An OTFS Channel Estimation Method, System, Device and Medium

Abstract

The present invention discloses an OTFS channel estimation method, system, device and medium, relating to the field of communication technologies. By introducing a regression network model, after obtaining the channel matrix originally, it is necessary to input the channel matrix into the regression network model for denoising processing in the model manner, so as to improve the anti-interference ability of the channel matrix for the transmitted signal of the first signal during the reconstruction process, thereby constructing a fitting relationship between the original time-delay Doppler domain channel estimation matrix and the true channel matrix and improving the estimation accuracy. By utilizing the powerful learning and fitting ability of the neural network, the initially estimated result input is further refined and optimized, and finally a more accurate channel estimation result is output, which can significantly improve the performance of the entire communication system.

Description

Technical Field

[0001] The present invention relates to the field of communication technologies, and in particular, to an OTFS channel estimation method, system, device, and medium. Background Art

[0002] During the communication process, the transmitter and the receiver communicate through a channel. The Doppler frequency shift in the Orthogonal Time Frequency Space (OTFS) modulation and demodulation technology is sampled in integer form, resulting in a reduced accurate representation of the actual fractional Doppler frequency shift during the sampling process, thereby causing Inter-Doppler Interference (IDI). As a result, the signal amplitude will be changed at the receiver, leading to an incorrect evaluation of the channel parameters by the system and reducing the communication quality.

[0003] Therefore, how to reduce channel interference to improve communication quality is an urgent problem to be solved by those skilled in the art. Summary of the Invention

[0004] The purpose of the present invention is to provide an OTFS channel estimation method, system, device, and medium to solve the problems of reduced accurate representation of conventional fractional Doppler frequency shift and incorrect evaluation of channel parameters by the system, which reduces the communication quality.

[0005] To solve the above technical problems, the present invention provides an OTFS channel estimation method, which is applied to the receiver and includes:

[0006] Receiving a first signal sent by a transmitter; and performing crosstalk analysis on the first signal to obtain a received signal in the time-delay Doppler domain;

[0007] Determining a target path according to the amplitude of the received signal in the time-delay Doppler domain and a threshold value; and constructing a channel matrix according to the path channel state information of the target path;

[0008] Inputting the channel matrix into a regression network model to obtain a denoised channel matrix, and reconstructing the transmitted signal corresponding to the first signal.

[0009] On the one hand, the training process of the regression network model includes:

[0010] Obtaining an initial regression network model, training data in a multipath channel scenario, and an actual channel matrix;

[0011] Determining a first target path according to the amplitude corresponding to the training data in the multipath channel scenario and the threshold value;

[0012] Constructing a first channel matrix according to the path channel state information of the first target path;

[0013] Use the first channel matrix as the input signal of the initial regression network model to obtain the first actual channel matrix of the initial regression network model;

[0014] Record the current iteration number;

[0015] When the difference between the first actual channel matrix and the actual channel matrix does not meet the preset condition, or when the current iteration number does not reach the preset iteration number, fine-tune the model parameters of the initial regression network model to obtain the fine-tuned initial regression network model; and return to the step of using the first channel matrix as the input signal of the initial regression network model until the difference between the first actual channel matrix and the actual channel matrix meets the preset condition, or the current iteration number reaches the preset iteration number.

[0016] On the other hand, input the channel matrix into the regression network model to obtain a denoised signal matrix, including:

[0017] Extract the features of the data of the channel matrix to obtain first feature data;

[0018] Perform feature learning residual processing on the first feature data to obtain second feature data; wherein, the feature learning residual processing of the regression network model is completed in a feature learning residual module, and the feature learning residual module includes multiple enhanced attention modules, and there is a cascading relationship between the multiple enhanced attention modules, and the enhanced attention module includes multiple convolutional kernels and a merging convolutional layer;

[0019] Perform reconstruction processing on the second feature data to obtain a denoised signal matrix.

[0020] On the other hand, the sizes of the multiple convolutional kernels in the enhanced attention module are different.

[0021] On the other hand, the difference between the first actual channel matrix and the actual channel matrix not meeting the preset condition includes:

[0022] Obtain the channel capacity and the rank corresponding to the eigenvalue of the first actual channel matrix and the actual channel matrix respectively;

[0023] If the difference between the respective channel capacities is greater than or equal to the first threshold, then compare the ranks corresponding to the respective eigenvalues;

[0024] If the rank corresponding to the respective eigenvalue is greater than or equal to the second threshold, it is determined that the difference between the first actual channel matrix and the actual channel matrix does not meet the preset condition.

[0025] On the other hand, the difference between the first actual channel matrix and the actual channel matrix does not meet the preset condition, including:

[0026] Obtain the respective correlations and respective estimation errors corresponding to the first actual channel matrix and the actual channel matrix;

[0027] If the difference between the respective correlations is not within the first preset range, then compare the respective estimation errors;

[0028] If the difference between the respective estimation errors is not within the second preset range, then determine that the difference between the first actual channel matrix and the actual channel matrix does not meet the preset condition.

[0029] On the other hand, the crosstalk analysis of the first signal to obtain the received signal in the time-delay Doppler domain includes:

[0030] Perform SFFT transformation on the first signal to obtain the first time-delay Doppler domain signal;

[0031] Obtain the first target row and the second target row of the first time-delay Doppler domain signal;

[0032] Intercept the signal from the first target row to the second target row of the first time-delay Doppler domain signal according to the first target row and the second target row as the received signal in the time-delay Doppler domain.

[0033] To solve the above technical problems, the present invention also provides an OTFS channel estimation system, including a receiving end and a transmitting end:

[0034] Control the transmitting end to transmit the first signal;

[0035] Control the receiving end to perform crosstalk analysis on the first signal to obtain the received signal in the time-delay Doppler domain; determine the target path according to the amplitude of the received signal in the time-delay Doppler domain and the threshold value; construct a channel matrix according to the path channel state information of the target path; perform denoising processing on the channel matrix to obtain the denoised channel matrix, so as to perform reconstruction processing on the transmitted signal corresponding to the first signal.

[0036] To solve the above technical problems, the present invention also provides an OTFS channel estimation device, including:

[0037] A memory for storing a computer program;

[0038] A processor for implementing the steps of the OTFS channel estimation method when executing the computer program.

[0039] To solve the above technical problems, the present invention further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the OTFS channel estimation method as described above are implemented.

