Multipath Delay Determination Method, Apparatus, Device, and Storage Medium

Abstract

The present application relates to a multipath delay determination method, apparatus, device, and storage medium. The method includes: obtaining a received signal received by a receiving end, and determining a delay spectrum vector of the received signal according to the received signal; determining a vector to be input according to the real part and the imaginary part of the delay spectrum vector; inputting the vector to be input into a target model to obtain a delay spectrum prediction vector of the received signal; the target model is obtained by training an initial model according to the real delay spectrum vectors of received signal samples and transmitted signal samples, the received signal samples are samples obtained by superimposing noises with different signal-to-noise ratios on initial received signal samples, the received signal samples are signal samples received by the receiving end after the transmitting end transmits the transmitted signal samples; and determining the multipath delays of each cluster included in the received signal according to the delay spectrum prediction vector. Using this method can improve the efficiency of multipath delay determination.

Description

Technical Field

[0001] The present application relates to the field of communication technologies, and in particular, to a multipath delay determination method, apparatus, device, and storage medium. Background Art

[0002] Multipath delay refers to the situation where after electromagnetic waves propagate through different paths, the components of the fields arrive at the receiving end at different times, and they are superimposed on each other according to their respective phases, causing interference, distorting the original signal, or generating errors. Multipath delay is an important parameter in the study of wireless channels and is also one of the key indicators reflecting the communication quality of wireless channels. The estimation of multipath delay parameters is a research topic with wide applications in modern digital signal processing, involving important fields such as wireless communication, radar, sonar, and wireless local positioning systems.

[0003] In traditional technologies, the estimation of multipath delay first establishes a forward mapping model from the delay values of multipath signals to the received signal or frequency response at the receiving end, and then uses preset formulas and algorithms for derivation to estimate the multipath delay. Since the signal is sent through subcarriers, the more subcarriers there are, the larger the amount of received signal data, and thus the process of deriving the multipath delay through preset formulas and algorithms becomes more complex, resulting in a lower efficiency of multipath delay estimation. Summary of the Invention

[0004] Based on this, it is necessary to provide a multipath delay determination method, apparatus, device, and storage medium that can improve the efficiency of multipath delay estimation for the above technical problems.

[0005] In a first aspect, the present application provides a multipath delay determination method. The method includes:

[0006] Obtain the received signal received by the receiving end, and determine the delay spectrum vector of the received signal according to the received signal;

[0007] Determine the vector to be input according to the real part and the imaginary part of the delay spectrum vector;

[0008] Input the vector to be input into the target model to obtain the delay spectrum prediction vector of the received signal; the target model is obtained by training an initial model according to the real delay spectrum vectors of the received signal samples and the transmitted signal samples, the received signal samples are samples obtained by superimposing noises with different signal-to-noise ratios on the initial received signal samples, and the received signal samples are the signal samples received by the receiving end after the transmitting end transmits the transmitted signal samples;

[0009] Determine the multipath delays of each cluster included in the received signal according to the delay spectrum prediction vector.

[0010] In one embodiment, the method further includes:

[0011] Input the delay spectrum vector of the received signal sample and the true delay spectrum vector of the transmitted signal sample into the initial model to obtain a predicted delay spectrum vector sample;

[0012] Train the initial model according to the predicted delay spectrum vector sample and the true delay spectrum vector to obtain the target model.

[0013] In one embodiment, the method further includes:

[0014] Input the received signal sample and the true delay spectrum vector of the transmitted signal sample into the initial model to obtain a predicted delay spectrum vector sample;

[0015] Train the initial model according to the predicted delay spectrum vector sample and the true delay spectrum vector to obtain the target model.

[0016] In one embodiment, training the initial model according to the predicted delay spectrum vector sample and the true delay spectrum vector to obtain the target model includes:

[0017] Determine the corresponding loss value according to the predicted delay spectrum vector sample and the true delay spectrum vector;

[0018] Train the initial model according to the loss value to obtain the target model.

[0019] In one embodiment, the transmitted signal sample is determined according to the preset delay range of each cluster, the preset total number of each cluster, and the preset delay interval between two adjacent clusters in each cluster.

[0020] In one embodiment, determining the delay spectrum vector of the received signal according to the received signal includes:

[0021] Obtain the transmitted signal corresponding to the received signal sent by the transmitting end;

[0022] Perform time-frequency transformation processing on the received signal to obtain a first frequency-domain signal, and perform time-frequency transformation processing on the transmitted signal to obtain a second frequency-domain signal;

[0023] Determine a delay covariance matrix according to the first frequency-domain signal and the second frequency-domain signal;

[0024] Determine the delay spectrum vector of the received signal according to the delay covariance matrix.

[0025] In one embodiment, the initial model includes a preset number of convolutional layers. Each convolutional layer uses a rectified linear unit as an activation function, and each convolutional layer is used to adjust the delay spectrum vector of the received signal.

