Beam forming method based on delay Doppler domain channel parameter prediction in high-speed mobile communication system

Abstract

The invention discloses a beam forming method based on delay Doppler domain channel parameter prediction in a high-speed mobile communication system, and relates to the technical field of wireless communication, in particular to a beam forming method. The invention aims to solve the problem of wave beam mismatch at the future moment caused by CSI aging in high-speed mobile communication. The method comprises the following steps: 1, obtaining a channel response matrix and a power spectrum in a DD domain; 2, obtaining an estimated value of an angle of arrival, Doppler and time delay; 3, obtaining a final cluster set and an adaptive threshold; 4, calculating a space vector; calculating an angle of arrival based on the spatial vector; calculating a cluster direction vector based on the angle of arrival; calculating a sub-modal Doppler center and a sub-modal space vector; calculating sub-modal equivalent gain estimation of each group in each cluster; and 5, obtaining a prediction channel based on the complex gain propulsion prediction of each group in each cluster, and obtaining a prediction beam based on the prediction channel.

Description

Technical Field

[0001] This invention relates to the field of wireless communication technology, specifically to a beamforming method based on orthogonal time-frequency spatial modulation (OTFS) delay-Doppler (DD) domain observation, which includes peak detection, adaptive peak clustering, cluster-level Doppler sub-component extraction, and channel extrapolation prediction in high mobility scenarios (such as high-speed rail / vehicle-to-everything). Background Technology

[0002] In high-speed mobile communication scenarios (such as high-speed rail and vehicle-to-everything communication), wireless channels exhibit rapidly changing time-varying characteristics, with large Doppler frequency shifts and short channel coherence times. Traditional beamforming techniques based on instantaneous channel state information (CSI) face the "channel aging" problem: due to delays in signal processing and feedback, the beam weights used by the transmitter become mismatched with the current actual channel, leading to a significant decrease in array gain. Therefore, to maintain high-quality connections, it is essential to predict future channel states.

[0003] Existing channel prediction methods typically operate directly in the time or frequency domain, extrapolating from the channel's temporal correlation. However, in high-speed mobile scenarios, the channel's Doppler spread is large, and its temporal correlation is weak, limiting the performance of direct prediction. OTFS modulation maps the signal to the DD domain, where the response of time-varying channels exhibits sparsity and quasi-staticity, with each distinguishable path possessing relatively stable delay, Doppler, and spatial angle parameters. This enables long-term, robust channel prediction using path physical parameters. Existing prediction methods based on recurrent neural networks, long short-term memory networks, or attention mechanisms require extensive offline testing and suffer from high complexity and poor timeliness. Existing methods struggle to simultaneously balance prediction accuracy and real-time complexity in high-speed mobile scenarios, thus requiring further improvement.

[0004] Currently, how to efficiently and robustly extract multipath parameters from the received signal in the DD domain, and then reasonably model and cluster them to achieve fast and effective channel prediction and beamforming is a technical problem that urgently needs to be solved. Summary of the Invention

[0005] The purpose of this invention is to solve the problem of future beam mismatch caused by CSI aging in high-speed mobile communication, and to propose a beamforming method based on delayed Doppler domain channel parameter prediction in high-speed mobile communication systems.

[0006] The specific process of beamforming based on delayed Doppler domain channel parameter prediction in high-speed mobile communication systems is as follows:

[0007] Step 1: Obtain the channel response matrix in the DD domain and power spectrum ;

[0008] The DD domain is a time-delay-Doppler domain;

[0009] Step 2: Based on power spectrum Design threshold ; set the threshold Convert to linear threshold Based on threshold Peak set obtained by nonmaximum suppression ; For sets Each grid point According to the power spectrum Grid points from largest to smallest Sort the peaks to obtain the peak set. For peak set Each grid point Filter and retain grid points that meet the criteria. Grid points that meet the conditions Constitutes the final peak set ; For the final peak set Each grid point Take space vector ;based on calculate ;calculate The maximum value corresponds to the estimated angle of arrival. At the same time, Doppler was obtained With delay ;

[0010] Step 3: Based on the estimated angle of arrival obtained in Step 2, and Doppler... Delay Perform peak adaptive clustering to obtain the final cluster set and adaptive threshold. ;

[0011] Step 4: Calculate the cluster set medium cluster spatial vectors Based on spatial vectors Computational cluster set medium cluster Angle of arrival Based on cluster set medium cluster Angle of arrival Calculate cluster set medium cluster Cluster direction vector ;

[0012] Computational cluster set medium cluster The The group power weighted mean is the sub-mode Doppler center. and cluster sets medium cluster The Submode space vector ;

[0013] Based on cluster set medium cluster Cluster direction vector and cluster set medium cluster The Submode space vector Calculate the submodal equivalent gain estimate for each group within each cluster. ;

[0014] Step 5: Estimating the equivalent submodal gain of each group within each cluster. Obtain the complex gain advance prediction for each group within each cluster. Based on the complex gain advance prediction of each group within each cluster Obtain the prediction channel Based on predictive channel Obtain the predicted beam .

[0015] The beneficial effects of this invention are as follows:

[0016] This invention proposes a channel prediction and beamforming method based on time-delay-Doppler domain channel parameter extraction, clustering, and modeling. This method can robustly estimate the physical parameters of the channel from the received signal, predict the channel state at future times, and calculate the optimal precoding or beamforming weights, thereby effectively combating channel aging, solving beam mismatch problems at future times, and improving system performance.

