Channel estimation method, device, equipment, storage medium and computer program product

Abstract

The invention relates to the technical field of satellite communication, and discloses a channel estimation method, device and equipment, a storage medium and a computer program product, and the method comprises the steps: obtaining received signal information, and inputting the received signal information into a preset channel estimation model; selecting a current layer of cascading layer from the plurality of cascading layers, and obtaining a previous layer of cascading layer corresponding to the current layer of cascading layer; according to the channel estimation value and the received signal information which are obtained through calculation of the previous cascading layer, a channel updating value is obtained through calculation of a gradient descent module; according to the channel update value, a current channel estimation value is calculated through a near-end mapping module; and if the current channel estimation value does not meet the preset stop condition, returning to the step of selecting the current cascade layer from the plurality of cascade layers until the current channel estimation value meets the condition, and taking the current channel estimation value as a target channel estimation value. The channel characteristics are automatically learned through iteration instead of depending on an inaccurate preset model, and accurate estimation of the complete channel state is achieved.

Description

Technical Field

[0001] The present application relates to the field of satellite communication technology, and in particular to a channel estimation method, apparatus, device, storage medium and computer program product. Background Art

[0002] To address the channel estimation problem in massive MIMO LEO (Multiple-Input Multiple-Output Low Earth Orbit) satellite communication systems, the existing technology mainly adopts a channel estimation method based on compressed sensing. However, this compressed sensing method has a priori dependence problem. Due to the high-speed movement of satellites and the complexity of the channel propagation environment, the actual channel characteristics are often significantly different from the preset sparse model, which will lead to a significant decrease in channel reconstruction accuracy. In some cases, the algorithm will fail to converge, making it difficult to guarantee the performance of the satellite communication system. Summary of the invention

[0003] The main purpose of this application is to provide a channel estimation method, apparatus, device, storage medium and computer program product, aiming to solve the technical problem that the existing channel estimation method based on compressed sensing has a significant decrease in channel reconstruction accuracy due to the high-speed movement of satellites and the complexity of the channel propagation environment, and cannot achieve accurate estimation of the complete channel state.

[0004] To achieve the above object, the present application proposes a channel estimation method, which includes: Acquire received signal information, and input the received signal information into a preset channel estimation model, wherein the preset channel estimation model includes a plurality of cascade layers, and the cascade layers include a gradient descent module and a proximal mapping module; Select a current cascade layer from the multiple cascade layers, and obtain a previous cascade layer corresponding to the current cascade layer; A channel update value is calculated by the gradient descent module according to the channel estimation value calculated by the previous cascade layer and the received signal information; Obtaining a current channel estimation value by calculating through a near-end mapping module according to the channel update value; If the current channel estimation value does not meet the preset stop condition, the process returns to the step of selecting the current cascade layer from the multiple cascade layers until the current channel estimation value meets the condition, and the current channel estimation value is used as the target channel estimation value.

[0005] Optionally, the step of calculating and obtaining a current channel estimation value through a proximal mapping module according to the channel update value includes: Obtaining the channel features generated by the previous cascade layer; The channel features are fused based on a spatial attention mechanism to obtain a target fused feature; A current channel estimation value is obtained by calculation according to the target fusion feature and the channel update value.

[0006] Optionally, the step of fusing the channel features based on a spatial attention mechanism to obtain a target fused feature includes: Connecting the channel features along the channel dimension to obtain an initial feature map; Processing the initial feature map based on a preset convolutional layer to generate a corresponding number of feature matrices, wherein the feature matrices represent different features; Assigning weights to different spatial positions of the feature matrix, and fusing the feature matrix based on the weights to obtain initial fused features; The size of the initial fusion feature is adjusted, and the adjustment result is added element by element to the encoding feature output by the current layer of the encoder residual block in the channel feature to obtain the target fusion feature.

[0007] Optionally, the characteristic matrix includes a first matrix, a second matrix and a third matrix; The step of assigning weights to different spatial positions of the feature matrix and fusing the feature matrix based on the weights to obtain initial fused features includes: Multiplying the transpose of the first matrix by the second matrix to obtain a product matrix; Calculating the product matrix based on a normalized exponential function to generate a spatial attention map, wherein each element in the spatial attention map represents an attention weight of a spatial position corresponding to the feature matrix; The third matrix is ​​weightedly aggregated according to the spatial attention map to obtain an initial fusion feature.

[0008] Optionally, the step of calculating a channel update value by the gradient descent module according to the channel estimation value calculated by the previous cascade layer and the received signal information includes: Obtaining a step length training parameter, wherein the step length training parameter is used to learn an optimal step length strategy; The channel estimation value and the received signal information are taken as input based on a gradient descent function to calculate a channel update value.

[0009] Optionally, before the step of acquiring received signal information and inputting the received signal information into a preset channel estimation model, the method further includes: Initializing trainable parameters in the preset channel estimation model, wherein the trainable parameters include step size training parameters and parameters in the proximal mapping module; The trainable parameters are optimized based on a preset training set and a loss function, and the optimized results are updated to the preset channel estimation model.