[0026] In a second aspect, the present application further provides a multipath delay determination device. The device includes:

[0027] An acquisition module, configured to acquire a received signal received by a receiving end, and determine a delay spectrum vector of the received signal according to the received signal;

[0028] A first determination module, configured to determine an input vector to be input according to a real part and an imaginary part of the delay spectrum vector;

[0029] A first input module, configured to input the input vector to be input into a target model to obtain a delay spectrum prediction vector of the received signal; the target model is obtained by training an initial model according to real delay spectrum vectors of received signal samples and transmitted signal samples, the received signal samples are samples obtained by superimposing noises with different signal-to-noise ratios on initial received signal samples, and the received signal samples are signal samples received by the receiving end after the transmitting end transmits the transmitted signal samples;

[0030] A second determination module, configured to determine multipath delays of each cluster included in the received signal according to the delay spectrum prediction vector.

[0031] In a third aspect, the present application further provides a computer device. The computer device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0032] Acquire a received signal received by a receiving end, and determine a delay spectrum vector of the received signal according to the received signal;

[0033] Determine an input vector to be input according to a real part and an imaginary part of the delay spectrum vector;

[0034] Input the input vector to be input into a target model to obtain a delay spectrum prediction vector of the received signal; the target model is obtained by training an initial model according to real delay spectrum vectors of received signal samples and transmitted signal samples, the received signal samples are samples obtained by superimposing noises with different signal-to-noise ratios on initial received signal samples, and the received signal samples are signal samples received by the receiving end after the transmitting end transmits the transmitted signal samples;

[0035] Determine multipath delays of each cluster included in the received signal according to the delay spectrum prediction vector.

[0036] In a fourth aspect, the present application further provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, and when the computer program is executed by a processor, the following steps are implemented:

[0037] Acquire a received signal received by a receiving end, and determine a delay spectrum vector of the received signal according to the received signal;

[0038] Determine the vector to be input according to the real part and the imaginary part of the time delay spectrum vector;

[0039] Input the vector to be input into the target model to obtain the time delay spectrum prediction vector of the received signal; the target model is obtained by training the initial model according to the real time delay spectrum vectors of the received signal samples and the transmitted signal samples, the received signal samples are the samples obtained by adding noises with different signal-to-noise ratios to the initial received signal samples, and the received signal samples are the signal samples received by the receiving end after the transmitting end transmits the transmitted signal samples;

[0040] Determine the multipath time delays of each cluster included in the received signal according to the time delay spectrum prediction vector.

[0041] In a fifth aspect, the present application further provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0042] Obtain the received signal received by the receiving end, and determine the time delay spectrum vector of the received signal according to the received signal;

[0043] Determine the vector to be input according to the real part and the imaginary part of the time delay spectrum vector;

[0044] Input the vector to be input into the target model to obtain the time delay spectrum prediction vector of the received signal; the target model is obtained by training the initial model according to the real time delay spectrum vectors of the received signal samples and the transmitted signal samples, the received signal samples are the samples obtained by adding noises with different signal-to-noise ratios to the initial received signal samples, and the received signal samples are the signal samples received by the receiving end after the transmitting end transmits the transmitted signal samples;

[0045] Determine the multipath time delays of each cluster included in the received signal according to the time delay spectrum prediction vector.

[0046] For the above multipath time delay determination method, device, equipment, and storage medium, obtain the received signal received by the receiving end, determine the time delay spectrum vector of the received signal according to the received signal, and determine the vector to be input according to the real part and the imaginary part of the time delay spectrum vector; input the vector to be input into the target model to obtain the time delay spectrum prediction vector of the received signal; the target model is obtained by training the initial model according to the real time delay spectrum vectors of the received signal samples and the transmitted signal samples, the received signal samples are the samples obtained by adding noises with different signal-to-noise ratios to the initial received signal samples, and the received signal samples are the signal samples received by the receiving end after the transmitting end transmits the transmitted signal samples; determine the multipath time delays of each cluster included in the received signal according to the time delay spectrum prediction vector. In the present application, the time delay spectrum prediction vector of the received signal is determined by the target model, and the multipath time delay is determined by the prediction vector, avoiding the process of determining the multipath time delay by formula and improving the efficiency of determining the multipath time delay. Brief Description of the Drawings

[0047] Figure 1 is an application environment diagram of a multipath delay determination method provided by an embodiment of the present application;

[0048] Figure 2 is a schematic flowchart of a multipath delay determination method provided by an embodiment of the present application;

[0049] Figure 3 is one of the schematic flowcharts of a target model training method provided by an embodiment of the present application;

[0050] Figure 4 is another schematic flowchart of a target model training method provided by an embodiment of the present application;

[0051] Figure 5 is a schematic flowchart of a delay spectrum vector determination method provided by an embodiment of the present application;

[0052] Figure 6 is a performance comparison diagram provided by an embodiment of the present application;

[0053] Figure 7 is a structural block diagram of a multipath delay determination device provided by an embodiment of the present application;

[0054] Figure 8 is an internal structure diagram of a computer device provided by an embodiment of the present application. Detailed Description of the Embodiments

[0055] In order to make the objectives, technical solutions and advantages of the present application more clear and understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0056] The multipath delay determination method provided by the embodiments of the present application can be applied to an application environment as shown in Figure 1 the following figure. Figure 1 is an application environment diagram of a multipath delay determination method provided by an embodiment of the present application. The application environment includes a signal sending end 101, a signal receiving end 102, and a computer device 103. The signal sending end 101 is used to send a signal to the signal receiving end 102. The signal receiving end 102 receives the signal sent by the signal sending end 101, and then sends the received signal to the computer device 103. The computer device 103 acquires the received signal of the signal receiving end 102, determines the delay spectrum vector of the received signal according to the received signal, further processes the delay spectrum vector of the received signal by using a target model to obtain a delay spectrum prediction vector of the received signal, and finally determines the multipath delays of each cluster included in the received signal according to the delay spectrum prediction vector.

