrix.
[0108] The noise power mentioned above is the statistical average power of additive white Gaussian noise at the receiver. Noise power is used to quantify the background interference level in the communication system. The higher the noise power, the more conservative the precoding strategy needs to be to maintain link stability.
[0109] The downlink transmit power mentioned above is the total transmit power limit used by the base station to transmit downlink data. Downlink transmit power is used to constrain the norm of the precoding matrix. Its role is to meet the linear range requirements of the power amplifier in the RF front-end, and it serves as a normalization factor when determining the downlink precoding matrix. The downlink transmit power allocation strategy affects the interference level among multiple users, and it needs to maximize user rates while satisfying power constraints. Any remaining downlink transmit power can be used for other control signaling or reference signals, helping to reduce energy consumption and extend equipment lifespan.
[0110] The aforementioned downlink precoding matrix is a complex matrix used for spatial preprocessing of multi-user downlink data streams. Its function is to process data streams from multiple users simultaneously. By linearly combining user data, the downlink precoding matrix achieves spatial interference cancellation or enhancement. It is typically constructed based on the minimum mean square error criterion, enabling it to adapt to complex channel conditions in multi-user scenarios. The column vectors of the downlink precoding matrix correspond to the precoding vectors of different users. The design of the downlink precoding matrix must meet power constraints and interference management requirements.
[0111] In one optional embodiment, the signal processing unit first performs eigenvalue decomposition on the downlink channel covariance matrix. The signal processing unit extracts the largest eigenvalue and its corresponding eigenvector from the downlink channel covariance matrix. After normalizing the eigenvector corresponding to the largest eigenvalue, the signal processing unit determines it as the downlink beamforming vector, which is then directly assigned to the downlink precoding parameters. The signal processing unit then uses the downlink precoding parameters to precode the downlink data. If the system is configured in multi-user mode, the signal processing unit obtains the downlink channel covariance matrix, noise power, and downlink transmit power. The signal processing unit constructs an improved problem based on the minimum mean square error criterion, solves this improved problem to obtain the downlink precoding matrix, and determines the downlink precoding matrix as the downlink precoding parameters. The signal processing unit then uses the downlink precoding parameters to precode the multi-user downlink data stream, completing the precoding parameter determination process.
[0112] The embodiments of this application determine the downlink precoding parameters through two optional methods, which not only ensures better beamforming in single-user scenarios, but also meets the interference management requirements in multi-user scenarios, thereby improving the flexibility and performance of the system.
[0113] The technical solution of this application embodiment can be divided into six parts: the first part is the system model and signal reception, the second part is the improvement problem construction, the third part is the angle power spectrum estimation, the fourth part is the phase deviation calibration, the fifth part is the alternating iteration, and the sixth part is the result output.
[0114] The first part (System Model and Signal Reception) establishes a physical model considering antenna phase deviation and derives the uplink and downlink channel covariance matrices, providing a fundamental mathematical expression for subsequent steps. Building upon this, the second part (Improved Problem Construction) utilizes the model derived in the first part to transform the complex non-convex joint estimation problem into a framework aimed at minimizing covariance matrix mismatch, clarifying the core variables to be solved: the angular power spectrum and the phase deviation vector. Subsequently, the third part (Angle Power Spectrum Estimation) and the fourth part (Phase Deviation Calibration) propose specific solution strategies for the two key variables in this framework: an iterative adaptive method to estimate the angular power spectrum when the phase deviation is fixed, and a coordinate descent method to calibrate the phase deviation step-by-step when the angular power spectrum is fixed, with each being a prerequisite for the other. Finally, the fifth part (Alternating Iteration) organically combines the solution steps of the third and fourth parts, alternately performing spectrum estimation and phase calibration until the objective function converges, achieving the search for a globally optimal solution. Finally, the sixth part (output results) uses the high-precision angular power spectrum and phase deviation vector obtained after convergence in the fifth part, combined with the downlink array response to reconstruct the downlink channel covariance matrix, thus completing the complete technical chain from uplink observation to downlink channel information acquisition. Each part is interconnected and together realizes high-precision channel covariance estimation under non-ideal hardware conditions.
[0115] In the system modeling and signal reception phase, the uplink training phase receives pilot signals independently transmitted by users, establishes a channel model considering antenna phase deviation, and derives the uplink and downlink channel covariance matrices.
[0116] Base station equipped A uniform linear array (ULA) of antennas serves multiple single-antenna user equipment. During the uplink training phase, users independently transmit pilot signals, and the base station receives the signals. for:
[0117] ;
[0118] in, Indicates the first The original received signal vectors of the base station corresponding to each uplink training sample, and the pilot signals have not yet been removed or normalized. This represents the uplink channel vector corresponding to the sample. Indicates the known pilot symbols transmitted by the user equipment. This represents the additive noise vector before pilot removal. The training sample number. This represents the total number of training samples in the uplink. The adjacency of a vector and a scalar indicates scalar multiplication. "" indicates that the signal vector and the noise vector are added together.
