A Hybrid Precoding Method for Shipborne Base Stations Based on CKM Angle Tracking
By constructing an angle-domain CKM-assisted precoding design, the hybrid precoding method for shipborne base stations solves the problems of high real-time channel estimation overhead and difficulty in tracking high-speed moving targets, achieving efficient channel information acquisition and dynamic beamforming, and improving the transmission efficiency and robustness of maritime communication systems.
Patent Information
- Application Number
- CN202611601206.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-10-16
- Publication Date
- 2026-11-17
AI Technical Summary
Shipborne base station large-scale MIMO communication systems suffer from high real-time channel estimation overhead and difficulty in tracking high-speed moving targets, making it difficult for the communication system to withstand the problems of channel estimation overhead and rapid channel changes.
A hybrid precoding method for shipborne base stations based on CKM angle tracking is adopted. By constructing an angle domain CKM, angle information is obtained from the CKM to assist in precoding design, reduce real-time channel estimation overhead, optimize user ship trajectory tracking accuracy, and achieve forward-looking beamforming.
It effectively reduces the real-time channel estimation overhead of large-scale MIMO communication systems, improves the transmission efficiency and robustness of maritime communication systems, and enhances dynamic target tracking and continuous coverage capabilities.
Smart Images

Figure CN123122599A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of wireless communication technology, and in particular to a hybrid precoding method for shipborne base stations based on CKM angle tracking. Background Technology
[0002] Driven by the national strategy of building a maritime power, my country's marine information industry is booming, and the demand for broadband communication in maritime areas is increasing daily. Massive MIMO hybrid precoding helps achieve dynamic beam coverage and improve information transmission efficiency while reasonably controlling the cost of communication antenna hardware, making it a key technology for realizing wireless broadband communication in maritime areas. Traditional hybrid precoding typically relies on real-time channel estimation to design optimal beams; however, shipborne base stations face challenges in adopting hybrid precoding technology due to both high channel estimation overhead and difficulty in tracking high-speed moving targets. Specifically, hybrid precoding design usually requires prior knowledge of the target's channel state information (CSI). However, due to the large number of shipborne massive MIMO antennas (tens, even hundreds or thousands), existing pilot-measure-based CSI feedback schemes face enormous channel estimation overhead, making the system unsustainable. On the other hand, the relative motion between the shipborne base station and the user vessel (speeds typically reaching 10-30 knots) and the rapid changes in channel state also pose a severe challenge to instantaneous channel estimation. In high-speed moving scenarios, existing instantaneous CSI feedback schemes struggle to track channel changes in real time. Summary of the Invention
[0003] To address the technical challenges of high real-time channel estimation overhead and difficulty in tracking high-speed moving targets in current shipborne base station hybrid precoding designs, this invention proposes a shipborne base station hybrid precoding method based on CKM angle tracking. Aiming to reduce system overhead and enhance continuous tracking coverage, the method constructs and utilizes CKM to acquire angle information to assist in the shipborne base station hybrid precoding design. This effectively reduces the real-time channel estimation overhead of large-scale MIMO communication systems and optimizes the accuracy of user ship trajectory tracking, thereby improving the transmission efficiency and robustness of maritime communication systems.
[0004] The inventive concept of this invention is: a hybrid precoding method for shipborne base stations based on CKM angle tracking. Its core idea lies in establishing a closed-loop architecture of "CKM construction → user trajectory tracking → channel matrix reconstruction → precoding design → feedback update" (e.g., Figure 1As shown in the figure, the CKM (Channel Code Measuring) is used as a bridge to communicate user location and hybrid beamforming design, achieving forward-looking beamforming. This scheme uses a large-scale MIMO antenna array on the shipborne base station side to cover the target sea area. First, the angle domain CKM is established through sampling measurement; then, the movement trajectory of the shipborne target user is predicted and the prediction result is corrected by combining it with the CKM; then, the channel angle parameters are obtained in advance by querying the CKM based on the trajectory prediction result to design a hybrid precoding scheme.
[0005] The specific steps of the plan are as follows: Figure 2 As shown, the process includes: First, sampling and measuring angle data within the coverage area of the shipborne base station, and constructing an angle domain CKM based on the measurement results; Second, using an optimal estimation algorithm to predict the target user's movement trajectory at the next moment, and timely correcting the trajectory prediction results by combining the position-angle relationship in the CKM; Next, querying the angle domain CKM to obtain channel angle parameters in advance based on the prediction results, and reconstructing the channel matrix based on the query results; Then, calculating and successively approximating the optimal all-digital precode through the singular value decomposition and orthogonal matching pursuit algorithm of the channel matrix, thereby generating a hybrid analog precode; Finally, comparing the actual measured values of the channel angle parameters with the predicted values of the CKM in real time, and dynamically updating the CKM.
[0006] To achieve the above-mentioned objectives, the present invention employs the following technical solution: a hybrid precoding method for shipborne base stations based on CKM angle tracking, comprising the following steps:
[0007] 1) Channel Sampling Measurement and CKM Construction: First, sparse grid sampling and wireless channel measurements are performed in the target sea area covered by the shipborne base station. Information such as the angles and signal strengths of the L strongest paths at different coordinates is recorded to construct the initial angle domain CKM corresponding to the base station coverage area. Utilizing the correspondence between the CKM's stored location and channel-related information, channel angle information can be quickly queried and obtained to guide hybrid precoding design, thereby reducing real-time channel estimation overhead. Specifically, the CKM can be obtained through sampling measurements and interpolation prediction, and its data storage structure can be represented as follows:
[0008] ;
[0009] In the formula, and Representing maps The number of rows and columns of grid points, and , and They represent the numbers on the map. OK The grid points correspond to the user's location, including the Angle of Arrival (AoA), Angle of Departure (AoD), and signal strength. , If a shipborne base station uses a uniform planar array, its AoA and AoD typically each include components in two dimensions: elevation and azimuth.