[0057] In one embodiment, as Figure 2 shown, Figure 2 FIG. is a schematic flowchart of a multipath delay determination method provided by an embodiment of the present application. Taking the method applied to Figure 1 the computer device 103 therein as an example, the method includes the following steps:

[0058] S201, obtain the received signal received by the receiving end, and determine the delay spectrum vector of the received signal according to the received signal.

[0059] Among them, the received signal is composed of multiple clusters, each cluster contains a preset number of sub-paths, the delay spectrum of the received signal reflects the time when each cluster in the received signal arrives at the receiving end. Specifically, each spectral peak in the delay spectrum represents each cluster in the received signal, and the time corresponding to each spectral peak represents the multipath delay of each cluster in the received signal. Since the delay spectrum vector can reflect the delay spectrum, the delay spectrum of the received signal can be determined through the delay spectrum vector of the received signal.

[0060] Exemplarily, an orthogonal frequency division multiplexing signal (Orthogonal Frequency Division Multiplexing, OFDM) can be used as the transmitted signal. Let the transmitted signal be S(t), set the number of subcarriers of the signal to N, obtain the received signal X(t) in the time domain received by the receiving end, and determine the delay spectrum vector of the received signal according to the received signal X(t)

[0061] S202, determine the vector to be input according to the real part and the imaginary part of the delay spectrum vector.

[0062] Exemplarily, the delay spectrum vector contains a real part and an imaginary part. Before inputting it into the target model, it is necessary to extract the real part and the imaginary part of and form a new vector to be input with the real part and the imaginary part in

[0063] S203, input the vector to be input into the target model to obtain the delay spectrum prediction vector of the received signal. The target model is obtained by training the initial model according to the real delay spectrum vectors of the received signal samples and the transmitted signal samples. The received signal samples are samples obtained by adding noises with different signal-to-noise ratios to the initial received signal samples, and the received signal samples are the signal samples received by the receiving end after the transmitting end transmits the transmitted signal samples.

[0064] Among them, the target model is a neural network model, and the prediction vector is the vector output by the target model.

[0065] Exemplarily, a Deep Convolutional Networks (DCN) is used as the initial model. The target model is a model obtained by training the initial model according to the true delay spectrum vectors of the received signal samples and the transmitted signal samples. The received signal samples are samples obtained by superimposing noises with different signal-to-noise ratios on the initial received signal samples, and the received signal samples are the signal samples received by the receiving end after the transmitting end transmits the signal samples; the vector to be input is input into the target model to obtain the delay spectrum prediction vector η of the received signal.

[0066] Optionally, other neural network models can also be used as the initial model, which is not limited in this embodiment.

[0067] S204. Determine the multipath delays of each cluster included in the received signal according to the delay spectrum prediction vector.

[0068] Exemplarily, according to the delay spectrum prediction vector η, the delay spectrum is determined, and according to the time corresponding to each spectral peak in the delay spectrum, the multipath delays of each cluster included in the received signal are determined. For example, there are two spectral peaks in the delay spectrum obtained according to η, namely spectral peak 1 and spectral peak 2. The cluster corresponding to spectral peak 1 is cluster 1, and the cluster corresponding to spectral peak 2 is cluster 2. If the time corresponding to spectral peak 1 is 1 ns and the time corresponding to spectral peak 2 is 10 ns, then the multipath delay of cluster 1 is 1 ns and the multipath delay of cluster 2 is 10 ns.

[0069] In the above multipath delay determination method, the received signal received by the receiving end is obtained, and the delay spectrum vector of the received signal is determined according to the received signal. The vector to be input is determined according to the real part and the imaginary part of the delay spectrum vector; the vector to be input is input into the target model to obtain the delay spectrum prediction vector of the received signal; the target model is obtained by training the initial model according to the true delay spectrum vectors of the received signal samples and the transmitted signal samples. The received signal samples are samples obtained by superimposing noises with different signal-to-noise ratios on the initial received signal samples, and the received signal samples are the signal samples received by the receiving end after the transmitting end transmits the transmitted signal samples; according to the delay spectrum prediction vector, the multipath delays of each cluster included in the received signal are determined. In this embodiment, the delay spectrum prediction vector of the received signal is determined through the target model, and the multipath delay is determined through this prediction vector, improving the efficiency of multipath delay determination.

[0070] Figure 3 is one of the schematic flowcharts of the target model training method provided by the embodiments of the present application. As Figure 3 shown, the above multipath delay determination method further includes:

[0071] S301. Input the delay spectrum vector of the received signal sample and the true delay spectrum vector of the transmitted signal sample into the initial model to obtain the predicted delay spectrum vector sample.

[0072] Exemplarily, let the time-delay spectrum vector of the received signal samples be Input into the initial model to obtain the predicted time-delay vector samples

[0073] S302. Train the initial model according to the predicted time-delay spectrum vector samples and the true time-delay spectrum vectors to obtain the target model.