[0040] An OTFS channel estimation method provided by the present invention is applied to a receiving end, which receives a first signal sent by a sending end; performs crosstalk analysis on the first signal to obtain a received signal in the time-delay Doppler domain; determines a target path according to the amplitude of the received signal in the time-delay Doppler domain and a threshold value; constructs a channel matrix according to the path channel state information of the target path; inputs the channel matrix into a regression network model to obtain a denoised channel matrix, and performs reconstruction processing on the sending signal corresponding to the first signal. The present invention introduces a regression network model. After obtaining the channel matrix originally, the channel matrix needs to be input into the regression network model for denoising processing in the model manner, so as to improve the anti-interference ability of the channel matrix for the sending signal of the first signal during the reconstruction processing, and to construct a fitting relationship between the original time-delay Doppler domain channel estimation matrix and the real channel matrix, thereby improving the estimation accuracy. By using the powerful learning and fitting ability of the neural network, the preliminary estimation result input is further refined and optimized, and finally a more accurate channel estimation result is output, which can significantly improve the performance of the entire communication system.

[0041] In addition, the present invention further provides an OTFS channel estimation system, device, and medium, which have the same beneficial effects as the above OTFS channel estimation method. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] To more clearly illustrate the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings without creative efforts.

[0043] Figure 1 A schematic diagram of the transmitted and received signals of an integer Doppler system provided by an embodiment of the present invention;

[0044] Figure 2 A schematic diagram of the transmitted and received signals of a fractional Doppler system provided by an embodiment of the present invention;

[0045] Figure 3 A flowchart of an OTFS channel estimation method provided by an embodiment of the present invention;

[0046] Figure 4 A schematic diagram of the change of OTFS data frames between a sending end and a receiving end provided by an embodiment of the present invention;

[0047] Figure 5 Schematic diagram of a regression network model provided by an embodiment of the present invention;

[0048] Figure 6 Schematic diagram of the principle of a regression network model provided by an embodiment of the present invention;

[0049] Figure 7 Schematic diagram of the operation process of a pooling layer provided by an embodiment of the present invention;

[0050] Figure 8 Another schematic diagram of the operation process of a pooling layer provided by an embodiment of the present invention;

[0051] Figure 9 Schematic diagram of the normalized mean square error curve of RIDChanNet under a five-path channel provided by an embodiment of the present invention;

[0052] Figure 10 Schematic diagram of the normalized mean square error curve of RIDChanNet under an eight-path channel provided by an embodiment of the present invention;

[0053] Figure 11 Schematic diagram of the normalized mean square error curve of RIDChanNet under a ten-path channel provided by an embodiment of the present invention;

[0054] Figure 12 Schematic diagram of the bit error rate curve of RIDChanNet under a five-path channel provided by an embodiment of the present invention;

[0055] Figure 13 Schematic diagram of the bit error rate curve of RIDChanNet under an eight-path channel provided by an embodiment of the present invention;

[0056] Figure 14 Schematic diagram of the bit error rate curve of RIDChanNet under a ten-path channel provided by an embodiment of the present invention;

[0057] Figure 15 Structure diagram of an OTFS channel estimation device provided by an embodiment of the present invention;

[0058] Figure 16 Structure diagram of an OTFS channel estimation apparatus provided by an embodiment of the present invention. Detailed implementation manners

[0059] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying 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. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the protection scope of the present invention.

[0060] The core of the present invention is to provide an OTFS channel estimation method, system, device, and medium to solve the problems of the reduction in the accurate representation of conventional fractional Doppler frequency shift and the incorrect evaluation of channel parameters by the system, which reduces the communication quality.

[0061] To enable those skilled in the art to better understand the solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0062] In an OTFS system, the Doppler frequency shift is used to describe the signal frequency change caused by the relative motion between the transmitter and the receiver. According to the numerical characteristics of the Doppler frequency shift, the OTFS system can be divided into an integer Doppler system and a fractional Doppler system. The ratio of the Doppler frequency shift of each path to the sampling resolution is defined as: ; where is the Doppler frequency shift of the th path, is the sampling resolution; if the corresponding to P paths are all integers, then the system is an integer Doppler system; otherwise, it is a fractional Doppler system. In the integer Doppler system, Figure 1 is a schematic diagram of the transmitted and received signals of an integer Doppler system provided by an embodiment of the present invention. As Figure 1 shown, the sampling point mapping follows the law of one-to-one mapping. The representation of the signal in the time-delay Doppler domain is relatively simple, and the complexity of channel estimation and equalization technologies is relatively low. Commonly used estimation methods include a threshold-based channel estimation algorithm and a compressive sensing-based channel estimation algorithm. In addition, the integer Doppler system can usually better utilize the sparsity characteristics of the OTFS technology to improve the transmission efficiency and performance of the system. In the fractional system, Figure 2 is a schematic diagram of the transmitted and received signals of a fractional Doppler system provided by an embodiment of the present invention. As Figure 2 shown, the Doppler frequency shift is sampled in the form of an integer in the time-delay Doppler domain, which results in the sampling unit being unable to accurately represent the actual fractional Doppler frequency shift, thereby causing IDI interference. IDI interference refers to the crosstalk received by the receiving unit from the sampling units corresponding to different Doppler parameters at the same time delay within the OTFS data frame. This interference will change the signal amplitude of the receiving unit, resulting in an incorrect estimation of the channel parameters by the system and reducing the communication quality. The OTFS channel estimation method provided by the present invention can solve the above technical problems.

[0063] Figure 3 is a flowchart of an OTFS channel estimation method provided by an embodiment of the present invention. As Figure 3 shown, this method is applied to the receiving end and includes:

[0064] S11: Receive the first signal sent by the sending end; and perform crosstalk analysis on the first signal to obtain the received signal in the time-delay Doppler domain;

[0065] S12: Determine the target path according to the amplitude of the received signal in the time-delay Doppler domain and the threshold value; and construct a channel matrix according to the path channel state information of the target path;

[0066] S13: Input the channel matrix into the regression network model to obtain the denoised channel matrix, and reconstruct the transmitted signal corresponding to the first signal.

[0067] Figure 4 The following is a schematic diagram of the change of the OTFS data frame between the sending end and the receiving end provided by the embodiment of the present invention. As Figure 4 shown, during the OTFS signal demodulation process at the sending end, the complete time-delay Doppler domain signal is subjected to the Inverse Symplectic Finite Fourier Transform (ISFFT) to obtain the time-frequency (TF) domain signal , where the definition of ISFFT is:

[0068] ;

[0069] Among them, , are the number of sampling points in the Doppler domain and the time-delay domain respectively, , are the indexes in the Doppler domain and the time-delay domain respectively; represents the index in the frequency domain dimension, is the index in the time dimension.

[0070] Perform the Heisenberg transform on to obtain the parallel time-domain signal, where is the transmitted waveform of the signal:

[0071] ;

[0072] Among them, is the time variable, is the sampling period of the OTFS system, is the subcarrier spacing of the system.

[0073] After goes through serial-to-parallel conversion, it is sent as the transmitted signal into the channel. The signal is superimposed with Gaussian white noise during the transmission process. Therefore, the received signal at the receiving end can be expressed as:

[0074] ;

[0075] wherein, is the Gaussian white noise superimposed during the transmission process. Assume that the channel has paths, and the channel fading gain, time delay, and Doppler shift corresponding to each path are , , ; is the index of the path.