[0017] This invention improves beamforming gain: By predicting future channels and designing beamformers accordingly, the received signal power can be increased compared to methods using outdated channel information, and further array gain can be improved.

[0018] This invention enhances robustness by employing robust peak detection and adaptive parameter clustering based on median absolute deviation (MAD) to improve the identification probability of true multipath components and reduce the impact of spurious peaks on modeling.

[0019] This invention has strong physical interpretability: Based on the sparse physical parameters of the channel (delay, Doppler, angle), it performs complex gain analysis on sub-components. Phase advancement enables prediction, the model is clear, and the prediction process conforms to the physical laws of channel changes.

[0020] This invention features low computational complexity and ease of real-time implementation: Instead of performing time-series prediction on the complete high-dimensional channel matrix, it first extracts sparse physical parameters of the channel in the delay-Doppler domain and then performs low-order modeling and phase-advancing prediction only on the Doppler parameters of each cluster based on clustering. Compared to prediction methods based on recurrent neural networks, long short-term memory networks, or attention mechanisms, this invention avoids large-scale matrix multiplication and high-dimensional state updates. Its prediction process involves only a small number of complex exponential operations and vector superposition operations, significantly reducing computational complexity and storage overhead, making it suitable for real-time deployment in high-speed mobile scenarios. Attached Figure Description

[0021] Figure 1 This is an overall flowchart of the method described in this invention;

[0022] Figure 2 A schematic diagram of the sparse structure of the delay-Doppler domain channel and peak detection;

[0023] Figure 3 This is a schematic diagram of channel cluster orientation estimation based on delay-Doppler domain parameter clustering;

[0024] Figure 4 A comparison chart of absolute array gain between the aging baseline for future moments, the predicted beam, and the ideal upper limit;

[0025] Figure 5 A comparison chart of relative performance gains between the aging baseline, predicted beam, and ideal upper limit for future moments. Detailed Implementation

[0026] Specific Implementation Method 1: The specific process of the beamforming method based on delayed Doppler domain channel parameter prediction in the high-speed mobile communication system of this implementation method is as follows:

[0027] Step 1: Obtain the channel response matrix in the DD domain and power spectrum ;

[0028] The DD domain is a time-delay-Doppler domain;

[0029] Step 2: Based on power spectrum Design threshold ; set the threshold Convert to linear threshold Based on threshold Peak set obtained by nonmaximum suppression ; For sets Each grid point According to the power spectrum Grid points from largest to smallest Sort the peaks to obtain the peak set. For peak set Each grid point Filter and retain grid points that meet the criteria. Grid points that meet the conditions Constitutes the final peak set ; For the final peak set Each grid point Take space vector ;based on calculate ;calculate The maximum value corresponds to the estimated angle of arrival. At the same time, Doppler was obtained With delay ;

[0030] Step 3: Based on the estimated angle of arrival obtained in Step 2, and Doppler... Delay Perform peak adaptive clustering to obtain the final cluster set and adaptive threshold. ;

[0031] Step 4: Calculate the cluster set medium cluster spatial vectors Based on spatial vectors Computational cluster set medium cluster Angle of arrival Based on cluster set medium cluster Angle of arrival Calculate cluster set medium cluster Cluster direction vector ;

[0032] Computational cluster set medium cluster The The group power weighted mean is the sub-mode Doppler center. and cluster sets medium cluster The Submode space vector ;

[0033] Based on cluster set medium cluster Cluster direction vector and cluster set medium cluster The Submode space vector Calculate the submodal equivalent gain estimate for each group within each cluster. ;

[0034] Step 5: Estimating the equivalent submodal gain of each group within each cluster. Obtain the complex gain advance prediction for each group within each cluster. Based on the complex gain advance prediction of each group within each cluster Obtain the prediction channel Based on predictive channel Obtain the predicted beam .

[0035] Specific Implementation Method Two: This implementation method differs from Specific Implementation Method One in that, in step one, the channel response matrix is ​​obtained in the DD domain. and power spectrum The DD domain is a time-delay-Doppler domain; the specific process is as follows:

[0036] Step 11: The receiving end (base station) obtains the time-delay-Doppler domain channel response matrix using orthogonal time-frequency spatial modulation (OTFS) or time-delay-Doppler domain signal processing techniques. ;

[0037] Steps 1 and 2: Based on channel response Calculate the power spectrum ; indicates as:

[0038]

[0039] in, Indicates the total number of antennas; Indicates the total number of Dopplers; Indicates the total number of delays;

[0040] Represents the channel response matrix The Middle The root receiving antenna at the delay index is Doppler index is The complex channel response value at the delay-Doppler domain grid cell;

[0041] The delay-Doppler domain is a two-dimensional discrete representation domain obtained through OTFS transformation;

[0042] The delay-Doppler grid consists of a discrete time delay set and a discrete Doppler set, with each index pair... This corresponds to a discrete delay-Doppler unit;

[0043] Delay Discretized into discrete time delay index Doppler shift Discretized into discrete Doppler indexes ;

[0044]

[0045]

[0046] Step 13: Analyze the power spectrum Perform a logarithmic transformation to obtain the logarithmic form of the power spectrum. ; indicates as:

[0047]

[0048] in, It represents extremely small positive numbers, avoiding logarithmic zero.

[0049] The other steps and parameters are the same as in Specific Implementation Method 1.