[0010] It is important to note that only a small number of user locations in the constructed CKM have multipath channel parameters obtained through measurement, while the channel parameters for the remaining locations are completed through spatial interpolation. Channel parameter completion can employ classic spatial interpolation algorithms such as K-Nearest Neighbors (KNN), Inverse Distance Weighting (IDW), or Kriging, or it can utilize deep learning-based methods such as Convolutional Neural Networks (CNN), Generative Adversarial Networks (GAN), or Graph Neural Networks (GNN). Since the number of measured user locations is much smaller than the number of interpolated user locations, the channel estimation overhead can be significantly reduced.
[0011] Since the channel angle information corresponding to all grid locations within the coverage area of the shipborne base station is pre-stored in the CKM, the channel angle parameters at the target location can be quickly obtained through map retrieval. Based on this, the angle-domain CKM essentially establishes a mapping relationship between location coordinates and multipath channel parameters, and can provide effective prior channel information for key aspects such as trajectory tracking and channel matrix reconstruction in subsequent hybrid precoding design.
[0012] 2) User Trajectory Tracking and CKM Correction. To obtain the precise position of the user vessel relative to the base station vessel, optimal estimation theory can be used to predict and track the user vessel's position. System state optimal estimation algorithms designed based on optimal estimation theory typically include Least Squares (LS), Kalman Filter (KF), and Extended Kalman Filter (EKF). Taking the EKF algorithm as an example, at the current time... The location prediction expression is as follows:
[0013] ;
[0014] In the formula, For nonlinear system models, For the target user's ship location The optimal estimate at time -1 Control quantities applied externally, This represents systematic error.
[0015] Since optimal estimation algorithms such as EKF rely on motion models, they are prone to error accumulation in complex and dynamic marine environments. Therefore, the user vessel trajectory tracking results predicted by optimal estimation algorithms can be corrected by combining the "position coordinates-multipath channel parameters" stored in the angle domain CKM. At each time step... The base station receives the uplink pilot signal transmitted by the user ship, extracts the multipath channel parameters, and inputs these parameters into the observations of the CKM query candidate positions. Simultaneously, based on the chi-square distribution of the squared Mahalanobis distance between the predicted value and the CKM query value (observation value), it determines whether the current predicted value of the optimal estimation algorithm is reliable (if the Mahalanobis distance does not exceed the chi-square threshold, the prediction value is considered reliable). If a line-of-sight (LoS) path exists and the predicted value of the optimal estimation algorithm is reliable, the user ship's position is updated using the predicted value of the optimal estimation algorithm; otherwise, if a line-of-sight (LoS) path does not exist or the predicted value of the optimal estimation algorithm is unreliable, the position prediction of the optimal estimation algorithm is corrected using the CKM query value.
[0016] Specifically, within the base station coverage area, an optimal estimation algorithm is used to track the position and status of user ships. When the uncertainty in the optimal estimation theory's prediction of the user ship's trajectory is large, the shipborne base station inputs the multipath channel parameters (such as AoA / AoD) measured in real time into the CKM as a query key, and the CKM returns the candidate location areas that may correspond to the multipath channel parameters. This is used as prior information to correct the prediction results of the optimal estimation algorithm, thereby improving the accuracy and robustness of the user ship trajectory prediction.
[0017] The shipborne base station system compares the Mahalanobis distance between the predicted value of the optimal estimation algorithm and the CKM query value in real time. If the Mahalanobis distance exceeds the chi-square threshold, the weight of the CKM query value is adaptively increased, thereby constraining the user location prediction result to the area indicated by the CKM, so as to avoid the problem of prediction results diverging in non-line-of-sight environments.
[0018] 3) CKM Query and Channel Matrix Reconstruction. The base station inputs the user's location coordinates, corrected by trajectory tracking, into the angle domain CKM. Based on its stored "location coordinates-multipath channel parameters" mapping, the CKM queries and outputs the expected angle and intensity information corresponding to that location. Then, it uses the angle and intensity information to reconstruct the channel matrix.
[0019] Specifically, the mathematical expression for reconstructing the channel matrix using the multipath channel parameters obtained from the CKM query is as follows:
[0020] ;
[0021] In the formula, and These represent the number of user receiving antennas and the number of base station transmitting antennas, respectively. This indicates the number of channel paths stored in CKM; , and They represent the first The signal strength (gain), AoA, and AoD of each path, with AoA and AoD each containing components of elevation and azimuth. , ; and These represent the array response vectors of the transmitting and receiving ends, respectively.
[0022] Taking the transmitting end as an example, assume that the shipborne base station uses a fixed, uniform planar array, and that the array antenna elements are regularly arranged in the vertical plane, with horizontal and vertical directions respectively having and One antenna element, satisfying Therefore, the response vector of the two-dimensional array can be expressed as:
[0023] ;
[0024] in:
[0025] ;
[0026] ;
[0027] In the formula, and These represent the elevation and azimuth angles of AoD, respectively. For carrier wavelength, The spacing between array antenna elements is typically taken as a value. , For Kronecker product.