[0076] The Delay-Doppler Domain (DD) channel in the ideal state can be expressed in the following form:

[0077] ;

[0078] wherein, is the channel fading gain of the th path, represents the Dirac delta function, which is an impulse function.

[0079] It should be noted that at the receiving end, the signal is demodulated to convert the serial received signal into a parallel time-domain signal , and then through sampling and the Wigner transform, the received symbol in the time-frequency domain, that is, the first signal, is obtained:

[0080] ;

[0081] wherein, is the time-frequency domain symbol of the waveform of the received pulse, with in the upper right corner indicating the conjugate complex number, is the receiving end; represents a variable in the convolution process and has no practical meaning.

[0082] In some embodiments, the crosstalk analysis is performed on the first signal to obtain the received signal in the delay-Doppler domain, including:

[0083] Performing an SFFT transform on the first signal to obtain a first delay-Doppler domain signal;

[0084] Obtaining the first target row and the second target row of the first delay-Doppler domain signal;

[0085] Intercepting the signal from the first target row to the second target row of the first delay-Doppler domain signal according to the first target row and the second target row as the received signal in the delay-Doppler domain.

[0086] Specifically, the time-frequency domain symbols are transformed to the delay-Doppler domain through the Sparse Fast Fourier Transform (SFFT) to obtain the received symbols in the DD domain:

[0087] .

[0088] Specifically, the first signal at the transmitting end maps the bit stream to data symbols through Quadrature Phase Shift Keying (QPSK), and arranges them in Figure 4 the data region of the DD domain signal . The remaining positions are called guard bands, which are distributed at the bottom of the data frame and are used to isolate the interference of data to pilot symbols and channel estimation. The transmitted symbols in this region take values as:

[0089] ;

[0090] where represents the pilot symbol, and are the coordinate parameters corresponding to the pilot point positions. It should be noted that guard intervals need to be set on both sides of the pilot symbol in the delay domain. The guard interval between the pilot and the data can prevent the pilot from aliasing with the data symbols affected by the delay; the guard interval on the other side can show the delay situation of the channel, and thus is used for channel estimation.

[0091] In the channel estimation process based on the threshold, the receiving end intercepts from row (the first target row) to row (the second target row) of symbols for channel estimation. The received symbols (delay-Doppler domain received signals) in this region can be expressed as:

[0092] ;

[0093] where indicates that there is a transmission path with a delay coefficient and a Doppler coefficient , , respectively represent the Doppler domain and delay domain coordinate information; , are the indices of the Doppler domain and delay domain respectively; is the additive white Gaussian noise in the DD domain. If is 0, it indicates that this path does not exist, and the received data only contains white noise.

[0094] The determination process of the received signal in the time-delay Doppler domain provided by this embodiment is obtained based on the data in the guard band to prevent the pilot from overlapping with the data symbols affected by time delay; the guard interval on the other side can show the time-delay situation of the channel and is thus used for channel estimation.

[0095] The time-delay Doppler domain equivalent channel matrix can be expressed as:

[0096] ;

[0097] Wherein, represents the initial phase of the th path, , represent the process variables in the signal convolution operation and have no practical significance; , are the indexes in the Doppler domain and the time-delay domain respectively; represents the equivalent channel Doppler response; the phase offset term and the equivalent channel Doppler response are given by the following formulas respectively, is the length of the cyclic prefix.

[0098] ;

[0099] ;

[0100] Wherein, the Doppler domain coordinates corresponding to the th path are , and represent the integer part and the fractional part respectively.

[0101] In the integer Doppler system, the Doppler domain coordinates corresponding to all paths can be expressed as , at this time ; while in the fractional Doppler system, , at this time regardless of taking any value, . Therefore, when the system changes from the integer Doppler system to the fractional Doppler system, many previous zeros turn into non-zeros due to the appearance, forming a new fractional Doppler equivalent channel matrix . consists of two parts, one is the original matrix under the integer system, and the other is the interference term . Therefore, the received signal in the time-delay Doppler domain under the fractional Doppler system can be re-expressed as:

[0102] ;

[0103] Wherein, For the accurate value of the received symbol, it is Doppler crosstalk. The occurrence of Doppler crosstalk is due to the pilot energy dissipating into each Doppler unit under a certain delay interval, changing the amplitude of the received symbol. Therefore, the occurrence of Doppler crosstalk will have an adverse impact on subsequent channel estimation and signal detection.

[0104] To distinguish the sampling points containing only noise from the sampling points corresponding to the paths, the data in the guard band is screened with a threshold, and generally the threshold is taken, where is the power spectral density of the noise. If the amplitude of the received signal , it is considered that there is a path corresponding to the Doppler-delay parameter of the sampling point (i.e., the target path), then is 1; otherwise, it is considered that there is no corresponding path, is 0, and at this time only contains the noise amplitude. According to the number, position, and amplitude of the screened sampling points, the number of multipaths and the delays , Doppler frequency shifts and channel fading gains of such path channel state information can be obtained, and then a complete DD-domain channel matrix can be constructed.

[0105] In addition to the Gaussian white noise component, Doppler crosstalk is also introduced into the received signal. This interference term greatly increases the amplitude of the received signal, making it difficult for the receiving end to identify the true transmission path through the threshold, and affecting the estimation accuracy of the time-delay Doppler domain channel.

[0106] The conventional technical means is to use the channel matrix in step S12 to reconstruct the transmitted signal corresponding to the first signal, but the corresponding signal quality is low. Based on the determined channel matrix, the present invention inputs the channel matrix into a regression network model. The regression network model can be a conventional network model or a network model improved based on the current channel estimation, which is not limited herein and can be set according to the actual situation. Based on the output of the regression network model, a denoised channel matrix is obtained to reconstruct the transmitted signal corresponding to the first signal.

[0107] An OTFS channel estimation method provided by an embodiment of the present invention is applied to a receiving end, which receives a first signal sent by a sending end; performs crosstalk analysis on the first signal to obtain a received signal in the time-delay Doppler domain; determines a target path according to the amplitude value of the received signal in the time-delay Doppler domain and a threshold value; constructs a channel matrix according to the path channel state information of the target path; inputs the channel matrix into a regression network model to obtain a denoised channel matrix, and performs reconstruction processing on the sending signal corresponding to the first signal. The present invention introduces a regression network model. After obtaining the channel matrix originally, it is necessary to input the channel matrix into the regression network model for denoising processing in a model manner, so as to improve the anti-interference ability of the channel matrix for the sending signal of the first signal during the reconstruction processing, so as to construct a fitting relationship between the original time-delay Doppler domain channel estimation matrix and the real channel matrix, and improve the estimation accuracy. Utilize the powerful learning and fitting ability of the neural network to further refine and optimize the input preliminary estimation result, and finally output a more accurate channel estimation result, which can significantly improve the performance of the entire communication system.

