CN115242581A - Sub-6GHz auxiliary mmWave channel estimation method and device of convolutional neural network and electronic equipment - Google Patents
Sub-6GHz auxiliary mmWave channel estimation method and device of convolutional neural network and electronic equipment Download PDFInfo
- Publication number
- CN115242581A CN115242581A CN202210685894.7A CN202210685894A CN115242581A CN 115242581 A CN115242581 A CN 115242581A CN 202210685894 A CN202210685894 A CN 202210685894A CN 115242581 A CN115242581 A CN 115242581A
- Authority
- CN
- China
- Prior art keywords
- mmwave
- 6ghz
- sub
- neural network
- convolutional neural
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000013527 convolutional neural network Methods 0.000 title claims abstract description 67
- 238000000034 method Methods 0.000 title claims abstract description 53
- 238000001228 spectrum Methods 0.000 claims abstract description 66
- 239000011159 matrix material Substances 0.000 claims abstract description 46
- 238000004891 communication Methods 0.000 claims abstract description 32
- 238000012549 training Methods 0.000 claims abstract description 31
- 238000006243 chemical reaction Methods 0.000 claims description 15
- 230000005540 biological transmission Effects 0.000 claims description 9
- 230000006870 function Effects 0.000 claims description 7
- 238000003062 neural network model Methods 0.000 abstract description 8
- 230000000875 corresponding effect Effects 0.000 description 12
- 239000013598 vector Substances 0.000 description 12
- 238000010586 diagram Methods 0.000 description 8
- 238000013507 mapping Methods 0.000 description 7
- 238000012545 processing Methods 0.000 description 5
- 238000004590 computer program Methods 0.000 description 4
- 238000003491 array Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 125000004432 carbon atom Chemical group C* 0.000 description 1
- 150000001875 compounds Chemical class 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000013506 data mapping Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 210000004205 output neuron Anatomy 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L25/00—Baseband systems
- H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines
- H04L25/0202—Channel estimation
- H04L25/0204—Channel estimation of multiple channels
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B17/00—Monitoring; Testing
- H04B17/30—Monitoring; Testing of propagation channels
- H04B17/391—Modelling the propagation channel
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B17/00—Monitoring; Testing
- H04B17/30—Monitoring; Testing of propagation channels
- H04B17/391—Modelling the propagation channel
- H04B17/3913—Predictive models, e.g. based on neural network models
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B7/00—Radio transmission systems, i.e. using radiation field
- H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
- H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
- H04B7/06—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station
- H04B7/0613—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station using simultaneous transmission
- H04B7/0615—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station using simultaneous transmission of weighted versions of same signal
- H04B7/0617—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station using simultaneous transmission of weighted versions of same signal for beam forming
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B7/00—Radio transmission systems, i.e. using radiation field
- H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
- H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
- H04B7/08—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the receiving station
- H04B7/0837—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the receiving station using pre-detection combining
- H04B7/0842—Weighted combining
- H04B7/086—Weighted combining using weights depending on external parameters, e.g. direction of arrival [DOA], predetermined weights or beamforming
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L25/00—Baseband systems
- H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines
- H04L25/0202—Channel estimation
- H04L25/0224—Channel estimation using sounding signals
Abstract
The invention relates to the technical field of wireless communication, in particular to a sub-6GHz auxiliary mmWave channel estimation method, a device and electronic equipment based on a convolutional neural network, wherein the method comprises the following steps: obtaining a sub-6GHz channel matrix; inputting the sub-6GHz channel matrix into a preset convolution neural network model, and predicting to obtain an mmWave space spectrum, wherein the convolution neural network model is a model constructed on the basis of the reciprocity relation of the sub-6GHz channel matrix and the mmWave channel matrix; and inputting the mmWave spatial spectrum into a preset softmax classifier, and determining the optimal beam of the mmWave channel. According to the method, the device and the electronic equipment, the sub-6GHz channel matrix is obtained by utilizing the convolutional neural network model constructed based on the reciprocity relation between the sub-6GHz and the mmWave channel and obtained through pre-training, the mmWave space spectrum can be predicted and obtained only by inputting the sub-6GHz channel matrix into the convolutional neural network model, the mmWave space spectrum is input into the softmax classifier, the optimal wave beam can be determined, the mmWave channel estimation step is simplified, and the time for predicting the wave beam each time is shortened.
Description
Technical Field
The invention relates to the technical field of wireless communication, in particular to a sub-6GHz auxiliary mmWave channel estimation method and device based on a convolutional neural network and electronic equipment.