[0050] Specific Implementation Method Three: This implementation method differs from Specific Implementation Method One or Two in that step two is based on the power spectrum. Design threshold ; set the threshold Convert to linear threshold Based on threshold Peak set obtained by nonmaximum suppression ; For sets Each grid point According to the power spectrum Grid points from largest to smallest Sort the peaks to obtain the peak set. For peak set Each grid point Filter and retain grid points that meet the criteria. Grid points that meet the conditions Constitutes the final peak set ; For the final peak set Each grid point Take space vector ;based on calculate ;calculate The maximum value corresponds to the estimated angle of arrival. At the same time, Doppler was obtained With delay The specific process is as follows:

[0051] Step 21: Based on power spectrum Design threshold The specific process is as follows:

[0052]

[0053] in, Indicates the threshold coefficient;

[0054] This represents the mean. , ; Indicates taking the median value;

[0055] Indicates standard deviation, , ;

[0056] Step 22: Set the threshold Convert to linear threshold ; indicates as:

[0057]

[0058] Steps two and three: Based on threshold Peak set obtained by nonmaximum suppression The specific process is as follows:

[0059] Set an empty set ;

[0060] Set neighborhood radius For each grid point :

[0061] If the following conditions are met: and equal Neighborhood If the maximum value is found within the range, then... Add to collection ;

[0062] otherwise, Do not join the set ;

[0063] in, Represents the power spectrum; Indicates the first The neighborhood radius of a Doppler; Indicates the first The neighborhood radius of the delay;

[0064] Step 2.4: For the set Each grid point According to the power spectrum Grid points from largest to smallest Sort the peaks to obtain the peak set. ;

[0065] Step 25: For the peak set Each grid point Filter and retain grid points that meet the criteria. Grid points that meet the conditions Constitutes the final peak set ;

[0066] The specific process is as follows:

[0067] Set minimum retention threshold , retain satisfaction grid points This forms the final peak set. ;

[0068] in ;

[0069] Represents the set of peaks The power spectrum of the first grid point;

[0070] Represents the set of peaks Power spectrum of all grid points;

[0071] Indicates intermediate variables; represents the power spectrum of the first grid point. Power spectrum of all grid points The ratio;

[0072] To represent extremely small positive numbers, avoid logarithmic zero;

[0073] Step 26: For the final peak set Each grid point Take space vector ; indicates as:

[0074]

[0075] in,

[0076] Represents the final set of peak values The Middle grid points spatial vectors, ;

[0077] This indicates that for each grid point In all Complex vectors under one antenna;

[0078] Represents the final set of peak values Medium grid points Total number;

[0079] Represent real numbers;

[0080] Step 27: Define the peak set The Middle grid points The angle of arrival is , ;

[0081] Define the set of angles ;

[0082] Denotes the first angle in the set of angles. This represents the second angle in the set of angles. Describes the first angle in the set of angles. One corner, Describes the first angle in the set of angles. An angle; for example: define an angle range from -30° to 30°, with a step size of 0.1°. that is The number of discrete angle points in the middle;

[0083] Step 28: Calculate the first angle in the angle set. The related angles ;

[0084] in,

[0085] Indicates the array direction vector; superscript This indicates finding the conjugate;

[0086] ;

[0087] Superscript This indicates the transpose; Represents the imaginary unit. ;

[0088] Step 29, Calculation The maximum value corresponds to the estimated angle of arrival. At the same time, Doppler was obtained With delay ;

[0089] Represented as:

[0090]

[0091] in,

[0092] Indicates related The maximum value corresponds to the estimated angle of arrival. Corresponding grid points Doppler;

[0093] Indicates related The maximum value corresponds to the estimated angle of arrival. Corresponding grid points The time delay.

[0094] Other steps and parameters are the same as in specific implementation method one or two.

[0095] Specific Implementation Method Four: This implementation method differs from Specific Implementation Methods One to Three in that the estimated angle of arrival obtained in Step Two is used in Step Three. Doppler Delay Perform peak adaptive clustering to obtain the final cluster set and adaptive threshold. The specific process is as follows:

[0096] Step 31: Calculate the final peak set Doppler resolution of each grid point and latency resolution ; indicates as:

[0097]

[0098] in, This represents the total number of Doppler grid points in step one; Indicates maximum Doppler;

[0099] This represents the total number of delay grid points in step one; This indicates the maximum latency, such as 5 μs;

[0100] Step 3.2, for the final peak set The Middle grid points The Euclidean distance is used to find the first grid points nearest neighbor grid point index , ; ;

[0101] Step 33, based on the first grid points corresponding angle of arrival and the grid points corresponding angle of arrival Calculate the nearest neighbor difference ;

[0102] Based on the grid points Corresponding Doppler and the grid points Corresponding Doppler Calculate the nearest neighbor difference ;

[0103] Based on the grid points Corresponding delay and the grid points Corresponding delay Calculate the nearest neighbor difference ;

[0104] Represented as:

[0105]

[0106] Steps 3 and 4: Based on the nearest neighbor difference Obtain the initial threshold ;

[0107] Based on nearest neighbor difference Obtain the initial threshold ;

[0108] Based on nearest neighbor difference Obtain the initial threshold ;

[0109] Represented as:

[0110]

[0111] in, Indicates the magnification factor; Indicates taking the median value;

[0112] Step 35: To avoid the threshold being too large or too small, this implementation method further squeezes the threshold;

[0113] Based on the initial threshold Obtain the Doppler threshold Based on the initial threshold Obtain the delay threshold Based on the initial threshold Obtain the arrival angle threshold ;