[0028] 4) Hybrid Precoding Design. Shipborne base stations can pre-design hybrid precoding schemes using the channel matrix reconstruction results, thereby directing the generated beam towards the target user. To establish the optimization objective of the hybrid precoding design scheme, it is necessary to calculate the optimal all-digital precoding matrix. Specifically, firstly, the channel matrix... Perform singular value decomposition to obtain ,in yes unitary matrix of order, It is semi-positive definite A diagonal matrix of order 1, and yes A unitary matrix of order; then a right singular matrix is selected. The former Column vector composition (here) Indicates the number of radio frequency chains. This represents the optimal precoding matrix design target that the antenna system can achieve without hardware constraints, while the optimization target of hybrid precoding is to approximate this target using a hybrid structure.
[0029] In a hybrid precoding architecture, to obtain analog precoding with better performance, a feasible approach is to use the Orthogonal Matching Pursuit (OMP) algorithm to successively approximate the optimal target (i.e., ...). Since the analog end can only use phase shifters to change the phase, and the phase shifter is a finite set of codebooks, the OMP algorithm is needed to select a set of column vectors from the codebook so that the analog precoding matrix formed by them is mathematically closest to... In other words, the task of the OMP algorithm is to extract the code from a pre-defined codebook. Selected from The most suitable "base vector" is selected, and the analog precoding matrix is gradually constructed through multiple iterations. .matrix Used to adjust the antenna array so that the main lobe of the beam points in the direction of the predicted angle.
[0030] After the analog precoding is determined, it is combined with the channel matrix. An equivalent low-dimensional channel can be obtained. At this point, the high-dimensional hybrid precoding design can be simplified to a low-dimensional digital precoding problem. During baseband processing, the minimum mean square error estimation criterion can be used to calculate the digital precoding matrix, and its calculation formula is as follows:
[0031] ;
[0032] In the formula Indicates noise power. Represents an identity matrix. Digital precoding matrix. Used for digital domain elimination of residual multi-user interference or inter-datastream interference. Based on And combined with previously obtained The hybrid precoding matrix can be calculated. .
[0033] 5) Actual Measurement Feedback and CKM Update. After receiving the actual signal at the next moment, measure the actual channel angle parameters and update the measured values. CKM query value Compare, if there is a deviation Exceeding the preset threshold (Right now If this occurs, a local update of the CKM is triggered (which can be achieved using Gaussian process regression or neural network interpolation). This allows the CKM to adapt to dynamic changes in the marine environment (tides, sea states, etc.). The CKM update result can be expressed as:
[0034] ;
[0035] In the formula, This represents the local update function of CKM, used for combining renew ; , , and They represent The Middle The user's location coordinates, angle of arrival (AoA), angle of departure (AoD), and signal strength; This represents a set of locally updated position coordinates.
[0036] Meanwhile, this invention proposes a shipborne base station hybrid precoding system based on CKM angle tracking. The system, applying the method described in this invention, includes the following steps:
[0037] The angle domain CKM construction module is configured to perform the following process: sample and measure angle data within the coverage area of the shipborne base station, and construct the angle domain CKM based on the results of the measured angle data;
[0038] The trajectory prediction result correction module is configured to perform the following process: predict the target user's movement trajectory at the next moment using the optimal estimation algorithm, and correct the trajectory prediction result in a timely manner by combining the position-angle relationship in the angle domain CKM;
[0039] The channel matrix reconstruction module is configured to perform the following process: query the angle domain CKM to obtain channel angle parameters in advance based on the trajectory prediction results, and reconstruct the channel matrix based on the query results;
[0040] The analog precoding module is configured to perform the following process: calculate the optimal all-digital precoding by singular value decomposition of the channel matrix, and successively approximate the all-digital precoding using the orthogonal matching pursuit (OMP) algorithm to generate analog precoding, and then design digital precoding by combining the channel matrix.
[0041] The CKM update module is configured to perform the following process: compare the actual measured values of the channel angle parameters with the CKM predicted values in real time, and dynamically update the CKM.
[0042] Meanwhile, the present invention proposes an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the computer program is executed, it implements the steps of the method described in the present invention.
[0043] Furthermore, the present invention proposes a computer-readable storage medium having a computer program stored thereon, the computer program being configured to implement the steps of the method described in the present invention when invoked by a processor.
[0044] Finally, the present invention provides a computer program product comprising a computer program / instructions that, when executed by a processor, implement the steps of the method described in the present invention.
[0045] Compared with the prior art, the beneficial effects of the present invention are as follows:
[0046] 1. This invention proposes a hybrid precoding method for shipborne base stations based on CKM angle tracking. By querying CKM to quickly obtain angle information to assist in the hybrid precoding design, the real-time channel estimation overhead of the shipborne base station large-scale MIMO communication system is greatly reduced, thereby effectively improving the system transmission efficiency.
[0047] 2. The hybrid precoding scheme proposed in this invention makes full use of the mapping relationship between multipath angle parameters and target position to improve the target tracking accuracy in dynamic marine scenarios. It effectively addresses the challenge of continuous dynamic beam coverage brought about by the bidirectional movement of base station ships and user ships, thereby improving dynamic target tracking capabilities and enhancing the robustness of marine communication. Attached Figure Description
[0048] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.
