Downlink channel estimation method, device, communication equipment and storage medium

Through the neural network model, the uplink channel and downlink channel state information is processed, which solves the problem of large downlink channel estimation overhead in the FDD system, and realizes accurate estimation of downlink channel state information and improves the stability of data transmission.

CN112039807BActive Publication Date: 2024-11-12ZTE CORP +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202010899719.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-08-31
Publication Date
2024-11-12
Estimated Expiration
2040-08-31

AI Technical Summary

Technical Problem

In the FDD system, the downlink channel estimation problem requires a lot of downlink training and uplink feedback, resulting in huge overhead, which is not suitable for large-scale MIMO systems, and the base station cannot obtain downlink channel status information through the uplink channel, affecting the data The accuracy of the sending.

Method used

通过使用预设的神经网络模型,将上行信道状态信息和部分下行信道状态信息进行处理,基于神经网络模型的输出结果完成下行信道估计,提高数据发送的准确性。

Benefits of technology

It realizes accurate estimation of downlink channel state information, reduces data transmission errors, improves the stability and accuracy of data transmission, and is suitable for large-scale MIMO systems.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN112039807B_ABST
    Figure CN112039807B_ABST
Patent Text Reader

Abstract

The embodiment of the present application provides a downlink channel estimation method, apparatus, communication equipment and storage medium, wherein the method comprises: inputting uplink channel state information and part of downlink channel state information into a preset neural network model; completing downlink channel estimation according to the output result of the neural network model; wherein the preset neural network model is generated based on uplink channel state information samples and part of downlink channel state information samples training. In the embodiment of the present application, the accuracy of downlink channel estimation is improved by predicting the downlink channel through uplink channel state information and part of downlink channel state information through a preset neural network model, which can improve the accuracy and stability of data transmission in the communication system.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of wireless communications, and in particular to a downlink channel estimation method, device, communication equipment and storage medium. Background Art

[0002] As a key technology in 5G or 6G communication systems, massive multi-input multi-output (MIMO) technology has advantages such as higher spectrum efficiency, energy efficiency, and high spatial resolution compared to traditional MIMO systems. Therefore, higher requirements are placed on the accuracy of downlink channel state information (CSI). In frequency division duplexing (FDD) systems, base stations cannot obtain downlink channels through uplink channels. The estimation problem of downlink channels in FDD systems becomes particularly important. For existing MIMO systems, downlink CSI is trained through a large number of downlink channel pilots, and the downlink CSI is fed back to the base station through the uplink. The overhead required for downlink training and uplink feedback is huge and is not suitable for massive MIMO systems. Summary of the invention

[0003] The embodiments of the present application provide a downlink channel estimation method, apparatus, communication device and storage medium to achieve accurate estimation of downlink channel state information and improve the accuracy of data transmission.

[0004] In order to achieve the above object, an embodiment of the present application provides a downlink channel estimation method, the method comprising:

[0005] Inputting the uplink channel state information and part of the downlink channel state information into a preset neural network model;

[0006] Complete downlink channel estimation according to the output result of the neural network model;

[0007] The preset neural network model is generated through training based on uplink channel state information samples and part of downlink channel state information samples.

[0008] In order to achieve the above object, an embodiment of the present application provides a downlink channel estimation device, which includes:

[0009] A model testing module, used for inputting uplink channel state information and part of downlink channel state information into a preset neural network model;

[0010] A channel estimation module, used to complete downlink channel estimation according to the output result of the neural network model;

[0011] The preset neural network model is generated through training with uplink channel state information samples and part of downlink channel state information samples.

[0012] In order to achieve the above object, an embodiment of the present application provides a communication device, which includes:

[0013] one or more processors;

[0014] A memory for storing one or more programs;

[0015] When the one or more programs are executed by the one or more processors, the one or more processors implement the downlink channel estimation method as described in any one of the embodiments of the present application.

[0016] In order to achieve the above-mentioned purpose, an embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, a downlink information estimation method as described in any one of the embodiments of the present application is implemented.

[0017] In an embodiment of the present application, a preset neural network model is used to process input uplink channel state information and part of the downlink channel state information, wherein the preset neural network model is generated through training of uplink channel state information samples and part of the downlink channel state information samples, and downlink channel estimation is completed based on the output result of the neural network model, thereby achieving accurate estimation of the downlink channel state information and improving the accuracy of data transmission. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is a flow chart of a downlink channel estimation method provided by an embodiment of the present application;

[0019] Figure 2 is an example diagram of a preset neural network model provided in an embodiment of the present application;

[0020] Figure 3 is a flowchart of another downlink channel estimation method provided by an embodiment of the present application;

[0021] Figure 4 is a flowchart of another downlink channel estimation method provided by an embodiment of the present application;

[0022] Figure 5 is a flowchart of another downlink channel estimation method provided by an embodiment of the present application;

[0023] Figure 6 is an example diagram of a preset neural network model provided in an embodiment of the present application;

[0024] Figure 7It is a schematic structural diagram of a downlink channel estimation device provided by an embodiment of the present application;

[0025] Figure 8 It is a schematic structural diagram of a communication device provided by an embodiment of the present application. Specific embodiments

[0026] To make the objectives, technical solutions, and advantages of the present application clearer and more understandable, the embodiments of the present application will be described in detail below with reference to the accompanying drawings. It should be noted that, without conflict, the embodiments and features in the embodiments of the present application can be combined arbitrarily with each other.

