A method, device and electronic device for identifying channel propagation status

Abstract

The present invention provides a method, device and electronic device for identifying the channel propagation state. The method comprises: extracting the CIR components of the channel impulse response CIR received by the signal receiving end in different polarization directions; obtaining the statistical characteristics between the multipaths according to the CIR components; performing data preprocessing on the CIR components and the statistical characteristics to obtain input data; processing the input data based on a convolutional neural network (CNN) model to obtain the channel propagation state. In an embodiment of the present application, the CIR components between the transmitting and receiving antennas in different polarization directions and the statistical characteristics between the multipaths are used as input data of the CNN model, and the channel propagation state is obtained based on the CNN model. By combining the CIR characteristics of the dual-polarization antenna with CNN, the channel propagation state can be accurately distinguished without relying on prior information.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, and in particular to a method, device and electronic equipment for identifying a channel propagation state. Background Art

[0002] In the field of indoor wireless positioning, many wireless positioning technologies are based on time of arrival (ToA), time difference of arrival (TDoA), direction of arrival (DoA) and received signal strength (RSS). Positioning technologies based on parameters such as ToA and DoA are very sensitive to the reliability of line of sight (LoS) signals. When the LoS signal is unavailable, that is, the system is in a non-line of sight (NLoS) state, a large number of positioning errors will be introduced into the entire system. For such applications in complex multipath propagation scenarios, relevant research is needed to mitigate the impact of NLoS propagation, and a key step is to identify the state of channel propagation.

[0003] Existing methods for identifying channel propagation states can be divided into two categories: one is based on statistical information identification, which is mainly based on the changes in the statistical characteristics of the signal caused by two different propagation states; the other is based on propagation path loss: this type of method extracts channel characteristic parameters, and realizes channel propagation state identification based on the distribution of channel characteristic parameters and hypothesis testing. Both of the above methods have high requirements for threshold division, and the threshold is sensitive to the environment. Sufficient prior information is needed to determine the appropriate threshold in order to achieve a high distinction accuracy. Summary of the invention

[0004] The purpose of the present invention is to provide a method, device and electronic device for identifying channel propagation status, so as to solve the problem that the existing method for identifying channel propagation status needs to rely on prior information.

[0005] An embodiment of the present invention provides a method for identifying a channel propagation state, comprising:

[0006] Extracting the CIR components of the channel impulse response (CIR) received by the receiving end in different polarization directions;

[0007] According to the CIR component, obtaining statistical characteristics between multipaths;

[0008] Performing data preprocessing on the CIR component and the statistical characteristics to obtain input data;

[0009] The input data is processed based on a convolutional neural network (CNN) model to obtain a channel propagation state.

[0010] Optionally, the CIR component includes:

[0011] The first CIR component generated by the vertically polarized array elements at the transmitting end and the receiving end;

[0012] The second CIR component generated by the array element with vertical polarization at the transmitting end and horizontal polarization at the receiving end;

[0013] The third CIR component generated by the array elements with horizontal polarization at the transmitting end and vertical polarization at the receiving end;

[0014] The fourth CIR component generated by the horizontally polarized array elements at the transmitting end and the receiving end.

[0015] Optionally, the statistical characteristics include: a mean and a variance of the CIR component.

[0016] Optionally, obtaining a statistical characteristic between multipaths according to the CIR component includes:

[0017] The average of the CIR components between the L multipaths is calculated. The calculation formula of the average is as follows:

[0018]

[0019] in, represents the mean value of the CIR component; |H rxu,txv | represents the CIR component generated by the array element with the receiving antenna polarization direction u and the transmitting antenna polarization direction v; L represents the number of multipaths, and l represents the lth path;

[0020] The variance of the CIR components between the L multipaths is calculated. The calculation formula of the variance is as follows:

[0021]

[0022] in, represents the variance of the CIR component; |H rxu,txv | represents the CIR component generated by the array element with the receiving antenna polarization direction u and the transmitting antenna polarization direction v; L represents the number of multipaths, and l represents the lth path.

[0023] Optionally, performing data preprocessing on the CIR component and the statistical characteristic to obtain input data includes:

[0024] The CIR components of different polarization directions and the statistical characteristics are divided into dimensions to obtain M×N dimensional input data.

[0025] Optionally, performing data preprocessing on the CIR component and the statistical characteristic to obtain input data includes:

[0026] The CIR components of different polarization directions, the statistical characteristics and the auxiliary features of the signal are divided into dimensions to obtain M×N dimensional input data.

[0027] Optionally, the auxiliary feature includes at least one of the following:

[0028] Rice Factor;

[0029] Signal rise time;

[0030] RMS delay spread;

[0031] Kurtosis of received power.

[0032] Optionally, the processing the input data based on a convolutional neural network (CNN) model to obtain a channel propagation state includes:

[0033] Inputting the input data into a convolutional neural network (CNN) model to obtain the probability of the channel propagation state;

[0034] The channel propagation state includes: a line-of-sight propagation state or a non-line-of-sight propagation state.

[0035] An embodiment of the present invention further provides an electronic device, comprising: a memory, a transceiver, and a processor:

[0036] A memory for storing a computer program; a transceiver for transmitting and receiving data under the control of the processor; and a processor for reading the computer program in the memory and performing the following operations:

[0037] Extracting CIR components of the channel impulse response CIR received by the signal receiving end in different polarization directions;

[0038] According to the CIR component, obtaining statistical characteristics between multipaths;

[0039] Performing data preprocessing on the CIR component and the statistical characteristics to obtain input data;

[0040] The input data is processed based on a convolutional neural network (CNN) model to obtain the channel propagation state.

[0041] Optionally, the CIR component includes:

[0042] The first CIR component generated by the vertically polarized array elements at the transmitting end and the receiving end;

[0043] The second CIR component generated by the array element with vertical polarization at the transmitting end and horizontal polarization at the receiving end;

[0044] The third CIR component generated by the array elements with horizontal polarization at the transmitting end and vertical polarization at the receiving end;

[0045] The fourth CIR component generated by the horizontally polarized array elements at the transmitting end and the receiving end.

[0046] Optionally, the statistical characteristics include: a mean and a variance of the CIR component.