[0119] This represents the total number of antennas configured in the uniform linear array of the base station, which will be used subsequently in this paper. and The meaning is the same. The equivalent received signal vector after removing the top pilot. The calculation formula is as follows:
[0120] ;
[0121] in, Indicates the first The equivalent received signal vector after removing the known pilot signal from each sample. Indicates the first The uplink channel vector of each sample. This represents the equivalent additive noise vector after pilot normalization. For the training sample index, " "" indicates vector addition.
[0122] This application's embodiments consider a wide-range stationary uncorrelated scattering model for the channel, which assumes stationary second-order statistics. Furthermore, the phase deviation matrix is considered. Channel modeling The calculation formula is as follows:
[0123] ;
[0124] in, To indicate carrier frequency place 3D channel vector, This indicates the uplink or downlink carrier frequency being examined. The angular power spectrum characterizes the received or transmitted power per unit angle in a given physical direction. The amplitude corresponding to the angular power spectrum is represented by θ, which represents the physical angle of arrival or departure. , This represents the lower limit of the angle integration interval. This represents the upper limit of the angle integration interval. This represents a complex stochastic small-scale fading process with unit power in the angular domain. Angular reciprocity is modeled by assuming that the angular power spectra (APS) of the uplink (UL) and downlink (DL) are identical. Let M×M be the diagonal matrix composed of the phase deviation factors of each antenna, where... The set representing antenna phase deviations, Represents the imaginary unit and ² = -1, The base of the natural exponential function is represented by M, and the number of base station antennas is represented by M. This represents the integral element of the angle. Furthermore, This represents the frequency-dependent antenna array response. The elements are:
[0125] ;
[0126] in, Represents the array response vector The One element, This is the antenna index, with values ranging from 1 to... , Represents the base of the natural exponential function. The imaginary unit, Pi For carrier frequency, The physical spacing between adjacent antennas. Let be the physical angle of the signal, and c be the speed of light in a vacuum. for The sine function, Indicates the first The position of the first antenna element is offset relative to the first antenna element.
[0127] This application's embodiments consider a uniform linear array, where the antenna exhibits a non-negligible phase deviation. Therefore, the antenna array response vector is... Modified to an antenna phase offset matrix, diagonal elements This represents the phase deviation vector composed of the unit mode phase factors of each antenna, where express Phase deviation of each antenna, For the first Phase deviation of the root antenna, For antenna index, The total number of antennas, The base of the natural exponential function, The imaginary unit is used. Assuming the uplink and downlink angular power spectra are reciprocal in a frequency division duplex system, the channel covariance matrix... It is expressed as follows:
[0128] ;
[0129] ;
[0130] ;
[0131] in, Represents frequency The M×M channel space covariance matrix at that location, Indicates statistical expectation. For frequency The channel vector at that location, express The conjugate transpose of . It represents the outer product of the channel vectors. This is the diagonal matrix of antenna phase deviation. For angular power spectrum, For array response vectors, for Conjugate transpose.
[0132] The phase deviation vector, For the outer product, Hadamard product of corresponding element-wise multiplication of matrices of the same dimension, with upper and lower limits of integration. , Give the effective angle range. θ is the angle variable. For angle integral infinitesimal element, This refers to the carrier frequency. (Bold) This represents the phase deviation vector, while In These represent differential notation, and their meanings differ. Equivalently, they represent the uplink and downlink channel covariance matrices. The calculation formula is as follows:
[0133] ;
[0134] in, This represents the channel covariance matrix at the uplink carrier frequency, including the effects of antenna phase deviation. Indicates the uplink carrier frequency. This represents the uplink array response vector. This represents the Hadamard product of corresponding elements of matrices of the same dimension.
[0135] ;
[0136] in, This represents the channel covariance matrix at the downlink carrier frequency, including the effects of antenna phase offset. Indicates the downlink carrier frequency. This represents the downlink array response vector. This represents the transpose of the downlink array response vector. This represents the Hadamard product of corresponding elements of matrices of the same dimension.
[0137] In the problem-solving phase, a joint improvement model that minimizes the covariance matrix mismatch is constructed based on the received signal, while simultaneously handling phase offset compensation and spectral estimation.
[0138] This application proposes an alternating improvement framework that simultaneously handles phase offset compensation and spectral estimation. During the uplink training phase, from the received signal... Iterative estimation of angular power spectrum And the phase deviation vector. Then estimate the downlink channel covariance matrix.