[0074] Exemplarily, let The true time-delay spectrum vector of the corresponding transmitted signal be η1. If there are m received signal samples in total, the set of time-delay spectrum vectors corresponding to the received signal samples is The set of true time-delay spectrum vectors is D2 = {η1, η2, η3, ……, η m}, and the set of predicted time-delay spectrum vector samples is Train the initial model according to each predicted time-delay spectrum vector sample in the set D3 of predicted time-delay spectrum vector samples and each true time-delay spectrum vector in the set D2 of true time-delay spectrum vectors to obtain the target model.

[0075] In the embodiments of the present application, the time-delay spectrum vector of the received signal samples and the true time-delay spectrum vector of the transmitted signal samples are input into the initial model to obtain the predicted time-delay spectrum vector samples. The initial model is trained according to the predicted time-delay spectrum vector samples and the true time-delay spectrum vectors to obtain the target model. By training the initial model according to the predicted time-delay spectrum vector samples and the true time-delay spectrum vectors, the target model is obtained, improving the accuracy of determining the time-delay spectrum prediction vector using the target model.

[0076] Figure 4 This is the second flowchart of the target model training method provided by the embodiments of the present application. This embodiment relates to a possible implementation manner of how to train the initial model according to the predicted time-delay spectrum vector samples and the true time-delay spectrum vectors to obtain the target model. On the basis of the above embodiments, as Figure 4 shown, the above S302 includes:

[0077] S401. Determine the corresponding loss value according to the predicted time-delay spectrum vector samples and the true time-delay spectrum vectors.

[0078] Exemplarily, the loss value function is expressed by the following formula:

[0079]

[0080] Each predicted time-delay spectrum vector sample in the set D3 of predicted time-delay spectrum vector samples and each true time-delay spectrum vector η in the set D2 of true time-delay spectrum vectors mSubstitute into the above loss value function to determine the loss value loss.

[0081] S402. Train the initial model according to the loss value to obtain the target model.

[0082] Exemplarily, train the initial model DCN according to the loss value loss, continuously adjust the model parameters of the initial model DCN to minimize the loss value loss, and determine the model corresponding to the minimum loss value as the target model.

[0083] In the embodiments of the present application, the corresponding loss value is determined according to the predicted delay spectrum vector sample and the true delay spectrum vector, and the initial model is trained according to the loss value to obtain the target model. The accuracy of determining the delay spectrum prediction vector by using the target model is improved.

[0084] In one of the embodiments, the transmitted signal sample is determined according to the preset delay range of each cluster, the preset total number of each cluster, and the preset delay interval between two adjacent clusters in each cluster.

[0085] The transmitted signal sample is determined according to the preset delay range of each cluster, the preset total number of each cluster, and the preset delay interval between two adjacent clusters in each cluster.

[0086] Exemplarily, an OFDM signal is used as the transmitted signal sample. Here, 80 MHz is used as the signal bandwidth, and the corresponding number of frequency points N is 256. To facilitate the training of the initial model, the preset delay of each cluster in the transmitted signal can be set between 0 ns and 120 ns, the preset total number of clusters is set to 2, and the preset delay interval between two adjacent clusters in each cluster is set between 2 ns and 40 ns. According to the above preset conditions, the transmitted signal sample is set. The signal-to-noise ratio of the received signal sample is set to 5 different levels [-10 dB, -5 dB, 0 dB, 5 dB, 10 dB], and different types of random noise are set for the same received signal sample at the same signal-to-noise ratio to learn to resist the influence of noise. For example, if the received signal sample is X, random noise w1 is set for X at the signal-to-noise ratio level of -10 dB, and the obtained received signal sample is X1; random noise w2 is set for X at the signal-to-noise ratio level of -10 dB, and the obtained received signal sample is X2.

[0087] Optionally, other signals can also be used as the transmitted signal sample. The preset delay of each cluster in the transmitted signal can also be greater than 120 ns, the preset total number of clusters can be greater than 2, and the preset delay interval between two adjacent clusters in each cluster can be greater than 40 ns. This embodiment does not limit this here.

[0088] Figure 5It is a schematic flowchart of a method for determining a time-delay spectrum vector provided by an embodiment of the present application. The embodiment of the present application relates to a possible implementation manner of how to determine the time-delay spectrum vector of a received signal. On the basis of the above embodiment, as Figure 5 shown, the above S201 includes:

[0089] S501, obtain the transmitted signal corresponding to the received signal sent by the transmitter.

[0090] Exemplarily, an OFDM signal is used as the transmitted signal S(t), and the number of subcarriers corresponding to the transmitted signal S(t), that is, the number of frequency points, is N.

[0091] S502, perform time-frequency transformation processing on the received signal to obtain a first frequency-domain signal, and perform time-frequency transformation processing on the transmitted signal to obtain a second frequency-domain signal.

[0092] Exemplarily, it is assumed that there are a total of L clusters and K sub-paths in the received signal x(t), and the number of sub-paths in each cluster is denoted as K l , so the received signal x(t) can be expressed as:

[0093]

[0094] where α l,k represents the complex amplitude of the k-th sub-path in the l-th cluster, τ l,k represents the multi-path time delay of the k-th sub-path in the l-th cluster, s(t - τ l,k ) represents the transmitted signal, and v(t) represents the Gaussian white noise.