[0047] Optionally, the processor is configured to read the computer program in the memory and perform the following operations:

[0048] The average of the CIR components between the L multipaths is calculated. The calculation formula of the average is as follows:

[0049]

[0050] in, represents the mean value of the CIR component; |H rxu,txv | represents the CIR component generated by the array element with the receiving antenna polarization direction u and the transmitting antenna polarization direction v; L represents the number of multipaths, and l represents the lth path;

[0051] The variance of the CIR components between the L multipaths is calculated. The calculation formula of the variance is as follows:

[0052]

[0053] in, represents the variance of the CIR component; |H rxu,txv | represents the CIR component generated by the array element with the receiving antenna polarization direction u and the transmitting antenna polarization direction v; L represents the number of multipaths, and l represents the lth path.

[0054] Optionally, the processor performs data preprocessing on the CIR component and the statistical characteristic to obtain input data, including:

[0055] The CIR components of different polarization directions and the statistical characteristics are divided into dimensions to obtain M×N dimensional input data.

[0056] Optionally, the processor performs data preprocessing on the CIR component and the statistical characteristic to obtain input data, including:

[0057] The CIR components of different polarization directions, the statistical characteristics and the auxiliary features of the signal are divided into dimensions to obtain M×N dimensional input data.

[0058] Optionally, the auxiliary feature includes at least one of the following:

[0059] Rice Factor;

[0060] Signal rise time;

[0061] RMS delay spread;

[0062] Kurtosis of received power.

[0063] Optionally, the processor processes the input data based on a convolutional neural network (CNN) model to obtain a channel propagation state, including:

[0064] Inputting the input data into a convolutional neural network (CNN) model to obtain the probability of the channel propagation state;

[0065] The channel propagation state includes: a line-of-sight propagation state or a non-line-of-sight propagation state.

[0066] An embodiment of the present invention further provides a device for identifying a channel propagation state, comprising:

[0067] An information extraction unit, used to extract CIR components of a channel impulse response CIR received by a signal receiving end in different polarization directions;

[0068] A first processing unit, configured to obtain a statistical characteristic between multipaths according to the CIR component;

[0069] A second processing unit, configured to perform data preprocessing on the CIR component and the statistical characteristics to obtain input data;

[0070] The state recognition unit is used to process the input data based on a convolutional neural network (CNN) model to obtain a channel propagation state.

[0071] An embodiment of the present invention further provides a processor-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned channel propagation state identification method are implemented.

[0072] The beneficial effects of the above technical solution of the present invention are:

[0073] In the embodiment of the present application, the CIR components between the transmitting and receiving antennas in different polarization directions and the statistical characteristics between the multipaths are used as the input data of the CNN model, and the channel propagation state is obtained based on the CNN model. By combining the CIR characteristics of the dual-polarized antenna with CNN, the channel propagation state can be accurately distinguished without relying on prior information. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 A schematic diagram showing a flow chart of a method for identifying a channel propagation state according to an embodiment of the present invention;

[0075] Figure 2 One of the schematic diagrams of the CNN model of an embodiment of the present invention is shown;

[0076] Figure 3 A schematic diagram showing a location configuration of a base station according to an embodiment of the present invention;

[0077] Figure 4 A second schematic diagram of a CNN model according to an embodiment of the present invention;

[0078] Figure 5 A second schematic diagram showing the location configuration of base stations according to an embodiment of the present invention;

[0079] Figure 6 A third schematic diagram of a CNN model showing an embodiment of the present invention;

[0080] Figure 7 A third schematic diagram showing the location configuration of a base station of the present invention;

[0081] Figure 8 A fourth schematic diagram of a CNN model showing an embodiment of the present invention;

[0082] Fig. 9 A schematic diagram showing the structure of a device for identifying a channel propagation state according to an embodiment of the present invention;

[0083] Fig.10 A schematic diagram showing the structure of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0084] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, the following will be described in detail in conjunction with the accompanying drawings and specific embodiments. In the following description, specific details such as specific configurations and components are provided only to help fully understand the embodiments of the present invention. Therefore, it should be clear to those skilled in the art that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present invention. In addition, for clarity and brevity, the description of known functions and structures is omitted.

[0085] It should be understood that the references to "one embodiment" or "an embodiment" throughout the specification mean that the specific features, structures, or characteristics associated with the embodiment are included in at least one embodiment of the present invention. Therefore, the references to "in one embodiment" or "in an embodiment" appearing throughout the specification do not necessarily refer to the same embodiment. In addition, these specific features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0086] In various embodiments of the present invention, it should be understood that the size of the serial numbers of the following processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0087] In the embodiments of the present invention, the term "and / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B may represent three situations: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the associated objects before and after are in an "or" relationship.

[0088] In the embodiments of the present application, the term "plurality" refers to two or more than two, and other quantifiers are similar.

[0089] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0090] When describing the embodiments of the present invention, some concepts used in the following description are first explained.

[0091] Specifically, the embodiments of the present invention provide a method, device and electronic device for identifying a channel propagation state, so as to solve the problem that the existing method for identifying a channel propagation state needs to rely on prior information.

[0092] like Figure 1 As shown, an embodiment of the present invention provides a method for identifying a channel propagation state, which specifically includes the following steps:

[0093] Step 101: extract CIR components of a channel impulse response CIR received by a signal receiving end in different polarization directions.

[0094] The receiving end is a signal receiver, which can be a base station or a terminal. The receiving end and the signal transmitting end are configured as dual-polarized antennas, and the dual-polarized antennas include horizontally polarized antennas and vertically polarized antennas. The horizontally polarized antenna and the vertically polarized antenna of the transmitting end send the same information in sequence, and the receiving end extracts the CIR components between the transmitting and receiving antennas in different polarization directions according to the CIRs received by its two polarized antennas.

[0095] Step 102: Obtain statistical characteristics between multipaths according to the CIR component.

[0096] The statistical characteristics may include the mean and variance of the CIR component, that is, the statistical characteristics between the multipaths are calculated for the CIR components extracted between the transmitting and receiving antennas of multiple different polarization directions. The number of multipaths can be estimated, for example, by using the minimum description length (MDL), Toeplitz matrix decomposition, etc. to detect the number of multipaths.