[0010] In addition, to achieve the above purpose, the present application also proposes a channel estimation device, the channel estimation device comprising: An information acquisition module, used to acquire received signal information and input the received signal information into a preset channel estimation model, wherein the preset channel estimation model includes a plurality of cascade layers, and the cascade layers include a gradient descent module and a proximal mapping module; A layer determination module, used to select a current cascade layer from multiple cascade layers, and obtain a previous cascade layer corresponding to the current cascade layer; A channel update module, configured to calculate a channel update value through the gradient descent module according to the channel estimation value calculated by the previous cascade layer and the received signal information; A channel estimation module, configured to obtain a current channel estimation value by calculating the channel update value through a near-end mapping module; The estimation and determination module is used to return to the step of selecting the current cascade layer from multiple cascade layers if the current channel estimation value does not meet the preset stop condition, until the current channel estimation value meets the condition, and use the current channel estimation value as the target channel estimation value.

[0011] In addition, to achieve the above-mentioned purpose, the present application also proposes a channel estimation device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the channel estimation method as described above.

[0012] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the channel estimation method described above are implemented.

[0013] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, wherein the computer program product includes a computer program, and when the computer program is executed by a processor, the steps of the channel estimation method described above are implemented.

[0014] The present application discloses obtaining received signal information and inputting the received signal information into a preset channel estimation model, wherein the preset channel estimation model includes multiple cascade layers, and the cascade layers include a gradient descent module and a proximal mapping module; selecting a current cascade layer from the multiple cascade layers, and obtaining the previous cascade layer corresponding to the current cascade layer; calculating a channel update value through the gradient descent module according to the channel estimation value calculated by the previous cascade layer and the received signal information; calculating the current channel estimation value through the proximal mapping module according to the channel update value; if the current channel estimation value does not meet the preset stop condition, returning to the step of selecting the current cascade layer from the multiple cascade layers until the current channel estimation value meets the condition, and taking the current channel estimation value as the target channel estimation value. Combining iterative optimization algorithms with deep learning enables the network to automatically learn channel features from data instead of relying on a preset model that may be inaccurate, fundamentally solving the problem that the traditional compressed sensing method relies heavily on channel prior information, resulting in performance degradation in complex channel environments, and achieving accurate estimation of the complete channel state. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0016] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0017] Figure 1 This is a flow chart of the first embodiment of the channel estimation method of the present application; Figure 2 Schematic diagram of the trainable proximal gradient descent network architecture; Figure 3 This is a flow chart of a second embodiment of the channel estimation method of the present application; Figure 4 This is a schematic diagram of the cross-layer feature attention fusion module structure; Figure 5 This is a flow chart of a third embodiment of the channel estimation method of the present application; Figure 6 This is a schematic diagram of the module structure of the channel estimation device according to an embodiment of the present application; Figure 7 Schematic diagram of the device structure of the hardware operating environment involved in the channel estimation method in the embodiment of the present application.

[0018] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0019] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0020] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0021] The main solution of the embodiment of the present application is: obtaining received signal information and inputting the received signal information into a preset channel estimation model, wherein the preset channel estimation model includes multiple cascade layers, and the cascade layers include a gradient descent module and a proximal mapping module; selecting a current cascade layer from the multiple cascade layers, and obtaining the previous cascade layer corresponding to the current cascade layer; calculating a channel update value through the gradient descent module based on the channel estimation value calculated from the previous cascade layer and the received signal information; calculating a current channel estimation value through the proximal mapping module based on the channel update value; if the current channel estimation value does not meet the preset stop condition, returning to the step of selecting the current cascade layer from the multiple cascade layers until the current channel estimation value meets the condition, and using the current channel estimation value as the target channel estimation value.

[0022] Since satellite communication systems are equipped with a large number of antenna arrays to improve system performance, this brings two key challenges: the increase in the number of antennas requires more pilot signals to obtain complete channel state information, which increases system overhead and reduces the effective data transmission rate; due to cost and complexity considerations, the number of RF links is far less than the number of antenna elements, resulting in insufficient channel observation dimensions. For the channel estimation problem in large-scale MIMO LEO satellite communication systems, the existing technology mainly adopts a channel estimation method based on compressed sensing. This method attempts to use the sparse characteristics of the channel for channel reconstruction, and recovers high-dimensional channel information from low-dimensional observation data through compressed sensing theory. However, this method has the following technical limitations: the compressed sensing method has a priori dependency problem. This method requires the prior information such as the sparsity and sparse basis of the channel to be determined in advance, and is heavily dependent on the basic assumption that the channel has sparse characteristics. In actual MIMO LEO satellite communication scenarios, due to the high-speed movement of satellites and the complexity of the channel propagation environment, the actual channel characteristics are often significantly different from the preset sparse model. This mismatch between the prior information and the actual channel model will lead to a significant decrease in the accuracy of channel reconstruction, and in some cases it will even cause the algorithm to fail to converge, making it difficult to guarantee system performance.

[0023] The present application provides an efficient channel estimation method, which can overcome the problem of insufficient channel observation dimension under the condition of a limited number of RF links, and reduce the dependence on pilot signals through an efficient channel estimation algorithm to achieve accurate estimation of the complete channel state.

[0024] It should be noted that the execution subject of this embodiment may be a computing service device with data processing, network communication and program running functions, such as a computer, or an electronic device capable of realizing the above functions. The following takes a satellite communication system as an example to illustrate this embodiment and the following embodiments.