[0114] Step 36: For the final peak set Each grid point According to the power spectrum matrix Grid points from largest to smallest Sort the peaks to obtain the final sorted set. ;

[0115] Set the first empty cluster, and then use the sorted final peak set. The first grid point Place it in the first empty cluster and calculate the cluster. The center of () ;

[0116] Process the sorted final peak set sequentially Each grid point For the current grid point Check existing clusters in sequence cluster center ;

[0117] If the current grid point Simultaneously satisfy the threshold (if there is 1 grid point) Satisfying the thresholds of multiple clusters, but without affecting the order of traversal, and placing it into the first cluster traversed: Then the current grid point Join cluster Cluster updates are performed using power weighting. Cluster center: ;

[0118] in,

[0119] The angle of arrival is obtained in step two. The Doppler obtained in step two, The delay is the one obtained in step two;

[0120] For cluster set Clusters in , For clusters The Middle Power spectrum of each grid point For the first indivual, , For clusters Total number of grid points;

[0121] For clusters The Middle grid points The corresponding Doppler, For clusters The Middle grid points The corresponding delay, For clusters The Middle grid points Corresponding angle of arrival;

[0122] For clusters The Doppler corresponding to the cluster center, For clusters The time delay corresponding to the cluster center, For clusters The angle of arrival corresponding to the cluster center;

[0123] If the current grid point If none of the existing clusters meet the threshold, then create a new cluster and set the current grid points... Add to the newly created cluster; the cluster center of the new cluster is equal to the current grid point. Corresponding angle of arrival, Doppler, and time delay; update the cluster set;

[0124] Step 37: Determine if two clusters exist. satisfy: ;

[0125] If it exists, proceed to step three eight;

[0126] If it does not exist, then the cluster set obtained in step three and six is ​​the final cluster set. ;

[0127] in,

[0128] Cluster The Doppler corresponding to the cluster center, Cluster The Doppler corresponding to the cluster center;

[0129] Cluster The time delay corresponding to the cluster center, Cluster The time delay corresponding to the cluster center;

[0130] Cluster The angle of arrival corresponding to the cluster center, Cluster The angle of arrival corresponding to the cluster center;

[0131] Step 38: Combine the two clusters They are merged into one cluster, and the new cluster center is recalculated based on power weighting.

[0132] Repeat step 37 until no two clusters remain. satisfy: ;

[0133] Obtain the final cluster set .

[0134] The other steps and parameters are the same as those in one of the specific implementation methods one to three.

[0135] Specific Implementation Method Five: This implementation method differs from Specific Implementation Methods One to Four in that, in steps three and five...

[0136] Based on the initial threshold Obtain the Doppler threshold Based on the initial threshold Obtain the delay threshold Based on the initial threshold Obtain the arrival angle threshold The specific process is as follows:

[0137] Step 3.51: Based on the initial threshold Obtain the Doppler threshold : ;

[0138] in, Represents the clipping function; Indicates the Doppler threshold; This indicates the set minimum Doppler value. This indicates the maximum Doppler value set.

[0139] ;

[0140] This means taking the minimum value. This indicates taking the maximum value;

[0141] Step 352: Based on the initial threshold Obtain the delay threshold : ;

[0142] in, Indicates the delay threshold; This indicates the minimum set latency. This indicates the maximum set delay value;

[0143] ;

[0144] Step 353: Based on the initial threshold Obtain the arrival angle threshold : ;

[0145] in, Indicates reaching the angle threshold. This indicates the minimum angle of arrival set. This indicates the maximum set angle of arrival.

[0146] ;

[0147] Steps three, five, and four.

[0148] Doppler threshold satisfy: ;

[0149] Delay threshold satisfy: ;

[0150] Arrival angle threshold satisfy: ;

[0151] in, This represents the minimum value of the Doppler threshold. This represents the minimum delay threshold. This represents the minimum value of the arrival angle threshold.

[0152] The other steps and parameters are the same as those in one of the specific implementation methods one to four.

[0153] Specific Implementation Method Six: This implementation method differs from Specific Implementation Methods One to Five in that the calculation of the cluster set in step four is... medium cluster spatial vectors Based on spatial vectors Computational cluster set medium cluster Angle of arrival Based on cluster set medium cluster Angle of arrival Calculate cluster set medium cluster Cluster direction vector ;

[0154] Computational cluster set medium cluster The The group power weighted mean is the sub-mode Doppler center. and cluster sets medium cluster The Submode space vector ;

[0155] Based on cluster set medium cluster Cluster direction vector and cluster set medium cluster The Submode space vector Calculate the submodal equivalent gain estimate for each group within each cluster. ;

[0156] The specific process is as follows:

[0157] Step 41: Calculate the cluster set medium cluster spatial vectors ; indicates as:

[0158]

[0159] in,

[0160] For cluster set medium cluster Spatial vector;

[0161] For clusters The Middle The spatial vector of each grid point;

[0162] For clusters The Middle Power spectrum of each grid point;

[0163] Step 42: Based on spatial vectors Computational cluster set medium cluster Angle of arrival ; indicates as:

[0164]

[0165] in,

[0166] Indicates the array direction vector; superscript This indicates finding the conjugate;

[0167] Step 43: Based on cluster set medium cluster Angle of arrival Calculate cluster set medium cluster Cluster direction vector ; indicates as:

[0168]

[0169] in, It is the direction vector;

[0170] ;

[0171] Step 44: Define the cluster set medium cluster All grid points Doppler sets are ;

[0172] When cluster Inner grid points The number is less than the threshold , make cluster Inner grid points Group 1;

[0173] When cluster Inner grid points Number greater than or equal to threshold , make cluster Inner grid points They were divided into two groups, Group 1 and Group 2.