[0049] Figure 1 This is a schematic diagram illustrating the overall design concept of the hybrid precoding design for CKM-assisted shipborne base stations for dynamic coverage in marine areas, as described in this invention.
[0050] Figure 2 This is a flowchart illustrating the key technologies of the hybrid precoding design for CKM-assisted shipborne base stations for dynamic coverage of sea areas according to the present invention.
[0051] Figure 3 This is a schematic diagram of a scenario where a shipborne base station's large-scale MIMO array dynamically covers marine users, as described in Embodiment 1 of the present invention.
[0052] Figure 4 This is a schematic diagram of channel sampling measurement and angle domain CKM construction in Embodiment 2 of the present invention.
[0053] Figure 5This is a schematic diagram of hybrid precoding beamforming based on CKM multipath angle prediction in Embodiment 3 of the present invention.
[0054] Figure 6 This is a performance curve of the spectral efficiency as a function of SNR for different precoding schemes in Embodiment 4 of the present invention.
[0055] Figure 7 This is a performance curve of the spectral efficiency of different precoding schemes in Embodiment 5 of the present invention as a function of relative ship speed. Detailed Implementation
[0056] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. Of course, the specific embodiments described herein are merely illustrative and not intended to limit the invention.
[0057] Example 1: Specific implementation scenarios of the present invention include channel measurement and CKM construction, user ship trajectory tracking, angle prediction and channel matrix reconstruction, hybrid precoding design, experimental feedback and CKM update, such as... Figure 1 As shown.
[0058] See Figure 2 This embodiment provides a technical solution: a hybrid precoding method for shipborne base stations based on CKM angle tracking, comprising the following steps:
[0059] First, angle data is sampled and measured within the coverage area of the shipborne base station, and an angle domain CKM is constructed based on the measurement results. Second, the optimal estimation algorithm is used to predict the movement trajectory of the shipborne target user at the next moment, and the trajectory prediction results are corrected in a timely manner by combining the position-angle relationship in the CKM.
[0060] Next, based on the prediction results, the channel angle parameters are obtained in advance by querying the CKM in the angle domain, and the channel matrix is reconstructed based on the query results;
[0061] Then, using singular value decomposition of the channel matrix and orthogonal matching pursuit (CKM) algorithm, the optimal all-digital precoder is calculated and successively approximated, thereby generating the hybrid analog precoder. Finally, the actual measured values of the channel angle parameters are compared with the predicted values of CKM in real time, and the CKM is dynamically updated. A detailed explanation of the above steps is as follows:
[0062] 1) Channel Sampling Measurement and CKM Construction. First, sparse grid sampling and wireless channel measurements are performed in the target sea area covered by the shipborne base station. Information such as the angles and signal strengths of the L strongest paths at different coordinates is recorded, constructing the initial angle domain CKM corresponding to the base station coverage area. Utilizing the correspondence between the CKM's stored location and channel-related information, channel angle information can be quickly queried and obtained to guide hybrid precoding design, thereby reducing real-time channel estimation overhead. Specifically, the CKM can be obtained through sampling measurements and interpolation prediction, and its data storage structure can be represented as follows:
[0063] ;
[0064] In the formula, and Representing maps The number of rows and columns of grid points, and , and They represent the numbers on the map. OK The AoA, AoD, and signal strength of the user location corresponding to each grid point, and , If a shipborne base station uses a uniform planar array, its AoA and AoD typically each include components in two dimensions: elevation and azimuth.
[0065] It is important to note that only a small number of user locations in the constructed CKM have multipath channel parameters obtained through measurement, while the channel parameters for the remaining locations are completed through spatial interpolation. This channel parameter completion can employ classic spatial interpolation algorithms such as KNN, IDW, or Kriging, or it can utilize deep learning-based methods such as CNN, GAN, or GNN. Since the number of measured user locations is much smaller than the number of interpolated user locations, the channel estimation overhead can be significantly reduced.
[0066] Since the channel angle information corresponding to all grid locations within the coverage area of the shipborne base station is pre-stored in the CKM, the channel angle parameters at the target location can be quickly obtained through map retrieval. Based on this, the angle-domain CKM essentially establishes a mapping relationship between location coordinates and multipath channel parameters, and can provide effective prior channel information for key aspects such as trajectory tracking and channel matrix reconstruction in subsequent hybrid precoding design.
[0067] 2) User Trajectory Tracking and CKM Correction. To obtain the precise position of the user vessel relative to the base station vessel, optimal estimation theory can be used to predict and track the user vessel's position. System state optimal estimation algorithms designed based on optimal estimation theory typically include LS, KF, and EKF. Taking the EKF algorithm as an example, the expression for the current position prediction is as follows:
[0068] ;
[0069] In the formula, For nonlinear system models, For the target user's ship location The optimal estimate at time -1 Control quantities applied externally, This represents systematic error.
