Millimeter wave MIMO sparse channel estimation method based on dynamic path verification mechanism

Abstract

The invention discloses a millimeter wave MIMO sparse channel estimation method based on a dynamic path verification mechanism, belongs to the technical field of wireless communication, and solves the technical problems in the prior art that adaptive adjustment cannot be performed according to the real channel complexity and the estimation precision is low. According to the millimeter wave MIMO sparse channel estimation method based on the dynamic path verification mechanism provided by the invention, the dynamic path verification estimation based on actually measured data is carried out, namely training and testing are carried out based on real millimeter wave channel measurement data, and the dynamic path verification mechanism is introduced in each iteration; in combination with a sparseness adaptive mechanism, dynamically controlling iteration termination according to training statistical information, and outputting an estimation vector and reconstructing a channel based on verification of actual measurement data; the invention relates to a sparse channel estimation method, in particular to a sparse channel estimation method combining a dynamic path verification mechanism and a sparseness adaptive strategy, which is suitable for low-complexity robust estimation in an actual measurement channel environment, and is especially suitable for a low-SNR or deep fading system.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technologies, and specifically to a millimeter-wave MIMO sparse channel estimation method based on a dynamic path verification mechanism. Background Art

[0002] Millimeter-wave communication is regarded as one of the key enabling technologies for 5G and future 6G due to its wideband spectrum resources. Millimeter-wave MIMO systems typically adopt large-scale arrays and hybrid beamforming architectures. Due to the severe path loss and occlusion effects during the propagation of high-frequency signals, millimeter-wave channels usually exhibit sparsity in the angular domain or delay domain. In recent years, many sparse channel estimation methods have been proposed, such as Orthogonal Matching Pursuit (OMP), Compressive Sampling Matching Pursuit (CoSaMP), Weighted OMP (WOMP), Sparse Bayesian Learning (SBL), etc. However, these methods generally have the following problems: (1) Most methods use fixed sparsity parameters and cannot be adaptively adjusted according to the true channel complexity; (2) Lack of a path verification mechanism, which is prone to introducing false paths and affecting the estimation accuracy; (3) Although methods such as SBL have better performance, their computational complexity is high and it is not conducive to the implementation of real-time systems; (4) Most studies are based on ideal geometric channel models and lack the adaptation and verification of real measured channels.

[0003] Therefore, there is an urgent need for a sparse channel estimation scheme that can adapt to complex actual channel structures, has robustness and low complexity, and a channel estimation method for millimeter-wave multiple-input multiple-output (mmWave MIMO) systems. Summary of the Invention

[0004] Aiming at the shortcomings and deficiencies in the prior art, the present invention provides a millimeter-wave MIMO sparse channel estimation method based on a dynamic path verification mechanism. It is trained and tested based on real millimeter-wave channel measurement data, which is close to engineering practice; a dynamic path verification mechanism is introduced in each iteration to improve the path selection accuracy; combined with a sparsity adaptive mechanism, the iteration termination is dynamically controlled according to the training statistical information; a channel estimation algorithm with high accuracy, low complexity, and strong generalization ability is realized.

[0005] To achieve the above objectives, the present invention is realized through the following technical solutions: The millimeter-wave MIMO sparse channel estimation method based on a dynamic path verification mechanism provided by the present invention is a dynamic path verification estimation based on measured data, that is, it is trained and tested based on real millimeter-wave channel measurement data, and a dynamic path verification mechanism is introduced in each iteration; combined with a sparsity adaptive mechanism, the iteration termination is dynamically controlled according to the training statistical information, and the specific steps are as follows:

[0006] S1. Construct a channel observation model: Based on the OFDM frame structure with a hybrid analog-digital architecture, establish a channel observation model for the millimeter-wave MIMO system; sparsely model the channel in the angular domain and construct a measurement matrix;

[0007] S2. Dynamic path verification weighted OMP algorithm: The inputs include a measurement vector, a measurement matrix, a dictionary matrix, as well as an upper bound on sparsity, the number of path verification candidates, and a threshold parameter; The algorithm steps include initialization, calculation of correlation, path candidate screening, dynamic path verification, path selection and support set update, and termination condition judgment;

[0008] Among them, the dynamic path verification mechanism is specifically as follows: In each iteration, select multiple candidate paths for verification, perform least squares estimation and residual calculation on each candidate path, and select the optimal path to update the support set;

[0009] The adaptive termination mechanism is: Set a residual descent rate threshold and an absolute residual threshold as double judgment conditions; Adaptively control the number of iterations according to the channel quality conditions;

[0010] S3. Output an estimated vector and reconstruct the channel in the verification based on measured data.