[0174] in,

[0175] For clusters The first grid point inside Doppler; For clusters The second grid point inside Doppler; For clusters Inner grid points Doppler; For clusters Inner grid points Doppler;

[0176] Steps four and five:

[0177] Computational cluster set medium cluster The The group power weighted mean is the sub-mode Doppler center. ; indicates as:

[0178]

[0179] Computational cluster set medium cluster The Submode space vector ; indicates as:

[0180]

[0181] in,

[0182] For clusters The Group 1 The spatial vector of each grid point;

[0183] For cluster set medium cluster The Group;

[0184] Step 46: Based on cluster set medium cluster Cluster direction vector and cluster set medium cluster The Submode space vector Calculate the submodal equivalent gain estimate for each group within each cluster. ; indicates as:

[0185]

[0186] in,

[0187] Superscript This indicates finding the conjugate; Indicates the current time;

[0188] For cluster set medium cluster Cluster direction vector.

[0189] The other steps and parameters are the same as those in one of the specific implementation methods one to five.

[0190] Specific Implementation Method Seven: This implementation method differs from Specific Implementation Methods One through Six in that step five involves estimating the sub-mode equivalent gain of each group within each cluster. Obtain the complex gain advance prediction for each group within each cluster. Based on the complex gain advance prediction of each group within each cluster Obtain the prediction channel Based on predictive channel Obtain the predicted beam The specific process is as follows:

[0191] Step 51: Estimating the equivalent submodal gain of each group within each cluster. Obtain the complex gain advance prediction for each group within each cluster. ; indicates as:

[0192]

[0193] in,

[0194] express Time Cluster The Submodal equivalent gain estimation of the group;

[0195] express Time Cluster The Submodal equivalent gain estimation of the group;

[0196] Represents a cluster set medium cluster The The group power weighted mean is the sub-mode Doppler center;

[0197] Indicates the prediction time interval;

[0198] The imaginary unit, ;

[0199] Step 5.2: Estimation of submodal equivalent gain for each group within each cluster. Get the current Time-based channel estimation ;

[0200] Step 53, based on Time Cluster The Submode equivalent gain estimation of the group ,get Time Prediction Channel ;

[0201] Step 54, regarding the current situation Time-based channel estimation Normalization is performed to obtain the beam. .

[0202] The other steps and parameters are the same as those in one of the specific implementation methods one to six.

[0203] Specific Implementation Method Eight: This implementation method differs from Specific Implementation Methods One through Seven in that: in step five-two, the sub-mode equivalent gain estimation based on each group within each cluster is performed... Get the current Time-based channel estimation ; indicates as:

[0204]

[0205] in, Represents a cluster set medium cluster The total number of groups;

[0206] For cluster set medium cluster Cluster direction vector.

[0207] The other steps and parameters are the same as those in any of the specific implementation methods one to seven.

[0208] Specific Implementation Method Nine: This implementation method differs from Specific Implementation Methods One through Eight in that: in step five-three, based on... Time Cluster The Submode equivalent gain estimation of the group ,get Time Prediction Channel ; indicates as:

[0209] .

[0210] The other steps and parameters are the same as those in one of the specific implementation methods one to eight.

[0211] Specific Implementation Method Ten: This implementation method differs from Specific Implementation Methods One through Nine in that: in step five-four, the current... Time-based channel estimation Normalization is performed to obtain the beam. ; indicates as:

[0212]

[0213] in,

[0214] To represent extremely small positive numbers, avoid logarithmic zero; Represents the 2-norm;

[0215] right Time Prediction Channel Normalization is performed to obtain the predicted beam. ; indicates as:

[0216]

[0217] in,

[0218] To represent extremely small positive numbers, avoid logarithmic zero; It represents the 2-norm.

[0219] The other steps and parameters are the same as those in any of the specific implementation methods one to nine.

[0220] Example:

[0221] 1) Simulation parameter settings:

[0222] This embodiment simulates the communication process between a base station and a mobile terminal in a high-speed railway scenario. The system carrier frequency is set to 3.5 GHz, the terminal moves at a speed of 350 km / h, and the base station uses a 64-antenna uniform linear array. The channel is modeled in the delay-Doppler (DD) domain with a grid size of 64 (Doppler dimension) × 32 (delay dimension), a maximum delay of 5 μs, and a maximum Doppler frequency offset of approximately 1134 Hz. The prediction time interval Δ = 4 × 10⁻³ s, and the threshold coefficient is... ,radius Energy retention threshold Intra-cluster splitting threshold To closely approximate the real high-speed rail propagation environment, the channel is constructed using a cluster-ray model, comprising one strong line-of-sight cluster and two non-line-of-sight multipath clusters. Each cluster contains multiple rays, each with a certain extension in time delay, Doppler frequency shift, and angle of arrival. Intra-cluster power is distributed according to an exponential decay pattern, and complex Gaussian noise is introduced to simulate actual reception conditions. In DD domain power map processing, robust statistical thresholding is used for peak detection. The threshold is determined by the median and median absolute deviation of the power distribution, with a threshold coefficient of 6. The neighborhood radius for non-maximum suppression is one grid cell in both the Doppler and time delay dimensions. During peak energy screening, candidate peaks with normalized energy not lower than 0.03 are retained. In the intra-cluster Doppler modeling stage, a single cluster is split into a maximum of two Doppler sub-components, and no splitting is performed when the number of peaks in the cluster is less than three, to avoid overfitting.