[0108] In some embodiments, the training process of the regression network model includes:

[0109] Obtain an initial regression network model, training data in a multipath channel scenario, and an actual channel matrix;

[0110] Determine a first target path according to the amplitude value and the threshold value corresponding to the training data in the multipath channel scenario;

[0111] Construct a first channel matrix according to the path channel state information of the first target path;

[0112] Take the first channel matrix as the input signal of the initial regression network model to obtain the first actual channel matrix of the initial regression network model;

[0113] Record the current iteration number;

[0114] When the difference between the first actual channel matrix and the actual channel matrix does not meet the preset condition, or when the current iteration number does not reach the preset iteration number, fine-tune the model parameters of the initial regression network model to obtain a fine-tuned initial regression network model; and return to the step of taking the first channel matrix as the input signal of the initial regression network model until the difference between the first actual channel matrix and the actual channel matrix meets the preset condition, or the current iteration number reaches the preset iteration number.

[0115] Figure 5 This is a schematic diagram of a regression network model provided by an embodiment of the present invention, as Figure 5As shown, in order to effectively reduce the negative impact of Doppler crosstalk and complex Gaussian white noise on the channel estimation results and improve the performance and accuracy of the communication system, a deep learning network architecture that performs excellently in the field of image denoising is introduced. This network incorporates an image self-attention mechanism, which can enhance the model's ability to capture complex channel characteristics and accurately establish the fitting relationship between the original time-delay Doppler domain channel estimation matrix and the true channel matrix, improving the estimation accuracy and laying a solid foundation for the performance optimization and practical application of the OTFS communication system.

[0116] Taking the result of the traditional threshold estimation algorithm as the input of the deep neural network model, and using the powerful learning and fitting ability of the neural network, the preliminary estimation result of the input is further refined and optimized, and finally a more accurate channel estimation result is output, which can significantly improve the performance of the entire communication system. The originally complex problem of fitting the time-delay Doppler domain channel response matrix is transformed into a two-channel image denoising problem, and the real and imaginary parts of the channel matrix are regarded as the two channels of the network input respectively, greatly facilitating the implementation of subsequent processing steps. Figure 5 In this, the model is divided into an offline stage and an online stage, and the tasks of the two stages are network training and real-time channel estimation respectively. With its unique network structure and powerful learning ability, the network can efficiently learn the non-linear mapping relationship between the channel estimation result and the true channel response. Especially the introduction of the self-attention module enables the network to more accurately focus on the key feature information in the channel response, while combating Doppler crosstalk and Gaussian white noise interference, retaining and enhancing the effective components of the channel response.

[0117] In the offline stage, training data in a multipath channel scenario and the known actual channel matrix are obtained. The first target path is determined according to the amplitude and threshold value corresponding to the training data in the multipath channel scenario; a first channel matrix is constructed according to the path channel state information of the first target path. The implementation manner of constructing the first channel matrix is the same as that of the above embodiment, and reference can be made to the above embodiment. The first channel matrix is input into the initial regression network model to output a first actual channel matrix. Based on the comparison between the first actual channel matrix and the actual channel matrix, when the difference between the first actual channel matrix and the actual channel matrix does not meet the preset condition, or when the current iteration number does not reach the preset iteration number, the model parameters of the initial regression network model are fine-tuned to obtain a fine-tuned initial regression network model. Here, the fine-tuning is based on the fine-tuning of the current model parameters, which can be set according to the actual situation and are not limited here. The fine-tuned initial regression network model continues to be trained until the difference between the first actual channel matrix and the actual channel matrix meets the preset condition, or the current iteration number reaches the preset iteration number, and then the training is ended to obtain the final regression network model.

[0118] The training process of the regression network model provided in this embodiment can enhance the model's ability to capture complex channel characteristics, accurately construct the fitting relationship between the original time-delay Doppler domain channel estimation matrix and the real channel matrix, and improve the estimation accuracy.

[0119] In some embodiments, inputting the channel matrix into the regression network model to obtain the denoised signal matrix includes:

[0120] Performing feature extraction processing on the data of the channel matrix to obtain first feature data;

[0121] Performing feature learning residual processing on the first feature data to obtain second feature data; wherein, the feature learning residual processing of the regression network model is completed in the feature learning residual module, and the feature learning residual module includes multiple enhancement attention modules, and there is a cascading relationship between the multiple enhancement attention modules. The enhancement attention module includes multiple convolutional kernels and a merging convolutional layer;

[0122] Performing reconstruction processing on the second feature data to obtain the denoised signal matrix.

[0123] Figure 6 Schematic diagram of the principle of a regression network model provided for an embodiment of the present invention, as Figure 6 shown, mainly includes the following three modules, namely the feature extraction module, the feature learning residual module, and the reconstruction module. The three modules are introduced separately below.

[0124] 1. Feature extraction module:

[0125] The feature extraction module consists of a convolutional layer and an activation layer, and is used to extract initial features from the input noisy image, and can capture the basic texture and edge information of the image. In the feature extraction module of RIDChanNet, the convolutional layer (Conv) extracts initial features from the network input :

[0126] ;

[0127] In the formula, represents the convolution operation, is the input parameter of the network input.

[0128] 2. Feature learning residual module:

[0129] The feature learning residual module is the core part of RIDChanNet (regression network model), and is composed of multiple enhancement attention modules (Enhancement Attention Module, EAM) cascaded (in Figure 6In the left figure, it is formed by using Enhanced Attention Module 1', Enhanced Attention Module 2' and Enhanced Attention Module 3'. Each EAM module learns the difference between the disturbed image and the original image through a specific structural design. In the feature extraction module, the network passes the initial features to the feature learning residual module, and the calculation process of this module is described as :

[0130] .

[0131] Traditional EAM modules mainly consist of a feature extraction sub-module and a feature attention sub-module. The feature extraction sub-module extracts and learns image features through two dilated convolution layers and a merging convolution layer. The dilated convolution layer expands the receptive field by increasing the dilation of the convolution kernel to skip the convolution operation on some pixels, thereby expanding the receptive field without increasing the number of parameters and the amount of computation, which helps to capture more extensive image context information. To avoid overfitting problems caused by an overly complex model and reduce the training and inference costs of the model, the two dilated convolution layers in the feature extraction sub-module are changed to convolution kernels. This modification enables the network to better adapt to the arrangement relationship of Doppler crosstalk on the channel matrix, enhancing the network's ability to suppress Doppler crosstalk; at the same time, it does not need to handle the complex hole filling and receptive field adjustment problems in dilated convolution, reducing the implementation difficulty of the network.

[0132] In some embodiments, the sizes of multiple convolution kernels within the enhanced attention module are different.

[0133] It should be noted that the modification in this embodiment is that the sizes of the convolution kernels are different, so as to increase the arrangement range on the channel matrix and enhance the network's ability to suppress Doppler crosstalk.