[0011] Preferably, the channel observation model in step S1 is specifically as follows:

[0012] The millimeter-wave MIMO system adopts a hybrid analog-digital architecture. Based on the OFDM frame structure, the received signal of the k-th subcarrier is modeled as follows:

[0013]

[0014] Among them, H[k] is the frequency-domain channel matrix on the k-th subcarrier; respectively represent the transmit and receive hybrid beamforming matrices in the k-th time slot of the training frame. The superscript H represents the conjugate transpose; q (m) represents the pilot vector of the m-th training frame, with a dimension of N s × 1, where N s is the number of data streams; represents Gaussian white noise;

[0015] The channel is represented under sparse modeling in the angular domain as:

[0016]

[0017] In the above formula, A t , A r respectively represent the dictionary matrices of the transmit and receive arrays, and their column vectors are response vectors in the discrete angular domain directions;

[0018] H v[k] is the k-th subcarrier angular domain sparse channel matrix. Vectorize it and construct the measurement matrix to get

[0019] y[k] = Φh v [k] + n c [k] (3)

[0020] where in the measurement matrix denotes the matrix A t 's complex conjugate, denotes the transpose of the matrix, denotes the Kronecker product; y[k] is the received measurement vector corresponding to the k-th subcarrier, and h v [k] = vec(H) is the angular domain sparse channel vector under the k-th subcarrier.

[0021] Preferably, in step S2, the dynamic path verification weighted OMP algorithm is as follows:

[0022] Input:

[0023] The measurement vector y[k], the measurement matrix Φ, and the dictionary matrix Ψ;

[0024] The sparsity upper limit L, the number of path verification candidates L val ;

[0025] The threshold parameters τ, ε;

[0026] Output:

[0027] The sparse channel estimation vector The channel estimation matrix

[0028] The specific algorithm steps are as follows:

[0029] S21. Initialization: The residual r (0) = y[k], the support set S (0) is an empty set, and the iteration number t = 1;

[0030] S22. Calculate the correlation:

[0031] Calculate the correlation of the current residual with all candidate paths:

[0032] c (t) = Φ H r (t-1)

[0033] where Φ is the measurement matrix, r (t-1) represents the residual vector of the previous iteration, and c (t) is the projection coefficient of all paths and the residual;

[0034] S23. Path candidate screening:

[0035] Sort the paths according to the magnitude of |c (t) |, and select the top L paths with the largest magnitude val to form a candidate set

[0036] Preferably, a dynamic path verification mechanism: By selecting multiple candidate paths for verification in each iteration, the specific steps are as follows:

[0037] a) Calculate the correlation between the current residual and all candidate paths;

[0038] b) Select the top K paths with the largest correlation as the candidate set;

[0039] c) Perform least squares estimation and residual calculation on each candidate path;

[0040] d) Select the path with the smallest residual as the optimal path and update the support set.

[0041] Preferably, the dynamic path verification mechanism includes the following specific algorithm:

[0042] For each path c in the set perform the following steps:

[0043] a) Set the temporary support set:

[0044] b) Solve the least squares problem and estimate the sparse coefficients:

[0045]

[0046] c) Calculate the residual:

[0047]

[0048] Path selection and support set update:

[0049] Select the optimal path:

[0050] and update the support set S (t) = S (t-1) ∪ {c *}; The current residual

[0051] Preferably, the termination condition in the adaptive termination mechanism is judged as follows:

[0052] If any of the following conditions is met, terminate the iteration:

[0053] a) The residual descent rate is too low:

[0054]

[0055] b) The absolute value of the current residual is too small: ||r (t) ||₂ < ε

[0056] where τ is the residual descent rate threshold, set to 0.01, and ε is the convergence threshold, set to 10 -4 .

[0057] Preferably, in the verification based on measured data, an estimated vector is output and the channel is reconstructed:

[0058] a) Construct a support subset Ψ of the sparse dictionary S , and obtain the channel vector:

[0059] b) Reconstruct the vector into a two-dimensional channel matrix:

[0060] where represents the sparse dictionary matrix, represents the sparse coefficient vector.