[0097] Step 103: perform data preprocessing on the CIR component and the statistical characteristic to obtain input data. The data preprocessing may be data shaping processing on the CIR component and the statistical characteristic to convert the original data into a matrix form as input data.

[0098] Step 104: Process the input data based on a convolutional neural network (CNN) model to obtain a channel propagation state.

[0099] The channel propagation state includes LoS and NLoS states. The CNN model is used to automatically extract input data features, and then the LoS\NLoS propagation state is identified. Among them, the CNN model is a training model obtained by using a large amount of M×N dimensional data with real LoS / NLoS labels as a training set, and continuously iterating so that the final loss function can converge to the minimum value. The input data obtained according to the CIR component and the statistical characteristics is input into the CNN model, and the recognition result of the channel propagation state between the receiving end and the transmitting end can be obtained.

[0100] In the prior art, when identifying the channel propagation state, it is necessary to extract sample features from multiple dimensions, set corresponding thresholds for different features, and judge the channel propagation state based on the thresholds. The setting of the thresholds needs to rely on sufficient prior information to determine the appropriate threshold value range, thereby achieving a higher distinction accuracy. In the embodiment of the present application, the CIR components between the transmitting and receiving antennas in different polarization directions and the statistical characteristics between the multipaths are used as input data of the CNN model, and the channel propagation state is obtained based on the CNN model. By combining the CIR characteristics of the dual-polarized antenna with CNN, the channel propagation state can be accurately distinguished without relying on prior information.

[0101] As an optional embodiment, the CIR component includes the following four CIR components of the transceiver antennas in different polarization directions:

[0102] (1) The first CIR component generated by the vertically polarized array elements at the transmitting end and the receiving end can be expressed as |H rxV,txV |;

[0103] (2) The second CIR component generated by the vertically polarized array element at the transmitting end and the horizontally polarized array element at the receiving end can be expressed as |H rxH,txV |;

[0104] (3) The third CIR component generated by the horizontally polarized array element at the transmitting end and the vertically polarized array element at the receiving end can be expressed as |H rxV,txH |;

[0105] (4) The fourth CIR component generated by the horizontally polarized array elements at the transmitting end and the receiving end can be expressed as |H rxH,txH |.

[0106] The statistical characteristics between the multipaths are calculated using the CIR components between the transmitting and receiving antennas in the four different polarization directions extracted above. The statistical characteristics may include: the mean and variance of the CIR components.

[0107] Specifically, obtaining the statistical characteristics between the multipaths according to the CIR components may include: calculating the mean of the CIR components between L multipaths, and the calculation formula of the mean is as follows:

[0108]

[0109] in, represents the mean value of the CIR component; |H rxu,txv | represents the CIR component generated by the array element with the receiving antenna polarization direction u and the transmitting antenna polarization direction v; L represents the number of multipaths, and l represents the lth path.

[0110] The variance of the CIR components between the L multipaths is calculated. The calculation formula of the variance is as follows:

[0111]

[0112] in, represents the variance of the CIR component; |H rxu,txv | represents the CIR component generated by the array element with the receiving antenna polarization direction u and the transmitting antenna polarization direction v; L represents the number of multipaths, and l represents the lth path.

[0113] As an optional embodiment, the performing data preprocessing on the CIR components and the statistical characteristics to obtain input data includes: dividing the CIR components and the statistical characteristics in different polarization directions into dimensions to obtain M×N dimensional input data.

[0114] In this embodiment, the data preprocessing is to convert the CIR components between the transmitting and receiving antennas in four different polarization directions, the statistical characteristics between the two multipaths (respectively and ) are sequentially reshaped into M×N dimensional data as the input of the CNN model.

[0115] As an optional embodiment, the data preprocessing of the CIR components and the statistical characteristics to obtain input data includes: dividing the CIR components of different polarization directions, the statistical characteristics and auxiliary features of the signal into dimensions to obtain M×N dimensional input data.

[0116] This embodiment adds auxiliary features when dividing the data dimensions. The auxiliary features may include at least one of the following:

[0117] Rice Factor;

[0118] Signal rise time;

[0119] RMS delay spread;

[0120] Kurtosis of received power.

[0121] In this embodiment, the input data of the CNN model can be divided into dimensions based on the CIR components of different polarization directions and the statistical characteristics, or based on the CIR components of different polarization directions, the statistical characteristics and auxiliary features. Among them, the channel information that can reflect the channel propagation state not only includes polarization information (i.e., the CIR components and the statistical characteristics), but also auxiliary information such as Rice factor, signal rise time, root mean square delay spread and kurtosis of received power. These auxiliary information can reflect the difference in statistical characteristics of LoS and NLoS paths to a certain extent, so when identifying the channel propagation state, the data is processed in combination with the auxiliary features as input data, and the input data containing the auxiliary features is processed based on the CNN model, so that a more accurate channel propagation state identification result can be obtained. In the simulation experiment, adding these auxiliary information and automatically extracting and identifying features through the CNN model can achieve a higher LoS / NLoS identification accuracy than simply using polarization information.

[0122] When dividing the data into dimensions, multiple data features are integrated and then divided into dimensions. The way of dividing the data into dimensions is, for example: each a*b dimension of the obtained CIR component is reshaped into 1*(a*b) dimension, and then arranged in a row in a certain order such as H11→H12→H21→H22 to obtain 1*(4*a*b) dimensional data, where H11, H12, H21, and H22 are CIR components of different polarization directions. When preprocessing the data, if there is other feature information for auxiliary identification (i.e., the auxiliary feature), it is reshaped into 1*c dimension and supplemented after the above-mentioned 1*(4*a*b) dimensional data. Finally, it is reshaped into M×N dimensional data, and the values ​​of M and N can be adjusted according to the results of multiple experiments.

[0123] As an optional embodiment, the processing of the input data based on the convolutional neural network (CNN) model to obtain the channel propagation state includes: inputting the input data into the convolutional neural network (CNN) model to obtain the probability of the channel propagation state; wherein the channel propagation state includes: line-of-sight propagation state or non-line-of-sight propagation state.

[0124] This embodiment uses the CNN model to automatically extract data features to identify the LoS / NLoS propagation state. Input data to the trained CNN model to obtain the recognition result of the channel propagation state, that is, the probability of the LoS\NLoS propagation state.