[0139] This application proposes an efficient channel covariance estimation method that considers antenna phase deviation, solving the following complex non-convex problem:
[0140] ;
[0141] ;
[0142] ;
[0143] in, Indicates in variable and Find the solution that minimizes the objective function. The angular power spectrum to be estimated, Let be the antenna phase deviation vector to be estimated. This represents the target or observation covariance matrix obtained from the uplink received samples before the antenna phase bias is superimposed.
[0144] in, This indicates that the following constraints are satisfied. The integral sign indicates that in to Integrating within the angular range, For uplink carrier frequency, This is the uplink array response vector. for The conjugate transpose of .
[0145] in, express The One element, is the base of the natural exponential function. The imaginary unit, For the first Phase deviation of the root antenna, The total number of antennas, and This indicates that the modulus of each phase factor is always 1.
[0146] The key to the calculation is the iterative estimation of the angular power spectrum. and antenna phase deviation vector .
[0147] In the angular power spectrum estimation stage, the phase deviation vector is fixed, and the discrete power spectrum is estimated using the Iterative Adaptive Approach (IAA) to obtain the initial uplink and downlink channel covariance matrices.
[0148] The embodiments of this application first utilize an iterative adaptive method algorithm to estimate... Uplink channel covariance matrix after phase calibration and downlink channel covariance matrix It can be discretized and represented as:
[0149] ;
[0150] in, This represents the uplink channel covariance matrix after phase calibration, reconstructed using a discrete angle grid. The summation sign indicates that... to Sum of all discrete angular components, For angular grid indexing, The total number of discrete angle grid points. For the first Power at each angle grid point. This is the corresponding uplink array response vector. for The conjugate transpose of . For the first A discrete angle.
[0151] in, It is the array manifold matrix composed of the column-wise array response vectors of the upper array. This is a discrete angular power diagonal matrix. This represents the covariance matrix formed after the discrete angular power is mapped onto the uplink array manifold. for The conjugate transpose of .
[0152] ;
[0153] in, This represents the downlink channel covariance matrix after phase calibration, reconstructed using a discrete angle grid. The summation sign indicates that... to Sum of all discrete angular components, For angular grid indexing, The total number of discrete angle grid points. For the first Power at each angle grid point. This is the corresponding downlink array response vector. for The conjugate transpose of . For the first Discrete angles This is the downlink carrier frequency.
[0154] in, Let be the array manifold matrix composed of all downlink array response vectors column-wise. This is a discrete angle power diagonal matrix shared with the uplink. This represents the downlink covariance matrix reconstructed using the reciprocity of the angular power spectrum. for .
[0155] ;
[0156] in, Denotes the upper array manifold matrix, symbol This means "defined as". Indicates the first Discrete angles Uplink frequency The array response column vector below, to For all discrete angle grid points, The total number of grid points is represented by square brackets, which indicate that the column vectors are concatenated into a matrix in sequence.
[0157] ;
[0158] in, Denotes the downlink array manifold matrix, symbol This means "defined as". Indicates the first Discrete angles Downlink frequency The array response column vector below, to For all discrete angle grid points, The total number of grid points is represented by square brackets, which indicate that the column vectors are concatenated into a matrix in sequence.
[0159] and, from[- , Uniform selection. , = . Represented by vector The elements are constructed using a power diagonal matrix with the main diagonal as the base. For diagonalization operators, = It is a discrete angular power vector. For the first Non-negative power of grid points at each angle.
[0160] ;
[0161] in, Indicates the received signal after phase calibration The model covariance matrix, For the upward array manifold matrix, This is a discrete angular power diagonal matrix. express The conjugate transpose of . This represents the variance of additive noise. Indicates the standard deviation of noise. It represents the identity matrix of the same order as the covariance matrix.
[0162] Received signal from UL The IAA algorithm is used to initially estimate the angular power, and then the uplink and downlink channel covariance matrices are obtained.
[0163] in, Indicates the first The received signal vector after pilot removal from each uplink training sample. For sample index, This represents the total number of training samples in the uplink. Indicates the complete range of values for the sample index. This represents the discrete angular power matrix initially estimated from these samples.
[0164] The IAA algorithm flow is as follows, where for the sake of simplicity, [the flow is simplified to] Abbreviation as .in, For the first Discrete angles Uplink frequency The array response vector under the following conditions Update the angle grid index for IAA.
[0165] First, perform the initialization, calculated using the following formula:
[0166] ;
[0167] in, Indicates the first Initial estimates of power at discrete angles, For the first Each uplink array response vector express The conjugate transpose of . This is the energy normalization factor for the response vector. The number of received samples used in this IAA estimation is compared with that in the embodiments of this application. Same meaning For sample index, For the first A received signal vector.
[0168] The summation sign indicates that the summation sign represents the sum of the sums ... to The summation of all samples, |·| represents the modulus of the complex scale; |·|² represents the square of the modulus. 1 / N represents the square of the normalization factor, and 1 / N represents the average of N samples.