[0027] Figure 1 It is a flowchart of a downlink channel estimation method provided by an embodiment of the present application. The embodiments of the present application can be applicable to the situation of predicting the downlink channel state information of the downlink channel. This method can be executed by a downlink channel estimation device and can be implemented in a software and / or hardware manner. Generally, it can be integrated in a base station. Refer to Figure 1 The method provided by the embodiments of the present application specifically includes the following steps:

[0028] Step 110: Input the uplink channel state information and partial downlink channel state information into a preset neural network model.

[0029] Among them, the uplink channel state information may be information on the channel state of the uplink communication link. The uplink channel state information may include a channel matrix, multipath delay, Doppler frequency offset, the rank of the MIMO channel, a beamforming vector, etc. The partial downlink channel state information may be channel information corresponding to the communication links of a part of the antennas in the base station's corresponding antennas. Specifically, the partial downlink channel state information may be a pilot. In the embodiments of the present application, the antenna used for communication is an array antenna, and the number of array antennas of this antenna is M; h(fd) is the downlink channel state information, and fd is the channel frequency of the downlink channel; h′(fd) is the channel state information of the partial downlink channel, and fd is the channel frequency of the downlink channel. len(h′(fd)) < len(h(fd)), and len(*) represents the length of the vector *. The partial downlink channel state information may be the channel state information of one or more downlink channels among all the downlink channels of the base station. The neural network model may be a neural network that has been pre-trained. After being trained with the uplink channel state information and the partial downlink channel state information, the neural network model can determine the numerical value of the corresponding downlink channel state information through the processing of the input information.

[0030] Specifically, the uplink channel state information may be attribute information of the user's data transmitted in the uplink channel obtained by the base station. When performing downlink channel estimation, the obtained uplink channel state information and part of the downlink channel state information may be input as input parameters into a preset neural network model, wherein the uplink channel state information and part of the downlink channel state information may be input into the input layer of the preset neural network model, and the amount of input data of the input layer may be adjusted by the degree of difference between the signal source and the array antenna in the wireless communication system. Furthermore, in an embodiment of the present application, the uplink channel state information input into the preset neural network model may be the channel state information of all uplink channels or the channel state information of part of the uplink channels. It can be understood that the more uplink channels corresponding to the uplink channel state channels input into the preset neural network model, the more accurate the downlink channel estimation can be.

[0031] Step 120: complete downlink channel estimation according to the output result of the neural network model; wherein the preset neural network model is generated based on the training of uplink channel state information samples and part of the downlink channel state information samples.

[0032] Among them, the output result can be the result of the neural network model processing the uplink channel state information and part of the downlink channel state information. The output result can correspond to the channel state information of all downlink channels. The output result can specifically be an identification number of the downlink channel state information or a specific value of the downlink channel state information.

[0033] In an embodiment of the present application, an output result of the neural network model output can be obtained, and the output result can be specifically a vector, and the output result can correspond to all downlink channels, and the output result can be composed of the channel state information of the downlink channel or the identification number of the downlink channel state information, and the downlink channel state information corresponding to the output result can be found for downlink channel transmission. Among them, the neural network model can be trained and generated by uplink channel state information samples and part of the downlink channel state information samples, and the uplink channel state information samples can be the uplink channel state information of a threshold number collected in advance, and the uplink channel state information in the uplink channel state information samples can correspond to all uplink channels of the base station, and part of the downlink channel state information samples can be the channel state information of the downlink channels collected in advance. After the neural network model is initially established, the neural network model can be trained according to a large number of training samples.

[0034] In an embodiment of the present application, input uplink channel state information and part of the downlink channel state information are processed through a preset neural network model, wherein the preset neural network model is generated through training of uplink channel state information samples and part of the downlink channel state information samples, and downlink channel estimation is completed based on the output result of the neural network model, thereby achieving accurate estimation of the downlink channel state information and improving the accuracy of data transmission.

[0035] Furthermore, based on the above-mentioned application embodiment, the neural network model at least includes a downlink module for processing the partial downlink channel state information, an uplink module for processing the uplink channel state information and a fusion module for fusing the feature vectors output by the uplink module and the downlink module.