[0070] Since optimal estimation algorithms such as EKF rely on motion models, they are prone to error accumulation in complex and dynamic marine environments. Therefore, the tracking results of user vessel trajectories predicted by optimal estimation algorithms can be corrected by combining the "position coordinates-multipath channel parameters" stored in the angle domain CKM. At each time t, the base station receives the uplink pilot signal transmitted by the user vessel, extracts the multipath channel parameters, and inputs the obtained multipath channel parameters into the CKM query for the observed values of candidate positions. Simultaneously, based on the chi-square distribution of the squared Mahalanobis distance between the predicted value of the optimal estimation algorithm and the CKM query value (observation value), the reliability of the current optimal estimation algorithm prediction value is determined (if the Mahalanobis distance does not exceed the chi-square threshold, the prediction value is considered reliable). If a line-of-sight (LoS) path exists and the optimal estimation algorithm prediction value is reliable, the user vessel position is updated using the optimal estimation algorithm prediction value; conversely, if a line-of-sight (LoS) path does not exist or the optimal estimation algorithm prediction value is unreliable, the position prediction of the optimal estimation algorithm is corrected using the CKM query value.
[0071] Specifically, within the base station coverage area, an optimal estimation algorithm is used to track the position and status of user ships. When the uncertainty in the optimal estimation theory's prediction of the user ship's trajectory is large, the shipborne base station inputs the multipath channel parameters (such as AoA / AoD) measured in real time into the CKM as a query key, and the CKM returns the candidate location areas that may correspond to the multipath channel parameters. This is used as prior information to correct the prediction results of the optimal estimation algorithm, thereby improving the accuracy and robustness of the user ship trajectory prediction.
[0072] The shipborne base station system compares the Mahalanobis distance between the predicted value of the optimal estimation algorithm and the CKM query value in real time. If the Mahalanobis distance exceeds the chi-square threshold, the weight of the CKM query value is adaptively increased, thereby constraining the user location prediction result to the area indicated by the CKM, so as to avoid the problem of prediction results diverging in non-line-of-sight environments.
[0073] 3) CKM Query and Channel Matrix Reconstruction. The base station inputs the user's location coordinates, corrected by trajectory tracking, into the angle domain CKM. Based on its stored "location coordinates-multipath channel parameters" mapping, the CKM queries and outputs the expected angle and intensity information corresponding to that location. Then, it uses the angle and intensity information to reconstruct the channel matrix.
[0074] Specifically, the mathematical expression for reconstructing the channel matrix using the multipath channel parameters obtained from the CKM query is as follows:
[0075] ;
[0076] In the formula, and These represent the number of user receiving antennas and the number of base station transmitting antennas, respectively. This indicates the number of channel paths stored in CKM; , and They represent the first The signal strength (gain), AoA, and AoD of each path, with AoA and AoD each containing components of elevation and azimuth. , ; and These represent the array response vectors of the transmitting and receiving ends, respectively.
[0077] Taking the transmitting end as an example, assume that the shipborne base station uses a fixed, uniform planar array, and that the array antenna elements are regularly arranged in the vertical plane, with horizontal and vertical directions respectively having and One antenna element, satisfying Therefore, the response vector of the two-dimensional array can be expressed as:
[0078] ;
[0079] in:
[0080] ;
[0081] ;
[0082] In the formula, and These represent the elevation and azimuth angles of AoD, respectively. For carrier wavelength, The spacing between array antenna elements is typically taken as a value. , For Kronecker product.
[0083] 4) Hybrid Precoding Design. Shipborne base stations can pre-design hybrid precoding schemes using the channel matrix reconstruction results, thereby directing the generated beam towards the target user. To establish the optimization objective of the hybrid precoding design scheme, it is necessary to calculate the optimal all-digital precoding matrix. Specifically, firstly, the channel matrix... Perform singular value decomposition to obtain ,in yes unitary matrix of order, It is semi-positive definite A diagonal matrix of order 1, and yes A unitary matrix of order; then a right singular matrix is selected. The former Column vector composition (here) Indicates the number of radio frequency chains. This represents the optimal precoding matrix design target that the antenna system can achieve without hardware constraints, while the optimization target of hybrid precoding is to approximate this target using a hybrid structure.
[0084] In a hybrid precoding architecture, to obtain analog precoding with better performance, a feasible approach is to use the OMP algorithm to successively approximate the optimal target (i.e., Since the analog end can only use phase shifters to change the phase, and the phase shifter is a finite set of codebooks, the OMP algorithm is needed to select a set of column vectors from the codebook so that the analog precoding matrix formed by them is mathematically closest to... In other words, the task of the OMP algorithm is to extract the code from a pre-defined codebook. Selected from The most suitable "base vector" is selected, and the analog precoding matrix is gradually constructed through multiple iterations. .matrix Used to adjust the antenna array so that the main lobe of the beam points in the direction of the predicted angle.
[0085] After the analog precoding is determined, it is combined with the channel matrix. An equivalent low-dimensional channel can be obtained. At this point, the high-dimensional hybrid precoding design can be simplified to a low-dimensional digital precoding problem. During baseband processing, the minimum mean square error estimation criterion can be used to calculate the digital precoding matrix, and its calculation formula is as follows:
[0086] ;
[0087] In the formula, Indicates noise power. Represents an identity matrix. Digital precoding matrix. Used for digital domain elimination of residual multi-user interference or inter-datastream interference. Based on And combined with previously obtained The hybrid precoding matrix can be calculated. .