[0169] ;
[0170] in, This represents the uplink model covariance matrix used in the current iteration of the IAA. For the upward array manifold matrix, This represents the set of all discrete angle grids. This is the current discrete angle power diagonal matrix. express The conjugate transpose of a matrix, and the contiguous writing of matrices, represent matrix multiplication.
[0171] Then, fork=1:Kdo, calculate .
[0172] in, Show the first The sample at the th Complex amplitude estimates at each angle grid point For the first Each uplink array response vector express The conjugate transpose of . Represents the current covariance matrix The inverse matrix, For the first A received signal vector, the numerator representing the adaptively weighted output, and the denominator representing the received signal vector. Represents the normalization coefficient. For the sample index range, For angular grid indexing.
[0173] Subsequently, the updated estimate of the discrete angle power is calculated, using the following formula:
[0174] ;
[0175] in, Indicates the first Updated estimates of power at discrete angles. To receive the number of samples, The sample index is used, and the summation sign indicates that the summation is performed on the sample index. to All samples summed For the first The sample at the th Complex amplitude estimation at each angle The square of the modulus of this complex amplitude. This indicates that the average of all samples is taken.
[0176] Check for convergence. If convergence is achieved, stop; otherwise, continue with the improved loop.
[0177] During the phase deviation calibration phase, with a fixed angle power spectrum, the optimal phase compensation value for each antenna is solved by deriving and simplifying the cost function using the coordinate descent method.
[0178] First, estimate the phase error. If fixed At this point, the problem degenerates into calculating the following formula:
[0179] ;
[0180] ;
[0181] in, Indicates the power spectrum at a fixed angle Then, only the phase deviation vector Find the minimum value. Represents the uplink observation covariance matrix. This represents the covariance matrix of the phase-uncalibrated model formed by the power spectrum at a fixed angle. The outer product of the phase deviation vector. Used to represent The conjugate transpose of , ⊙ denotes the Hadamah product. This represents the square of the norm.
[0182] in, Indicates constraints. for The One element, is the base of the natural exponential function. The imaginary unit, For the first Phase deviation of the root antenna, constraint express The modulus is 1.
[0183] The above problem can be solved using the coordinate descent method. The phase error... The segmentation is performed using the following formula:
[0184] ;
[0185] in, express 3D phase deviation vector, Indicates its first One element, Indicates the first The nth standard unit vector, i.e., the nth One element is 1, and the rest are 0. Indicates will The The vector obtained by setting one element to 0 and keeping the rest unchanged. The antenna index is updated for the current coordinates. The total number of antennas is represented by the vertical arrangement of column vectors. The formula above shows that the vector can be decomposed into the current coordinate components. With other coordinate components sum.
[0186] in It is a unit vector, its One term is 1, and all other terms are 0. This leads to the following formula:
[0187] ;
[0188] ;
[0189] in, The outer product of the phase deviation vectors is represented. express The conjugate transpose of . For the first One phase factor, For the first A standard unit vector, For the first The remaining phase vector after setting each element to zero, superscript The expression represents complex conjugation; writing vectors adjacently indicates an outer product or matrix multiplication. Because | |=1, | in the expansion |² It can be simplified to .
[0190] ;
[0191] Among them, the left end This represents the Hadamard product of the phase deviation outer product and the covariance matrix of the uncalibrated uplink model; ⊙ denotes the element-wise multiplication of the matrices. Indicates the remaining coordinate items, Current coordinates self-term, and These are a pair of intersecting terms that are conjugate transposes of each other. for The complex conjugate of , where the superscript H indicates the conjugate transpose. This is the index of the currently updated coordinates.
[0192] The cost function is rewritten as:
[0193] ;
[0194] in, Represents the uplink observation covariance matrix. This represents the model covariance matrix after superimposing the current phase deviation. This indicates that the update is in progress. When the coordinates are... Irrelevant residual matrix Indicates the relationship with the first A matrix of cross terms linearly related to each phase factor For the first One phase factor, For its conjugate, for The conjugate transpose of . Represents the square of the norm. For coordinate index.
[0195] ;
[0196] in, This indicates that the coordinates are being updated in descending order. The residual covariance matrix when there are phase factors, For the uplink observation covariance matrix, For the first The remaining phase vector after setting each element to zero For the first A standard unit vector, superscript This indicates the conjugate transpose. and For the corresponding outer product, ⊙ represents the Hadamard product. The covariance matrix of the uplink model with uncalibrated phase. This is the index of the current coordinates.
[0197] ;
[0198] in, Indicates the coordinate descent and the first The cross-term coefficient matrix corresponding to each phase factor For the first A standard unit vector, For the first The remaining phase vector after setting each element to zero, with the superscript H indicating the conjugate transpose. The circle represents the outer product, and the circle (⊙) represents the Hadamard product. This represents the covariance matrix of the uplink model without phase calibration. This is the index of the current coordinates.