[0036] Among them, the downlink module can be a neural network model that processes downlink channel state information, the uplink module can be a neural network model that processes uplink channel state information, the fusion module can be a neural network model that fuses the information output by the downlink module and the uplink module, the feature vector can be a vector composed of feature data, and the feature vector can be the output result of the uplink module and the downlink module.

[0037] In the embodiment of the present application, the preset neural network model may be composed of a downlink module, an uplink module and a fusion module. The modules constituting the preset neural network model may be independent neural network models or neural network models with associated relationships. For example, when the uplink module processes the uplink channel state information, the output result of the downlink module may be used as an input parameter of the uplink module when processing the uplink channel state information. Exemplarily, Figure 2 is an example diagram of a preset neural network model provided in an embodiment of the present application, see Figure 2 A neural network model may include an uplink module 201, a downlink module 202 and a fusion module 203. The uplink module 201 and the downlink module 202 may be relatively independent neural network models. The uplink module 201 does not use the information of the downlink module 202 when processing the uplink channel state information. Similarly, the downlink module 202 does not use the information of the uplink module 201 when processing the downlink channel state information. The feature vectors generated by the uplink module 201 and the downlink module 202 after processing the uplink channel state information and part of the downlink channel state information respectively are input to the fusion module 203, and the fusion module 203 fuses the input feature vectors. The fusion method may include using a deep neural network to process the feature vectors and performing weighted average processing on the feature vectors.

[0038] Figure 3is a flowchart of another downlink channel estimation method provided by an embodiment of the present application. The embodiment of the present application is a specific embodiment based on the above-mentioned embodiment of the application. When inputting a preset neural network model, uplink channel state information and part of downlink channel state information are converted. See Figure 3 , the method provided in the embodiment of the present application comprises the following steps:

[0039] Step 210: convert the uplink channel state information and part of the downlink channel state information according to a preset mapping relationship.

[0040] Among them, the preset mapping relationship can be a mapping relationship that converts part of the downlink channel state information and the uplink channel state information into an input vector. For example, the values ​​of each channel state information can be arranged as a feature vector in a preset order, or the values ​​of each channel state information can be arranged as a feature vector after being weighted.

[0041] Specifically, part of the downlink channel state information and uplink channel state information to be input into the preset neural network model can be converted into feature vectors according to a preset mapping relationship. It can be understood that part of the downlink channel state information and the uplink channel state information can correspond to one feature vector, or they can correspond to one feature vector respectively.

[0042] Step 220: Input the converted uplink channel state information and part of the downlink channel state information into the hidden layers of the uplink module and the downlink module respectively, so as to be processed by using the hidden layers and the activation functions corresponding to the hidden layers.

[0043] The hidden layer may be a layer in the neural network model that processes and calculates data, and the preset neural network model may include one or more hidden layers. The activation function may be a rule used in the hidden layer to process and calculate data, and the activation function may specifically include a sigmoid function, a tanh function, a relu function, and a leaky relu function.

[0044] Specifically, part of the downlink channel state information and the uplink channel state information converted by the preset mapping relationship can be input as feature vectors into the hidden layers of the uplink module and the downlink module respectively, and the uplink channel state information and part of the downlink channel state information can be processed by activation functions in the hidden layers of the uplink module and the downlink module.

[0045] Step 230: Input the processing results of each hidden layer into the fusion module through the output layers of the upstream module and the downstream module, and generate the output result in the fusion module.

[0046] In an embodiment of the present application, the uplink module and the downlink module may respectively include multiple hidden layers, the output layers of the uplink module and the downlink module may output their respective processing results, the fusion module may fuse the processing results corresponding to the uplink module and the downlink module respectively, and the processing results after fusion by the fusion module may be used as the output results of the preset neural network model.

[0047] Step 240: Use the channel state information corresponding to the output result as an estimate of the downlink channel.

[0048] In an embodiment of the present application, uplink channel state information and part of downlink channel state information are converted according to a preset mapping relationship, and the converted uplink channel state information and part of the downlink channel state information are input into a hidden layer of an uplink module and a hidden layer of a downlink module for processing, wherein a corresponding activation function is provided in each hidden layer, and the processing result of the hidden layer is input into a fusion module through the output layer of the uplink module and the downlink module for fusion to generate an output result, and the output result is used as the corresponding channel state information as an estimate of the downlink channel, thereby achieving accurate estimation of the downlink channel state information in the MIMO system and improving the stability of data transmission.

[0049] Further, based on the above application embodiment, the preset mapping relationship includes a corresponding relationship between a complex number and the real part and imaginary part of the complex number.

[0050] In an embodiment of the present application, the preset mapping relationship for converting the uplink channel state information and part of the downlink channel state information is the correspondence between a complex number and the real part and the imaginary part of the complex number. The numerical values ​​corresponding to the uplink channel state information and part of the downlink channel state information can be used as complex numbers to obtain the corresponding real part and imaginary part respectively. The real part and imaginary part corresponding to the uplink channel state information and part of the downlink channel state information can be used as the feature vector of the preset neural network model input.