[0088] 5) Actual Measurement Feedback and CKM Update. After receiving the actual signal at the next moment, measure the actual channel angle parameters and update the measured values. CKM query value Compare, if there is a deviation Exceeding the preset threshold (Right now If this occurs, a local update of the CKM is triggered (which can be achieved using Gaussian process regression or neural network interpolation); thus enabling the CKM to adapt to dynamic changes in the marine environment (tides, sea states, etc.). The CKM update result can be expressed as:
[0089] ;
[0090] In the formula, This represents the local update function of CKM, used for combining renew ; , , and They represent The Middle The user's location coordinates, angle of arrival (AoA), angle of departure (AoD), and signal strength; This represents a set of locally updated position coordinates.
[0091] Example 2:
[0092] To better illustrate the overall communication architecture of the shipborne base station dynamically serving user ships, this embodiment establishes a typical application scenario where the shipborne base station provides communication coverage for three user ships, such as... Figure 3 As shown. Figure 3 The upper left side of the shipborne base station deploys a large-scale MIMO antenna array to generate dynamic beams to continuously cover target user ships in the sea area, and connects to the CKM server to obtain prior information on channel angles; the lower right side describes the number and spatial distribution of user ships. The shipborne base station dynamically covers all user ships by adjusting the beam direction in real time. In this embodiment, the shipborne base station is equipped with... root antenna and One radio frequency link, and Each user vessel is equipped with One antenna. The system operates at a frequency of 4.9 GHz, and the relative speed between the base station and the user ship is in the range of 10 to 30 knots.
[0093] Example 3:
[0094] To better illustrate the CKM construction method, this embodiment presents the approach of constructing an angle-domain CKM through channel sampling measurements, such as... Figure 4 As shown, sparse grid sampling and wireless channel measurement are first performed in the target sea area covered by the shipborne base station, and the multipath AoA / AoD and other channel angle information at different location coordinates are recorded. Then, based on this, the angle information of other locations is predicted by interpolation, and the angle domain CKM is constructed by combining the channel measurement and interpolation results. Figure 4 In the CKM (Pilot Path Management) algorithm, the AoA and AoD of the top L strongest paths are extracted for each user location, and AoA and AoD include elevation and azimuth angles. Only a small number of user locations have their angles obtained through measurement, while the angles of the remaining locations are predicted through interpolation. Interpolation methods can be traditional methods such as KNN, IDW, and Kriging, or they can be deep learning-based methods. Since the number of measured locations is much smaller than the number of interpolated locations, the CKM construction process does not incur excessive pilot estimation overhead.
[0095] Example 4:
[0096] To further illustrate the process of angle-domain CKM-assisted hybrid precoding design, this embodiment presents the approach of using multipath angle prediction results queried by CKM to assist in hybrid precoding design for shipborne base stations, such as... Figure 5 As shown. Figure 5 In the angle domain CKM, different colors of the grid points represent channel angles of different magnitudes. The channel matrix can be reconstructed using the angles obtained from CKM queries, and the optimal all-digital precoding matrix is calculated through singular value decomposition of the channel matrix. Then, the OMP algorithm is used to approximate the optimal all-digital precoding matrix step by step, thereby obtaining the analog precoding matrix. Finally, an equivalent channel matrix is constructed by combining the channel matrix, and the digital precoding matrix is calculated based on this using the least mean square estimation criterion to eliminate interference between users and between data streams.
[0097] Example 5:
[0098] To verify the transmission efficiency of the proposed hybrid precoding method for shipborne base stations based on CKM angle tracking, this embodiment simulates the spectral efficiency of the hybrid precoding scheme under different signal-to-noise ratios (SNR). During the simulation, an angle-domain CKM is established in a uniform sampling manner within the coverage area of a large-scale MIMO array, with a sampling interval of 50 m. The relative speed between the base station ship and the user ship is set to 10 knots. The SNR of the sea channel propagation environment is set to -10, -5, 0, 5, 10, and 15 dB, respectively. A traditional hybrid precoding scheme based on pilot measurement and feedback is used as a benchmark comparison method. The relationship curves between the system spectral efficiency and SNR under the two hybrid precoding schemes are shown below. Figure 6 As shown. From Figure 6As can be seen, in the low SNR region, the hybrid precoding scheme based on CKM angle tracking proposed in this invention exhibits superior performance, and its generated beam achieves higher spectral efficiency. This is mainly because, under low SNR conditions, the pilot measurement and feedback-based scheme is significantly affected by pilot noise, leading to substantial distortion in channel estimation. In contrast, the channel information relied upon by the scheme in this invention primarily originates from pre-stored data in the CKM, making it less susceptible to noise disturbances and thus exhibiting stronger robustness. As shown in Table 1, as the SNR gradually increases, the noise impact weakens, improving the estimation accuracy of the pilot-based scheme and narrowing the performance gap between the two schemes. In fact, in the higher SNR region, the traditional scheme may even slightly outperform the scheme in this invention. However, considering that actual maritime communication scenarios typically involve complex channel environments and high levels of noise interference, the hybrid precoding method based on angle-domain CKM proposed in this invention is more advantageous in practical deployments and can effectively improve the spectral efficiency of maritime communication systems.