Background
High data rate communication is an important direction for the development of 5G and 6G communication, which puts higher requirements on communication bandwidth, and the development towards the frequency band of millimeter wave and above is a trend of mobile communication. However, millimeter waves (mmWave) have the characteristics of long pilot sequence, low beam precision, small base station coverage area and the like, so that the use scene of the millimeter waves is very limited, and the millimeter waves cannot become a new communication mode for replacing sub-6GHz in a short period. According to the current domestic standard, the microwave communication mainly based on sub-6GHz is still used, and the technical innovation cannot be fundamentally realized.
On the other hand, channel estimation has been an important component in communication systems, mainly through the process of receiving data to obtain Channel State Information (CSI) according to a channel model. In mmWave's communication system, due to the use of massive MIMO, its channel is not easily estimated using conventional methods.
The method for performing channel estimation based on sub-6GHz auxiliary mmWave is verified through experimental measurement, but although the correlation between sub-6GHz and mmWave is researched in the prior art, a comprehensive network model which can utilize the relation between sub-6GHz and mmWave, has higher reliability and better expression effect and meets the millimeter wave channel characteristics is lacked, and a network model which adapts to the reciprocity between sub-6GHz and mmWave to realize mmWave channel estimation is also lacked. In addition, in the existing technical scheme for performing sub-6 GHz-assisted mmWave channel estimation by using a neural network, for example, in the patent with the publication number of CN112073106a, which is named as a millimeter wave beam prediction method and apparatus, an electronic device, and a readable storage medium, received signals according to two different frequency bands need to be respectively input when millimeter wave beam prediction is performed, training data of two corresponding different channels are obtained, the two training data are input to a neural network model for fusion, and then beam prediction is performed, the prediction method is complex, and sub-6GHz cannot be directly used for performing channel estimation on mmWave.
Disclosure of Invention
In order to solve the technical problem, the invention provides a sub-6GHz auxiliary mmWave channel estimation method and device based on a convolutional neural network and electronic equipment.
The sub-6GHz auxiliary mmWave channel estimation method based on the convolutional neural network comprises the following steps:
obtaining a sub-6GHz channel matrix;
inputting the sub-6GHz channel matrix into a preset convolutional neural network model, and predicting to obtain an mmWave space spectrum, wherein the convolutional neural network model is a model constructed on the basis of the reciprocity relation of the sub-6GHz and the mmWave channels;
and inputting the mmWave spatial spectrum into a preset softmax classifier, and determining the optimal beam of the mmWave channel.
Further, the method further comprises:
and predicting to obtain the maximum transmission rate of the mmWave according to the optimal beam of the mmWave channel.
Further, inputting the sub-6GHz channel matrix into a preset convolutional neural network model, and predicting to obtain an mmWave spatial spectrum includes:
extracting sub-6GHz spatial spectrum information from a sub-6GHz channel matrix by using a convolutional neural network model;
performing frequency correlation conversion on the sub-6GHz spatial spectrum information, and converting low-dimensional characteristic information into high-dimensional characteristic information;
and reconstructing mmWave information according to the high-dimensional characteristic information to construct and obtain an mmWave spatial spectrum.
Further, the preset convolutional neural network model takes the minimum mean square error as a loss function, and the loss function is expressed as:
where θ represents a model estimation parameter set, H s For incoming sub-6GHz channel information, F (H) s θ) is the result of the generation, H m And n is the number of sample sets, and is the corresponding mmWave channel information.
Further, the method also comprises training the softmax classifier, and the training method comprises the following steps:
according to the relation of user u under all beam codebooksTo obtainAs a training set, the softmax classifier is trained.
The invention also provides a sub-6GHz auxiliary mmWave channel estimation device based on a convolutional neural network, which comprises an acquisition module, an mmWave spatial spectrum prediction module and an optimal beam determination module, wherein:
the acquisition module is connected with the mmWave spatial spectrum prediction module and is used for acquiring a sub-6GHz channel matrix;
the mmWave spatial spectrum prediction module is connected with the optimal wave beam determination module and used for inputting the sub-6GHz channel matrix into a preset convolution neural network model and predicting to obtain an mmWave spatial spectrum, wherein the convolution neural network model is a model constructed on the basis of the reciprocity relation of the sub-6GHz channel and the mmWave channel;
and the optimal beam determining module is used for inputting the mmWave spatial spectrum to a preset softmax classifier and determining the optimal beam of the mmWave channel.