[0095] Sample the received signal x(t) at N points to obtain the received signal in the discrete time domain, which is expressed as:

[0096] x(t) = [x(t1), x(t2), x(t3), ……, x(t N )]

[0097] where N represents the number of frequency points.

[0098] Perform the fast Fourier transform algorithm on the received signal in the time domain to obtain the first frequency-domain signal X:

[0099] X = [x(f1), x(f2),..., x(f N )]

[0100]

[0101] where the noise z(f i ) is additive Gaussian white noise with a mean of 0 and a variance of Similarly, the transmission signal S(t) can be subjected to time-frequency transformation processing to obtain a second frequency-domain signal.

[0102] S503. Determine the time-delay covariance matrix according to the first frequency-domain signal and the second frequency-domain signal.

[0103] Exemplarily, the ratio obtained by dividing the first frequency-domain signal by the second frequency-domain signal is the channel response H

[0104] H can be expressed by the following relational expression:

[0105] H = Va + w

[0106] H = [H(f1), H(f2),..., H(f N )] Τ

[0107] V = [v(τ1), v(τ2),..., v(τ K )]

[0108]

[0109] a = [α1, α2,...α K Τ

[0110] w = [w(1), …, w(N)]

[0111] Among them, V represents the relational expression matrix with respect to τ k a represents the parameter matrix, and w represents the matrix composed of noise signals, and the elements in this matrix are all noise signals.

[0112] Let Γ = [Γ1, Γ2, … Γ P represent the discrete time-delay set obtained by sampling, ΔΓ represent the sampling interval, the true time-delay value is included in Γ, and has a medium-small quantization error. Then H can be re-expressed in an over-complete form:

[0113]

[0114] Among them In the case of Γ p = τ k the quantization error is small, and in other cases the corresponding time-delay covariance matrix is defined as R

[0115]

[0116] Among them represents the signal power on the p-th time-delay grid. Similarly, η = [η1, η2, …, η Pis the time-delay spectrum vector of the received signal, and I represents the identity matrix.

[0117] S504. Determine the time-delay spectrum vector of the received signal according to the time-delay covariance matrix.

[0118] Exemplarily, the n-th column of R can be expressed by the following relational expression:

[0119]

[0120] where A n = [v(Γ1)v Η (Γ1)e n , v(Γ2)v Η (Γ2)e n , …, v(Γ P )v Η (Γ P )e n , and e n is an N×1 vector.

[0121] Furthermore, R can be vectorized into y, which is expressed by the following relational expression:

[0122]

[0123] where

[0124] Optionally, in this solution, the time-delay covariance matrix can be obtained using a single snapshot, that is, a single measurement. Then the time-delay covariance matrix can be obtained through the following relational expression:

[0125]

[0126] After vectorizing we get which can be expressed by the following relational expression:

[0127]

[0128] where Δy is the estimation error of y, and the time-delay spectrum vector η of the received signal can be obtained according to the above relational expression.

[0129] In the embodiments of the present application, a transmission signal corresponding to a received signal sent by a transmitting end is obtained, the received signal is subjected to time-frequency transformation processing to obtain a first frequency-domain signal, and the transmission signal is subjected to time-frequency transformation processing to obtain a second frequency-domain signal. According to the first frequency-domain signal and the second frequency-domain signal, a time-delay covariance matrix is determined, and according to the relationship between the time-delay covariance matrix and the time-delay spectrum vector of the received signal, the time-delay spectrum vector of the received signal is determined. The conversion from the received signal to the time-delay spectrum vector of the received signal is realized.

[0130] In one of the embodiments, the initial model includes a preset number of convolutional layers, each convolutional layer uses a rectified linear unit as an activation function, and each convolutional layer is used to adjust the time-delay spectrum vector of the received signal.

[0131] The initial model includes a preset number of convolutional layers, each convolutional layer uses a rectified linear unit as an activation function, and each convolutional layer is used to adjust the time-delay spectrum vector of the received signal.

[0132] Exemplarily, the number of convolutional layers can be set to 6, and each convolutional layer uses a rectified linear unit (ReLU) as an activation function. The expression of this activation function is:

[0133] f(x) = max(0, x)

[0134] Set the input vector of the target model as Then the output of the k-th convolutional layer of the initial model is:

[0135] c = ReLU(Pa(W k *c k +b k ), k = 1, 2,... 6, where W k represents the convolution kernel of the k-th convolutional layer, and b k represents the bias of the k-th convolutional layer. * represents the convolution operator, and ReLU represents the activation function f(x).

[0136] The output of the fully connected layer can be represented by the following relational expression:

[0137]

[0138] where is the weight of the neuron of the fully connected layer, and is the bias of the neuron of the fully connected layer.