[0051] In an exemplary embodiment, R[x] and S[x] represent the real part and imaginary part of x respectively, and the preset mapping relationship can be expressed as γ: x→(R(x T ), S(x T )) T .

[0052] Further, based on the above application embodiment, the processing results of each hidden layer are input into the fusion module through the output layer of the uplink module and the downlink module, and the output result is generated in the fusion module, including:

[0053] In the fusion module, the processing results output by the uplink module and the downlink module are spliced ​​according to preset rules; the spliced ​​processing results are processed through at least one hidden layer in the fusion module and the activation function corresponding to the hidden layer to generate an output result.

[0054] Among them, the preset rule can be a rule for splicing the processing results of the uplink module and the processing results of the downlink module. For example, the processing results are weighted averaged, the calculation result can be used as the spliced ​​processing result, and the numerical values ​​corresponding to the same channel attributes in the processing results of the uplink module and the processing results of the downlink module can be spliced ​​in the same row or column.

[0055] In an embodiment of the present application, a preset rule for splicing the processing results of the uplink module and the processing results of the downlink module may be pre-set in the fusion module. The processing results output by the uplink module and the downlink module are spliced ​​according to the preset rules, and the activation function in the fusion module processes the spliced ​​processing results to generate an output result corresponding to the downlink channel state information.

[0056] Figure 4 is a flowchart of another downlink channel estimation method provided in an embodiment of the present application. The embodiment of the present application is a specific embodiment based on the above embodiment. Figure 4 , the method provided in the embodiment of the present application comprises the following steps:

[0057] Step 310: Generate a neural network model based on uplink channel state information samples and part of downlink channel state information samples.

[0058] The sources of the uplink channel state information samples and part of the downlink channel state information samples may include a local base station, all base stations within the cell where the local base station is located, or any base station.

[0059] In an embodiment of the present application, uplink channel state information and part of downlink channel state information, as well as downlink channel state information labels corresponding to each uplink channel state information and downlink channel state information labels corresponding to each part of downlink channel state information can be collected in advance, wherein the downlink channel state information label can be the downlink channel state information actually used by the base station under the uplink channel state information, and the downlink channel state information actually used by the base station under the part of downlink channel state information. The neural network model is trained by the acquired uplink channel state information samples and part of the downlink channel state information samples, and the neural network model may include an input layer, an output layer, a hidden layer, etc.

[0060] Step 320: Input the uplink channel state information and part of the downlink channel state information into a preset neural network model.

[0061] Step 330: complete downlink channel estimation according to the output result of the neural network model; wherein the preset neural network model is generated based on uplink channel state information samples and part of the downlink channel state information samples training.

[0062] Figure 5 is a flowchart of another downlink channel estimation method provided in an embodiment of the present application. The embodiment of the present application specifies the process of training and generating a neural network model based on the above embodiment. Figure 5 , the method provided in the embodiment of the present application comprises the following steps:

[0063] Step 410: Collect some downlink channel state information samples and uplink channel state information samples and corresponding downlink channel state information labels.

[0064] The downlink channel state information label may be the downlink channel state information actually used by the base station under the uplink channel state information, and the downlink channel state information actually used by the base station under the partial downlink channel state information.

[0065] The sources of the uplink channel state information samples and part of the downlink channel state information samples may include channel state information collected by a local base station, all base stations within a cell where the local base station is located, or any base station.

[0066] Specifically, uplink channel state information and channel state information of the downlink channel determined by the uplink channel state information can be collected in the base station, and the channel state information can be used as the corresponding downlink channel state information label. Similarly, part of the downlink channel state information and channel state information of the downlink channel determined by the part of the downlink channel state information can be collected in the base station, and the channel state information can be used as the corresponding downlink channel state information label.

[0067] Step 420: Convert each portion of downlink channel state information in the partial downlink channel state information sample and each uplink channel state information in the uplink channel state information into an uplink feature vector and a downlink feature vector respectively according to a preset mapping relationship.

[0068] In an embodiment of the present application, the collected partial downlink channel state information samples and the uplink channel state information samples may be preprocessed to convert each channel state information into a feature vector, the feature vector corresponding to the uplink channel state information may be an uplink feature vector, and the feature vector corresponding to the partial downlink channel state information may be a downlink feature vector. The preset mapping relationship may include a correspondence between a complex value and its real part and its imaginary part.

[0069] Step 430: Combine each uplink feature vector, each downlink feature vector and the corresponding downlink channel state information label into a training set; and train the neural network model according to the training set.

[0070] In an embodiment of the present application, each uplink feature vector and the corresponding downlink channel state information label, as well as each downlink feature vector and the corresponding downlink channel state information label can be used as a training set, and the training set can be used to train and generate a neural network model.