[0025] Based on this, the embodiment of the present application provides a channel estimation method, referring to Figure 1 , Figure 1 This is a flow chart of the first embodiment of the channel estimation method of the present application.

[0026] In this embodiment, the channel estimation method includes: Step S10, obtaining received signal information, and inputting the received signal information into a preset channel estimation model, wherein the preset channel estimation model includes a plurality of cascade layers, and the cascade layers include a gradient descent module and a proximal mapping module.

[0027] It should be noted that in the satellite communication system, the receiving end receives the signal from the satellite through the antenna array, which can be represented by the received signal matrix. The preset channel estimation model can be a trainable proximal gradient descent network (TPGD-Net). By expanding the iterative proximal gradient descent (PGD) algorithm into a layered network structure, it not only retains the theoretical basis of the PGD algorithm, but also introduces the adaptive ability of deep learning. Through this algorithm expansion method, TPGD-Net not only has interpretability and theoretical guarantee, but also can make full use of the powerful feature extraction and adaptive learning capabilities of deep learning.

[0028] It should be understood that the design of multiple cascade layers enables iterative optimization of the channel estimation model. Each layer further optimizes and adjusts the signal estimation value based on the previous layer. The gradient descent module of the cascade layer corresponds to the gradient descent step of the PGD algorithm, and the proximal mapping module corresponds to the proximal mapping step in the PGD algorithm.

[0029] Step S20: Select a current cascade layer from the multiple cascade layers, and obtain an upper cascade layer corresponding to the current cascade layer.

[0030] It should be noted that in the multi-layer cascade layer structure, the current cascade layer is the one that is currently being processed or focused on, and operations such as obtaining input and calculating output will be performed on it during the operation. The previous cascade layer is located before the current layer in the connection order of the cascade layer relative to the current cascade layer. Among them, the output of the previous cascade layer is usually used as the input of the current cascade layer.

[0031] Step S30, calculating a channel update value through the gradient descent module according to the channel estimation value calculated by the previous cascade layer and the received signal information.

[0032] It should be noted that the channel estimation value calculated by the previous cascade layer is the estimation result of the channel state calculated in the previous cascade layer during the iteration process of the preset channel estimation model. It is an approximate representation of the channel state based on the existing information (such as the received signal matrix, the pilot matrix and the parameters of the preset channel estimation model itself) in the previous round of iteration. It will be used as important input information in the calculation of the current cascade layer for further optimizing and updating the channel estimation results.

[0033] It should be understood that the channel update value is an intermediate value for optimizing the channel estimation value, which can reflect the basic principle of gradient descent, that is, updating the current channel estimation value along the negative direction of the objective function gradient.

[0034] In one example, a massive MIMO LEO channel It can be estimated by solving the following optimization problem:

[0035] In this optimization problem, is the data fidelity term that ensures the consistency between the estimated channel and the observed measurement value. This data fidelity term is crucial to ensure the reliability of the estimation result because it directly reflects the difference between the estimated channel and the actual received signal. Regularization function The prior information of the channel is introduced to normalize the solution of the equation. The choice of regularization function has an important impact on the final estimation result because it determines how to use the prior knowledge of the channel to improve the estimation performance. It is used to balance the data fidelity term and the regularization function, and the choice of parameters needs to weigh the degree of fit of the model to the observed data and the degree of dependence on prior information. The value will make the model more inclined to rely on prior information, while smaller A lower value will make the model more dependent on the observed data.

[0036] In this optimization model, represents the received signal matrix (which can be determined by the received signal information), The pilot matrix is ​​a known matrix designed in advance and known in advance. It is usually composed of a series of specific training sequences. It is usually used to insert pilot signals (the signal obtained by multiplying the pilot matrix and the transmitted information) into the transmitted signal and estimate the channel characteristics by using the known pilot matrix and the received signal containing the pilot signal. is the channel matrix to be estimated, Represents the product of the pilot matrix and the channel matrix.

[0037] An effective method to solve the above optimization problem is the proximal gradient descent algorithm. This algorithm generates the estimated channel by iteratively updating the following steps. In the gradient descent module of TPGD-Net, Layer Channel estimate of the layer is the input and output channel update value The formula is as follows:

[0038] in, It is The step size of gradient descent in the layer, Represents the product of the pilot matrix and the channel estimate.

[0039] Furthermore, in order to accelerate the convergence speed of the preset channel estimation model and improve the channel estimation efficiency, the step S20 may include: A step length training parameter is obtained, wherein the step length training parameter is used to learn an optimal step length strategy; based on a gradient descent function, the channel estimation value and the received signal information are used as input to calculate a channel update value.

[0040] It should be noted that in order to enhance the flexibility of the preset channel estimation model and speed up the convergence, the step size is set as a trainable parameter in each layer, allowing the preset channel estimation model to learn the optimal step size strategy during the training process, which may achieve faster convergence than the fixed step size method. By changing the traditional fixed step size to a trainable parameter, the optimization process can be adaptively adjusted to speed up the convergence. This design not only improves the efficiency of channel estimation, but also reduces the computational burden of the system, making this method more suitable for actual deployment.

[0041] Step S40: obtaining a current channel estimation value by calculation through a near-end mapping module according to the channel update value.