[0223] 2) Simulation content and result analysis:

[0224] Simulation generates a DD domain channel power distribution map, such as Figure 3 Multiple energy concentration regions, corresponding to different propagation paths, can be clearly observed. After automatic peak detection and adaptive clustering, these scattered points are rationally grouped into three physical clusters, such as... Figure 2 As shown, consistent with the preset cluster structure, this verifies the ability of the proposed method to extract effective paths in complex channels.

[0225] In terms of performance, three beamforming strategies were compared: the traditional method directly uses the current channel estimation result; the method of this invention extrapolates the cluster-level Doppler parameters based on the current estimation result to predict the channel at future times before beamforming; the theoretical upper limit assumes that the actual channel at future times is known and is only used as a reference for the upper bound of performance. Simulation results show that, as Figure 4 , Figure 5As shown, the method of this invention brings an array gain improvement of approximately 1.07 dB compared to traditional methods, and the difference from the theoretical upper limit is approximately 2.01 dB. The results demonstrate that the proposed channel prediction method based on DD domain clustering and Doppler phase advancement can effectively track channel evolution in high-speed moving scenarios, improving the alignment accuracy of downlink beamforming and system performance without significantly increasing computational complexity.

[0226] This invention may have other embodiments. Without departing from the spirit and essence of this invention, those skilled in the art can make various corresponding changes and modifications according to this invention, but these corresponding changes and modifications should all fall within the protection scope of the appended claims.

Claims

1. A beamforming method based on delayed Doppler domain channel parameter prediction in a high-speed mobile communication system, characterized in that: The specific process of the method is as follows: Step 1: Obtain the channel response matrix in the DD domain and power spectrum ; The DD domain is a time-delay-Doppler domain; Step 2: Based on power spectrum Design threshold ; set the threshold Convert to linear threshold Based on threshold Peak set obtained by nonmaximum suppression ; For sets Each grid point According to the power spectrum Grid points from largest to smallest Sort the peaks to obtain the peak set. For peak set Each grid point Filter and retain grid points that meet the criteria. Grid points that meet the conditions Constitutes the final peak set ; For the final peak set Each grid point Take space vector ;based on calculate ; calculate The maximum value corresponds to the estimated angle of arrival. At the same time, Doppler was obtained With delay ; Step 3: Based on the estimated angle of arrival obtained in Step 2, and Doppler... Delay Perform peak adaptive clustering to obtain the final cluster set and adaptive threshold. ; Step 4: Calculate the cluster set medium cluster spatial vectors ; Based on space vectors Computational cluster set medium cluster Angle of arrival Based on cluster set medium cluster Angle of arrival Calculate cluster set medium cluster Cluster direction vector ; Computational cluster set medium cluster The The group power weighted mean is the sub-mode Doppler center. and cluster sets medium cluster The Submode space vector ; Based on cluster set medium cluster Cluster direction vector and cluster set medium cluster The Submode space vector Calculate the submodal equivalent gain estimate for each group within each cluster. ; Step 5: Estimating the equivalent submodal gain of each group within each cluster. Obtain the complex gain advance prediction for each group within each cluster. Based on the complex gain advance prediction of each group within each cluster Obtain the prediction channel Based on predictive channel Obtain the predicted beam .

2. The beamforming method based on delayed Doppler domain channel parameter prediction in a high-speed mobile communication system according to claim 1, characterized in that: In step one, the channel response matrix is ​​obtained in the DD domain. and power spectrum ; The DD domain is a time-delay-Doppler domain; The specific process is as follows: Step 11: The receiver obtains the delay-Doppler domain channel response matrix. ; Steps 1 and 2: Based on channel response Calculate the power spectrum ; indicates as: in, Indicates the total number of antennas; Indicates the total number of Dopplers; Indicates the total number of delays; Represents the channel response matrix The Middle The root receiving antenna at the delay index is Doppler index is The complex channel response value at the delay-Doppler domain grid cell; Step 13: Analyze the power spectrum Perform a logarithmic transformation to obtain the logarithmic form of the power spectrum. ; indicates as: in, It represents a very small positive number.