[0134] The feature attention sub-module first uses global average pooling to reduce the feature map from to to capture global context information. Subsequently, through an adaptive threshold control mechanism, two convolution layers and a Sigmoid activation function are used to generate a feature attention map. The first convolution layer is used to reduce the number of channels, and the second convolution layer restores the original number of channels. Finally, the feature weights of each channel are obtained through the Sigmoid activation function. This module dynamically adjusts the feature weights of each channel through the feature attention mechanism, enabling the network to pay more attention to features beneficial to the denoising task while suppressing useless or harmful features, which helps to improve the network's anti-interference performance and feature expression ability.

[0135] ;

[0136] In the formula, is the output of the global pooling layer, is a pooling layer, 、 represent the height and width of the feature map; here, 、 represent the indices of the feature image pixel points; in the formulas describing the network, both the feature map and the feature values are extracted from the network input matrix. is the feature value at the position in the feature map. In addition, in order to further extract useful features in the image, RIDChanNet introduces an adaptive threshold control mechanism to capture channel dependencies from the descriptors retrieved by global average pooling, learn the non-linear cooperative effects and mutual exclusive relationships between channels, and uses a soft shrinkage and Sigmoid function to implement the gating mechanism. and are the soft shrinkage operator and the Sigmoid-type operator respectively, where, In the OTFS signal processing process of the above embodiment, this variable represents the impulse function, and in the network, its meaning is the soft shrinkage operator, then the gating mechanism can be expressed as:

[0137] ;

[0138] In the formula, and are the channel reduction and channel upsampling operators respectively. The output of the global pooling layer enters the downsampling convolutional layer as input for feature extraction operations. In order to distinguish channel features, the global pooling layer is activated by a soft shrinkage function. Subsequently, it is input into the upsampling convolutional layer for Sigmoid activation. In order to calculate the statistic, the output of the Sigmoid activation function is adaptively rescaled by the input of the channel features as:

[0139] ;

[0140] The operation of the th module of EAM can be expressed as:

[0141] ;

[0142] In the formula, is the output of two sub-modules in the th EAM, which is equivalent to . Considering the existence of the residual structure, the output of each EAM module is , and at the same time, the output of the previous EAM module will be used as the input of the next EAM, and multiple EAM modules form a cascade relationship.

[0143] The pooling layer is an important part of the EAM module. Its function is to downsample the input data. The specific operation is to slide a pooling window on the input feature map, and according to the pooling type, calculate the statistical value of the data within the window as the output. Figure 7 FIG. is a schematic diagram of the operation process of a pooling layer provided by an embodiment of the present invention. Figure 8 FIG. is another schematic diagram of the operation process of a pooling layer provided by an embodiment of the present invention. As Figure 7 、 8 shown, the pooling layer reduces the size of the feature map through downsampling, thereby reducing the data dimension, reducing redundant information, enabling the model to focus more on important features, and at the same time reducing the input size of the next layer, thereby reducing the amount of calculation and the number of parameters. The pooling operation can also smooth the feature map to reduce the interference of noise, making the model more robust to small changes in the input data. In addition, the pooling layer makes each output pixel point correspond to a larger area of the input image through downsampling, expanding the receptive field of the model, which helps the model capture features in a larger range.

[0144] The area corresponding to the zero element is marked as the low-frequency area, and the non-zero element corresponds to the high-frequency area. And, in the fitting process of the time-delay Doppler domain channel matrix, the Doppler crosstalk in a specific scenario is represented as local information. The traditional pooling layer mainly uses local information rather than global information, which may cause interference from local information during channel matrix estimation. In order to reduce the interference effect of local information, a global average pooling strategy is adopted in the pooling layer. By retaining global features, the robustness of the network system can be improved.

[0145] The EAM module also introduces two activation functions, ReLU and Sigmoid. The expression of the Sigmoid function is:

[0146] 。

[0147] The Sigmoid function converts the input value to the range of (0, 1), and the obtained result can be interpreted as a probability, representing the variable nodes in the network, which is suitable for building models with predicted probabilities as outputs. In addition, the gradient of the Sigmoid function shows a smooth characteristic in most domains, which can effectively avoid sudden changes in the output value and ensure the smoothness of the training process. Moreover, the derivative expression of the Sigmoid function is concise and clear, which greatly facilitates the calculation steps in the backpropagation algorithm.

[0148] The ReLU function only needs to perform a simple threshold determination: if the input value exceeds 0, it is retained, otherwise it is set to 0, significantly reducing the amount of calculation and improving the operation speed of the neural network. The expression of ReLU is: 。

[0149] In the positive number range, the derivative of the ReLU function remains constant at 1, and the gradient does not weaken due to the accumulation of the number of network layers, which helps to alleviate the phenomenon of vanishing gradients in deep neural networks. ReLU enables the neural network to simulate a non-linear mapping relationship, which helps to enhance the fitting ability of the network.

[0150] 3. Reconstruction module:

[0151] The reconstruction module consists of a convolutional layer and a long skip connection operation, which is used to convert the learned residual features back to the image space and add them to the input image to generate a denoised image.

[0152] is the feature output by the feature learning residual module. Input it into the convolutional layer of the reconstruction module to obtain: ; in the formula. However, simply cascading the residual modules cannot obtain better performance. For this reason, RIDChanNet stacks the initial feature on the output of the convolutional layer of the reconstruction module to obtain the final network output: . In the formula, represents the convolution operation. The stacking operation is called the Long Skip Connection (LSC), which can simplify the information flow between groups; is the output of the convolutional layer of the reconstruction module. Finally, the network outputs the same number of feature maps as the input end of the network to complete a forward propagation.

[0153] The structure of the regression network model provided in this embodiment enables the network to more effectively separate and suppress the interference components in the channel impulse response matrix by focusing on the learning and processing of interference features. Thus, while retaining the details and structural information of the channel matrix, it achieves higher-quality restoration, not only improving the processing efficiency but also greatly enhancing the robustness of the model.

[0154] In some embodiments, the difference between the first actual channel matrix and the actual channel matrix does not meet the preset conditions, including:

[0155] Obtain the channel capacity and the rank corresponding to the eigenvalues of the first actual channel matrix and the actual channel matrix respectively;

[0156] If the difference between their respective channel capacities is greater than or equal to the first threshold, then compare the ranks corresponding to their respective eigenvalues;

[0157] If the ranks corresponding to their respective eigenvalues are greater than or equal to the second threshold, it is determined that the difference between the first actual channel matrix and the actual channel matrix does not meet the preset conditions.