[0199] Therefore, the above equation can be simplified to:
[0200] ;
[0201] Where Tr(·) represents the trace of the matrix, which is the sum of the elements of the main diagonal. For the first The residual covariance matrix corresponding to each coordinate. For the first The cross-term coefficient matrix of each phase factor, For the first One phase factor, It is its complex conjugate, and the superscript H indicates the conjugate transpose. The square of the norm is expressed using parentheses and square brackets to indicate the preceding matrix operation, where adjacent matrices or scalars are multiplied by their matrix representations. This equation writes the square of the norm in trace form and follows... and Expand.
[0202] If fixed The embodiments of this application can be improved in the following ways. ,Right now:
[0203] ;
[0204] in, Indicates a unit modulus constraint Select the option that maximizes the objective. , For the first One phase factor, is the base of the natural exponential function. The imaginary unit, For the first Phase deviation of the root antenna, To take the real part of a complex number, In order to be with the first The complex scalar coefficients related to each coordinate and They represent and The complex conjugate of the operator is indicated by the curly braces, which denote the scope of the real part operator.
[0205] in ,but The better solution is .
[0206] in, This indicates that the coordinates are being updated in descending order. The complex scalar coefficients are calculated from the current residual matrix and the cross term matrix at each phase, where Tr[·] represents the trace of the matrix. For the first The residual covariance matrix corresponding to each coordinate. This is the cross term coefficient matrix, where the superscript H denotes the conjugate transpose. For the first The optimal phase deviation of the antenna, ∠(·) represents the complex phase angle. express The complex conjugate, For antenna or coordinate index.
[0207] During the alternating iteration phase, the process continues until the improved objective function converges, yielding the final angular power spectrum estimate and phase deviation vector estimate.
[0208] Based on the proposed alternating improvement framework, during the uplink training phase, the angular power spectrum and antenna phase deviation vector are iteratively estimated starting from the received signal. The system continuously performs two operations alternately: first, under the condition of a fixed antenna phase deviation vector, an iterative adaptive method is used to estimate the angular power spectrum; second, under the condition of a fixed angular power spectrum, a coordinate descent method is used to improve the antenna phase deviation vector.
[0209] When the improved objective function reaches the convergence condition, the alternating iterative algorithm stops, and the system extracts and outputs the final high-precision angular power spectrum estimate and the actual antenna phase deviation vector estimate.
[0210] Finally, the results are output, and the downlink channel covariance matrix is calculated based on the downlink array response using the final angular power spectrum estimate and phase deviation vector estimate.
[0211] Figure 2 This is a schematic diagram of an optional channel data transmission method according to an embodiment of the present invention, as shown below. Figure 2 As shown, the method includes: system model and signal reception, improvement problem construction, angle power spectrum estimation, phase deviation calibration, alternating iteration and result output.
[0212] The system model and signal reception are as follows: during the uplink training phase, pilot signals independently transmitted by users are received to establish a channel model that considers antenna phase deviation and the uplink and downlink channel covariance matrices are derived.
[0213] The problem is improved by constructing a joint improved model based on the received signal that minimizes the covariance matrix mismatch, while simultaneously handling phase offset compensation and spectral estimation.
[0214] For angle power spectrum estimation, the fixed phase deviation vector is estimated by using an iterative adaptive method to perform discrete power spectrum estimation to obtain the initial uplink and downlink channel covariance matrices.
[0215] Phase deviation calibration: For fixed-angle power spectrum, the optimal phase compensation value for each antenna is solved by deriving and simplifying the cost function through coordinate descent method.
[0216] Alternate iterations are performed until the improved objective function converges to obtain the final angular power spectrum estimate and phase deviation vector estimate.
[0217] The results output uses the final angular power spectrum estimate and phase deviation vector estimate to calculate the downlink channel covariance matrix based on the downlink array response and outputs the performance comparison results.
[0218] Figure 3 This is a schematic diagram of an optional channel data transmission method according to an embodiment of the present invention, such as... Figure 3 As shown, the method includes: system initialization and uplink pilot reception, improvement problem construction and preprocessing, fixed-phase angle power spectrum estimation, fixed-spectrum antenna phase deviation calibration, alternating iteration and global convergence determination, and downlink channel covariance reconstruction and application.
[0219] System initialization and uplink pilot reception: The base station system completes initialization, and the antenna array enters receive mode. During the uplink training phase of the system, multiple user equipment within the service area independently transmit uplink pilot signals to the base station according to a predetermined scheduling scheme. The large-scale uniform linear array antenna equipped with the base station synchronously receives these radio frequency signals, which contain channel state and environmental noise.