[0042] For ease of understanding, the following examples are provided, but are not intended to limit this solution. Figure 2 , Figure 2Schematic diagram of the trainable proximal gradient descent network architecture.

[0043] The traditional proximal gradient descent algorithm generates an estimated channel by iteratively updating the following steps:

[0044] in, It depends on the regularization function The proximal operator encapsulates the influence of the regularization function and enables the algorithm to effectively handle non-smooth regularization terms.

[0045] In existing millimeter wave channel estimation, most studies usually adopt The norm is used as a regularizer to promote the sparsity of beamspace channels. The norm, as a proxy for sparsity, has proven its effectiveness in many applications. However, using such a hand-designed regularization function to express the true prior information of the channel may be oversimplified, which may lead to degraded estimation performance. This is because The norm assumes that the channel coefficients are independent, while actual mmWave channels may have more complex structures and correlations.

[0046] To address this problem, this proposal unfolds the PGD algorithm into a deep neural network and equips it with a data-driven proximal operator. This approach allows the deep neural network to extract prior information from the channel data and perform implicit proximal operations. By learning the complex patterns and structures in the data, this approach can overcome the limitations of hand-designed regularization functions, thereby providing more accurate and robust channel estimates.

[0047] like Figure 2 As shown. TPGD-Net consists of Each layer is composed of a gradient descent module (GDM) and a proximal mapping module (PMM). GDM corresponds to the gradient descent step in the PGD algorithm. The output of layer GDM is As input, the output channel update value is as follows:

[0048] PMM corresponds to the proximal mapping step in the PGD algorithm, which aims to transform the input Estimated Channel The regularization function obtains the estimated channel by introducing prior information , and only appears in the proximal operator. Therefore, it is proposed to replace the proximal operator with a neural network instead of learning a regularization function and explicitly computing its proximal operator, thereby learning complex prior information directly from the data without pre-specifying the form of the regularization function.

[0049] The neural network in PMM is a U-shaped network with encoder-decoder blocks and skip connections. This U-shaped network architecture design can capture and utilize multi-scale feature information, thereby more effectively handling the complexity of the channel estimation problem. The neural network uses a channel attention block (CAB) to extract important features of the GDM output, and an encoder-decoder block for downsampling (Down) and upsampling (Up) to extract the encoder output of the previous layer. and decoder output , to obtain multi-scale feature maps. The channel attention mechanism used allows the neural network to adaptively emphasize the most relevant feature channels, thereby improving the efficiency and effectiveness of feature extraction. The residual block (RB) is used to extract features at each scale. At the same time, the use of residual connections can alleviate the gradient vanishing problem in deep network training and allow identity mapping to be learned, retaining input information. Downsampling and upsampling use maximum pooling and bilinear interpolation respectively, which can process information at different spatial resolutions, thereby capturing the multi-scale characteristics of the channel. The final output channel estimate And the optimized features . In addition, in order to solve the problem of information transfer in deep networks, the Cross-layer Feature Attention Fusion (CLFAF) module is introduced. This module allows information to flow freely between different depth levels by establishing direct connections between different network layers. Through the attention mechanism, the CLFAF module can intelligently identify and select the most relevant and important features, greatly improving the utilization efficiency of feature information. This design not only effectively alleviates the information bottleneck problem in deep networks, but also significantly enhances the adaptability to complex channel environments.

[0050] Therefore, the proposed neural network bypasses the explicit hand-designed regularization function, explores the inherent prior information of the mmWave channel, and implicitly performs the near-end operation. The advantage of this approach is that it can learn complex prior structures from the training data without being restricted by the form of a predefined regularization function.

[0051] Step S50: If the current channel estimation value does not meet the preset stop condition, return to the step of selecting the current cascade layer from multiple cascade layers until the current channel estimation value meets the condition, and use the current channel estimation value as the target channel estimation value.

[0052] It should be noted that the preset stop condition may be that the number of iterations of the channel estimation value reaches a maximum value, or the current number of iterations exceeds the number of preset cascade layers. The preset stop condition may also be that the accuracy of the channel estimation value reaches a certain threshold.

[0053] It can be understood that the channel estimation value is gradually optimized through multi-layer iteration. Each layer is improved on the basis of the previous layer, so that the channel estimation value is closer and closer to the true value. The number of iteration layers can also be reasonably controlled to prevent the preset channel estimation model from being over-iterated and causing overfitting problems, thereby reducing unnecessary computational overhead. When the current channel estimation value does not meet the preset stop condition, the calculated channel estimation value needs to be used as input and transmitted to the next layer of the channel estimation preset channel estimation model. When the current channel estimation value meets the preset stop condition, it indicates that the preset channel estimation model has learned and adapted to the characteristics of the channel to a certain extent. The last updated channel estimation value can be considered to be a relatively stable and reliable channel estimation result that can be obtained under the current network structure and training conditions.