Further, the apparatus further comprises a maximum rate prediction module, wherein:
and the maximum rate prediction module is connected with the optimal beam determination module and used for predicting and obtaining the maximum transmission rate of the mmWave according to the optimal beam of the mmWave channel.
Further, the mmWave spatial spectrum prediction module inputs the sub-6GHz channel matrix into a preset convolutional neural network model, and predicting to obtain the mmWave spatial spectrum includes:
extracting sub-6GHz spatial spectrum information from a sub-6GHz channel matrix by using a convolutional neural network model;
performing frequency correlation conversion on the sub-6GHz spatial spectrum information, and converting low-dimensional characteristic information into high-dimensional characteristic information;
and reconstructing mmWave information according to the high-dimensional characteristic information to construct and obtain an mmWave spatial spectrum.
Further, the training method of the softmax classifier in the optimal beam determination module comprises the following steps:
according to the relation of the user u under all beam codebooksTo obtainAnd as a training set, training the softmax classifier.
The invention also provides electronic equipment which comprises a processor, a communication interface, a memory and a communication bus, wherein the processor and the communication interface are used for finishing mutual communication by the memory through the communication bus; a memory for storing processor-executable instructions; and the processor is used for realizing the steps of the method when executing the instructions stored in the memory.
The sub-6GHz auxiliary mmWave channel estimation method, the device and the electronic equipment based on the convolutional neural network have the following beneficial effects that:
the method comprises the steps of establishing a convolutional neural network model in advance based on the reciprocity relation of sub-6GHz and mmWave channels, inputting the sub-6GHz channel matrix to the convolutional neural network model after obtaining the sub-6GHz channel matrix, and obtaining an mmWave space spectrum, so that direct conversion of the sub-6GHz and the mmWave is achieved, the prediction process is simplified, and the optimal wave beam of the mmWave channel is predicted by utilizing a softmax classifier subsequently, so that the training cost brought by wave beam training is greatly saved, the original information rate is improved, a structural model is directly used, and the time for predicting the wave beam each time is shortened.
Drawings
For a clearer explanation of the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts;
FIG. 1 is a flowchart of a sub-6GHz auxiliary mmWave channel estimation method based on a convolutional neural network in an embodiment of the present invention;
FIG. 2 is a schematic diagram of a convolutional neural network prediction in an embodiment of the present invention;
FIG. 3 is a diagram of a data mapping relationship in an embodiment of the invention;
FIG. 4 is a flow chart of convolutional neural network prediction in an embodiment of the present invention;
FIG. 5 is a schematic diagram of a softmax classifier in one embodiment of the invention;
FIG. 6 is a diagram illustrating the results of softmax classification in an embodiment of the present invention;
FIG. 7 is a flowchart of a sub-6GHz auxiliary mmWave channel estimation method based on a convolutional neural network according to another embodiment of the present invention;
FIG. 8 is a flowchart III of a sub-6GHz auxiliary mmWave channel estimation method based on a convolutional neural network according to another embodiment of the present invention;
FIG. 9 is a schematic diagram of the basic architecture of the CNN model;
FIG. 10 is a schematic diagram of a sub-6GHz auxiliary mmWave channel estimation device based on a convolutional neural network in an embodiment of the present invention;
fig. 11 is a schematic diagram of a sub-6GHz auxiliary mmWave channel estimation apparatus based on a convolutional neural network according to another embodiment of the present invention;
FIG. 12 is a diagram of an electronic device in an embodiment of the invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be described clearly and completely with reference to the accompanying drawings in the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In an embodiment of the present invention, a sub-6GHz assisted mmWave channel estimation method based on a convolutional neural network is provided, as shown in fig. 1, the method includes:
step S101: obtaining a sub-6GHz channel matrix;
step S102: and inputting the sub-6GHz channel matrix into a preset convolution neural network model, and predicting to obtain an mmWave spatial spectrum.
In the step, the convolutional neural network model is a model constructed based on the reciprocity relation between sub-6GHz and mmWave channels.
The reciprocity of sub-6GHz and mmWave channels is a theoretical basis for realizing sub-6GHz auxiliary mmWave communication, and a mapping relation existing between the sub-6GHz and the mmWave channels needs to be deduced, so that the sub-6GHz auxiliary mmWave communication is realized.
In the present invention, channel reciprocity means that reverse Channel State Information (CSI) can be obtained by using a forward channel signal in a TDD communication system, and it means that channel information of the two is the same. Channel reciprocity refers to the difference in spatial characteristics between different frequency bands and between transmission channels with different numbers of transmit and receive antennas. At present, it has been proved that amplitude and phase compensation methods can be used to achieve reciprocity between adjacent frequency channels in FDD mode.