[0220] To improve the problem formulation and preprocessing, the base station's signal processing unit removes pilot signals from the received uplink signal. Based on the physical assumption of reciprocity between the uplink and downlink angular power spectra in a frequency division duplex system, the processing unit constructs a joint improvement model. The core objective of this model is to construct an improvement problem that minimizes the covariance matrix mismatch based on the received signal, while incorporating antenna phase deviation and spatial angular power distribution as core variables.
[0221] For fixed-phase angular power spectrum estimation, the processing unit initiates an alternating iterative improvement algorithm. In the current iteration, the system first temporarily fixes the estimated value of the antenna phase deviation and calls the iterative adaptive method module to estimate the angular power spectrum. This module initializes the angular power spectrum and, using the inverse of the current covariance matrix, updates the angular power spectrum components one by one for each discrete angle term until the convergence condition of this submodule is met.
[0222] Following the steps above, the processing unit fixes the recently updated angular power spectrum and then focuses on improving the phase deviation of the antenna array. At this point, the system uses a coordinate descent method to mathematically segment the phase deviation variable of the entire array, performing individual improvements and calculations for the hardware phase deviation of each antenna to obtain the most accurate phase compensation value for the current round.
[0223] The system control unit continuously and alternately executes the two calculation steps of angle power spectrum estimation and phase deviation improvement, using alternating iterations and global convergence determination. During this alternating cycle, the system monitors the changing trend of the overall improvement objective function in real time until convergence. After algorithm convergence, the system extracts and outputs the final high-precision angle power spectrum estimate and the actual phase deviation vector estimate.
[0224] Downlink channel covariance reconstruction and application. Finally, the base station uses the extracted final angular power spectrum estimate and phase deviation vector estimate, combined with the system's predicted downlink array response characteristics, to directly calculate the accurate downlink channel covariance matrix. This embodiment organically integrates phase calibration and channel estimation processes, improving channel estimation accuracy and the overall robustness of the communication system under non-ideal hardware environments without requiring an independent phase measurement step.
[0225] Figure 4 This is a schematic diagram of an optional channel data transmission simulation result according to an embodiment of the present invention, such as... Figure 4 As shown, Figure 4 As shown, the horizontal axis represents the number of sampling points, and the vertical axis represents the Normalized Mean Square Error (NMSE) performance. As the number of sampling points increases, the NMSE of the uplink sampling covariance and the uplink true covariance gradually decreases until it approaches zero. The uplink covariance estimation performance gradually stabilizes after the number of sampling points reaches 100. Similarly, the downlink covariance transformation performance follows the same trend as the uplink covariance estimation performance, gradually stabilizing after the number of sampling points reaches 100.
[0226] Figure 5 This is a schematic diagram of an optional channel data transmission simulation result according to an embodiment of the present invention, such as... Figure 5As shown, the horizontal axis represents the number of sampling points, and the vertical axis represents the vector similarity (…). As the number of sampling points increases, the performance of the uplink covariance estimation and the conversion performance of the downlink covariance gradually improve, meaning the vector similarity increases. When the number of sampling points is 100, the change in vector similarity gradually stabilizes. With an increase in the number of sampling points, the vector similarity between the uplink sampled covariance estimated using the method of this application and the uplink true covariance overlaps with the vector similarity between the uplink sampled covariance and the downlink true covariance.
[0227] According to an embodiment of the present invention, an apparatus embodiment for a channel data transmission method is provided. It should be noted that the apparatus can be used to execute the above-described channel data transmission method. Figure 6 This is a schematic diagram of a channel data transmission device according to an embodiment of the present invention, such as... Figure 6 As shown, the device includes: a receiving device 602, an updating device 604, a determining device 606, and a transmitting device 608.
[0228] The receiving device 602 is used to receive the uplink pilot signal transmitted by the user equipment, and determine the observed sample value of the uplink channel covariance matrix based on the uplink pilot signal. The uplink channel covariance matrix is used to represent the statistical estimation matrix of the uplink channel space under the condition of antenna phase deviation and noise interference. The updating device 604 is used to update the initial antenna phase deviation vector and the initial angle power spectrum parameter based on the observed sample value to obtain the target antenna phase deviation vector and the target angle power spectrum parameter. The determining device 606 is used to process the target antenna phase deviation vector and the target angle power spectrum parameter using the antenna array response matrix of the downlink band to determine the downlink channel covariance matrix. The downlink channel covariance matrix is used to represent the statistical estimation matrix of the reconstructed downlink channel space. The transmitting device 608 is used to determine the downlink precoding parameters based on the downlink channel covariance matrix and transmit downlink data to the user equipment using the downlink precoding parameters.