[0125] Among them, the CNN model is for example Figure 2As shown in the figure, the CNN structure consists of multiple convolutional layers, pooling layers, fully connected layers, and SoftMax layers. The convolutional layer performs convolution operations with the input data through the convolution kernel, extracts features and generates feature maps as the input of the next layer. It can be expressed by the following formula:

[0126]

[0127] Among them, M j represents the input feature map of the jth neuron, w is the convolution kernel connecting the (g-1)th layer to the gth layer, is the bias of the jth neuron in the gth layer, represents the characteristics of the i-th neuron in the (g-1)-th layer, Represents the characteristics of the jth neuron in the gth layer. f is the linear rectification function (ReLU) activation function, and the activation function formula is as follows:

[0128]

[0129] When x>0, the gradient is always 1, which solves the gradient dissipation problem of the back propagation algorithm when optimizing deep neural networks and converges quickly.

[0130] The convolution layer is followed by the pooling layer, which aggregates and counts the feature information of adjacent regions, uses probabilistic statistical features to reduce the dimension and reduce the amount of calculation. At the same time, the use of pooling reduces overfitting and the propagation of noise. After being processed by the convolution layer and the pooling layer, deeper features are extracted from the input data.

[0131] The fully connected layer performs weighted summation of the features of the previous layer (i.e., the pooling layer) and converts the output into a one-dimensional vector; the last fully connected layer uses the SoftMax activation function when outputting. Its main function is to combine the features extracted from the previous layers for classification. The output result can be interpreted as the probability of each category. The probability formula is as follows:

[0132]

[0133] Among them, X is the output of the previous layer (i.e., the output of the pooling layer), T represents the transpose, ω represents the weight of the neuron; R is the number of labels, r is the rth label, in this embodiment, R = 2, and the two labels are "LoS" and "NLoS" respectively; s represents the subscript of the predicted class, s = 1, 2; P represents the probability of LoS or NLoS classification.

[0134] In this embodiment, the CNN model can adopt: (1) dropout optimization. Introducing the dropout function in the fully connected layer can reduce the occurrence of overfitting; (2) Adam optimization. The setting of the learning rate determines whether the neural network can converge to the global minimum. Selecting a higher learning rate cannot converge to the global minimum. Selecting a smaller learning rate helps the neural network converge to the global minimum, but it takes longer to converge. The Adam optimization algorithm is an extension of the stochastic gradient descent method. It designs independent adaptive learning rates for different parameters by calculating the first-order moment estimate and the second-order moment estimate of the gradient, so that CNN can achieve good convergence results in a relatively short time. In addition, since there is no standard network architecture that can bring high accuracy in all test sets, different architectures can be tried through experiments to achieve better classification performance of CNN by adjusting the number of convolutional layers and pooling layers, and changing the size and number of convolution kernels.

[0135] The training of the CNN model requires the collection of a large amount of M×N dimensional data with real LoS / NLoS labels. A part of the collected M×N dimensional data is used as a training set, and it is continuously iterated so that the final loss function can converge to the minimum value; the remaining part of the data is used as a test set to test whether the trained model achieves the desired effect. The required CNN model can be obtained by saving the weights of each layer of the trained CNN network.

[0136] Optionally, after the input data is processed based on the convolutional neural network CNN model and the channel propagation state is obtained, the accuracy of the recognition result can be determined. For example: the receiving end CIR of the known actual channel propagation state is used as the input data of the CNN model. After obtaining the recognition result of the channel propagation state, it is compared with the actual channel propagation state to determine the accuracy. Taking the channel propagation state recognition result as LoS as an example, B CIRs with the largest power can be selected from the CIRs of A base stations, and the probability of the CNN model predicting LoS is sorted, and C base stations (BaseStation, BS) with the highest LoS probability are selected to be judged as BSs containing LoS, where A≥B≥C.

[0137] The implementation process of the channel propagation state identification method of the present application is described below through specific embodiments.

[0138] Example 1:

[0139] (1) Taking the signal transmitting end as a base station as an example, the base station location is configured. The base station location is configured as an indoor factory (Indoor Office, IOO) scene with sparse clusters and a relatively high base station height; IOO: L = 120m, W = 50m, H = 3m, such as Figure 3 shown.

[0140] (2) The configuration of the receiving and transmitting antennas is as follows:

[0141] The transmitter is configured as a 4×4 dual-polarized antenna with a polarization angle of + / -45°;

[0142] The receiving end is configured as a dual-polarized antenna with 2×2 vertical / horizontal polarization angles.

[0143] (3) The structure of the CNN model is as follows Figure 4 As shown:

[0144] A 5-layer CNN structure is used, which includes 2 convolutional layers superimposed on the pooling layer, 2 fully connected layers and a SoftMax function output layer, such as Figure 4 As shown. The convolution layer uses a 5×5 convolution kernel with a step size of 1, and the pooling kernel uses a 2×2 maximum pooling kernel with a step size of 2, both of which adopt a zero padding strategy. The ReLu nonlinear activation function is used after each convolution layer. The first convolution layer contains 32 convolution kernels, the input data dimension is 8×9, the output dimension is 8×9×32, and the dimension after the maximum pooling layer is 4×5×32; the second convolution layer contains 64 convolution kernels, the output dimension is 4×5×64, and the output dimension after the maximum pooling is 2×3×64; the first layer of the fully connected layer includes 384 neurons, the second layer includes 64 neurons, and the output layer includes 2 neurons. The Softmax activation function normalizes the output so that it ranges between (0, 1).

[0145] (4) The steps of identifying the channel propagation state include:

[0146] Step 1: The horizontal polarization antenna and the vertical polarization antenna of the transmitter send the same information in sequence. The receiver extracts the CIR components of four different polarization directions based on the information received by its own two polarization antennas: |H rxV,txV |、|H rxH,txV |、|H rxV,txH |、|H rxH,txH |; Calculate the statistical characteristics of multipath for the CIR components extracted between the transmitting and receiving antennas in four different polarization directions: mean and variance The calculation formula is not described in detail.

[0147] Step 2: Data preprocessing: The CIR components between the transmitting and receiving antennas in four different polarization directions and the statistical characteristics between the two multipaths ( and ) are sequentially reshaped into 8×9 dimensions as the input of the CNN model.