[0071] Step 440: Input the uplink channel state information and part of the downlink channel state information into a preset neural network model.

[0072] Step 450, complete downlink channel estimation according to the output result of the neural network model; wherein the preset neural network model is generated based on the training of uplink channel state information samples and part of the downlink channel state information samples.

[0073] Further, based on the above application embodiment, the training of the neural network model according to the training set includes:

[0074] The uplink feature vector and the downlink feature vector in the training set are input into the neural network model, and the predicted value of the output downlink channel state information is obtained; the loss value of the predicted value and the corresponding downlink channel state information label is determined according to the preset loss function; the training of the neural network model is completed when the loss value is determined to be less than the preset threshold, otherwise, the uplink feature vector and the downlink feature vector in the training set are continued to be input into the neural network module for training.

[0075] Specifically, the uplink feature vector and the downlink feature vector are respectively input into the neural network model for processing, and the processing result output by the neural network model can be obtained. The processing result can be a predicted value of the downlink channel state information, and the loss value between the predicted value and the corresponding downlink state information label is determined. The calculation process of the loss value can be determined by a preset loss function. When the loss value is less than a preset threshold, it can be considered that the neural network training is completed. Otherwise, continue to obtain the channel uplink feature vector and the downlink feature vector and input them into the neural network model for training. The training process can repeat the above process until the loss value is less than the preset threshold.

[0076] In an exemplary embodiment, Figure 6 is an example diagram of a preset neural network model provided in an embodiment of the present application, Figure 6 The structure of the preset deep neural network model is given. Among them, Respectively represent the real and imaginary parts of x, Indicates that a mapping operation is performed.

[0077] Specifically, the working process of the downlink module is as follows:

[0078] First, Recorded as xP, xP is input into the downstream module through the input layer, and the feature vector processed by the hidden layer is obtained by using the activation function corresponding to each hidden layer. The downstream module has a total of L P Layers, including 1 input layer, 1 output layer and L P -2 hidden layers. L P It can be set to 5, which means it includes three hidden layers. The feature vector of the hidden layer can be expressed as:

[0079]

[0080] in, are the parameters that the network needs to train. is the nonlinear transformation function of the lth layer, which can be further improved as follows:

[0081]

[0082] in, and are the weight matrix, bias parameter and nonlinear activation function of the corresponding layer respectively. The activation function is the LeakyReLU function, that is, [g re (z)] p =max{[z] p / a, [z] p}, a>1where [z] p represents the p-th element of vector z, where p=1, 2, ..., len(z), and len(z) represents the length of vector z.

[0083] The feature vector processed by the hidden layer is input to the output layer of the downstream module, and the output feature vector is obtained by using the activation function of the output layer.

[0084] Based on the content of the above embodiment, the function of the uplink module is the same as the processing process of the downlink module, and the function of the uplink module is to process the uplink channel state information and obtain the output feature vector corresponding to the uplink channel state information, including: performing a preset mapping operation on the vector containing the uplink channel state information to obtain the corresponding feature vector; inputting the feature vector into the uplink module to obtain the corresponding output feature vector.

[0085] Optionally, the preset mapping in the embodiment is to convert the uplink channel information h(f U ) is mapped to The network structure of the uplink module is similar to that of the downlink module. The input vector is The specific working process will not be described here.

[0086] Based on the content of the above embodiment, the function of the fusion module is to fuse the output feature vectors of the downlink module and the uplink module and then process them to obtain the output vector. The parameters of the downlink channel state information can be predicted based on the output vector to complete the downlink channel estimation. Specifically, the output feature vector of the downlink module and the output feature vector of the uplink module can be fused using a preset fusion function to obtain a preset feature vector; the feature vector is input into a preset fusion module network to obtain the output vector, and the downlink channel parameters are predicted based on the output vector to complete the downlink channel estimation.

[0087] Specifically, the fusion module can be implemented in the following manner:

[0088] The output feature vectors of the downstream module and the upstream module are fused:

[0089]

[0090] where F P (x P ) is the output feature vector of the downstream module; is the connection function; F U (x U ) is the output feature vector of the upstream module; x Fus is the fused feature vector. Fus After being input into the fusion module network, it passes through L Fus After the hidden layer, we get the vector z Fus , z Fus The output vector is obtained after the input is sent to the output layer and the output layer activation function is applied. The downlink channel parameters are predicted based on the output vector to complete the downlink channel estimation.

[0091] Based on the content of the above embodiment, before the uplink channel state information and part of the downlink channel state information are input into the preset deep neural network model, it also includes: obtaining multiple information samples of the uplink channel state information and part of the downlink channel state information, and the downlink channel state information label corresponding to each information sample; after performing the above preset mapping operation on each uplink channel state information and part of the downlink channel state information, the feature vector corresponding to the uplink channel state information and part of the downlink channel state information is obtained; the combination of the feature vector corresponding to each uplink channel state information and part of the downlink channel state information and the downlink channel state information label is used as a training sample, thereby obtaining multiple training samples, and the above deep neural network model is trained using the multiple training samples.