[0061] The present invention provides a millimeter-wave MIMO sparse channel estimation method based on a dynamic path verification mechanism. It has the following beneficial effects:

[0062] (1) The present invention uses real millimeter-wave channel measurement data for training and testing, which is close to engineering practice; a dynamic path verification mechanism is introduced in each iteration to improve the path selection accuracy; combined with a sparsity adaptive mechanism, the iteration termination is dynamically controlled according to the training statistical information; a channel estimation algorithm with high precision, low complexity and strong generalization ability is realized. It has the following advantages:

[0063] Introducing a dynamic path verification mechanism: Compared with the traditional OMP-like algorithm that only selects the maximum correlation path in each round, the proposed scheme performs residual verification among multiple candidate paths, effectively avoiding the error problem caused by early misselected paths and significantly improving the estimation accuracy and robustness.

[0064] The adaptive termination mechanism is reasonably designed: The present invention can adaptively control the number of iterations according to different channel quality conditions (such as low SNR or deep fading) through a dual judgment mechanism of the residual descent rate threshold and the absolute residual, avoiding overfitting and unnecessary computational overhead, and taking into account both accuracy and efficiency.

[0065] The implementation complexity is controllable: Although a multi-path verification mechanism is introduced, since only a limited number of paths are used for residual estimation in each round, the overall computational complexity is still significantly lower than that of the global search algorithms.

[0066] Robustness is better than traditional OMP: The evaluation results through simulation experiments and public measured data sets (AI for 5G Challenge channel data set of North Carolina State University) show that the algorithm of the present invention has lower NMSE and more stable convergence performance than the baseline algorithm (such as weighted OMP) under the measured channel conditions. Description of the Drawings

[0067] Figure 1 It is a diagram of the process of obtaining the measured data channel based on Embodiment 2 of the present invention;

[0068] Figure 2 It is a curve diagram comparing the NMSE performance of the present invention and different algorithms under different SNR conditions;

[0069] Figure 3 It is a curve diagram comparing the NMSE performance of the present invention and different algorithms under different numbers of training frames. Detailed Embodiments

[0070] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments.

[0071] To further illustrate the practical operability of the present invention, two specific embodiments are provided.

[0072] Embodiment 1:

[0073] As Figure 1 shown, the millimeter-wave MIMO sparse channel estimation method based on the dynamic path verification mechanism of the present invention includes the following contents:

[0074] S1: Channel observation model

[0075] The millimeter-wave MIMO system adopts a hybrid analog-digital structure. Based on the OFDM frame structure, the received signal of the k-th subcarrier is modeled as follows:

[0076]

[0077] where H[k] is the frequency-domain channel matrix on the k-th subcarrier; respectively represent the transmit and receive hybrid beamforming matrices in the k-th time slot of the training frame, and the superscript H represents the conjugate transpose; q (m) represents the pilot vector of the m-th training frame, with a dimension of N s ×1, where N s is the number of data streams; represents white Gaussian noise.

[0078] The channel is sparsely modeled in the angular domain as:

[0079]

[0080] In the above formula, A t , A r respectively represent the dictionary matrices of the transmitting and receiving arrays, and their column vectors are the response vectors in the discrete angular domain direction. H v [k] is the angular domain sparse channel matrix of the k-th subcarrier.

[0081] Vectorize it and construct the measurement matrix to obtain

[0082] y[k] = Φh v [k] + n c [k] (3)

[0083] where in the measurement matrix represents the complex conjugate of matrix A t , represents the transpose of the matrix, represents the Kronecker product. y[k] is the received measurement vector corresponding to the k-th subcarrier, and h v [k] = vec(H) is the angular domain sparse channel vector under the k-th subcarrier.

[0084] S2: Dynamic Path Verification Weighted OMP Algorithm

[0085] Input:

[0086] Measurement vector y[k], measurement matrix Φ, dictionary matrix Ψ

[0087] Sparsity upper limit L, path verification candidate number L val

[0088] Threshold parameters τ, ε

[0089] Output:

[0090] Sparse channel estimation vector Channel estimation matrix

[0091] The specific algorithm steps are as follows:

[0092] 1. Initialization: Residual r (0) = y[k], support set S (0) is an empty set, and the iteration number t = 1;

[0093] 2. Calculate the correlation:

[0094] Calculate the correlation of the current residual with all candidate paths:

[0095] c(t) = Φ H r (t-1)

[0096] where Φ is the measurement matrix, r (t-1) represents the residual vector of the previous iteration, and c (t) is the projection coefficient of all paths and the residual.