[0099] Table 1. Spectral efficiency as a function of SNR (unit: bps / Hz)
[0100]
[0101] Example 6:
[0102] To verify the robustness of the proposed hybrid precoding method for shipborne base stations based on CKM angle tracking in high-speed moving scenarios, this embodiment simulates the spectral efficiency of the hybrid precoding scheme under different relative ship speeds. During the simulation, an angle-domain CKM was established in a uniform sampling manner within the coverage area of a large-scale MIMO array, with a sampling interval of 50 m. The SNR of the sea channel propagation environment was set to 10 dB. The relative speeds of the base station ship and the user ship were set to 10, 15, 20, 25, and 30 knots, respectively. A traditional hybrid precoding scheme based on pilot measurement and feedback was used as a benchmark comparison method. The relationship curves between the system spectral efficiency and relative ship speed under the two hybrid precoding schemes are shown below. Figure 7As shown in the figure, at higher relative ship speeds, the proposed hybrid precoding scheme based on CKM angle tracking exhibits superior performance, achieving higher spectral efficiency in its generated beam. This is because, under high relative ship speeds, the pilot measurement and feedback-based scheme is significantly affected by pilot delay, leading to CSI expiration and increased estimation noise. In contrast, the channel information relied upon by this invention primarily originates from pre-stored data in the CKM, making it less affected by pilot delay and thus exhibiting stronger robustness. As shown in Table 2, as the relative ship speed gradually decreases, the adverse effects of pilot delay weaken, and the estimation accuracy of the pilot-based scheme improves accordingly. The performance gap between the two schemes narrows, and in lower ship speed regions, the traditional scheme may even slightly outperform this invention. However, considering that ships typically operate at high speeds in real-world maritime scenarios, the proposed hybrid precoding method based on angle-domain CKM is more advantageous in practical deployments, effectively enhancing dynamic beam coverage and improving the robustness of maritime communication systems.
[0103] Table 2. Numerical values of spectral efficiency as a function of relative ship speed (unit: bps / Hz)
[0104]
[0105] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A hybrid precoding method for shipborne base stations based on CKM angle tracking, characterized in that, Includes the following steps: S1. Sample and measure angle data within the coverage area of the shipborne base station, and construct the angle domain CKM based on the results of the angle measurement data; S2. The optimal estimation algorithm is used to predict the target user's movement trajectory at the next moment, and the trajectory prediction results are corrected in a timely manner by combining the position-angle relationship in the angle domain CKM. S3. Based on the trajectory prediction results, query the angle domain CKM to obtain the channel angle parameters in advance, and reconstruct the channel matrix based on the query results; S4. Calculate the optimal all-digital precode by singular value decomposition of the channel matrix, and use the orthogonal matching pursuit (OMP) algorithm to approximate the all-digital precode successively to generate analog precode, and then design digital precode by combining the channel matrix. S5. Real-time comparison of actual measured values of channel angle parameters with predicted values of CKM, and dynamic update of CKM.
2. The hybrid precoding method for shipborne base stations based on CKM angle tracking according to claim 1, characterized in that, In step S1, sparse grid sampling and wireless channel measurement are first performed in the target sea area covered by the shipborne base station, and the strongest power is recorded at different location coordinates. The path angle and signal strength information are used to construct the initial angle domain CKM corresponding to the base station coverage area; By leveraging the correspondence between the CKM storage location and channel-related information, channel angle information can be quickly queried and obtained to guide hybrid precoding design. The CKM is obtained through sampling measurements and interpolation prediction, and its data storage structure is represented as follows: ; In the formula, and Representing maps The number of rows and columns of grid points. , and They represent the numbers on the map. OK The arrival angle AoA, departure angle AoD, and signal strength of the user location corresponding to each grid point, and , If a shipborne base station adopts a uniform planar array, its AoA and AoD each contain components of two dimensions: elevation angle and azimuth angle.
3. The hybrid precoding method for shipborne base stations based on CKM angle tracking according to claim 2, characterized in that, In step S2, optimal estimation theory is used to predict and track the user ship's position. Based on optimal estimation theory, the optimal estimation algorithm for the position state adopts the Extended Kalman Filter (EKF) algorithm. Based on the EKF algorithm, the current time... The location prediction expression is as follows: ; In the formula, For nonlinear system models, For the target user's ship location The optimal estimate at time -1 Control quantities applied externally, This represents systematic error; for the user ship trajectory tracking results predicted by the optimal estimation algorithm, at each time step... The base station receives the uplink pilot signal sent by the user ship, extracts the multipath channel parameters, and inputs the obtained multipath channel parameters into the observation values of the CKM query candidate position. Based on the chi-square distribution of the signal interval according to the square of the Mahalanobis distance between the predicted value of the optimal estimation algorithm and the CKM query value, it determines whether the current prediction value of the optimal estimation algorithm is reliable. If the line-of-sight propagation LosS path exists and the prediction value of the optimal estimation algorithm is reliable, the position of the user ship is updated according to the prediction value of the optimal estimation algorithm. Otherwise, if the line-of-sight propagation LosS path does not exist or the prediction value of the optimal estimation algorithm is unreliable, the position prediction of the optimal estimation algorithm is corrected using the CKM query value.
4. The hybrid precoding method for shipborne base stations based on CKM angle tracking according to claim 3, characterized in that, In S2, the shipborne base station inputs the multipath channel parameters measured in real time into the CKM as a query key, and the CKM returns the candidate location area corresponding to the multipath channel parameters, which is used as prior information to correct the prediction result of the optimal estimation algorithm. The shipborne base station system compares the Mahalanobis distance between the predicted value of the optimal estimation algorithm and the CKM query value in real time. If the Mahalanobis distance exceeds the chi-square threshold, the weight of the CKM query value is adaptively increased to constrain the user location prediction result to the area indicated by the CKM.