[0092] First, obtain multiple partial downlink channel state information samples And, a plurality of uplink channel state information samples h(f U), and obtain the corresponding downlink channel state information h(f D ),Will As the corresponding part of the downlink channel information and uplink channel information h(f U ) tag.

[0093] Secondly, h(f U ) and the corresponding The combination of labels is taken as a sample, thus obtaining multiple training samples. h(f U ) is mapped and input into the constructed deep neural network model, and the relevant parameters of the deep neural network model are adjusted according to the output results to complete the training of the deep neural network model, thereby obtaining the above-mentioned preset deep data network model.

[0094] Based on the content of the above embodiment, the deep neural network model is trained using the multiple training samples, including: inputting any part of the downlink channel information and the feature vector of the uplink channel information sample into the deep neural network model, and outputting the predicted value of the downlink channel state information parameter; using a preset loss function to calculate the loss value according to the predicted value of the downlink channel state information parameter and the downlink channel state information label corresponding to the sample; if the loss value is less than the preset threshold, the deep neural network model training is completed.

[0095] First, select any part of the downlink channel state information parameter from the pilot signal and the uplink channel information sample and uplink channel state information parameter h(f U ),Will h(f U ) After completing the corresponding mapping, it is input into the deep neural network model, and is output from the output layer after nonlinear transformation of each hidden layer and its activation function. The predicted value of The predicted values ​​are respectively and respectively the labels corresponding to the training samples, and the loss value corresponding to the loss function is calculated.

[0096] For example, for a regression network, the loss function can be the L2 norm function, that is:

[0097]

[0098] Where V is the number of samples in a single batch, v represents the sequence number of the sample in the batch, is the prediction vector output by the network, and y(v) represents the label corresponding to the training sample, that is,

[0099] In the training phase, the deep neural network model performs layer-by-layer nonlinear transformation on the input sample data to obtain the predicted output, and calculates the loss value corresponding to the loss function according to the corresponding label. The loss function can be gradually optimized through the adaptive matrix estimation (ADAM) algorithm, thereby continuously optimizing and updating the network parameters until the loss function converges.

[0100] After the training is completed, the parameters of the deep neural network model remain unchanged, and the predicted values ​​​​are obtained using the input of the test set. The performance is evaluated by calculating the error between the predicted value and the label value.

[0101] Before training the deep neural network model, it is necessary to train the network of the downlink module, the network of the uplink module, and the network of the fusion module respectively. The training process of the downlink module network, the uplink module network, and the fusion module network is similar to the training process of the deep neural network model, and will not be repeated here. Among them, the input value of the downlink module network is Tagged with ; The input values ​​and labels of the upstream module network are ; The input value of the fusion module network is F P (x P ), F U (x U ), tagged The training order of all the above-mentioned deep neural network models is: first complete the training process of the downlink module network, the uplink module network, and the fusion module network, and then complete the training process of the neural network model.

[0102] Figure 7 is a structural diagram of a downlink channel estimation device provided in an embodiment of the present application, Figure 7 The device shown in the figure can execute the downlink channel estimation method provided in the embodiment of the present application, and specifically executes the functional modules and beneficial effects corresponding to the method. The device can be implemented by software and / or hardware, and specifically includes: a model testing module 501 and a channel estimation module 502.

[0103] The model testing module 501 is used to input the uplink channel state information and part of the downlink channel state information into a preset neural network model.

[0104] The channel estimation module 502 is used to complete the downlink channel estimation according to the output result of the neural network model; wherein the preset neural network model is generated by training the uplink channel state information samples and part of the downlink channel state information samples.

[0105] In an embodiment of the present application, the input uplink channel state information and part of the downlink channel state information are processed through the neural network model preset by the model testing module, wherein the preset neural network model is generated through training of uplink channel state information samples and part of the downlink channel state information samples, and the channel estimation module completes the downlink channel estimation based on the output result of the neural network model, thereby achieving accurate estimation of the downlink channel state information and improving the accuracy of data transmission.

[0106] Furthermore, based on the above-mentioned application embodiment, the neural network model in the downlink channel estimation device includes at least a downlink module for processing the partial downlink channel state information, an uplink module for processing the uplink channel state information, and a fusion module for fusing the feature vectors output by the uplink module and the downlink module.

[0107] Further, based on the above application embodiment, the model testing module 501 includes:

[0108] The parameter conversion unit is used to convert the uplink channel state information and the part of the downlink channel state information according to a preset mapping relationship.

[0109] The model processing unit is used to input the converted uplink channel state information and the part of the downlink channel state information into the hidden layers of the uplink module and the downlink module respectively, so as to process them using the hidden layers and the activation functions corresponding to the hidden layers.