[0229] The updating device is further configured to, in the current iteration, fix the initial antenna phase deviation vector, update the initial angular power spectrum parameters using an iterative adaptive algorithm based on the observed sample values to obtain updated angular power spectrum parameters, fix the updated angular power spectrum parameters, update the initial antenna phase deviation vector using the coordinate descent method based on the observed sample values to obtain updated antenna phase deviation vector; and determine the target antenna phase deviation vector and target angular power spectrum parameters based on the updated angular power spectrum parameters and the updated antenna phase deviation vector.
[0230] The updating device is further configured to, in response to the fact that the updated angle power spectrum parameter and the updated antenna phase deviation vector do not meet the preset iteration conditions, determine the updated angle power spectrum parameter as the initial angle power spectrum parameter and the updated antenna phase deviation vector as the initial antenna phase deviation vector, and repeat the execution steps of the current iteration round until the updated angle power spectrum parameter and the updated antenna phase deviation vector meet the preset iteration conditions; and in response to the fact that the updated angle power spectrum parameter and the updated antenna phase deviation vector meet the preset iteration conditions, determine the updated angle power spectrum parameter as the target angle power spectrum parameter and determine the updated antenna phase deviation vector as the target antenna phase deviation vector.
[0231] The updating device is also used to construct an initial estimation model of the uplink channel covariance matrix based on the initial antenna phase deviation vector and the observed sample values; determine the signal decorrelation estimate of the discrete spatial angle point based on the initial angle power spectrum parameters and the initial estimation model; and update the initial angle power spectrum parameters based on the signal decorrelation estimate to obtain the updated angle power spectrum parameters.
[0232] The updating device is further configured to traverse multiple antennas in the base station array, using the initial antenna phase deviation vector corresponding to the currently traversed antenna as the current variable to be updated, and the initial antenna phase deviation vectors corresponding to the other antennas as fixed variables; construct a target cost function based on the current variable to be updated, the fixed variables, and the observed sample values; convert the target cost function into a univariate function of the current variable to be updated; determine the target phase deviation vector of the currently traversed antenna based on the extremum condition of the univariate function; and determine the updated antenna phase deviation vector based on the target phase deviation vectors of multiple antennas.
[0233] The determining device is further used to perform phase compensation correction on the antenna array response matrix of the downlink frequency band using the target antenna phase deviation vector to obtain the corrected downlink array response; and to determine the downlink channel covariance matrix based on the corrected downlink array response and the target angular power spectrum parameters.
[0234] The device is also used to: construct the antenna array response matrix of the downlink band based on the center frequency of the downlink band and the geometry of the base station antenna array.
[0235] The receiving device is further configured to receive uplink pilot signals transmitted by user equipment via a uniform linear array; and to transmit downlink data to user equipment using downlink precoding parameters, including transmitting downlink data to user equipment via a uniform linear array using downlink precoding parameters.
[0236] The transmitting device is further configured to extract the downlink beamforming vector corresponding to the largest eigenvalue based on the characteristic distribution of the downlink channel covariance matrix, and determine the downlink precoding parameters based on the downlink beamforming vector; or, using the minimum mean square error, determine the downlink precoding matrix based on the downlink channel covariance matrix, noise power, and downlink transmission power, and determine the downlink precoding matrix as the downlink precoding parameters.
[0237] Embodiments of this application also provide an electronic device, including: a memory storing an executable program; and a processor for running the program, wherein the program executes the methods in various embodiments of the present invention during runtime.
[0238] Embodiments of this application also provide a computer-readable storage medium including a stored executable program, wherein, when the executable program is running, it controls the device where the computer-readable storage medium is located to perform the methods of various embodiments of the present invention.
[0239] Embodiments of this application also provide a computer program product, including a computer program that, when executed by a processor, implements the methods of various embodiments of the present invention.
[0240] Embodiments of this application also provide a computer program product, including a non-volatile computer-readable storage medium for storing a computer program that, when executed by a processor, implements the methods in various embodiments of the present invention.
[0241] Embodiments of this application also provide a computer program that, when executed by a processor, implements the methods described in the various embodiments of the present invention.
[0242] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.
[0243] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.
[0244] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0245] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.
[0246] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.
[0247] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.
Claims
1. A channel data transmission method, characterized in that, include: Receive uplink pilot signals sent by user equipment, and determine the observed sample values of the uplink channel covariance matrix based on the uplink pilot signals, wherein the uplink channel covariance matrix is used to represent the statistical estimation matrix of the uplink channel space under the condition of antenna phase deviation and noise interference; Based on the observed sample values, the initial antenna phase deviation vector and the initial angular power spectrum parameters are updated respectively to obtain the target antenna phase deviation vector and the target angular power spectrum parameters; The target antenna phase deviation vector and the target angular power spectrum parameters are processed using the antenna array response matrix of the downlink frequency band to determine the downlink channel covariance matrix, wherein the downlink channel covariance matrix is used to represent the statistical estimation matrix of the reconstructed downlink channel space; Downlink precoding parameters are determined based on the downlink channel covariance matrix, and downlink data is transmitted to the user equipment using the downlink precoding parameters.