[0158] Specifically, the channel capacity is an important indicator to measure the transmission capacity of a channel, and each channel's capacity is calculated to compare them. By comparing the channel capacities corresponding to different channel matrices, the differences in their transmission efficiencies can be evaluated. If the difference between their respective channel capacities is greater than or equal to the first threshold, then the ranks corresponding to their respective eigenvalues are compared. The eigenvalues of the channel matrix can reflect the rank and diversity of the channel. The higher the rank of the channel matrix, the more independent paths the channel has, and the greater the channel capacity. By comparing the eigenvalues of different channel matrices, the differences in their spatial diversity can be understood. If the ranks corresponding to their respective eigenvalues are greater than or equal to the second threshold, it is determined that the difference between the first actual channel matrix and the actual channel matrix does not meet the preset condition.

[0159] In this embodiment, determining that the difference between the first actual channel matrix and the actual channel matrix does not meet the preset condition based on the channel capacity and the rank corresponding to the eigenvalue improves the accuracy of the comparison relationship between different channel matrices.

[0160] In some other embodiments, the difference between the first actual channel matrix and the actual channel matrix not meeting the preset condition includes:

[0161] Obtain the respective correlations and respective estimation errors corresponding to the first actual channel matrix and the actual channel matrix;

[0162] If the difference between their respective correlations is not within the first preset range, then compare their respective estimation errors;

[0163] If the difference between their respective estimation errors is not within the second preset range, it is determined that the difference between the first actual channel matrix and the actual channel matrix does not meet the preset condition.

[0164] Specifically, the correlation of the channel matrix can affect the transmission quality of the signal. A channel matrix with a lower correlation can provide better signal demodulation performance. By analyzing the correlation of the channel matrix, the performance of different channel matrices can be compared. If the difference between their respective correlations is not within the first preset range, then compare their respective estimation errors. The channel estimation error is an important indicator to measure the performance of the channel estimation algorithm. By comparing the estimation errors corresponding to different channel matrices, the performance of the channel estimation algorithm under different channel conditions can be evaluated. For example, the Minimum Mean Square Error (MMSE) algorithm takes into account the influence of noise, and its estimation error can be compared with the Least Squares (LS) algorithm to evaluate the performance difference under different channel conditions. If the difference between their respective estimation errors is not within the second preset range, it is determined that the difference between the first actual channel matrix and the actual channel matrix does not meet the preset condition.

[0165] The difference between the first actual channel matrix and the actual channel matrix determined based on the correlation and their respective estimated errors in this embodiment does not meet the preset condition, improving the accuracy of the comparison relationship between different channel matrices.

[0166] The signal-to-noise ratio, as an important indicator for measuring signal quality, physically represents the ratio of signal power to noise power. In a wireless communication environment with multipath propagation and mobility, the change in the signal-to-noise ratio will directly affect the estimation accuracy of the channel matrix. Therefore, in the scenario of collecting the channel matrix as training data, this measure can help the model better understand and simulate the law of the channel matrix changing with the signal-to-noise ratio, thereby improving the accuracy and robustness of channel estimation.

[0167] In a wireless communication environment with multipath propagation and mobility, the number of multipaths and the moving speed are important factors affecting the change of the channel matrix. To simplify the problem complexity and enhance the model's ability to reconstruct a certain type of mathematical relationship, two parameters, namely the number of multipaths and the maximum moving speed, are fixed, and training data is generated in the scenario where the number of multipaths is 5 and the maximum moving speed is 1000 km / h. This can make the training data more targeted, help the model learn the optimal solution fitted in this state, deepen the model's understanding of the multipath fast time-varying channel scenario, and thus show transferability in the same type of problems.

[0168] Quantitatively, 5000 groups of estimated channel matrices and true channel matrices are collected as training outputs and inputs respectively. The training set and the validation set are divided at a ratio of 9:1, and these data are used to train the channel estimation model, enabling the communication system to have stronger adaptive capabilities. When the channel environment changes, the system can quickly and accurately adjust the channel estimation parameters, thereby ensuring the stability and reliability of communication. This is of great significance for improving the overall performance and user experience of the communication system. To prevent unreasonable training data from affecting the model, the L1 loss function is selected when training the RIDChanNet-based channel estimation network. This loss function penalizes errors linearly, reduces the impact of large errors, and lowers the training difficulty, which is beneficial to enhancing the robustness of the model.

[0169] To compare the performance of RIDChanNet with that of Convolutional Neural Network (CNN), Residual Network (ResNet), and the threshold-based channel estimation algorithm, and to observe the estimation performance of the four detection algorithms under different signal-to-noise ratios and the number of multipaths, Matlab Monte Carlo simulations were carried out. Table 1 is the simulation parameter table of OTFS channel estimation based on RIDChanNet. As shown in Table 1, the ratio of data symbol power to noise power is defined as: . The ratio of pilot power to noise power is defined as: . The relationship between the pilot signal-to-noise ratio and the data signal-to-noise ratio can be described as: .

[0170] is the ratio of pilot power to data symbol power. Since the power of pilot symbols is much higher than that of data symbols, is also much higher than .

[0171] Table 1

[0172]

[0173] To test the estimation accuracy of the RIDChanNet estimation scheme and evaluate the performance of RIDChanNet in practical applications, the normalized mean square error was used as a measure of channel estimation accuracy. At the same time, three comparison schemes were selected, namely the threshold-based channel estimation algorithm given in the above embodiments, the existing residual neural network-based estimation algorithm, and the existing CNN-based channel estimation algorithm, corresponding to "Conventional", "ResNet", and "SR" in the legend respectively. The legend corresponding to the proposed algorithm is "RID".

[0174] Figure 9 This is a schematic diagram of the normalized mean square error curve of RIDChanNet in a five-path channel provided by an embodiment of the present invention. As Figure 9 shown, in a five-path fast time-varying channel, thanks to the powerful feature extraction and data fitting capabilities of RIDChanNet, the proposed channel estimation scheme exhibits better performance than several other optimization schemes. This advantage is mainly attributed to the powerful feature extraction and noise suppression capabilities of the EAM module. By learning and optimizing the network parameters, the proposed scheme can more accurately fit the characteristics of the real DD-domain channel, greatly improving the channel estimation accuracy and significantly reducing the communication bit error rate.

[0175] Compared with several other estimation schemes, the traditional channel estimation scheme performs the worst because it does not handle IDI interference and noise, seriously affecting the estimation accuracy of the channel response. Compared with the traditional estimation scheme, the CNN-assisted OTFS channel estimation method has improved performance. However, due to the lack of residual connections between layers in this network, it is prone to overfitting during training, so the improvement is relatively small. The ResNet estimation method shows relatively superior performance. However, its simple structural characteristics make it difficult to extract the edge information of the image, so the estimation accuracy is slightly inferior to RIDChanNet.