5. The hybrid precoding method for shipborne base stations based on CKM angle tracking according to claim 1, characterized in that, In S3, the base station inputs the user's position coordinates after trajectory tracking correction into the angle domain CKM; CKM queries and outputs the expected angle and intensity information corresponding to the location based on its stored "location coordinates-multipath channel parameters" mapping relationship, and reconstructs the channel matrix using the angle and intensity information; The mathematical expression for reconstructing the channel matrix using the multipath channel parameters obtained from the CKM query is as follows: ; In the formula, and These represent the number of user receiving antennas and the number of base station transmitting antennas, respectively. This indicates the number of channel paths stored in CKM; , and They represent the first The signal strength, AoA, and AoD of each path, with AoA and AoD each containing components of elevation and azimuth. , ; and These represent the array response vectors at the transmitting and receiving ends, respectively; Assume the shipborne base station uses a fixed, uniform planar array, and the array antenna elements are regularly arranged in the vertical plane, with horizontal and vertical directions respectively having and One antenna element, satisfying The response vector of a two-dimensional array is represented as: ; in: ; ; In the formula, and These represent the elevation and azimuth angles of AoD, respectively. For carrier wavelength, The spacing between the array antenna elements is denoted as , and is taken as . , For Kronecker product.
6. The hybrid precoding method for shipborne base stations based on CKM angle tracking according to claim 1, characterized in that, In step S4, the shipborne base station utilizes the channel matrix reconstruction results, including the following steps: S41. Calculate the all-digital precoding matrix. To establish the optimization objective of the hybrid precoding design scheme, the optimal all-digital precoding matrix is calculated. First, the channel matrix... Perform singular value decomposition to obtain ,in, yes unitary matrix of order, It is semi-positive definite Diagonal matrix of order 1 yes Unitary matrix of order; select right singular matrix The former Column vector composition , Indicates the number of radio frequency chains; This represents the optimal precoding matrix design target that the antenna system can achieve without hardware constraints. The optimization target of hybrid precoding is approximated by a hybrid structure. S42. Iterative solution of the analog precoding matrix In a hybrid precoding architecture, to obtain analog precoding with good performance, a feasible approach is to use the OMP algorithm to successively approximate the optimal target. Since the analog end uses a phase shifter to change the phase, and the phase shifter is a finite set of codebooks, the OMP algorithm is used to select a set of column vectors from the codebook so that the analog precoding matrix formed by them is mathematically close to... The task of the OMP algorithm is to extract from a pre-defined codebook. Selected from A "basic vector" is used to construct an analog precoding matrix through multiple iterations. Used to adjust the antenna array so that the main lobe of the beam points in the direction of the predicted angle; S43. Calculate the hybrid precoding matrix. After the analog precoding is determined, it is combined with the channel matrix. Obtaining an equivalent low-dimensional channel The high-dimensional hybrid precoding design is a low-dimensional digital precoding. During baseband processing, the minimum mean square error estimation criterion is used to calculate the digital precoding matrix, and its calculation formula is as follows: ; In the formula, Indicates noise power. Represents an identity matrix, a digital precoding matrix. Used for digital domain elimination of residual multi-user interference or inter-datastream interference, based on And combined with previously obtained Calculate the required hybrid precoding matrix .
7. The hybrid precoding method for shipborne base stations based on CKM angle tracking according to claim 1, characterized in that, In step S5, after receiving the actual signal at the next moment, the actual channel angle parameter is measured, and the measured value is recorded. CKM query value Compare, if there is a deviation Exceeding the preset threshold ,Right now This triggers a local update of the CKM, allowing it to adapt to the dynamic changes in the marine environment. The CKM update result is represented as follows: ; In the formula, This represents the local update function of CKM, used for combining renew ; , , and They represent The Middle The user's location coordinates, angle of arrival (AoA), angle of departure (AoD), and signal strength; This represents a set of locally updated position coordinates.
8. A shipborne base station hybrid precoding system based on CKM angle tracking, characterized in that, The system comprising the steps of applying the method according to any one of claims 1 to 7, wherein the system includes: The angle domain CKM construction module is configured to perform the following process: sample and measure angle data within the coverage area of the shipborne base station, and construct the angle domain CKM based on the results of the measured angle data; The trajectory prediction result correction module is configured to perform the following process: predict the target user's movement trajectory at the next moment using the optimal estimation algorithm, and correct the trajectory prediction result in a timely manner by combining the position-angle relationship in the angle domain CKM; The channel matrix reconstruction module is configured to perform the following process: query the angle domain CKM to obtain channel angle parameters in advance based on the trajectory prediction results, and reconstruct the channel matrix based on the query results; The analog precoding module is configured to perform the following process: calculate the optimal all-digital precoding by singular value decomposition of the channel matrix, and successively approximate the all-digital precoding using the orthogonal matching pursuit (OMP) algorithm to generate analog precoding, and then design digital precoding by combining the channel matrix. The CKM update module is configured to perform the following process: compare the actual measured values of the channel angle parameters with the CKM predicted values in real time, and dynamically update the CKM.
9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the computer program is executed, it implements the steps of the method as described in any one of claims 1 to 7.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that, The computer program is configured to implement the steps of the method according to any one of claims 1 to 7 when invoked by a processor.