H can be obtained according to the same frequency of the channel in the traditional TDD system T uplink =H downlink The sub-6GHz auxiliary mmWave communication mode refers to that in an FDD system, sub-6GHz information obtained by uplink pilot frequency is used for completing conversion of mmWave downlink Channel State Information (CSI). In the case of FDD system, the two systems have different performances due to frequency selectivity, so that the uplink and downlink channels are not simply transposed. And the high-frequency antenna adopts the design of massive MIMO due to factors such as power compensation, antenna size and the like, so that the high-frequency antenna adopts the design of massive MIMOThe mapping relation between sub-6GHz and mmWave should satisfy:
whereinDenoted as user u at frequency f 1 The number of antennas is N 1 In order to implement a mapping relationship from sub-6GHz to mmWave, the relationship between the user position movement, the change of the number of antenna arrays and the change of the frequency all affect the relationship between the two, and we mainly discuss the change caused by the frequency change.
Based on the spatial correlation between the sub-6GHz system and the millimeter wave system, channel estimation can be performed by extracting sub-6GHz spatial angle information (such as AoA and AoD) and using out-of-band information (OOB) to assist the millimeter waves. The CSI of the millimeter waves is determined by the low-frequency spatial information through frequency correlation, so that the conversion of channel parameters is completed on the basis of determining a sub-6GHz channel, and the related information of mmWave is determined.
Suppose that at the sub-6GHz codebook training stage, a transmitting end has a beam forming vectorBeam forming vector for receiving endAnd processing the signal, and adopting a millimeter wave channel model in the formula to represent the received signal:
wherein the content of the first and second substances,representing the transmitted symbol vector for the sub-6GHz system,representing the sub-6GHz system noise vector. Thus, the received signals are combined
In the above-mentioned formula, the compound has the following structure,refers to a dictionary matrix of the transmitting end, anReceiving end channel dictionary matrixAnd is Which represents white gaussian noise, is generated,the element with the largest absolute value corresponds to the optimal beam pair of the sub-6GHz system and the optimal arrival angle (AoA) and departure angle (AoD) of the channel.
If vectorizedDefinition ofThenI.e. representing the optimal AoA and AoD indices. In particular, if the optimal AoD index is recorded asThe index of the optimal AoA can be expressed as i s * =r s * -(j s * -1)M MS . After the optimal AoA and AoD directions of the sub-6GHz channel are determined, the better spatial angle information corresponding to the millimeter wave channel can be obtained preliminarily by utilizing the spatial consistency of the millimeter wave channel and the sub-6GHz channel. The sub-6GHz space spectrum is as follows:
in the above formula, the first and second carbon atoms are,namely, the out-of-band information used in the millimeter wave channel estimation, and the millimeter wave spatial spectrum | E | can be obtained in the same way.
Therefore, the sub-6GHz space spectrum can be obtainedThe method is directly used for recovering the millimeter-wave band sparse signals. Because the optimal beamforming codebook for realizing the millimeter wave hybrid MIMO system is obtained through the dictionary matrix, then the beam training is carried out, and the dictionary matrix can be constructed corresponding to the millimeter wave band according to the indexes of the optimal AoA and AoD of the sub-6GHz channel. Index values of AoA and AoD of sub-6GHz channels can be obtained through calculation of a spatial spectrum, and then indexes of AoA and AoD of the selected channels are respectively stored in a set I s ={i s * And set J s ={j s * In (c) }. After the low-frequency spatial angle index is found, the spatial angle information of the millimeter wave can be further determined by utilizing the spatial correlation between the sub-6GHz and the millimeter wave channel, and then the AoA and AoD index set I of the millimeter wave channel is obtained m ={i m * } and J m ={j m * Thus, a coarse estimation of the millimeter wave channel can be made.
The derivation process derives the reciprocity relationship between the sub-6GHz and the mmWave, and then a convolution neural network model can be constructed based on the reciprocity relationship between the sub-6GHz and the mmWave in the invention.
Specifically, in this step, as shown in fig. 2, 3, and 4, the sub-6GHz channel matrix is input into a preset convolutional neural network model, and predicting to obtain the mmWave spatial spectrum includes:
step S1021: and extracting sub-6GHz spatial spectrum information from the sub-6GHz channel matrix by using a convolutional neural network model.
Step S1022: and performing frequency correlation conversion on the sub-6GHz spatial spectrum information, and converting the low-dimensional characteristic information into high-dimensional characteristic information.