[0054] In this embodiment, the received signal information is obtained and input into a preset channel estimation model, wherein the preset channel estimation model includes multiple cascade layers, and the cascade layers include a gradient descent module and a proximal mapping module; the current cascade layer is selected from the multiple cascade layers, and the previous cascade layer corresponding to the current cascade layer is obtained; the channel estimation value calculated by the previous cascade layer and the received signal information are used to calculate the channel update value through the gradient descent module; the current channel estimation value is calculated by the proximal mapping module according to the channel update value; if the current channel estimation value does not meet the preset stop condition, the step of selecting the current cascade layer from the multiple cascade layers is returned until the current channel estimation value meets the condition, and the current channel estimation value is used as the target channel estimation value. Combining the iterative optimization algorithm with deep learning enables the preset channel estimation model to automatically learn channel features from data instead of relying on a preset model that may be inaccurate, fundamentally solving the problem that the traditional compressed sensing method relies heavily on channel prior information, resulting in performance degradation in complex channel environments, and achieving accurate estimation of the complete channel state.

[0055] Reference Figure 3 , Figure 3This is a flow chart of a second embodiment of a channel estimation method of the present application. Based on the above first embodiment, a second embodiment of a channel estimation method of the present application is proposed. In the second embodiment, the step S40 includes: Step S401, obtaining the channel features generated by the previous cascade layer.

[0056] It should be noted that channel characteristics are relevant information that can reflect channel characteristics. Each cascade layer will optimize the channel characteristics generated by the previous layer for further analysis and estimation of channel status, etc., to help better restore the original signal of the transmitter or accurately grasp the true situation of the channel.

[0057] It should be understood that when the current cascade layer generates the current channel estimation value, the optimized channel features are also input into the next layer.

[0058] Step S402: fusing the channel features based on a spatial attention mechanism to obtain target fused features.

[0059] It should be understood that since the PGD algorithm produces estimated channels in each iteration, the TPGD-Net in the preset channel estimation model is designed to accept these channels as input and output estimated channels at each layer. This design allows TPGD-Net to utilize and improve previous estimation results at each step. However, the limited transmission of feature information between layers in TPGD-Net leads to a degradation in channel estimation performance. This is because each layer mainly relies on the output of the previous layer without fully utilizing the information of all previous layers. To address this limitation and taking into account the characteristics of the mmWave channel (characterized by sparsity and concentration), a spatial attention mechanism can be used to selectively fuse features from different layers, helping the preset channel estimation model to focus on the most relevant spatial regions and features.

[0060] In one example, a cross-layer feature attention fusion module (CLFAF) based on a spatial attention mechanism is proposed in TPGD-Net. The introduction of this module aims to enhance the information flow between the layers of the preset channel estimation model, thereby improving the overall estimation performance. The core idea of ​​the CLFAF module is to use the spatial attention mechanism to selectively fuse features from different layers. This approach is particularly suitable for dealing with the sparse and concentrated characteristics of millimeter wave channels because it can help the preset channel estimation model focus on the most relevant spatial regions and features. Specifically, the CLFAF module may perform the following operations: collect feature maps from multiple previous layers; use the spatial attention mechanism to assign weights to different spatial positions of each feature map based on these weights; fuse features from different layers; combine the fused features with the features of the current layer to produce a richer and more informative representation.

[0061] Furthermore, in order to enable the preset channel estimation model to focus on the key feature area, enhance the characterization capability of the channel features, and improve the feature fusion quality, the step S402 may include: The channel features are connected along the channel dimension to obtain an initial feature map; the initial feature map is processed based on a preset convolutional layer to generate a corresponding number of feature matrices, wherein the feature representations represented by the feature matrices are different; weights are assigned to different spatial positions of the feature matrices, and the feature matrices are fused based on the weights to obtain an initial fused feature; the size of the initial fused feature is adjusted, and the adjustment result is added element by element to the encoding feature output by the current layer of the encoder residual block in the channel feature to obtain a target fused feature.

[0062] It should be noted that the channel dimension is the number of channels in the feature map represented by the convolutional neural network, and each channel can be regarded as a specific feature representation of the input data. The feature matrix refers to the output result after the convolution layer processing, which represents the new feature information extracted or converted from the initial feature map in the form of a matrix. Different feature matrices carry different aspects of feature representation, that is, each feature matrix reflects the characteristics of different dimensions and angles of the channel features contained in the initial feature map. For example, some feature matrices may focus on the spatial distribution characteristics of the channel, while others may focus on frequency-related characteristics.

[0063] It should be understood that the number of feature matrices generated by connecting the channel features along the channel dimension can be determined according to the type of features or by the number of convolution kernels.

[0064] It can be understood that the spatial attention mechanism is used to assign weights to different spatial positions of each feature map, and based on these weights, features from different layers are fused, and the fused features are combined with the features of the current layer. In this way, the preset channel estimation model can more effectively utilize the information accumulated throughout the estimation process, allowing the preset channel estimation model to dynamically determine the importance of spatial regions and features of previous layers to the current estimation, thereby potentially improving the accuracy and robustness of the estimation.

[0065] Furthermore, in order to allow the preset channel estimation model to more effectively utilize and propagate information in the entire estimation process, the step of assigning weights to different spatial positions of the feature matrix and fusing the feature matrix based on the weights to obtain the initial fused features may include: The transpose of the first matrix is ​​multiplied by the second matrix to obtain a product matrix; the product matrix is ​​calculated based on a normalized exponential function to generate a spatial attention map, wherein each element in the spatial attention map represents the attention weight of the spatial position corresponding to the feature matrix; the third matrix is ​​weightedly aggregated according to the spatial attention map to obtain an initial fusion feature.