[0139] The performance of the target model obtained in this application when determining the multipath delay can be compared with the performance of the traditional technology when determining the multipath delay. As Figure 6 shown,Figure 6 This is a performance comparison diagram provided by an embodiment of the present application. The Root Mean Squared Error (RMSE) is used to estimate the performance comparison, which can be expressed by the following relational expression:

[0140]

[0141] Among them, represents the multipath delay value obtained by the target model DCN during the m c -th Monte Carlo simulation, τ represents the true delay value, L represents the number of clusters, and M c represents the number of Monte Carlo simulations. For the convenience of comparison, the delay interval between two clusters is selected to be greater than 12.5 ns, the search step of the Smoothing Multiple Signal Classification (SMUSIC) algorithm is set to 0.1 ns, the interpolation accuracy of the target model DCN is also 0.1 ns, and the cubic method is used for one-dimensional interpolation to reconstruct the delay spectrum. The SNR ranges from -10 dB to 10 dB, with a step of 1 dB. 1000 Monte Carlo simulations are performed under each SNR condition to eliminate random effects. The results are as Figure 6 shown. The target model DCN constructed in the present application is superior to the smooth MUSIC algorithm and the smooth ESPRIT algorithm in terms of performance, and is superior to the Maximum Likelihood (ML) algorithm when the signal-to-noise ratio exceeds 3 dB, obtaining the best estimation accuracy.

[0142] Compare the time used in the multipath delay determination process of the target model obtained in the present application with the time used in the traditional techniques of Smoothing Estimating signal parameters via rotational invariance techniques (Smoothing ESPRIT), the SMUSIC algorithm, and the Maximum Likelihood (ML) algorithm when determining the multipath delay of signals with different bandwidths, as shown in Table 1:

[0143] Table 1

[0144]

[0145] It should be understood that although the steps in the flowcharts involved in the above embodiments are sequentially shown according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear indication in this article, there is no strict order limit for the execution of these steps, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be executed alternately or in turn with at least a part of other steps or steps or stages in other steps.

[0146] Based on the same inventive concept, an embodiment of the present application further provides a multipath delay determination device for implementing the multipath delay determination method involved above. The implementation solution provided by this device to solve the problem is similar to the implementation solution described in the above method. Therefore, the specific limitations in one or more embodiments of the multipath delay determination device provided below can refer to the limitations on the multipath delay determination method in the above text, and will not be repeated here.

[0147] In one embodiment, as Figure 7 shown, a multipath delay determination device 700 is provided, including: an acquisition module 701, a first determination module 702, a first input module 703, and a second determination module 704, where:

[0148] The acquisition module 701 is configured to acquire the received signal received by the receiving end, and determine the delay spectrum vector of the received signal according to the received signal;

[0149] The first determination module 702 is configured to determine the vector to be input according to the real part and the imaginary part of the delay spectrum vector;

[0150] The first input module 703 is configured to input the vector to be input into the target model to obtain the delay spectrum prediction vector of the received signal; the target model is obtained by training the initial model according to the real delay spectrum vectors of the received signal samples and the transmitted signal samples. The received signal samples are the samples obtained by superimposing noises with different signal-to-noise ratios on the initial received signal samples. The received signal samples are the signal samples received by the receiving end after the transmitting end transmits the transmitted signal samples;

[0151] The second determination module 704 is configured to determine the multipath delays of each cluster included in the received signal according to the delay spectrum prediction vector.

[0152] In one of the embodiments, the multipath delay determination device 700 further includes:

[0153] The first input module is configured to input the time-delay spectrum vector of the received signal samples and the true time-delay spectrum vector of the transmitted signal samples into the initial model to obtain predicted time-delay spectrum vector samples;

[0154] The training module is configured to train the initial model according to the predicted time-delay spectrum vector samples and the true time-delay spectrum vectors to obtain a target model.

[0155] In one embodiment, the training module includes:

[0156] The first determination unit is configured to determine the corresponding loss value according to the predicted time-delay spectrum vector samples and the true time-delay spectrum vectors;

[0157] The training unit is configured to train the initial model according to the loss value to obtain a target model.

[0158] In one embodiment, the transmitted signal samples are determined according to the preset time-delay range of each cluster, the preset total number of each cluster, and the preset time-delay interval between two adjacent clusters in each cluster.

[0159] In one embodiment, the acquisition module 701 includes:

[0160] The first acquisition unit is configured to acquire the transmitted signal corresponding to the received signal sent by the transmitter;

[0161] The processing unit is configured to perform time-frequency transformation processing on the received signal to obtain a first frequency-domain signal, and perform time-frequency transformation processing on the transmitted signal to obtain a second frequency-domain signal;

[0162] The second determination unit is configured to determine the time-delay covariance matrix according to the first frequency-domain signal and the second frequency-domain signal;

[0163] The third determination unit is configured to determine the time-delay spectrum vector of the received signal according to the time-delay covariance matrix.

[0164] In one embodiment, the initial model includes a preset number of convolutional layers. Each convolutional layer uses a rectified linear unit as the activation function, and each convolutional layer is used to adjust the time-delay spectrum vector of the received signal.

[0165] Each module in the above multipath time-delay determination device can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in the processor of the computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to the above modules.

[0166] In one embodiment, a computer device is provided. The computer device can be a terminal, and its internal structure diagram can be as Figure 8As shown in the figure. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected via a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a mobile cellular network, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it realizes a method for determining the path delay.