Specifically, the conversion from low-dimensional to high-dimensional characteristic information is performed through the above mentioned reciprocity relationship between sub-6GHz and mmWave.
Step S1023: and reconstructing mmWave information according to the high-dimensional characteristic information to construct an mmWave space spectrum.
Step S103: and inputting the mmWave spatial spectrum into a preset softmax classifier, and determining the optimal beam of the mmWave channel.
As shown in fig. 5 and fig. 6, the output values of the softmax classifier are correlated, the probability sum is 1, and the result is very good for a single feature. The method is used for predicting probability vectors by utilizing a deep neural network model according to specific characteristic vectors, and selecting the wave beam corresponding to the maximum probability value in the probability vectors as the optimal wave beam of the millimeter wave channel.
The convolutional neural network model obtains predicted output by carrying out layer-by-layer nonlinear conversion on input data, the number len (p) of output neurons of the classifier is consistent with the number of elements of a predefined millimeter wave codebook, the output is a classification vector corresponding to each predicted value, and the size of an element in the vector represents the probability that a beam corresponding to the element is the optimal beam of a channel.
In this embodiment, the sub-6GHz channel matrix is obtained, the processes of extracting corresponding spatial spectrum information, performing frequency correlation conversion, and reconstructing mmWave information are completed through a convolutional neural network, a mapping relationship from a low dimension to a high dimension is realized, and then the spatial spectrum of mmWave is utilized to complete the whole process of selecting an optimal beam through a softmax classifier.
Generally, due to the fact that dictionary matrix dimensions adopted by sub-6GHz and mmWave are different, beams of mmWave are narrower, and resolution is higher. And the channel is generally expressed in a complex form, and the real part information and the imaginary part information can be extracted and respectively put into two channels for processing. Firstly, carrying out normalization processing on a low-dimensional sub-6GHz channel and coding by using a dictionary matrix with low resolution to obtain sparse correlation coefficients. According to the comparison of the spatial spectrum information, the relative information of mmWave can be known to exist in the sub-6GHz spatial spectrum. Therefore, by replacing the codebook with mmWave high resolution, the low-dimensional sparse relation is converted into the high-dimensional sparse coefficient through the nonlinear mapping relation, and finally the mmWave space spectrum is constructed.
In the sub-6GHz auxiliary mmWave channel estimation method based on the convolutional neural network provided in the embodiment, a convolutional neural network model is established in advance based on the reciprocity relationship between sub-6GHz and mmWave channels, after a sub-6GHz channel matrix is obtained, the sub-6GHz channel matrix is input to the convolutional neural network model, and an mmWave space spectrum is obtained, so that direct conversion between sub-6GHz and mmWave is realized, the prediction process is simplified, and the optimal beam of the mmWave channel is predicted by using a softmax classifier subsequently, so that not only is the training cost brought by beam training greatly saved, and the original information rate is improved, but also a structural model is directly used, and the time for predicting the beam each time is reduced.
In yet another embodiment of the present invention, as shown in fig. 7, the method further comprises:
step S104: and predicting to obtain the maximum transmission rate of the mmWave according to the optimal wave beam of the mmWave channel.
Specifically, after the optimal beamforming vector pair and the corresponding channel direction are found, the maximum transmission rate can be obtained according to the optimal beamforming vector pair and the corresponding channel direction.
WhereinDenotes the channel matrix at the kth carrier of the u-th user, f p Represents the p-th beamforming vector in the codebook in order to hopefully find the largest f in each carrier p To achieve the maximum transmission rate.
In yet another embodiment of the present invention, as shown in fig. 8, the method further comprises:
step S105: the convolutional neural network model is trained in advance.
Specifically, the training method comprises the following steps: the mapping rules between the input and output data are automatically extracted by the neural network in the learning phase and are stored in all the connections of the network in a distributed manner. Firstly, carrying out normalization processing on a low-dimensional sub-6GHz channel and coding by using a dictionary matrix with low resolution to obtain sparse correlation coefficients. And extracting and converting the codebook with mmWave high resolution from the input features through a convolution layer and a pooling layer by model feature extraction, namely converting the low-dimensional sparse relationship into a high-dimensional sparse coefficient through a nonlinear mapping relationship, and finally constructing an mmWave spatial spectrum. Based on the features extracted in the process, a new classifier is trained, namely, a corresponding convolutional neural network model is pre-trained for the conversion between subsequent channels.
The CNN model basic architecture is shown in fig. 9.