[0148] Step 3: Input the data into the trained CNN model to identify the LoS\NLoS propagation status.

[0149] The training of the CNN model includes: the terminal randomly scatters 600 points within the range of 12 base stations, collects 7200 8×9 dimensional data described in step 2, and uses the corresponding real LoS / NLoS propagation state information as the data label, and then uses 6000 of the collected data as the training set, inputs them into the CNN network for continuous iteration, so that the final loss function can converge to the minimum value. The remaining 1200 data are used as a test set to test whether the trained model achieves the desired effect. The CNN model required for step 3 can be obtained by saving the weights of each layer of the trained CNN network.

[0150] Step 4 (optional): Data post-processing: Select the 10 CIRs with the highest power from the CIRs of the 12 base stations, sort the probability results predicted by CNN as LOS, select the 3 BSs with the highest LoS probability as the BSs containing LoS, and calculate their accuracy.

[0151] Example 2:

[0152] (1) Taking the signal transmitter as a base station as an example, the base station location is first configured. The base station location is configured as an indoor factory with dense clutter and high basestation height (InF-DH) scenario; InF-DH: L = 120m W = 60m D = 20m, such as Figure 5 shown.

[0153] (2) The configuration of the receiving antenna and the transmitting antenna is as follows:

[0154] The transmitter is configured as a 4×4 dual-polarized antenna with a polarization angle of + / -45°;

[0155] The receiving end is configured as a dual-polarized antenna with 2×2 vertical / horizontal polarization angles.

[0156] (3) CNN model structure, such as Figure 6 shown.

[0157] This example uses a 6-layer CNN structure, which includes 3 convolutional layers, 2 fully connected layers, and a SoftMax function output layer. Figure 6As shown. The convolutional kernel used in the convolutional layer is a 3×3 convolutional kernel with a step size of 1, and the pooling kernel is a 2×2 maximum pooling kernel with a step size of 2, both of which adopt the zero padding strategy. The ReLu nonlinear activation function is used after each convolutional layer. The first convolutional layer contains 32 convolutional kernels, the input data dimension is 9×9, the output dimension is 9×9×32, and the dimension after the maximum pooling layer is 5×5×32; the second convolutional layer contains 64 convolutional kernels, the output dimension is 5×5×64, and the output dimension after the maximum pooling is 3×3×64; the third convolutional layer contains 64 convolutional kernels, and the pooling output dimension is 2×2×64; the first layer of the fully connected layer has 256 neurons, the second layer has 64 neurons, and the output layer has 2 neurons. The Softmax activation function normalizes the output so that its range is between (0, 1).

[0158] (4) Steps for channel propagation state identification:

[0159] Step 1: The horizontally polarized antenna and the vertically polarized antenna at the transmitting end send the same information in sequence. The receiving end extracts the CIR components between the transmitting and receiving antennas in four different polarization directions based on the information received by its own two polarized antennas: |H rxV,txV |、|H rxH,txV |、|H rxV,txH |、|H rxH,txH |; Then calculate the statistical characteristics of multipath for the CIR components extracted between the transmitting and receiving antennas in four different polarization directions: mean and variance The calculation formula is not described in detail.

[0160] Step 2: Data preprocessing: The CIR components between the transmitting and receiving antennas in four different polarization directions and the statistical characteristics between the two multipaths ( and ) and the Rice factors are sequentially reshaped into 9×9 dimensional data as the input of the CNN model.

[0161] Step 3: Input the data into the trained CNN model to identify the LoS\NLoS propagation status.

[0162] The training of the CNN model includes: the terminal randomly scatters 600 points within the range of 18 base stations, collects 10,800 9×9 dimensional data described in step 2, and uses the corresponding real LoS / NLoS propagation state information as the data label, and then uses 9,000 of the collected data as the training set, inputs them into the CNN network for continuous iteration, so that the final loss function can converge to the minimum value. The remaining 1,800 data are used as a test set to test whether the trained model achieves the desired effect. The CNN model required for step 3 can be obtained by saving the weights of each layer of the trained CNN network.

[0163] Step 4 (optional): Data post-processing: Select the 10 CIRs with the highest power from the 18 base station CIRs, sort the probability results predicted by CNN as LoS, select the 3 BSs with the highest LoS probability as the BSs containing LoS, and calculate their accuracy.

[0164] The simulation results show that in this embodiment, the accuracy of selecting the three BSs with the highest LoS probabilities as the BSs containing the LoS path can reach 98.3%.

[0165] Example 3:

[0166] (1) Taking the signal transmitter as a base station as an example, the base station location is first configured. The base station location is configured as an indoor factory InF-SH scenario with sparse clusters and a relatively high base station height; InF-SH: L = 300 mW = 150 m D = 50 m, such as Figure 7 shown.

[0167] (2) The configuration of the receiving antenna and the transmitting antenna is as follows:

[0168] The transmitter is configured as a 4×4 dual-polarized antenna with a polarization angle of + / -45°;

[0169] The receiving end is configured as a dual-polarized antenna with 2×2 vertical / horizontal polarization angles.

[0170] (3) CNN model structure, such as Figure 8 shown.

[0171] This example uses a 6-layer CNN structure, which includes 3 convolutional layers, 2 fully connected layers, and a SoftMax function output layer. Figure 8As shown. The convolutional kernel used in the convolutional layer is a 5×5 convolution kernel with a step size of 1, and the pooling kernel is a 2×2 maximum pooling kernel with a step size of 2, both of which adopt the zero padding strategy. The ReLu nonlinear activation function is used after each convolutional layer. The first convolutional layer contains 64 convolutional kernels, the input data dimension is 9×9, the output dimension is 9×9×64, and the dimension after the maximum pooling layer is 5×5×64. The second convolutional layer contains 64 convolutional kernels, the output dimension is 5×5×64, and the output dimension after the maximum pooling is 3×3×64; the third convolutional layer contains 64 convolutional kernels, and the pooling output dimension is 2×2×64; the first layer of the fully connected layer has 256 neurons, the second layer has 64 neurons, and the output layer has 2 neurons. The Softmax activation function normalizes the output so that its range is between (0, 1).