3. The beamforming method based on delayed Doppler domain channel parameter prediction in a high-speed mobile communication system according to claim 2, characterized in that: Step two is based on power spectrum Design threshold ; set the threshold Convert to linear threshold Based on threshold Peak set obtained by nonmaximum suppression ; For sets Each grid point According to the power spectrum Grid points from largest to smallest Sort the peaks to obtain the peak set. For peak set Each grid point Filter and retain grid points that meet the criteria. Grid points that meet the conditions Constitutes the final peak set ; For the final peak set Each grid point Take space vector ;based on calculate ;calculate The maximum value corresponds to the estimated angle of arrival. At the same time, Doppler was obtained With delay ; The specific process is as follows: Step 21: Based on power spectrum Design threshold The specific process is as follows: in, Indicates the threshold coefficient; This represents the mean. , ; Indicates taking the median value; Indicates standard deviation, , ; Step 22: Set the threshold Convert to linear threshold ; indicates as: Steps two and three: Based on threshold Peak set obtained by nonmaximum suppression The specific process is as follows: Set an empty set ; Set neighborhood radius For each grid point : If the following conditions are met: and equal Neighborhood If the maximum value is found within the range, then... Add to collection ; otherwise, Do not join the set ; in, Represents the power spectrum; Indicates the first The neighborhood radius of a Doppler; Indicates the first The neighborhood radius of the delay; Step 2.4: For the set Each grid point According to the power spectrum Grid points from largest to smallest Sort the peaks to obtain the peak set. ; Step 25: For the peak set Each grid point Filter and retain grid points that meet the criteria. Grid points that meet the conditions Constitutes the final peak set ; The specific process is as follows: Set minimum retention threshold , retain satisfaction grid points This forms the final peak set. ; in ; Represents the set of peaks The power spectrum of the first grid point; Represents the set of peaks Power spectrum of all grid points; Indicates intermediate variables; Represents a very small positive number; Step 26: For the final peak set Each grid point Take space vector ; indicates as: in, Represents the final set of peak values The Middle grid points spatial vectors, ; This indicates that for each grid point In all Complex vectors under one antenna; Represents the final set of peak values Medium grid points Total number; Represent real numbers; Step 27: Define the peak set The Middle grid points The angle of arrival is , ; Define the set of angles ; Denotes the first angle in the set of angles. This represents the second angle in the set of angles. Describes the first angle in the set of angles. One corner, Describes the first angle in the set of angles. Each corner; Step 28: Calculate the first angle in the angle set. The related angles ; in, Indicates the array direction vector; superscript This indicates finding the conjugate; ； Superscript This indicates the transpose; Represents the imaginary unit. ; Step 29, Calculation The maximum value corresponds to the estimated angle of arrival. At the same time, Doppler was obtained With delay ; Represented as: in, Indicates related The maximum value corresponds to the estimated angle of arrival. Corresponding grid points Doppler; Indicates related The maximum value corresponds to the estimated angle of arrival. Corresponding grid points The time delay.

4. The beamforming method based on delayed Doppler domain channel parameter prediction in a high-speed mobile communication system according to claim 3, characterized in that: The angle of arrival estimate obtained in step two is used in step three. Doppler Delay Perform peak adaptive clustering to obtain the final cluster set and adaptive threshold. ; The specific process is as follows: Step 31: Calculate the final peak set Doppler resolution of each grid point and latency resolution ; indicates as: in, This represents the total number of Doppler grid points in step one; Indicates maximum Doppler; This represents the total number of delay grid points in step one; Indicates the maximum delay; Step 3.2, for the final peak set The Middle grid points The Euclidean distance is used to find the first... grid points nearest neighbor grid point index , ; ; Step 33 Based on the grid points corresponding angle of arrival and the grid points corresponding angle of arrival Calculate the nearest neighbor difference ; Based on the grid points Corresponding Doppler and the grid points Corresponding Doppler Calculate the nearest neighbor difference ; Based on the grid points Corresponding delay and the grid points Corresponding delay Calculate the nearest neighbor difference ; Represented as: Steps three and four Based on nearest neighbor difference Obtain the initial threshold ; Based on nearest neighbor difference Obtain the initial threshold ; Based on nearest neighbor difference Obtain the initial threshold ; Represented as: in, Indicates the magnification factor; Indicates taking the median value; Step 35: Based on the initial threshold Obtain the Doppler threshold Based on the initial threshold Obtain the delay threshold Based on the initial threshold Obtain the arrival angle threshold ; Step 36: For the final peak set Each grid point According to the power spectrum matrix Grid points from largest to smallest Sort the peaks to obtain the final sorted set. ; Set the first empty cluster, and then use the sorted final peak set. The first grid point Place it in the first empty cluster and calculate the cluster. center ; Process the sorted final peak set sequentially Each grid point For the current grid point Check existing clusters in sequence cluster center ; If the current grid point Simultaneously satisfy the threshold: Then the current grid point Join cluster Cluster updates are performed using power weighting. Cluster center: ; in, The angle of arrival is obtained in step two. The Doppler obtained in step two, The delay is the one obtained in step two; For cluster set Clusters in , For clusters The Middle Power spectrum of each grid point For the first indivual, , For clusters Total number of grid points; For clusters The Middle grid points The corresponding Doppler, For clusters The Middle grid points The corresponding delay, For clusters The Middle grid points Corresponding angle of arrival; For clusters The Doppler corresponding to the cluster center, For clusters The time delay corresponding to the cluster center, For clusters The angle of arrival corresponding to the cluster center; If the current grid point If none of the existing clusters meet the threshold, then create a new cluster and set the current grid points... Add to the newly created cluster; the cluster center of the new cluster is equal to the current grid point. Corresponding angle of arrival, Doppler, and time delay; update the cluster set; Step 37: Determine if two clusters exist. satisfy: ; If it exists, proceed to step three-eight; If it does not exist, then the cluster set obtained in step three and six is ​​the final cluster set. ; in, Cluster The Doppler corresponding to the cluster center, Cluster The Doppler corresponding to the cluster center; Cluster The time delay corresponding to the cluster center, Cluster The time delay corresponding to the cluster center; Cluster The angle of arrival corresponding to the cluster center, Cluster The angle of arrival corresponding to the cluster center; Step 38: Combine the two clusters They are merged into one cluster, and the new cluster center is recalculated based on power weighting. Repeat step 37 until no two clusters remain. satisfy: ; The final cluster set is obtained. .