[0176] To verify the generality of RIDChanNet, the simulation scenario was extended to fast time-varying channels with eight paths and ten paths. Figure 10 This is a schematic diagram of the normalized mean square error curve of RIDChanNet in an eight-path channel provided by an embodiment of the present invention. As Figure 10 shown, in the eight-path channel, the IDI crosstalk of the system becomes more serious. The estimation accuracy of all schemes decreases at low signal-to-noise ratios; in a more complex channel environment, the feature extraction ability of the residual network decreases slightly; at this time, the performance of RIDChanNet still shows absolute superiority compared with several other schemes. Figure 11 This is a schematic diagram of the normalized mean square error curve of RIDChanNet in a ten-path channel provided by an embodiment of the present invention. As Figure 11 shown, in the ten-path channel, it becomes more difficult to extract channel features. Among them, CNN is prone to overfitting due to the lack of residual connections between network layers, and its generalization ability is poor. Although the estimation accuracy of ResNet and RIDChanNet decreases, they still maintain good stability. Thanks to the unique feature extraction mechanism and self-attention module of RIDChanNet, it has stronger feature extraction ability and generalization ability compared with ResNet, and can well fit the mathematical expression of IDI interference. Therefore, it has superior and stable performance in different channel environments.

[0177] To further evaluate the impact of the proposed scheme on the entire communication system, the bit error rate (BER) is used as an index to measure the performance of the optimization scheme. Figure 12 This is a schematic diagram of the bit error rate curve of RIDChanNet in a five-path channel provided by an embodiment of the present invention. As Figure 12 shown, in the five-path channel, the traditional estimation scheme approaches the bit error rate floor at high signal-to-noise ratios. The reason is that at high signal-to-noise ratios, IDI interference replaces noise as the main factor affecting communication performance, and this part of the impact cannot be removed only by increasing the signal-to-noise ratio. Neural networks can extract the hidden features of the interference part and use the non-linear mapping relationship between a large number of parameters to fit IDI and noise interference. InFigure 12 Among them, RIDChanNet maintains relatively superior performance with its variable network architecture and strong feature extraction ability. Figure 13 It is a schematic diagram of the bit error rate curve of RIDChanNet under an eight-path channel provided by an embodiment of the present invention. As Figure 13 shown, under an eight-path channel, the difficulty of channel estimation further increases. Comparing Figure 12 with Figure 13 the bit error rate curve of the ideal channel matrix, it can be seen that the error of signal detection also increases. At this time, RIDChanNet still maintains relatively stable estimation performance, showing the same change trend as the curve under the ideal state, further proving the stability and accuracy of this method. Figure 14 It is a schematic diagram of the bit error rate curve of RIDChanNet under a ten-path channel provided by an embodiment of the present invention. As Figure 14 shown, when the number of paths further increases to 10, the performance gap between RIDChanNet and several other network structures becomes larger. The reason is that the feature extraction module effectively extracts the deep features of IDI interference, so it can maintain stable estimation performance in channel environments with different numbers of paths. The structures of the other two networks are relatively simple compared to RIDChanNet and it is difficult to maintain the original performance in a more complex environment.

[0178] Furthermore, the present invention also provides an OTFS channel estimation system, including a receiving end and a transmitting end:

[0179] Control the transmitting end to send a first signal;

[0180] Control the receiving end to perform crosstalk analysis on the first signal to obtain a received signal in the time-delay Doppler domain; determine the target path according to the amplitude of the received signal in the time-delay Doppler domain and a threshold value; construct a channel matrix according to the path channel state information of the target path; perform denoising processing on the channel matrix to obtain a denoised channel matrix, so as to reconstruct the transmitted signal corresponding to the first signal.

[0181] For the introduction of an OTFS channel estimation system provided by the present invention, please refer to the above method embodiment. The present invention will not elaborate here, and it has the same beneficial effects as the above OTFS channel estimation method.

[0182] The above details each embodiment corresponding to the OTFS channel estimation method. On this basis, the present invention also discloses an OTFS channel estimation device corresponding to the above method. Figure 15 It is a structural diagram of an OTFS channel estimation device provided by an embodiment of the present invention. As Figure 15 shown, the OTFS channel estimation device includes:

[0183] A receiving module 11, configured to receive a first signal sent by a sending end; and perform crosstalk analysis on the first signal to obtain a received signal in the time-delay Doppler domain;

[0184] A determining module 12, configured to determine a target path according to the amplitude of the received signal in the time-delay Doppler domain and a threshold value; and construct a channel matrix according to the path channel state information of the target path;

[0185] A reconstruction processing module 13, configured to input the channel matrix into a regression network model to obtain a denoised channel matrix, and perform reconstruction processing on the transmitted signal corresponding to the first signal.

[0186] Since the embodiments of the device part correspond to the above embodiments, the embodiments of the device part are described with reference to the embodiments of the above method part and will not be elaborated herein.

[0187] For the introduction of an OTFS channel estimation device provided by the present invention, please refer to the above method embodiments. The present invention will not be elaborated herein, and it has the same beneficial effects as the above OTFS channel estimation method.

[0188] Figure 16 The following is a structural diagram of an OTFS channel estimation device provided by an embodiment of the present invention. As Figure 16 shown, the device includes:

[0189] A memory 21, configured to store a computer program;

[0190] A processor 22, configured to implement the steps of the OTFS channel estimation method when executing the computer program.

[0191] The OTFS channel estimation device provided in this embodiment may include, but is not limited to, a smart phone, a tablet computer, a notebook computer, or a desktop computer, etc.

[0192] Among them, the processor 22 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 22 may be implemented in at least one hardware form of a Digital Signal Processor (DSP), a Field-Programmable Gate Array (FPGA), or a Programmable Logic Array (PLA). The processor 22 may also include a main processor and a coprocessor. The main processor is a processor for processing data in the wake state, also known as the Central Processing Unit (CPU); the coprocessor is a low-power processor for processing data in the standby state. In some embodiments, the processor 22 may be integrated with a Graphics Processing Unit (GPU), and the GPU is responsible for rendering and drawing the content to be displayed on the display screen. In some embodiments, the processor 22 may further include an Artificial Intelligence (AI) processor, and the AI processor is used to process computational operations related to machine learning.

[0193] The memory 21 may include one or more computer-readable storage media, and the computer-readable storage media may be non-transitory. The memory 21 may further include high-speed random access memory and non-volatile memory, such as one or more disk storage devices and flash storage devices. In this embodiment, the memory 21 is at least used to store the following computer program 211. After the computer program is loaded and executed by the processor 22, it can implement the relevant steps of the OTFS channel estimation method disclosed in any of the foregoing embodiments. In addition, the resources stored in the memory 21 may further include an operating system 212 and data 213, etc., and the storage method may be transient storage or permanent storage. Among them, the operating system 212 may include Windows, Unix, Linux, etc. The data 213 may include, but is not limited to, the data involved in the OTFS channel estimation method, etc.

[0194] In some embodiments, the OTFS channel estimation device may further include a display screen 23, an input / output interface 24, a communication interface 25, a power supply 26, and a communication bus 27.

[0195] Those skilled in the art can understand that Figure 16 the structure shown in

[0196] The processor 22 implements the OTFS channel estimation method provided in any of the above embodiments by invoking the instructions stored in the memory 21.