[0097] 3. Path candidate screening:

[0098] Sort the paths according to the magnitude of |c (t) |, and select the top L val paths with the largest magnitude to form the candidate set

[0099] 4. Dynamic path verification:

[0100] For each path c in the set perform the following steps:

[0101] a) Set the temporary support set:

[0102] b) Solve the least squares problem to estimate the sparse coefficients:

[0103]

[0104] c) Calculate the residual:

[0105]

[0106] 5. Path selection and support set update:

[0107] Select the optimal path:

[0108] and update the support set S (t) = S (t-1) ∪ {c *}; The current residual

[0109] 6. Termination condition judgment:

[0110] If any of the following conditions is satisfied, terminate the iteration:

[0111] a) The residual decrease rate is too low:

[0112]

[0113] b) The absolute value of the current residual is too small: ||r (t) ||2 < ε

[0114] Among them, τ is the residual descent rate threshold, set to 0.01, and ε is the convergence threshold, set to 10 -4 。

[0115] 7. Output the estimated vector and reconstruct the channel:

[0116] a) Construct the support subset Ψ of the sparse dictionary S , and obtain the channel vector:

[0117] b) Reconstruct the vector into a two-dimensional channel matrix:

[0118] Among them, represents the sparse dictionary matrix, represents the sparse coefficient vector.

[0119] Embodiment 2:

[0120] The dynamic path verification estimation based on measured data of the present invention. In this embodiment, the real-scenario channel measurement samples provided in the public dataset are used to construct a typical single-user millimeter-wave MMO channel estimation scenario, and the method proposed by the present invention is used to complete the channel estimation.

[0121] System parameter setting:

[0122] 1. Use the real-scenario channel measurement samples in the AI for 5G Challenge channel dataset of North Carolina State University. Each channel has multiple paths and contains information such as angle, gain, and delay;

[0123] 2. Number of transmit antennas: N t = 16; Number of receive antennas N r = 64, both are uniform linear arrays;

[0124] 3. Number of RF links: N RF = 4, carrier frequency is 28 GHz, and the number of subcarriers K = 64;

[0125] 4. Analog beamforming matrices F tr , W tr Adopt the directional beam based on the discrete Fourier transform;

[0126] 5. The number of training frames used for each estimation is 20;

[0127] 6. Sparsity upper limit L = 6, number of path verification candidates L val = 2.

[0128] Estimation process:

[0129] 1. Use the channel data H[k] in the measured scenario to construct the observation vector y[k].

[0130] 2. Use the method of the present invention to perform dynamic path iteration for each sub - carrier k = 1,..., K.

[0131] Initialize the residual and the support set.

[0132] Calculate the correlation between the current residual and all candidate paths.

[0133] Select the top K paths with the largest correlation as the candidate set.

[0134] Perform least - squares estimation and residual calculation for each candidate path.

[0135] Select the path with the smallest residual to update the support set.

[0136] Judge the termination condition. If not satisfied, continue the iteration.

[0137] 3. Finally, obtain the channel estimation

[0138] 4. Calculate the normalized mean - square error index to verify the algorithm performance.

[0139] Verification result:

[0140] The method of the present invention is superior to the traditional weighted OMP and its improved algorithms under low signal - to - noise ratio and limited training frame conditions. The simulation results show that the dynamic path verification mechanism can effectively screen out mis - selected paths and improve the estimation accuracy.

[0141] Example 3:

[0142] The comparison of the millimeter - wave MIMO sparse channel estimation method based on the dynamic path verification mechanism of the present invention with the measured data results of the existing millimeter - wave MIMO sparse channel simulation estimation methods is as Figure 2 and Figure 3 shown in the simulation performance comparison results.

[0143] Using Matlab simulation, the system and channel parameters are the same as those in Example 2. The normalized mean - square error is defined as:

[0144] The algorithm of the present invention is compared with the following traditional sparse channel estimation algorithms: weighted OMP, improved weighted OMP, improved weighted CoSaMP, weighted OMP with multiple paths.