[0110] Further, based on the above application embodiment, the preset mapping relationship in the parameter conversion unit includes a corresponding relationship between a complex number and the real part and imaginary part of the complex number.

[0111] Further, based on the above application embodiment, the channel estimation module 502 includes:

[0112] The result fusion unit is used to input the processing results of each hidden layer into the fusion module through the output layers of the upstream module and the downstream module, and generate an output result in the fusion module.

[0113] The channel estimation unit is used to use the channel state information corresponding to the output result as an estimate of the downlink channel.

[0114] Further, based on the above application embodiment, the result fusion unit is specifically used for:

[0115] In the fusion module, the processing results output by the uplink module and the downlink module are spliced ​​according to preset rules; the spliced ​​processing results are processed through at least one hidden layer in the fusion module and the activation function corresponding to the hidden layer to generate an output result.

[0116] Furthermore, based on the above application embodiment, it also includes:

[0117] The training module is used to train and generate the neural network model based on the uplink channel state information samples and part of the downlink channel state information samples.

[0118] Further, based on the above application embodiment, the training module includes:

[0119] The information collection unit is used to collect part of the downlink channel state information samples and uplink channel state information samples and the corresponding downlink channel state information labels.

[0120] The sample conversion unit is used to convert each part of downlink channel state information in the partial downlink channel state information sample and each uplink channel state information in the uplink channel state information into an uplink feature vector and a downlink feature vector respectively according to a preset mapping relationship.

[0121] A sample generating unit, used for combining each of the uplink feature vectors and each of the downlink feature vectors and corresponding downlink channel state information labels into a training set;

[0122] A training execution unit is used to train the neural network model according to the training set.

[0123] Furthermore, based on the above-mentioned application embodiment, the training execution unit is specifically used to: input the uplink feature vector and the downlink feature vector in the training set into the neural network model, and obtain the predicted value of the output downlink channel state information; determine the loss value of the predicted value and the corresponding downlink channel state information label according to a preset loss function; when it is determined that the loss value is less than a preset threshold, the training of the neural network model is completed, otherwise continue to use the uplink feature vector and the downlink feature vector in the training set to input into the neural network module for training.

[0124] Figure 8 is a schematic diagram of the structure of a communication device provided in an embodiment of the present application, such as Figure 8 As shown, the communication device includes a processor 80, a memory 81, an input device 82 and an output device 83; the number of processors 80 in the communication device can be one or more. Figure 8 A processor 80 is taken as an example; the communication device processor 80, the memory 81, the input device 82 and the output device 83 can be connected by a bus or other means. Figure 8 The example of connecting through bus is taken in the following.

[0125] The memory 81, as a computer-readable storage medium, can be used to store software programs, computer executable programs and modules, such as the modules corresponding to the downlink channel estimation device in the embodiment of the present application (model testing module 501 and channel estimation module 502). The processor 80 executes various functional applications and data processing of the communication device by running the software programs, instructions and modules stored in the memory 81, that is, implements the above method.

[0126] The memory 81 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created according to the use of the terminal, etc. In addition, the memory 81 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 81 may further include a memory remotely arranged relative to the processor 80, and these remote memories may be connected to the communication device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0127] The input device 82 may be used to receive input digital or character information and generate key signal input related to user settings and function control of the communication device. The output device 83 may include a display device such as a display screen.

[0128] The embodiment of the present application also provides a computer-readable storage medium, wherein the computer-executable instructions are used to execute a function remote customization method when executed by a computer processor, the method comprising:

[0129] Determine the customized function according to the business requirements, and divide the customized function into at least one minimum functional unit; determine the remote function operation instruction of the minimum functional unit; send the remote function operation instruction to the device terminal so that the device terminal implements the customized function. Or,

[0130] Receive a remote function operation instruction sent by a remote server; obtain basic functions in a pre-stored minimum function unit set according to the remote function operation instruction; and construct a customized function according to the basic function.

[0131] Of course, the storage medium containing computer executable instructions provided in an embodiment of the present application, whose computer executable instructions are not limited to the method operations described above, can also execute related operations in the downlink channel estimation method provided in any embodiment of the present application.

[0132] The above description is merely an exemplary embodiment of the present application and is not intended to limit the protection scope of the present application.

[0133] It will be appreciated by those skilled in the art that the term user terminal covers any suitable type of wireless user equipment, such as a mobile phone, a portable data processing device, a portable web browser or a vehicle-mounted mobile station.

[0134] In general, various embodiments of the present application can be implemented in hardware or dedicated circuits, software, logic or any combination thereof. For example, some aspects can be implemented in hardware, while other aspects can be implemented in firmware or software that can be executed by a controller, microprocessor or other computing device, although the present application is not limited thereto.

[0135] Embodiments of the present application may be implemented by executing computer program instructions by a data processor of a mobile device, for example in a processor entity, or by hardware, or by a combination of software and hardware. The computer program instructions may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages.