[0197] For the introduction of an OTFS channel estimation device provided by the present invention, please refer to the above method embodiments. The present invention will not be elaborated herein, and it has the same beneficial effects as the above OTFS channel estimation method.

[0198] Furthermore, the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by the processor 22, the steps of the above OTFS channel estimation method are implemented.

[0199] It can be understood that if the method in the above embodiments is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and executes all or part of the steps of the methods in the various embodiments of the present invention. The aforementioned storage media include: USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs, etc., which can store program codes.

[0200] For the introduction of a computer-readable storage medium provided by the present invention, please refer to the above method embodiments. The present invention will not be elaborated herein, and it has the same beneficial effects as the above OTFS channel estimation method.

[0201] The above has provided a detailed introduction to an OTFS channel estimation method, system, device, and medium provided by the present invention. The various embodiments in the specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The same or similar parts among the various embodiments can be referred to each other. For the device disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the description in the method part. It should be noted that for those of ordinary skill in the art in this technical field, without departing from the principle of the present invention, several improvements and modifications can still be made to the present invention, and these improvements and modifications also fall within the protection scope of the present invention.

[0202] It should also be noted that in this specification, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent in such process, method, article or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or device comprising the element.

Claims

1. An OTFS channel estimation method, characterized in that: Applied to the receiving end, including: receiving a first signal sent by a transmitting end; and performing crosstalk analysis according to the first signal to obtain a delay Doppler domain received signal; Determine a target path according to the amplitude and threshold value of the delay Doppler domain received signal; and construct a channel matrix according to the path channel state information of the target path; The channel matrix is ​​input into a regression network model to obtain a denoised channel matrix, and a transmission signal corresponding to the first signal is reconstructed; wherein the training process of the regression network model stops training when the difference between the first actual channel matrix and the actual channel matrix meets a preset condition, or when the current number of iterations reaches a preset number of iterations; the first actual channel matrix is ​​obtained by an initial regression network model; The difference between the first actual channel matrix and the actual channel matrix does not meet a preset condition, including: Obtaining the channel capacities and ranks corresponding to the eigenvalues ​​corresponding to the first actual channel matrix and the actual channel matrix respectively; if the difference between the respective channel capacities is greater than or equal to the first threshold, comparing the ranks corresponding to the respective eigenvalues; if the ranks corresponding to the respective eigenvalues ​​are greater than or equal to the second threshold, determining that the difference between the first actual channel matrix and the actual channel matrix does not meet the preset condition; Alternatively, obtain the correlation and estimation error corresponding to the first actual channel matrix and the actual channel matrix respectively; if the difference between the corresponding correlations is not within the first preset range, compare the corresponding estimation errors respectively; if the difference between the corresponding estimation errors is not within the second preset range, determine that the difference between the first actual channel matrix and the actual channel matrix does not meet the preset conditions.

2. The OTFS channel estimation method according to claim 1, characterized in that: The training process of the regression network model includes: Obtain the initial regression network model, training data in a multipath channel scenario, and the actual channel matrix; Determine a first target path according to the amplitude corresponding to the training data in the multipath channel scenario and the threshold value; constructing a first channel matrix according to the path channel state information of the first target path; Using the first channel matrix as an input signal of the initial regression network model to obtain a first actual channel matrix of the initial regression network model; Record the current number of iterations; When the difference between the first actual channel matrix and the actual channel matrix does not meet the preset conditions, or the current number of iterations does not reach the preset number of iterations, the model parameters of the initial regression network model are fine-tuned to obtain the fine-tuned initial regression network model; and the method returns to the step of using the first channel matrix as the input signal of the initial regression network model until the difference between the first actual channel matrix and the actual channel matrix meets the preset conditions, or the current number of iterations reaches the preset number of iterations.

3. The OTFS channel estimation method according to claim 2, characterized in that: Inputting the channel matrix into the regression network model to obtain a denoised signal matrix includes: Performing feature extraction processing on the data of the channel matrix to obtain first feature data; The first feature data is subjected to feature learning residual processing to obtain second feature data; wherein the feature learning residual processing of the regression network model is completed in a feature learning residual module, and the feature learning residual module includes a plurality of enhanced attention modules, and there is a cascade relationship between the plurality of enhanced attention modules, and the enhanced attention module includes a plurality of convolution kernels and a merged convolution layer; The second characteristic data is reconstructed to obtain a denoised signal matrix.

4. The OTFS channel estimation method according to claim 3, characterized in that: The sizes of multiple convolution kernels in the enhanced attention module are different.

5. The OTFS channel estimation method according to claim 1, characterized in that: The performing crosstalk analysis according to the first signal to obtain a delay Doppler domain received signal includes: Performing SFFT transformation on the first signal to obtain a first delay Doppler domain signal; Acquire a first target row and a second target row of the first delay-Doppler domain signal; The first delay-Doppler domain signal is intercepted according to the first target row and the second target row, and the signal from the first target row to the second target row is used as the delay-Doppler domain received signal.

6. An OTFS channel estimation system, characterized in that: Including the receiving end and the sending end: Controlling the sending end to send a first signal; Controlling the receiving end to perform crosstalk analysis according to the first signal to obtain a delay Doppler domain received signal; determining a target path according to the amplitude and threshold value of the delay Doppler domain received signal; Constructing a channel matrix according to the path channel state information of the target path; Inputting the channel matrix into a regression network model to obtain a denoised channel matrix to reconstruct a transmission signal corresponding to the first signal; wherein the training process of the regression network model stops training when the difference between the first actual channel matrix and the actual channel matrix meets a preset condition or when the current number of iterations reaches a preset number of iterations; the first actual channel matrix is ​​obtained by an initial regression network model; The difference between the first actual channel matrix and the actual channel matrix does not meet a preset condition, including: Obtaining the channel capacities and ranks corresponding to the eigenvalues ​​corresponding to the first actual channel matrix and the actual channel matrix respectively; if the difference between the respective channel capacities is greater than or equal to the first threshold, comparing the ranks corresponding to the respective eigenvalues; if the ranks corresponding to the respective eigenvalues ​​are greater than or equal to the second threshold, determining that the difference between the first actual channel matrix and the actual channel matrix does not meet the preset condition; Alternatively, obtain the correlation and estimation error corresponding to the first actual channel matrix and the actual channel matrix respectively; if the difference between the corresponding correlations is not within the first preset range, compare the corresponding estimation errors respectively; if the difference between the corresponding estimation errors is not within the second preset range, determine that the difference between the first actual channel matrix and the actual channel matrix does not meet the preset conditions.

7. An OTFS channel estimation device, characterized in that: include: Memory for storing computer programs; A processor, configured to implement the steps of the OTFS channel estimation method according to any one of claims 1 to 5 when executing the computer program.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the OTFS channel estimation method according to any one of claims 1 to 5 are implemented.

Images