[0167] Those skilled in the art can understand that Figure 8 the structure shown in the figure is only a block diagram of some structures related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0168] In one embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory. When the processor executes the computer program, the following steps are implemented:

[0169] Obtain the received signal received by the receiving end, and determine the delay spectrum vector of the received signal according to the received signal;

[0170] Determine the vector to be input according to the real part and the imaginary part of the delay spectrum vector;

[0171] Input the vector to be input into the target model to obtain the delay spectrum prediction vector of the received signal; the target model is obtained by training the initial model according to the real delay spectrum vectors of the received signal samples and the transmitted signal samples. The received signal samples are samples obtained by superimposing noises with different signal-to-noise ratios on the initial received signal samples. The received signal samples are the signal samples received by the receiving end after the transmitting end transmits the transmitted signal samples;

[0172] According to the delay spectrum prediction vector, determine the multipath delays of each cluster included in the received signal.

[0173] In one embodiment, when the processor executes the computer program, the following steps are also implemented:

[0174] Input the delay spectrum vector of the received signal sample and the real delay spectrum vector of the transmitted signal sample into the initial model to obtain a predicted delay spectrum vector sample;

[0175] Train the initial model according to the predicted delay spectrum vector sample and the real delay spectrum vector to obtain the target model.

[0176] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0177] Determine a corresponding loss value according to the predicted delay spectrum vector sample and the true delay spectrum vector;

[0178] Train the initial model according to the loss value to obtain the target model.

[0179] In one embodiment, the transmitted signal sample is determined according to the preset delay range of each cluster, the preset total number of each cluster, and the preset delay interval between two adjacent clusters in each cluster.

[0180] In one embodiment, when the processor executes the computer program, the following steps are further implemented:

[0181] Obtain the transmitted signal corresponding to the received signal sent by the transmitter;

[0182] Perform time-frequency transformation processing on the received signal to obtain a first frequency-domain signal, and perform time-frequency transformation processing on the transmitted signal to obtain a second frequency-domain signal;

[0183] Determine the delay covariance matrix according to the first frequency-domain signal and the second frequency-domain signal;

[0184] Determine the delay spectrum vector of the received signal according to the delay covariance matrix.

[0185] In one embodiment, the initial model includes a preset number of convolutional layers. Each convolutional layer uses a rectified linear unit as the activation function, and each convolutional layer is used to adjust the delay spectrum vector of the received signal.

[0186] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0187] Obtain the received signal received by the receiver, and determine the delay spectrum vector of the received signal according to the received signal;

[0188] Determine the vector to be input according to the real part and the imaginary part of the delay spectrum vector;

[0189] Input the vector to be input into the target model to obtain the delay spectrum prediction vector of the received signal; the target model is obtained by training the initial model according to the true delay spectrum vectors of the received signal samples and the transmitted signal samples. The received signal samples are samples obtained by superimposing noises with different signal-to-noise ratios on the initial received signal samples. The received signal samples are the signal samples received by the receiver after the transmitter sends the transmitted signal samples;

[0190] Determine the multipath delays of each cluster included in the received signal according to the delay spectrum prediction vector.

[0191] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0192] Input the time-delay spectrum vector of the received signal samples and the true time-delay spectrum vector of the transmitted signal samples into the initial model to obtain predicted time-delay spectrum vector samples;

[0193] Train the initial model according to the predicted time-delay spectrum vector samples and the true time-delay spectrum vectors to obtain a target model.

[0194] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0195] Determine the corresponding loss value according to the predicted time-delay spectrum vector samples and the true time-delay spectrum vectors;

[0196] Train the initial model according to the loss value to obtain a target model.

[0197] In one embodiment, the transmitted signal samples are determined according to the preset time-delay ranges of each cluster, the preset total number of each cluster, and the preset time-delay interval between two adjacent clusters in each cluster.

[0198] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented:

[0199] Obtain the transmitted signal corresponding to the received signal sent by the transmitter;

[0200] Perform time-frequency transformation processing on the received signal to obtain a first frequency-domain signal, and perform time-frequency transformation processing on the transmitted signal to obtain a second frequency-domain signal;

[0201] Determine the time-delay covariance matrix according to the first frequency-domain signal and the second frequency-domain signal;

[0202] Determine the time-delay spectrum vector of the received signal according to the time-delay covariance matrix.

[0203] In one embodiment, the initial model includes a preset number of convolutional layers. Each convolutional layer uses a rectified linear unit function as the activation function, and each convolutional layer is used to adjust the time-delay spectrum vector of the received signal.

[0204] In one embodiment, a computer program product is provided, including a computer program, which when executed by a processor, implements the following steps:

[0205] Obtain the received signal received by the receiver, and determine the time-delay spectrum vector of the received signal according to the received signal;

[0206] Determine the vector to be input according to the real part and the imaginary part of the time-delay spectrum vector;

[0207] Input the vector to be input into the target model to obtain the predicted vector of the delay spectrum of the received signal; the target model is obtained by training the initial model according to the true delay spectrum vectors of the received signal samples and the transmitted signal samples, the received signal samples are the samples obtained by superimposing noises with different signal-to-noise ratios on the initial received signal samples, and the received signal samples are the signal samples received at the receiving end after the transmitting end transmits the transmitted signal samples;

[0208] Determine the multipath delays of each cluster included in the received signal according to the predicted vector of the delay spectrum.

[0209] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented:

[0210] Input the delay spectrum vector of the received signal sample and the true delay spectrum vector of the transmitted signal sample into the initial model to obtain the predicted delay spectrum vector sample;

[0211] Train the initial model according to the predicted delay spectrum vector sample and the true delay spectrum vector to obtain the target model.