More specifically, in an implementation manner in this embodiment, the preset convolutional neural network model takes a minimum mean square error as a loss function, and the loss function is expressed as:
where θ represents a model estimation parameter set, H s Is input sub-6GHz channel information (i.e. sub-6GHz channel matrix), F (H) s θ) is the result of the generation, H m And n is the number of sample sets, and is the corresponding mmWave channel information (i.e. mmWave spatial spectrum).
In yet another embodiment of the present invention, the method further comprises training a softmax classifier, the training method comprising:
according to the relation of the user u under all beam codebooksTo obtainAs a training set, the softmax classifier is trained.
The invention also provides a sub-6GHz auxiliary mmWave channel estimation device based on a convolutional neural network, as shown in fig. 10, the device includes an obtaining module 901, an mmWave spatial spectrum prediction module 902, and an optimal beam determination module 903, wherein:
the acquisition module 901 is connected with the mmWave spatial spectrum prediction module 902 and used for acquiring a sub-6GHz channel matrix;
the mmWave spatial spectrum prediction module 902 is connected with the optimal beam determination module 903 and is used for inputting the sub-6GHz channel matrix into a preset convolutional neural network model and predicting to obtain an mmWave spatial spectrum, wherein the convolutional neural network model is a model constructed based on the reciprocity relation between the sub-6GHz channel and the mmWave channel;
and an optimal beam determining module 903, configured to input the mmWave spatial spectrum to a preset softmax classifier, and determine an optimal beam of the mmWave channel.
In yet another embodiment of the present invention, as shown in fig. 11, the apparatus further comprises a maximum rate prediction module 904, wherein:
the maximum rate predicting module 904 is connected to the optimal beam determining module 903, and configured to predict the maximum transmission rate of the mmWave according to the optimal beam of the mmWave channel.
In another embodiment of the present invention, the mmWave spatial spectrum prediction module 902 inputs the sub-6GHz channel matrix into a preset convolutional neural network model, and obtaining the mmWave spatial spectrum by prediction includes:
extracting sub-6GHz spatial spectrum information from the sub-6GHz channel matrix by using a convolutional neural network model;
performing frequency correlation conversion on the sub-6GHz spatial spectrum information, and converting low-dimensional characteristic information into high-dimensional characteristic information;
and reconstructing mmWave information according to the high-dimensional characteristic information to construct and obtain an mmWave spatial spectrum.
In another embodiment of the present invention, the training method of the softmax classifier in the optimal beam determination module 903 includes:
according to the relation of the user u under all beam codebooksTo obtainAnd as a training set, training the softmax classifier.
As shown in fig. 12, an embodiment of the present invention further provides an electronic device, which includes a processor 1101, a communication interface 1102, a memory 1103 and a communication bus 1104, where the processor 1101, the communication interface 1102 and the memory 1103 complete communication with each other through the communication bus 1104, and the memory 1103 is used for storing a computer program; the processor 1101 is configured to, when executing the program stored in the memory 1103, implement the sub-6GHz auxiliary mmWave channel estimation method based on the convolutional neural network in the foregoing embodiment.
The communication bus mentioned in the above electronic device may be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The communication bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one thick line is shown, but this does not mean that there is only one bus or one type of bus. The communication interface is used for communication between the electronic equipment and other equipment. The memory may include a Random Access Memory (RAM) or a Non-volatile memory (NVM), such as at least one disk memory. Optionally, the memory may also be at least one memory device located remotely from the processor.
The processor may be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), and the like; but also Digital Signal Processors (DSPs), application Specific Integrated Circuits (ASICs), field Programmable Gate Arrays (FPGAs) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components.
In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. The procedures or functions according to the embodiments of the invention are brought about in whole or in part when the computer program instructions are loaded and executed on a computer. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device.
The computer instructions may be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, from one website site, computer, server, or data center to another website site, computer, server, or data center via wire (e.g., coaxial cable, fiber optic, digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.).
The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device, such as a server, a data center, etc., that incorporates one or more of the available media. The usable medium may be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid State Disk (SSD)), among others.
According to the sub-6GHz auxiliary mmWave channel estimation method, device and electronic equipment based on the convolutional neural network, a convolutional neural network model is established in advance based on the reciprocity relation of sub-6GHz and mmWave channels, after a sub-6GHz channel matrix is obtained, the sub-6GHz channel matrix is input into the convolutional neural network model to obtain an mmWave space spectrum, so that direct conversion of sub-6GHz and mmWave is achieved, the prediction process is simplified, and the optimal wave beam of the mmWave channel is predicted by using a softmax classifier subsequently, so that the training cost brought by wave beam training is greatly saved, the original information rate is improved, the structural model is directly used, and the time for predicting the wave beam each time is shortened.