[0145] Result analysis: Figure 2The NMSE comparison curves of the above five algorithms under the condition of low signal-to-noise ratio from -15 dB to -5 dB are shown, and the results are as follows: (1) The algorithm of the present invention performs optimally in the whole range. The proposed algorithms all show the lowest NMSE values, verifying their robustness and anti-interference ability in high-noise environments. (2) Compared with the weighted OMP with static sparsity control and the multi-path extended version, by introducing a dynamic path verification mechanism and a sparsity adaptive update strategy, the present invention effectively reduces the risk of misselected paths under low SNR conditions, thus achieving better performance. (3) Although the improved weighted CoSaMP combines the path backtracking characteristics of the CoSaMP framework and has certain performance in some medium SNR scenarios, it is sensitive to noise in the low SNR region, and the overall NMSE performance is not as good as the verification result of the algorithm of the present invention. (4) As the SNR gradually increases, the NMSE of the algorithm of the present invention decreases smoothly, indicating that its estimation process has good numerical stability and generalization ability. It is proved that the overall trend of the method of the present invention is stable and the convergence is good.

[0146] This embodiment further analyzes the influence of the number of training frames (denoted as M) on the normalized mean square error performance of each algorithm. As Figure 3 shows the change trend of the estimation accuracy of different algorithms when the number of training frames increases from 20 to 100 under the same SNR condition. The results are as follows: The overall performance of the present invention is significantly improved as the number of training frames increases: as the number of training frames increases from a relatively small value (such as M = 20) to a larger value, the NMSE of all algorithms shows a downward trend. This is in line with the compressive sensing theory, that is, more observation samples can improve the recoverability of the sparse channel. The algorithm of the present invention performs optimally under all numbers of training frames. Figure 3 It verifies the important influence of the number of training frames on the sparse channel estimation performance, and especially shows the characteristics of strong adaptability, high efficiency and leading estimation accuracy of the algorithm proposed by the present invention under different training sample conditions. This algorithm does not rely on fixed sparsity setting and can automatically adjust the estimation process according to the observed data, reflecting higher practical application value.

[0147] In summary, the following conclusions can be drawn: Under the condition of low signal-to-noise ratio (-15 dB to -5 dB), the NMSE performance of the method of the present invention is better than that of other comparison algorithms, showing better robustness and anti-interference ability. As the number of training frames increases (from 20 to 100), the method of the present invention performs optimally under all numbers of training frames, showing good adaptability and estimation accuracy.

[0148] The above embodiments fully demonstrate the application effect of the method of the present invention in an actual millimeter-wave MIMO system and verify its superior performance in a complex channel environment.

[0149] Finally, although this specification is described in terms of embodiments, not every embodiment contains only one independent technical solution. This narrative style of the specification is only for clarity. Those skilled in the art should regard the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

[0150] As mentioned above, the above are only the preferred specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, making equivalent substitutions or changes, should be covered by the protection scope of the present invention.

Claims

1. A millimeter-wave MIMO sparse channel estimation method based on a dynamic path verification mechanism, characterized in that Dynamic path verification estimation based on measured data, that is, training and testing are carried out based on real millimeter-wave channel measurement data, and a dynamic path verification mechanism is introduced in each iteration; combined with a sparsity adaptive mechanism, the iteration termination is dynamically controlled according to the training statistical information. The specific steps are as follows: S1. Construct a channel observation model: Based on the OFDM frame structure of the hybrid analog-digital structure, establish a channel observation model for the millimeter-wave MIMO system; sparsely model the channel in the angular domain and construct a measurement matrix; S2. Dynamic path verification weighted OMP algorithm: The inputs include a measurement vector, a measurement matrix, a dictionary matrix, as well as an upper limit of sparsity, the number of path verification candidates, and a threshold parameter; the algorithm steps include initialization, calculation of correlation, path candidate screening, dynamic path verification, path selection and support set update, and termination condition judgment; Among them, the dynamic path verification mechanism is specifically as follows: In each iteration, multiple candidate paths are selected for verification. For each candidate path, least squares estimation and residual calculation are performed, and the optimal path is selected to update the support set; The adaptive termination mechanism is: Set the residual descent rate threshold and the absolute residual threshold as double judgment conditions; adaptively control the number of iterations according to the channel quality conditions; S3. Output the estimated vector and reconstruct the channel in the verification based on measured data.