[0136] The block diagram of any logic flow in the accompanying drawings of the present application can represent program steps, or can represent interconnected logic circuits, modules and functions, or can represent a combination of program steps and logic circuits, modules and functions. The computer program can be stored on a memory. The memory can have any type suitable for the local technical environment and can be implemented using any suitable data storage technology, such as but not limited to read-only memory (ROM), random access memory (RAM), optical memory device and system (digital versatile disc DVD or CD disc), etc. Computer-readable media may include non-transient storage media. The data processor can be any type suitable for the local technical environment, such as but not limited to a general-purpose computer, a special-purpose computer, a microprocessor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a programmable logic device (FGPA) and a processor based on a multi-core processor architecture.

[0137] By way of exemplary and non-limiting examples, a detailed description of exemplary embodiments of the present application has been provided above. However, various modifications and adjustments to the above embodiments will be apparent to those skilled in the art when considered in conjunction with the accompanying drawings and claims, but will not depart from the scope of the present invention. Therefore, the proper scope of the present invention will be determined according to the claims.

Claims

1. A downlink channel estimation method, characterized in that: The method comprises: Inputting the uplink channel state information and part of the downlink channel state information into a preset neural network model; Complete downlink channel estimation according to the output result of the neural network model; Wherein, the preset neural network model is generated by training based on uplink channel state information samples and part of downlink channel state information samples; The neural network model at least includes a downlink module for processing the partial downlink channel state information, an uplink module for processing the uplink channel state information, and a fusion module for fusing feature vectors output by the uplink module and the downlink module.

2. The method according to claim 1, characterized in that The step of inputting the uplink channel state information and part of the downlink channel state information into a preset neural network model includes: Converting the uplink channel state information and the part of the downlink channel state information according to a preset mapping relationship; The converted uplink channel state information and the part of the downlink channel state information are respectively input into the hidden layers of the uplink module and the downlink module, so as to be processed by using the hidden layers and the activation functions corresponding to the hidden layers.

3. The method according to claim 2, characterized in that The preset mapping relationship includes a corresponding relationship between a complex number and a real part and an imaginary part of the complex number.

4. The method according to claim 2, characterized in that The method of completing downlink channel estimation according to the output result of the neural network model includes: Inputting the processing results of each hidden layer into the fusion module through the output layers of the uplink module and the downlink module, and generating an output result in the fusion module; The channel state information corresponding to the output result is used as an estimate of the downlink channel.

5. The method according to claim 4, characterized in that The processing results of each hidden layer are input into the fusion module through the output layer of the uplink module and the downlink module, and the output result is generated in the fusion module, including: In the fusion module, the processing results output by the uplink module and the downlink module are spliced ​​according to preset rules; The spliced ​​processing results are processed through at least one hidden layer in the fusion module and the activation function corresponding to the hidden layer to generate an output result.

6. The method according to claim 1, characterized in that Also includes: The neural network model is generated based on training of uplink channel state information samples and part of downlink channel state information samples.

7. The method according to claim 6, characterized in that The training and generating of the neural network model based on uplink channel state information samples and part of downlink channel state information samples includes: Collecting some downlink channel state information samples and uplink channel state information samples and corresponding downlink channel state information labels; Convert each part of downlink channel state information in the partial downlink channel state information sample and each uplink channel state information in the uplink channel state information into an uplink feature vector and a downlink feature vector respectively according to a preset mapping relationship; Combining each of the uplink feature vectors, each of the downlink feature vectors, and corresponding downlink channel state information labels into a training set; The neural network model is trained according to the training set.

8. The method according to claim 7, characterized in that The training of the neural network model according to the training set comprises: Inputting the uplink feature vector and the downlink feature vector in the training set into the neural network model, and obtaining the predicted value of the output downlink channel state information; Determine the loss value of the predicted value and the corresponding downlink channel state information label according to a preset loss function; When it is determined that the loss value is less than a preset threshold, the training of the neural network model is completed; otherwise, the uplink feature vector and the downlink feature vector in the training set are continuously used to input into the neural network module for training.

9. A downlink channel estimation device, characterized in that: The device comprises: A model testing module, used for inputting uplink channel state information and part of downlink channel state information into a preset neural network model; A channel estimation module, used to complete downlink channel estimation according to the output result of the neural network model; Wherein, the preset neural network model is generated through training of uplink channel state information samples and part of downlink channel state information samples; The neural network model at least includes a downlink module for processing the partial downlink channel state information, an uplink module for processing the uplink channel state information, and a fusion module for fusing feature vectors output by the uplink module and the downlink module.

10. A communication device, characterized in that: The communication device comprises: one or more processors; A memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the downlink channel estimation method as described in any one of claims 1-8.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the downlink channel estimation method as described in any one of claims 1 to 8 is implemented.