It will be appreciated by those skilled in the art that changes could be made to the details of the above-described embodiments without departing from the underlying principles thereof. The scope of the invention is, therefore, indicated by the appended claims, in which all terms are intended to be interpreted in their broadest reasonable sense unless otherwise indicated.
Claims (10)
1. A sub-6GHz auxiliary mmWave channel estimation method based on a convolutional neural network is characterized by comprising the following steps:
obtaining a sub-6GHz channel matrix;
inputting the sub-6GHz channel matrix into a preset convolutional neural network model, and predicting to obtain an mmWave space spectrum, wherein the convolutional neural network model is a model constructed on the basis of the reciprocity relation of sub-6GHz and mmWave channels;
and inputting the mmWave spatial spectrum to a preset softmax classifier, and determining the optimal beam of the mmWave channel.
2. The convolutional neural network-based sub-6GHz auxiliary mmWave channel estimation method of claim 1, further comprising:
and predicting to obtain the maximum transmission rate of the mmWave according to the optimal beam of the mmWave channel.
3. The method of claim 1, wherein the sub-6GHz auxiliary mmWave channel estimation method based on the convolutional neural network is characterized in that the sub-6GHz channel matrix is input into a preset convolutional neural network model, and predicting to obtain an mmWave spatial spectrum comprises:
extracting sub-6GHz spatial spectrum information from the sub-6GHz channel matrix by using the convolutional neural network model;
performing frequency correlation conversion on the sub-6GHz spatial spectrum information, and converting low-dimensional characteristic information into high-dimensional characteristic information;
and reconstructing mmWave information according to the high-dimensional characteristic information to construct and obtain an mmWave space spectrum.
4. The sub-6GHz auxiliary mmWave channel estimation method based on convolutional neural network of claim 1, wherein the preset convolutional neural network model takes minimum mean square error as a loss function, and the loss function is expressed as:
5. The convolutional neural network-based sub-6GHz assisted mmWave channel estimation method of claim 1, further comprising training a softmax classifier, the training method comprising:
6. A sub-6GHz auxiliary mmWave channel estimation device based on a convolutional neural network is characterized by comprising an acquisition module, an mmWave spatial spectrum prediction module and an optimal beam determination module, wherein:
the acquisition module is connected with the mmWave spatial spectrum prediction module and is used for acquiring a sub-6GHz channel matrix;
the mmWave spatial spectrum prediction module is connected with the optimal beam determination module and is used for inputting the sub-6GHz channel matrix into a preset convolutional neural network model and predicting to obtain an mmWave spatial spectrum, wherein the convolutional neural network model is a model constructed on the basis of the reciprocity relation between the sub-6GHz channel and the mmWave channel;
and the optimal beam determining module is used for inputting the mmWave spatial spectrum to a preset softmax classifier and determining the optimal beam of the mmWave channel.
7. The convolutional neural network-based sub-6GHz auxiliary mmWave channel estimation device of claim 6, further comprising a maximum rate prediction module, wherein:
and the maximum rate prediction module is connected with the optimal beam determination module and used for predicting and obtaining the maximum transmission rate of the mmWave according to the optimal beam of the mmWave channel.
8. The convolutional neural network-based sub-6GHz auxiliary mmWave channel estimation device of claim 6, wherein the mmWave spatial spectrum prediction module inputs the sub-6GHz channel matrix into a preset convolutional neural network model, and predicting the mmWave spatial spectrum comprises:
extracting sub-6GHz spatial spectrum information from the sub-6GHz channel matrix by using the convolutional neural network model;
performing frequency correlation conversion on the sub-6GHz spatial spectrum information, and converting low-dimensional characteristic information into high-dimensional characteristic information;
and reconstructing mmWave information according to the high-dimensional characteristic information to construct and obtain an mmWave space spectrum.