[0172] (4) Steps for channel propagation state identification:

[0173] Step 1: The horizontal polarization antenna and vertical polarization antenna of the sender send the same information in sequence. The receiver extracts four CIR components in different polarization directions based on the information received by its own two polarization antennas: |H rxV,txV |、|H rxH,txV |、|H rxV,txH |、|H rxH,txH |; Then calculate the statistical characteristics of multipath for the CIR components extracted between the transmitting and receiving antennas in four different polarization directions: and variance The calculation formula is not described in detail.

[0174] Step 2: Data preprocessing: The CIR components between the transmitting and receiving antennas in four different polarization directions and the statistical characteristics between the two multipaths ( and ), Rice factor, and RMS delay spread are sequentially reshaped into 9×9 dimensional data as the input of the CNN model.

[0175] Step 3: Input the data into the trained CNN model to identify the LoS\NLoS propagation status.

[0176] The training of the CNN model includes: the terminal randomly scatters 600 points within the range of 18 base stations, collects 10,800 9×9 dimensional data described in step 2, and uses the corresponding real LoS / NLoS propagation state information as the data label, and then uses 9,000 of the collected data as the training set, inputs them into the CNN network for continuous iteration, so that the final loss function can converge to the minimum value. The remaining 1,800 data are used as a test set to test whether the trained model achieves the desired effect. The CNN model required for step 3 can be obtained by saving the weights of each layer of the trained CNN network.

[0177] Step 4 (optional): Data post-processing: Select the 10 CIRs with the highest power from the 18 base station CIRs, sort the probability results predicted by CNN as LoS, select the 3 BSs with the highest LoS probability as the BSs containing LoS, and calculate their accuracy.

[0178] The simulation results show that in this embodiment, the accuracy of selecting the three BSs with the highest LoS probabilities as the BSs containing the LoS path can reach 99%.

[0179] It should be noted that, when preprocessing the data, in addition to the above-mentioned Rice factor and RMS delay spread, one or more of the other auxiliary features can be added on the basis of the CIR component and statistical characteristics for processing as the input data of the CNN model, for example: using the CIR components of four different polarization directions, two statistical characteristics, Rice factor, signal rise time, RMS delay spread and kurtosis of received power to divide the data dimensions, and the method of dividing the data dimensions is not described here. Since the auxiliary features can reflect the difference in the statistical characteristics of the LoS and NLoS paths to a certain extent, when identifying the channel propagation state, adding these auxiliary information and automatically extracting and identifying features through the CNN model can achieve a higher LoS / NLoS identification accuracy than using polarization information alone.

[0180] In the embodiment of the present application, the CIR components between the transmitting and receiving antennas in different polarization directions and the statistical characteristics between the multipaths are used as the input data of the CNN model, and the channel propagation state is obtained based on the CNN model. By combining the CIR characteristics of the dual-polarized antenna with CNN, the channel propagation state can be accurately distinguished without relying on prior information.

[0181] The above embodiments introduce the positioning method of the present invention. The following embodiments will further illustrate the corresponding device in conjunction with the accompanying drawings.

[0182] Specifically, Fig. 9 As shown, an embodiment of the present invention provides a channel propagation state identification device 900, including:

[0183] The information extraction unit 910 is used to extract CIR components of the channel impulse response CIR received by the signal receiving end in different polarization directions;

[0184] A first processing unit 920 is used to obtain a statistical characteristic between multipaths according to the CIR component;

[0185] A second processing unit 930 is used to perform data preprocessing on the CIR component and the statistical characteristics to obtain input data;

[0186] The state recognition unit 940 is used to process the input data based on a convolutional neural network (CNN) model to obtain a channel propagation state.

[0187] Optionally, the CIR component includes:

[0188] The first CIR component generated by the vertically polarized array elements at the transmitting end and the receiving end;

[0189] The second CIR component generated by the array element with vertical polarization at the transmitting end and horizontal polarization at the receiving end;

[0190] The third CIR component generated by the array elements with horizontal polarization at the transmitting end and vertical polarization at the receiving end;

[0191] The fourth CIR component generated by the horizontally polarized array elements at the transmitting end and the receiving end.

[0192] Optionally, the statistical characteristics include: a mean and a variance of the CIR component.

[0193] Optionally, the first processing unit 920 is specifically configured to:

[0194] The average of the CIR components between the L multipaths is calculated. The calculation formula of the average is as follows:

[0195]

[0196] in, represents the mean value of the CIR component; |H rxu,txv | represents the CIR component generated by the array element with the receiving antenna polarization direction u and the transmitting antenna polarization direction v; L represents the number of multipaths, and l represents the lth path;

[0197] The variance of the CIR components between the L multipaths is calculated. The calculation formula of the variance is as follows:

[0198]

[0199] in, represents the variance of the CIR component; |Hrxu,txv | represents the CIR component generated by the array element with the receiving antenna polarization direction u and the transmitting antenna polarization direction v; L represents the number of multipaths, and l represents the lth path.

[0200] Optionally, the second processing unit 930 is specifically used to: divide the CIR components of different polarization directions and the statistical characteristics into dimensions to obtain M×N dimensional input data.

[0201] Optionally, the second processing unit 930 is specifically used to: divide the CIR components of different polarization directions, the statistical characteristics and auxiliary features of the signal into dimensions to obtain M×N dimensional input data.

[0202] Optionally, the auxiliary feature includes at least one of the following:

[0203] Rice Factor;

[0204] Signal rise time;

[0205] RMS delay spread;

[0206] Kurtosis of received power.

[0207] Optionally, the state recognition unit is specifically used to: input the input data into a convolutional neural network (CNN) model to obtain the probability of the channel propagation state;

[0208] The channel propagation state includes: a line-of-sight propagation state or a non-line-of-sight propagation state.

[0209] In the embodiment of the present application, the CIR components between the transmitting and receiving antennas in different polarization directions and the statistical characteristics between the multipaths are used as the input data of the CNN model, and the channel propagation state is obtained based on the CNN model. By combining the CIR characteristics of the dual-polarized antenna with CNN, the channel propagation state can be accurately distinguished without relying on prior information.

[0210] It should be noted here that the above-mentioned device provided in the embodiment of the present invention can implement all the method steps implemented in the above-mentioned method embodiment, and can achieve the same technical effect. The parts and beneficial effects that are the same as the method embodiment in this embodiment will not be described in detail here.