[0066] For ease of understanding, the following examples are provided, but are not intended to limit this solution. Figure 4 , Figure 4 The CLFAF module in TPGD-Net first concatenates the feature maps along the channel dimension. and , and obtain a size of The initial feature map of the CLFAF module. C, H, and W are the number of channels, height, and width of the feature map, respectively. This feature connection operation is the first step of the CLFAF module, allowing the CLFAF module to consider both the features from the current layer and the features from the previous layer. , the encoder output of the previous layer and decoder output In this way, the CLFAF module can comprehensively utilize information from different levels and different processing stages to provide rich and diverse inputs for subsequent feature fusion, while reflecting an important insight: the accuracy of channel estimation can be improved by integrating information from multiple processing stages.

[0067] Input the initial feature map obtained above into a Convolutional layer ( conv), the size is The intermediate characteristics . The use of convolution has several important functions: first, it allows cross-channel information fusion and realizes information interaction between different feature maps; second, it can adjust the number of channels of the feature map, while keeping the spatial dimension unchanged, controlling the computational complexity; finally, it introduces nonlinearity and enhances the expressiveness of the model. This step can be seen as a preliminary fusion and compression of the connected features, preparing a more compact and information-rich feature representation for the subsequent attention mechanism.

[0068] Intermediate feature map Further through three Convolutional layer ( conv) to obtain a size of Feature map and , and then reshape them to a size of The design draws on the idea of ​​self-attention mechanism, especially the concepts of query, key and value in Transformer. By generating three different feature representations, the module prepares the necessary components for subsequent attention calculations. This design allows TPGD-Net to learn complex feature interaction patterns, rather than just simple linear combinations.

[0069] Then the matrix The transpose of (transpose) and the matrix Multiply, and pass the multiplied matrix through the normalized exponential (softmax) function to obtain the spatial attention map :

[0070] in, Each element in represents the matrix and matrix The attention weight corresponding to the spatial position is calculated by and The similarity between them implements the spatial attention mechanism to dynamically determine which spatial positions are important. The application of the softmax function ensures the non-negativity and sum of the attention weights to 1, making the attention allocation more explicit and explainable. For millimeter wave channels, the energy of the channel is often concentrated in a specific spatial direction or region. The spatial attention mechanism effectively captures the sparse and clustered characteristics of the channel.

[0071] Then the matrix The transpose and matrix Multiply, the resulting matrix (initial fusion feature) is passed through the learnable parameters After scaling, it is reshaped into This step can be understood as weighted aggregation of matrix features according to attention weights. Learnable parameters Provides additional flexibility, allowing the preset channel estimation model to adjust the strength of the attention mechanism. This reshaped matrix is ​​similar to the feature map Add element by element to get the final feature fusion output :

[0072] The final fusion operation combines the attention-weighted features with the original features to form a residual structure. The use of residual connections helps alleviate the gradient vanishing problem in deep networks and allows the preset channel estimation model to retain the original feature information when necessary. The resulting features It is the selective fusion of features of the current layer and the previous layer. As the fusion feature of this layer, it participates in the feature optimization process of the current layer.

[0073] Therefore, the CLFAF module significantly improves the cross-layer feature information transmission by introducing cross-layer feature fusion and attention mechanism, thereby improving the channel estimation performance. This design not only allows the preset channel estimation model to more effectively utilize and propagate information throughout the estimation process, but also enables the preset channel estimation model to adaptively focus on the most relevant features and spatial locations. This capability is invaluable for complex tasks such as mmWave channel estimation, as it allows the preset channel estimation model to handle the time-varying and spatial selectivity of the channel.

[0074] Step S403: Calculate and obtain a current channel estimation value according to the target fusion feature and the channel update value.

[0075] In one example, the channel estimation value output by the near-end mapping module can be expressed as:

[0076] in Indicates Parameters of the layer PMM. is the proximal operator. and is in the encoder residual block and the decoder residual block, from the Layer Features extracted from different scales. This expression reflects how the proximal mapping module integrates information from different scales and layers to generate the final channel estimate, showing how TPGD-Net transforms the proximal mapping step in traditional optimization algorithms into a learnable neural network operation. The advantage of this approach is that it can adaptively learn complex prior information without explicitly defining a regularization function.

[0077] In this embodiment, the channel features generated by the previous cascade layer are obtained; the channel features are fused based on the spatial attention mechanism to obtain the target fused features; and the current channel estimation value is calculated based on the target fused features and the channel update value. The spatial attention mechanism is used to selectively fuse features of different layers, helping the preset channel estimation model to focus on the most relevant spatial regions and features, thereby improving the channel estimation performance.

[0078] Reference Figure 5 , Figure 5 This is a flow chart of a third embodiment of a channel estimation method of the present application. Based on the above second embodiment, a third embodiment of a channel estimation method of the present application is proposed.

[0079] In the third embodiment, before step S10, the method further includes: Step S01, initializing trainable parameters in the preset channel estimation model, wherein the trainable parameters include step size training parameters and parameters in the proximal mapping module.