2. The method according to claim 1, characterized in that, Based on the observed sample values, the initial antenna phase deviation vector and the initial angular power spectrum parameters are updated respectively to obtain the target antenna phase deviation vector and the target angular power spectrum parameters, including: In the current iteration, the initial antenna phase deviation vector is fixed, and the initial angular power spectrum parameter is updated based on the observed sample value using an iterative adaptive algorithm to obtain the updated angular power spectrum parameter. The updated angular power spectrum parameter is then fixed, and the initial antenna phase deviation vector is updated based on the observed sample value using a coordinate descent method to obtain the updated antenna phase deviation vector. Based on the updated angular power spectrum parameters and the updated antenna phase deviation vector, the target antenna phase deviation vector and the target angular power spectrum parameters are determined.
3. The method according to claim 2, characterized in that, Determining the target antenna phase deviation vector and the target angle power spectrum parameters based on the updated angle power spectrum parameters and the updated antenna phase deviation vector includes: In response to the fact that the updated angle power spectrum parameter and the updated antenna phase deviation vector do not meet the preset iteration conditions, the updated angle power spectrum parameter is determined as the initial angle power spectrum parameter, and the updated antenna phase deviation vector is determined as the initial antenna phase deviation vector. The execution steps of the current iteration round are repeated until the updated angle power spectrum parameter and the updated antenna phase deviation vector meet the preset iteration conditions. In response to the updated angle power spectrum parameter and the updated antenna phase deviation vector satisfying the preset iteration condition, the updated angle power spectrum parameter is determined to be the target angle power spectrum parameter, and the updated antenna phase deviation vector is determined to be the target antenna phase deviation vector.
4. The method according to claim 3, characterized in that, The initial angular power spectrum parameters are updated using an iterative adaptive algorithm based on the observed sample values to obtain the updated angular power spectrum parameters, including: Based on the initial antenna phase deviation vector and the observed sample values, an initial estimation model for the uplink channel covariance matrix is constructed. Based on the initial angular power spectrum parameters and the initial estimation model, determine the signal decorrelation estimate for discrete spatial angle points; The initial angular power spectrum parameters are updated based on the signal decorrelation estimate to obtain the updated angular power spectrum parameters.
5. The method according to claim 3, characterized in that, The initial antenna phase deviation vector is updated using the coordinate descent method based on the observed sample values to obtain the updated antenna phase deviation vector, including: Traverse multiple antennas in the base station array antenna, take the initial antenna phase deviation vector corresponding to the currently traversed antenna as the current variable to be updated, and take the initial antenna phase deviation vector corresponding to the other antennas in the multiple antennas as fixed variables; Based on the current variable to be updated, the fixed variable, and the observed sample values, construct the target cost function; The target cost function is converted into a univariate function of the variable to be updated. Based on the extreme value condition of the single-variable function, the target phase deviation vector of the currently traversed antenna is determined; The updated antenna phase deviation vector is determined based on the target phase deviation vectors of the multiple antennas.
6. The method according to any one of claims 1 to 5, characterized in that, The downlink channel covariance matrix is determined by processing the target antenna phase deviation vector and the target angular power spectrum parameters using the antenna array response matrix of the downlink band, including: The target antenna phase deviation vector is used to perform phase compensation correction on the antenna array response matrix of the downlink frequency band to obtain the corrected downlink array response; The downlink channel covariance matrix is determined based on the corrected downlink array response and the target angle power spectrum parameters.
7. The method according to any one of claims 1 to 5, characterized in that, Receive uplink pilot signals sent by user equipment, including: The uplink pilot signal transmitted by the user equipment is received through a uniform linear array; Sending downlink data to the user equipment using the downlink precoding parameters includes: The downlink data is transmitted to the user equipment via the uniform linear array using the downlink precoding parameters.
8. The method according to any one of claims 1 to 5, characterized in that, Determining downlink precoding parameters based on the downlink channel covariance matrix includes: Based on the characteristic distribution of the downlink channel covariance matrix, the downlink beamforming vector corresponding to the largest eigenvalue is extracted, and the downlink precoding parameters are determined based on the downlink beamforming vector. or, Using the minimum mean square error, the downlink precoding matrix is determined based on the downlink channel covariance matrix, noise power, and downlink transmit power, and the downlink precoding matrix is determined as the downlink precoding parameters.
9. An electronic device, characterized in that, include: Memory, which stores executable programs; A processor for running the program, wherein the program, when running, performs the method according to any one of claims 1 to 8.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored executable program, wherein, when the executable program is executed, it controls the device on which the storage medium is located to perform the method according to any one of claims 1 to 8.