[0211] It should be noted that the division of units in the embodiments of the present application is schematic and is only a logical function division. There may be other division methods in actual implementation. In addition, each functional unit in each embodiment of the present application may be integrated into a processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0212] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a processor-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) or a processor (processor) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), disk or optical disk and other media that can store program codes.

[0213] like Fig.10 As shown, an embodiment of the present invention further provides an electronic device, including: a memory 1020, a transceiver 1000, and a processor 1010; wherein the memory 1020 is used to store a computer program; the transceiver 1000 is used to send and receive data under the control of the processor 1010; the processor 1010 is used to read the computer program in the memory and perform the following operations:

[0214] Extracting CIR components of the channel impulse response CIR received by the signal receiving end in different polarization directions;

[0215] According to the CIR component, obtaining statistical characteristics between multipaths;

[0216] Performing data preprocessing on the CIR component and the statistical characteristics to obtain input data;

[0217] The input data is processed based on a convolutional neural network (CNN) model to obtain the channel propagation state.

[0218] Optionally, the CIR component includes:

[0219] The first CIR component generated by the vertically polarized array elements at the transmitting end and the receiving end;

[0220] The second CIR component generated by the array element with vertical polarization at the transmitting end and horizontal polarization at the receiving end;

[0221] The third CIR component generated by the array elements with horizontal polarization at the transmitting end and vertical polarization at the receiving end;

[0222] The fourth CIR component generated by the horizontally polarized array elements at the transmitting end and the receiving end.

[0223] Optionally, the statistical characteristics include: a mean and a variance of the CIR component.

[0224] Optionally, the processor 1010 is configured to read the computer program in the memory and perform the following operations:

[0225] The average of the CIR components between the L multipaths is calculated. The calculation formula of the average is as follows:

[0226]

[0227] in, represents the mean value of the CIR component; |H rxu,txv | represents the CIR component generated by the array element with the receiving antenna polarization direction u and the transmitting antenna polarization direction v; L represents the number of multipaths, and l represents the lth path;

[0228] The variance of the CIR components between the L multipaths is calculated. The calculation formula of the variance is as follows:

[0229]

[0230] in, represents the variance of the CIR component; |H rxu,txv | represents the CIR component generated by the array element with the receiving antenna polarization direction u and the transmitting antenna polarization direction v; L represents the number of multipaths, and l represents the lth path.

[0231] Optionally, the processor 1010 performs data preprocessing on the CIR component and the statistical characteristic to obtain input data, including:

[0232] The CIR components of different polarization directions and the statistical characteristics are divided into dimensions to obtain M×N dimensional input data.

[0233] Optionally, the processor 1010 performs data preprocessing on the CIR component and the statistical characteristic to obtain input data, including:

[0234] The CIR components of different polarization directions, the statistical characteristics and the auxiliary features of the signal are divided into dimensions to obtain M×N dimensional input data.

[0235] Optionally, the auxiliary feature includes at least one of the following:

[0236] Rice Factor;

[0237] Signal rise time;

[0238] RMS delay spread;

[0239] Kurtosis of received power.

[0240] Optionally, the processor processes the input data based on a convolutional neural network (CNN) model to obtain a channel propagation state, including:

[0241] Inputting the input data into a convolutional neural network (CNN) model to obtain the probability of the channel propagation state;

[0242] The channel propagation state includes: a line-of-sight propagation state or a non-line-of-sight propagation state.

[0243] In the embodiment of the present application, the CIR components between the transmitting and receiving antennas in different polarization directions and the statistical characteristics between the multipaths are used as the input data of the CNN model, and the channel propagation state is obtained based on the CNN model. By combining the CIR characteristics of the dual-polarized antenna with CNN, the channel propagation state can be accurately distinguished without relying on prior information.

[0244] Among them, Fig.10 In the embodiment, the bus architecture may include any number of interconnected buses and bridges, specifically linking together various circuits of one or more processors represented by processor 1000 and memory represented by memory 1020. The bus architecture may also link together various other circuits such as peripherals, voltage regulators, and power management circuits, which are well known in the art and are therefore not further described herein. The bus interface provides an interface. The transceiver 1010 may be a plurality of components, namely, a transmitter and a transceiver, providing a unit for communicating with various other devices on a transmission medium. The processor 1000 is responsible for managing the bus architecture and general processing, and the memory 1020 may store data used by the processor 1000 when performing operations.

[0245] The processor 1010 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or a complex programmable logic device (CPLD). The processor may also adopt a multi-core architecture.

[0246] It should be noted here that the above-mentioned electronic device provided by the embodiment of the present invention can implement all the method steps implemented by the above-mentioned method embodiment, and can achieve the same technical effect. The parts and beneficial effects that are the same as the method embodiment in this embodiment will not be described in detail here.

[0247] In addition, a specific embodiment of the present invention further provides a processor-readable storage medium on which a computer program is stored, wherein the program, when executed by the processor, implements the steps of the method for identifying the channel propagation state as described above. And the same technical effect can be achieved, so in order to avoid repetition, it will not be repeated here. Among them, the readable storage medium can be any available medium or data storage device that can be accessed by the processor, including but not limited to magnetic storage (such as floppy disk, hard disk, magnetic tape, magneto-optical disk (MO), etc.), optical storage (such as CD, DVD, BD, HVD, etc.), and semiconductor storage (such as ROM, EPROM, EEPROM, non-volatile memory (NAND FLASH), solid-state drive (SSD)), etc.

[0248] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) that contain computer-usable program code.

[0249] The present application is described with reference to the flowchart and / or block diagram of the method, device (system) and computer program product according to the embodiment of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, and the combination of the process and / or box in the flowchart and / or block diagram can be implemented by computer executable instructions. These computer executable instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device for implementing a process or multiple processes in the flowchart and / or a box or multiple boxes in the block diagram.

[0250] These processor-executable instructions may also be stored in a processor-readable memory that can direct a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the processor-readable memory produce a product including an instruction device that implements the functions specified in one or more processes in the flowchart and / or one or more boxes in the block diagram.

[0251] These processor-executable instructions may also be loaded onto a computer or other programmable data processing device so that a series of operational steps are performed on the computer or other programmable device to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more processes in the flowchart and / or one or more boxes in the block diagram.