[0080] It should be noted that trainable parameters are parameters in the channel estimation model that can be automatically adjusted during the training process based on the input data and the expected output results (such as accurate signal estimation values). By continuously optimizing these parameters, the model can better fit the data and improve the accuracy and performance of signal estimation.

[0081] It is understood that the trainable parameters may include step size training parameters in the gradient descent module and optimizations in the proximal mapping module, such as parameters for scaling the matrix in the cross-layer feature attention fusion module.

[0082] Step S02, optimizing the trainable parameters based on a preset training set and a loss function, and updating the optimized results into the preset channel estimation model.

[0083] It is understandable that the training process of optimizing the trainable parameters through the loss function and the preset training set can be regarded as a reinterpretation of the traditional iterative optimization algorithm under the deep learning framework. The interpretability and theoretical basis of the traditional method are retained, while the powerful expression ability and adaptability of deep learning are utilized to not only improve the performance of the algorithm, but also enhance its robustness and adaptability under various complex channel conditions.

[0084] In one example, TPGD-Net can optimize its network parameters by using the stochastic gradient descent algorithm. To train, to minimize the loss function :

[0085] in, represents the step size training parameter, Indicates The parameters of the layer PMM, this loss function measures the estimated channel With the true channel The Euclidean distance between all By minimizing this loss function, the preset channel estimation model can learn how to produce increasingly accurate channel estimates at each layer.

[0086] In this embodiment, the trainable parameters in the preset channel estimation model are initialized, wherein the trainable parameters include step size training parameters and parameters in the proximal mapping module; the trainable parameters are optimized based on a preset training set and a loss function, and the optimized results are updated to the preset channel estimation model. By changing the traditional fixed step size to a trainable parameter, the preset channel estimation model can adaptively adjust the optimization process, speed up the convergence speed, and reduce the computational burden of the system.

[0087] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the channel estimation method of the present application. More simple transformations based on this technical concept are all within the protection scope of the present application.

[0088] The present application also provides a channel estimation device, please refer to Figure 6 , the channel estimation device comprises: An information acquisition module 10, used to acquire received signal information and input the received signal information into a preset channel estimation model, wherein the preset channel estimation model includes a plurality of cascade layers, and the cascade layers include a gradient descent module and a proximal mapping module; A layer determination module 20, configured to select a current cascade layer from a plurality of cascade layers, and obtain a previous cascade layer corresponding to the current cascade layer; A channel updating module 30, configured to obtain a channel updating value by calculating the channel estimation value obtained by the previous cascade layer and the received signal information through the gradient descent module; A channel estimation module 40, configured to obtain a current channel estimation value by calculating the channel update value through a near-end mapping module; The estimation determination module 50 is used to return to the step of selecting the current cascade layer from multiple cascade layers if the current channel estimation value does not meet the preset stop condition, until the current channel estimation value meets the condition, and use the current channel estimation value as the target channel estimation value.

[0089] The channel estimation device provided by the present application adopts the channel estimation method in the above embodiment, which can solve the technical problem that the existing channel estimation method based on compressed sensing cannot accurately estimate the complete channel state due to the high-speed movement of the satellite and the complexity of the channel propagation environment, resulting in a significant decrease in the accuracy of channel reconstruction. Compared with the prior art, the beneficial effects of the channel estimation device provided by the present application are the same as the beneficial effects of the channel estimation method provided by the above embodiment, and the other technical features in the channel estimation device are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0090] The present application provides a channel estimation device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the channel estimation method in the above-mentioned embodiment 1.

[0091] Reference below Figure 7 , which shows a schematic diagram of the structure of a channel estimation device suitable for implementing an embodiment of the present application. The channel estimation device in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 7 The channel estimation device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0092] like Figure 7 As shown, the channel estimation device may include a processing device 1001 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of the channel estimation device are also stored. The processing device 1001, ROM1002, and RAM1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems can be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD: Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the channel estimation device to communicate with other devices wirelessly or by wire to exchange data. Although the figure shows a channel estimation device with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented or have alternatively.

[0093] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0094] The channel estimation device provided by the present application adopts the channel estimation method in the above embodiment, which can solve the technical problem that the existing channel estimation method based on compressed sensing cannot accurately estimate the complete channel state due to the high-speed movement of the satellite and the complexity of the channel propagation environment, resulting in a significant decrease in the accuracy of channel reconstruction. Compared with the prior art, the beneficial effects of the channel estimation device provided by the present application are the same as the beneficial effects of the channel estimation method provided by the above embodiment, and other technical features in the channel estimation device are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0095] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0096] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0097] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, and the computer-readable program instructions are used to execute the channel estimation method in the above-mentioned embodiment.

[0098] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM: Random Access Memory), a read-only memory (ROM: Read Only Memory), an erasable programmable read-only memory (EPROM: Erasable Programmable Read Only Memory or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM: CD-Read Only Memory), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency: Radio Frequency), etc., or any suitable combination of the above.

[0099] The computer-readable storage medium may be included in the channel estimation device; or may exist independently without being assembled into the channel estimation device.

[0100] The computer-readable storage medium carries one or more programs. When the one or more programs are executed by the channel estimation device, the channel estimation device executes the channel estimation method described above.