5. The beamforming method based on delayed Doppler domain channel parameter prediction in a high-speed mobile communication system according to claim 4, characterized in that: In step three and five, based on the initial threshold Obtain the Doppler threshold Based on the initial threshold Obtain the delay threshold Based on the initial threshold Obtain the arrival angle threshold ; The specific process is as follows: Step 3.51: Based on the initial threshold Obtain the Doppler threshold : ; in, Represents the clipping function; Indicates the Doppler threshold; This indicates the set minimum Doppler value. This indicates the maximum Doppler value set. ； This means taking the minimum value. This indicates taking the maximum value; Step 352: Based on the initial threshold Obtain the delay threshold : ; in, Indicates the delay threshold; This indicates the minimum set latency. This indicates the maximum set delay value; ； Step 353: Based on the initial threshold Obtain the arrival angle threshold : ; in, Indicates reaching the angle threshold. This indicates the minimum angle of arrival set. This indicates the maximum set angle of arrival. ； Steps three, five, and four. Doppler threshold satisfy: ; Delay threshold satisfy: ; Arrival angle threshold satisfy: ; in, This represents the minimum value of the Doppler threshold. This represents the minimum delay threshold. This represents the minimum value of the arrival angle threshold.

6. The beamforming method based on delayed Doppler domain channel parameter prediction in a high-speed mobile communication system according to claim 5, characterized in that: Step four involves calculating the cluster set. medium cluster spatial vectors Based on spatial vectors Computational cluster set medium cluster Angle of arrival Based on cluster set medium cluster Angle of arrival Calculate cluster set medium cluster Cluster direction vector ; Computational cluster set medium cluster The The group power weighted mean is the sub-mode Doppler center. and cluster sets medium cluster The Submode space vector ; Based on cluster set medium cluster Cluster direction vector and cluster set medium cluster The Submode space vector Calculate the submodal equivalent gain estimate for each group within each cluster. ; The specific process is as follows: Step 41: Calculate the cluster set medium cluster spatial vectors ; indicates as: in, For cluster set medium cluster Spatial vector; For clusters The Middle The spatial vector of grid points; For clusters The Middle Power spectrum of each grid point; Step 42: Based on spatial vectors Computational cluster set medium cluster Angle of arrival ; indicates as: in, Indicates the array direction vector; superscript This indicates finding the conjugate; Step 43: Based on cluster set medium cluster Angle of arrival Calculate cluster set medium cluster Cluster direction vector ; indicates as: in, It is the direction vector; ； Step 44: Define the cluster set medium cluster All grid points Doppler sets are ; When cluster Inner grid points The number is less than the threshold , make cluster Inner grid points Group 1; When cluster Inner grid points Number greater than or equal to threshold , make cluster Inner grid points They were divided into two groups, Group 1 and Group 2. in, For clusters The first grid point inside Doppler; For clusters The second grid point inside Doppler; For clusters Inner grid points Doppler; For clusters Inner grid points Doppler; Steps four and five: Computational cluster set medium cluster The The group power weighted mean is the sub-mode Doppler center. ; indicates as: Computational cluster set medium cluster The Submode space vector ; indicates as: in, For clusters The Group 1 The spatial vector of grid points; For cluster set medium cluster The Group; Step 46: Based on cluster set medium cluster Cluster direction vector and cluster set medium cluster The Submode space vector Calculate the submodal equivalent gain estimate for each group within each cluster. ; indicates as: in, Superscript This indicates finding the conjugate; Indicates the current time; For cluster set medium cluster Cluster direction vector.

7. The beamforming method based on delayed Doppler domain channel parameter prediction in a high-speed mobile communication system according to claim 6, characterized in that: Step five involves estimating the submodal equivalent gain of each group within each cluster. Obtain the complex gain advance prediction for each group within each cluster. Based on the complex gain advance prediction of each group within each cluster Obtain the prediction channel Based on predictive channel Obtain the predicted beam The specific process is as follows: Step 51: Estimating the equivalent submodal gain of each group within each cluster. Obtain the complex gain advance prediction for each group within each cluster. ; indicates as: in, express Time Cluster The Submodal equivalent gain estimation of the group; express Time Cluster The Submodal equivalent gain estimation of the group; Represents a cluster set medium cluster The The group power weighted mean is the sub-mode Doppler center; Indicates the prediction time interval; The imaginary unit, ; Step 5.2: Estimation of submodal equivalent gain for each group within each cluster. Get the current Time-based channel estimation ; Step 53, based on Time Cluster The Submode equivalent gain estimation of the group ,get Time Prediction Channel ; Step 54, regarding the current situation Time-based channel estimation Normalization is performed to obtain the beam. .

8. The beamforming method based on delayed Doppler domain channel parameter prediction in a high-speed mobile communication system according to claim 7, characterized in that: In step five two, the submodal equivalent gain estimation is based on each group within each cluster. Get the current Time-based channel estimation ; indicates as: in, Represents a cluster set medium cluster The total number of groups; For cluster set medium cluster Cluster direction vector.

9. The beamforming method based on delayed Doppler domain channel parameter prediction in a high-speed mobile communication system according to claim 8, characterized in that: Step 53 is based on Time Cluster The Submode equivalent gain estimation of the group ,get Time Prediction Channel ; indicates as: 。 10. The beamforming method based on delayed Doppler domain channel parameter prediction in a high-speed mobile communication system according to claim 9, characterized in that: In step five four, the current Time-based channel estimation Normalization is performed to obtain the beam. ; indicates as: in, Represents a very small positive number; Represents the 2-norm; right Time Prediction Channel Normalization is performed to obtain the predicted beam. ; indicates as: in, Represents a very small positive number; It represents the 2-norm.

Images