[0212] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented:

[0213] Determine the corresponding loss value according to the predicted delay spectrum vector sample and the true delay spectrum vector;

[0214] Train the initial model according to the loss value to obtain the target model.

[0215] In one embodiment, the transmitted signal sample is determined according to the preset delay range of each cluster, the preset total number of each cluster, and the preset delay interval between two adjacent clusters in each cluster.

[0216] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented:

[0217] Obtain the transmitted signal corresponding to the received signal sent by the transmitting end;

[0218] Perform time-frequency transformation processing on the received signal to obtain the first frequency-domain signal, and perform time-frequency transformation processing on the transmitted signal to obtain the second frequency-domain signal;

[0219] Determine the delay covariance matrix according to the first frequency-domain signal and the second frequency-domain signal;

[0220] Determine the delay spectrum vector of the received signal according to the delay covariance matrix.

[0221] In one embodiment, the initial model includes a preset number of convolutional layers, each convolutional layer uses the rectified linear unit function as the activation function, and each convolutional layer is used to adjust the delay spectrum vector of the received signal.

[0222] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0223] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the various embodiments provided in this application can include at least one of non-volatile and volatile memories. Non-volatile memory can include Read-Only Memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the various embodiments provided in this application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the various embodiments provided in this application can be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., without limitation.

[0224] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered to be within the scope described in this specification.

[0225] The above embodiments merely represent several implementation manners of the present application. The description thereof is relatively specific and detailed, but it should not be construed as a limitation to the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all fall within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.

Claims

1. A multipath delay determination method, comprising the following steps: Obtain the received signal received by the receiving end, and determine the delay spectrum vector of the received signal according to the received signal; Determine the vector to be input according to the real part and the imaginary part of the delay spectrum vector; Input the vector to be input into the target model to obtain the delay spectrum prediction vector of the received signal; The target model is obtained by training an initial model according to the real delay spectrum vectors of the received signal samples and the transmitted signal samples. The received signal samples are samples obtained by adding noises with different signal-to-noise ratios to the initial received signal samples. The received signal samples are the signal samples received by the receiving end after the transmitting end transmits the transmitted signal samples; Determine the multipath delays of each cluster included in the received signal according to the delay spectrum prediction vector; Wherein, the determining the delay spectrum vector of the received signal according to the received signal includes: Determine the transmitted signal corresponding to the received signal transmitted by the transmitting end; Perform time-frequency transformation processing on the received signal to obtain a first frequency-domain signal, and perform time-frequency transformation processing on the transmitted signal to obtain a second frequency-domain signal; Determine a delay covariance matrix according to the first frequency-domain signal and the second frequency-domain signal; Determine the delay spectrum vector of the received signal according to the delay covariance matrix.

2. The method according to claim 1, characterized in that, The method further includes: Input the delay spectrum vectors of the received signal samples and the real delay spectrum vectors of the transmitted signal samples into the initial model to obtain predicted delay spectrum vector samples; Train the initial model according to the predicted delay spectrum vector samples and the real delay spectrum vectors to obtain the target model.

3. The method according to claim 2, wherein The training the initial model according to the predicted delay spectrum vector samples and the real delay spectrum vectors to obtain the target model includes: Determine the corresponding loss value according to the predicted delay spectrum vector samples and the real delay spectrum vectors; Train the initial model according to the loss value to obtain the target model.

4. The method according to any one of claims 1-3, characterized in that, The transmitted signal samples are determined according to the preset delay ranges of each cluster, the preset total number of each cluster, and the preset delay intervals between adjacent clusters in each cluster.

5. The method according to any one of claims 1 to 3, characterized in that, The initial model includes a preset number of convolutional layers. Each convolutional layer uses a rectified linear unit as an activation function, and each convolutional layer is used to adjust the delay spectrum vector of the received signal.

6. A multipath delay determination device, characterized in that, The device includes: An acquisition module, configured to acquire the received signal received by the receiving end, and determine the delay spectrum vector of the received signal according to the received signal; A first determination module, configured to determine the vector to be input according to the real part and the imaginary part of the delay spectrum vector; A first input module, configured to input the vector to be input into the target model to obtain the delay spectrum prediction vector of the received signal; the target model is obtained by training an initial model according to the real delay spectrum vectors of the received signal samples and the transmitted signal samples. The received signal samples are samples obtained by adding noises with different signal-to-noise ratios to the initial received signal samples. The received signal samples are the signal samples received by the receiving end after the transmitting end transmits the transmitted signal samples; A second determination module, configured to determine the multipath delays of each cluster included in the received signal according to the delay spectrum prediction vector; Wherein, determining the delay spectrum vector of the received signal according to the received signal includes: Determining a transmitted signal corresponding to the received signal transmitted by the transmitting end; Performing time-frequency transformation processing on the received signal to obtain a first frequency-domain signal, and performing time-frequency transformation processing on the transmitted signal to obtain a second frequency-domain signal; Determining a delay covariance matrix according to the first frequency-domain signal and the second frequency-domain signal; Determining the delay spectrum vector of the received signal according to the delay covariance matrix.

7. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, the steps of the method according to any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

9. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images