[0101] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0102] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0103] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0104] The readable storage medium provided by the present application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned channel estimation method, and can solve the technical problem that the existing channel estimation method based on compressed sensing has a significant decrease in channel reconstruction accuracy due to the high-speed movement of satellites and the complexity of the channel propagation environment, and cannot achieve accurate estimation of the complete channel state. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by the present application are the same as the beneficial effects of the channel estimation method provided by the above-mentioned embodiment, which will not be repeated here.

[0105] The present application also provides a computer program product, including a computer program, which implements the steps of the channel estimation method as described above when the computer program is executed by a processor.

[0106] The computer program product provided by the present application can solve the technical problem that the existing channel estimation method based on compressed sensing cannot accurately estimate the complete channel state due to the high-speed movement of the satellite and the complexity of the channel propagation environment, resulting in a significant decrease in the accuracy of channel reconstruction. Compared with the prior art, the beneficial effects of the computer program product provided by the present application are the same as the beneficial effects of the channel estimation method provided by the above embodiment, which will not be described in detail here.

[0107] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A channel estimation method, characterized in that: The channel estimation method comprises: Acquire received signal information, and input the received signal information into a preset channel estimation model, wherein the preset channel estimation model includes a plurality of cascade layers, and the cascade layers include a gradient descent module and a proximal mapping module; Select a current cascade layer from the multiple cascade layers, and obtain a previous cascade layer corresponding to the current cascade layer; A channel update value is calculated by the gradient descent module according to the channel estimation value calculated by the previous cascade layer and the received signal information; Obtaining a current channel estimation value by calculating through a near-end mapping module according to the channel update value; If the current channel estimation value does not meet the preset stop condition, the process returns to the step of selecting the current cascade layer from the multiple cascade layers until the current channel estimation value meets the condition, and the current channel estimation value is used as the target channel estimation value.

2. The channel estimation method according to claim 1, characterized in that: The step of calculating and obtaining a current channel estimation value through a proximal mapping module according to the channel update value comprises: Obtaining the channel features generated by the previous cascade layer; The channel features are fused based on a spatial attention mechanism to obtain a target fused feature; A current channel estimation value is obtained by calculation according to the target fusion feature and the channel update value.

3. The channel estimation method according to claim 2, characterized in that: The step of fusing the channel features based on the spatial attention mechanism to obtain target fusion features includes: Connecting the channel features along the channel dimension to obtain an initial feature map; Processing the initial feature map based on a preset convolutional layer to generate a corresponding number of feature matrices, wherein the feature matrices represent different features; Assigning weights to different spatial positions of the feature matrix, and fusing the feature matrix based on the weights to obtain initial fused features; The size of the initial fusion feature is adjusted, and the adjustment result is added element by element to the encoding feature output by the current layer of the encoder residual block in the channel feature to obtain the target fusion feature.

4. The channel estimation method according to claim 3, characterized in that: The characteristic matrix includes a first matrix, a second matrix and a third matrix; The step of assigning weights to different spatial positions of the feature matrix and fusing the feature matrix based on the weights to obtain initial fused features includes: Multiplying the transpose of the first matrix by the second matrix to obtain a product matrix; Calculating the product matrix based on a normalized exponential function to generate a spatial attention map, wherein each element in the spatial attention map represents an attention weight of a spatial position corresponding to the feature matrix; The third matrix is ​​weightedly aggregated according to the spatial attention map to obtain an initial fusion feature.

5. The channel estimation method according to claim 1, wherein: The step of calculating the channel update value by the gradient descent module according to the channel estimation value calculated by the previous cascade layer and the received signal information comprises: Obtaining a step length training parameter, wherein the step length training parameter is used to learn an optimal step length strategy; The channel estimation value and the received signal information are taken as input based on a gradient descent function to calculate a channel update value.

6. The channel estimation method according to any one of claims 1 to 5, characterized in that: Before the step of obtaining the received signal information and inputting the received signal information into the preset channel estimation model, the method further includes: Initializing trainable parameters in the preset channel estimation model, wherein the trainable parameters include step size training parameters and parameters in the proximal mapping module; The trainable parameters are optimized based on a preset training set and a loss function, and the optimized results are updated to the preset channel estimation model.

7. A channel estimation device, characterized in that: The device comprises: An information acquisition module, used to acquire received signal information and input the received signal information into a preset channel estimation model, wherein the preset channel estimation model includes a plurality of cascade layers, and the cascade layers include a gradient descent module and a proximal mapping module; A layer determination module, used to select a current cascade layer from multiple cascade layers, and obtain a previous cascade layer corresponding to the current cascade layer; A channel update module, configured to calculate a channel update value through the gradient descent module according to the channel estimation value calculated by the previous cascade layer and the received signal information; A channel estimation module, configured to obtain a current channel estimation value by calculating the channel update value through a near-end mapping module; The estimation and determination module is used to return to the step of selecting the current cascade layer from multiple cascade layers if the current channel estimation value does not meet the preset stop condition, until the current channel estimation value meets the condition, and use the current channel estimation value as the target channel estimation value.

8. A channel estimation device, characterized in that: The device comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the channel estimation method according to any one of claims 1 to 6.

9. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the channel estimation method according to any one of claims 1 to 6 are implemented.

10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the steps of the channel estimation method according to any one of claims 1 to 6 are implemented.

Images