[0252] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for identifying a channel propagation state, characterized in that: include: Extracting CIR components of a channel impulse response CIR received by a signal receiving end in different polarization directions, wherein the signal receiving end is configured as a dual-polarization antenna; According to the CIR component, obtaining statistical characteristics between multipaths; Performing data preprocessing on the CIR component and the statistical characteristics to obtain input data; The input data is processed based on a convolutional neural network (CNN) model to obtain the channel propagation state.

2. The method according to claim 1, characterized in that The CIR components include: The first CIR component generated by the vertically polarized array elements at the transmitting end and the receiving end; The second CIR component generated by the array element with vertical polarization at the transmitting end and horizontal polarization at the receiving end; The third CIR component generated by the array elements with horizontal polarization at the transmitting end and vertical polarization at the receiving end; The fourth CIR component generated by the horizontally polarized array elements at the transmitting end and the receiving end.

3. The method according to claim 1, characterized in that The statistical characteristics include: a mean value and a variance of the CIR component.

4. The method according to claim 3, characterized in that The obtaining, according to the CIR component, a statistical characteristic between multipaths includes: The average of the CIR components between the L multipaths is calculated. The calculation formula of the average is as follows: in, represents the mean value of the CIR component; |H rxu,txv | represents the CIR component generated by the array element with the receiving antenna polarization direction u and the transmitting antenna polarization direction v; L represents the number of multipaths, and l represents the lth path; The variance of the CIR components between the L multipaths is calculated. The calculation formula of the variance is as follows: in, represents the variance of the CIR component; |H rxu,txv | represents the CIR component generated by the array element with the receiving antenna polarization direction u and the transmitting antenna polarization direction v; L represents the number of multipaths, and l represents the lth path.

5. The method according to claim 1, characterized in that: The performing data preprocessing on the CIR component and the statistical characteristic to obtain input data includes: The CIR components of different polarization directions and the statistical characteristics are divided into dimensions to obtain M×N dimensional input data.

6. The method according to claim 1, characterized in that The performing data preprocessing on the CIR component and the statistical characteristic to obtain input data includes: The CIR components of different polarization directions, the statistical characteristics and the auxiliary features of the signal are divided into dimensions to obtain M×N dimensional input data.

7. The method according to claim 6, characterized in that The auxiliary features include at least one of the following: Rice Factor; Signal rise time; RMS delay spread; Kurtosis of received power.

8. The method according to claim 1, characterized in that: The step of processing the input data based on a convolutional neural network (CNN) model to obtain a channel propagation state includes: Inputting the input data into a convolutional neural network (CNN) model to obtain the probability of the channel propagation state; The channel propagation state includes: a line-of-sight propagation state or a non-line-of-sight propagation state.

9. An electronic device, characterized in that: include: Memory, transceiver, processor: Memory for storing computer programs; a transceiver, for transmitting and receiving data under the control of the processor; A processor is configured to read the computer program in the memory and perform the following operations: Extracting CIR components of a channel impulse response CIR received by a signal receiving end in different polarization directions, wherein the signal receiving end is configured as a dual-polarization antenna; According to the CIR component, obtaining statistical characteristics between multipaths; Performing data preprocessing on the CIR component and the statistical characteristics to obtain input data; The input data is processed based on a convolutional neural network (CNN) model to obtain the channel propagation state.

10. The electronic device according to claim 9, characterized in that: The CIR components include: The first CIR component generated by the vertically polarized array elements at the transmitting end and the receiving end; The second CIR component generated by the array element with vertical polarization at the transmitting end and horizontal polarization at the receiving end; The third CIR component generated by the array elements with horizontal polarization at the transmitting end and vertical polarization at the receiving end; The fourth CIR component generated by the horizontally polarized array elements at the transmitting end and the receiving end.

11. The electronic device according to claim 9, characterized in that: The statistical characteristics include: a mean value and a variance of the CIR component.

12. The electronic device according to claim 11, characterized in that: The processor is configured to read the computer program in the memory and perform the following operations: The average of the CIR components between the L multipaths is calculated. The calculation formula of the average is as follows: in, represents the mean value of the CIR component; |H rxu,txv | represents the CIR component generated by the array element with the receiving antenna polarization direction u and the transmitting antenna polarization direction v; L represents the number of multipaths, and l represents the lth path; The variance of the CIR components between the L multipaths is calculated. The calculation formula of the variance is as follows: in, represents the variance of the CIR component; |H rxu,txv | represents the CIR component generated by the array element with the receiving antenna polarization direction u and the transmitting antenna polarization direction v; L represents the number of multipaths, and l represents the lth path.

13. The electronic device according to claim 9, characterized in that: The processor performs data preprocessing on the CIR component and the statistical characteristic to obtain input data, including: The CIR components of different polarization directions and the statistical characteristics are divided into dimensions to obtain M×N dimensional input data.

14. The electronic device according to claim 9, characterized in that: The processor performs data preprocessing on the CIR component and the statistical characteristic to obtain input data, including: The CIR components of different polarization directions, the statistical characteristics and the auxiliary features of the signal are divided into dimensions to obtain M×N dimensional input data.

15. The electronic device according to claim 14, characterized in that: The auxiliary features include at least one of the following: Rice Factor; Signal rise time; RMS delay spread; Kurtosis of received power.

16. The electronic device according to claim 9, characterized in that: The processor processes the input data based on a convolutional neural network (CNN) model to obtain a channel propagation state, including: Inputting the input data into a convolutional neural network (CNN) model to obtain the probability of the channel propagation state; The channel propagation state includes: a line-of-sight propagation state or a non-line-of-sight propagation state.

17. A device for identifying channel propagation status, characterized in that: include: An information extraction unit, used to extract CIR components of a channel impulse response CIR received by a signal receiving end in different polarization directions, wherein the signal receiving end is configured as a dual-polarization antenna; A first processing unit, configured to obtain a statistical characteristic between multipaths according to the CIR component; A second processing unit, configured to perform data preprocessing on the CIR component and the statistical characteristics to obtain input data; The state recognition unit is used to process the input data based on a convolutional neural network (CNN) model to obtain a channel propagation state.

18. A processor-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 for identifying the channel propagation state according to any one of claims 1 to 8 are implemented.

Images