2. The millimeter-wave MIMO sparse channel estimation method based on a dynamic path verification mechanism according to claim 1, characterized in that, The channel observation model in step S1 is specifically as follows: The millimeter-wave MIMO system adopts a hybrid analog-digital structure. Based on the OFDM frame structure, the received signal of the k-th subcarrier is modeled as follows: where \(H[k]\) is the frequency-domain channel matrix on the \(k\)-th subcarrier; respectively represent the transmit and receive hybrid beamforming matrices in the \(k\)-th time slot of the training frame, and the superscript \(H\) represents conjugate transpose; \(q\) (m) represents the pilot vector of the \(m\)-th training frame, with a dimension of \(N\) s \(\times1\), where \(N\) s is the number of data streams; represents additive white Gaussian noise; The channel is represented as sparse modeling in the angular domain: In the above formula, A t , A r respectively represent the dictionary matrices of the transmitting and receiving arrays, and their column vectors are the response vectors in the discrete angular domain direction; H v [k] is the k-th subcarrier angular domain sparse channel matrix. Vectorize it and construct a measurement matrix to obtain y[k] = Φh v [k] + n c [k] (3) Among them, in the measurement matrix represents the matrix A t is the complex conjugate of, represents the transpose of the matrix, represents the Kronecker product; y[k] is the received measurement vector corresponding to the k-th subcarrier, and h v [k] = vec(H) is the angular-domain sparse channel vector at the k-th subcarrier.

3. The millimeter-wave MIMO sparse channel estimation method based on a dynamic path verification mechanism according to claim 1, wherein In step S2, the dynamic path verification weighted OMP algorithm is specifically as follows: Input: Measurement vector y[k], measurement matrix Φ, dictionary matrix Ψ; Sparsity upper limit L, path verification candidate number L val ; Threshold parameters τ, ε; Output: Sparse channel estimation vector Channel estimation matrix The specific algorithm steps are as follows: S21. Initialization: residual r (0) = y[k], support set S (0) is an empty set, iteration number t = 1; S22. Calculate the correlation: Calculate the correlation of the current residual with all candidate paths: c( t ) = Φ H r( t-1 ) where Φ is the measurement matrix, and r (t-1) represents the residual vector of the previous iteration, and c (t) is the projection coefficient of all paths and the residual; S23. Path candidate screening: Sort the paths according to the magnitude of |c (t) |, and select the top L paths with the largest magnitude val to form a candidate set 4. The millimeter-wave MIMO sparse channel estimation method based on a dynamic path verification mechanism according to claim 1, wherein The dynamic path verification mechanism: By selecting multiple candidate paths for verification in each iteration, the specific steps are as follows: a) Calculate the correlation of the current residual with all candidate paths; b) Select the top K paths with the largest correlation as the candidate set; c) Perform least squares estimation and residual calculation on each candidate path; d) Select the path with the smallest residual as the optimal path and update the support set.

5. The millimeter-wave MIMO sparse channel estimation method based on a dynamic path verification mechanism according to claim 4, wherein The dynamic path verification mechanism includes the following specific algorithm: For the set perform the following steps for each path c in it: a) Set up a temporary support set: b) Solve the least squares problem and estimate the sparse coefficients: c) Calculate the residual: Path selection and support set update: Select the optimal path: And update the support set S (t) = S (t-1) ∪ {c *}; The current residual 6. The millimeter-wave MIMO sparse channel estimation method based on a dynamic path verification mechanism according to claim 5, characterized in that The termination condition judgment in the adaptive termination mechanism is as follows: If any of the following conditions is met, the iteration is terminated: a) The residual descent rate is too low: b) The absolute value of the current residual is too small: ||r (t) ||2 < ε Among them, τ is the residual descent rate threshold, set to 0.01, and ε is the convergence threshold, set to 10 -4 .

7. The millimeter-wave MIMO sparse channel estimation method based on a dynamic path verification mechanism according to claim 3, characterized in that Output the estimated vector and reconstruct the channel in the verification based on measured data: a) Construct the support subset Ψ of the sparse dictionary S , and obtain the channel vector: b) Reconstruct the vector into a two-dimensional channel matrix: Among them, is represented as a sparse dictionary matrix, represents a sparse coefficient vector.

Images