9. The convolutional neural network-based sub-6GHz auxiliary mmWave channel estimation device of claim 6, wherein the training method of the softmax classifier in the optimal beam determination module comprises:
10. The electronic equipment is characterized by comprising a processor, a communication interface, a memory and a communication bus, wherein the processor and the communication interface are used for realizing the communication between the processor and the memory through the communication bus;
a memory for storing processor-executable instructions;
a processor for implementing the method steps of any one of claims 1 to 5 when executing instructions stored on a memory.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210685894.7A CN115242581B (en) | 2022-06-16 | 2022-06-16 | Sub-6GHz auxiliary mmWave channel estimation method and device of convolutional neural network and electronic equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210685894.7A CN115242581B (en) | 2022-06-16 | 2022-06-16 | Sub-6GHz auxiliary mmWave channel estimation method and device of convolutional neural network and electronic equipment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN115242581A true CN115242581A (en) | 2022-10-25 |
CN115242581B CN115242581B (en) | 2023-11-03 |
Family
ID=83670181
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210685894.7A Active CN115242581B (en) | 2022-06-16 | 2022-06-16 | Sub-6GHz auxiliary mmWave channel estimation method and device of convolutional neural network and electronic equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115242581B (en) |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20180287822A1 (en) * | 2017-03-31 | 2018-10-04 | Mitsubishi Electric Research Laboratories, Inc. | System and Method for Channel Estimation in mmWave Communications Exploiting Joint AoD-AoA Angular Spread |
CN111435926A (en) * | 2019-01-11 | 2020-07-21 | 中国移动通信有限公司研究院 | MIMO system channel prediction method, device, medium and equipment |
CN111901258A (en) * | 2020-05-08 | 2020-11-06 | 中兴通讯股份有限公司 | Method for realizing reciprocity of uplink and downlink channels, communication node and storage medium |
CN112073106A (en) * | 2020-08-14 | 2020-12-11 | 清华大学 | Millimeter wave beam prediction method and device, electronic device and readable storage medium |
CN113260084A (en) * | 2021-05-18 | 2021-08-13 | 北京邮电大学 | Millimeter wave-based vehicle networking V2X communication link establishment method |
WO2022005052A1 (en) * | 2020-07-01 | 2022-01-06 | 엘지전자 주식회사 | Method and device for transmitting signal in wireless communication system |
-
2022
- 2022-06-16 CN CN202210685894.7A patent/CN115242581B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20180287822A1 (en) * | 2017-03-31 | 2018-10-04 | Mitsubishi Electric Research Laboratories, Inc. | System and Method for Channel Estimation in mmWave Communications Exploiting Joint AoD-AoA Angular Spread |
CN111435926A (en) * | 2019-01-11 | 2020-07-21 | 中国移动通信有限公司研究院 | MIMO system channel prediction method, device, medium and equipment |
CN111901258A (en) * | 2020-05-08 | 2020-11-06 | 中兴通讯股份有限公司 | Method for realizing reciprocity of uplink and downlink channels, communication node and storage medium |
WO2022005052A1 (en) * | 2020-07-01 | 2022-01-06 | 엘지전자 주식회사 | Method and device for transmitting signal in wireless communication system |
CN112073106A (en) * | 2020-08-14 | 2020-12-11 | 清华大学 | Millimeter wave beam prediction method and device, electronic device and readable storage medium |
CN113260084A (en) * | 2021-05-18 | 2021-08-13 | 北京邮电大学 | Millimeter wave-based vehicle networking V2X communication link establishment method |
Non-Patent Citations (1)
Title |
---|
修越 等: "基于K学习的Sub-6GHz辅助毫米波信道信息获取", 《电子科技大学学报》 * |
Also Published As
Publication number | Publication date |
---|---|
CN115242581B (en) | 2023-11-03 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111817768B (en) | Channel estimation method for intelligent reflecting surface wireless communication | |
CN113748614A (en) | Channel estimation model training method and device | |
CN114124623B (en) | Wireless communication channel estimation method and device | |
CN113315560B (en) | Beam forming method of compact planar array Massive MIMO system | |
CN102868432A (en) | Blind beam forming device and method under dual-stage neural network | |
CN114915523B (en) | Intelligent super-surface channel estimation method and system based on model driving | |
Elbir et al. | Robust hybrid beamforming with quantized deep neural networks | |
CN115606157A (en) | Channel estimation for antenna arrays | |
CN115242581A (en) | Sub-6GHz auxiliary mmWave channel estimation method and device of convolutional neural network and electronic equipment | |
WO2023087819A1 (en) | Beam training method, first node, second node, communication system, and medium | |
CN107070820B (en) | Path-by-path channel estimation method under multipath channel and codebook construction method | |
CN115022134A (en) | Millimeter wave large-scale MIMO system channel estimation method and system based on non-iterative reconstruction network | |
CN111555784B (en) | Precoding method and device and electronic equipment | |
CN113543017A (en) | Wave beam switching method of millimeter wave internet of vehicles and related equipment | |
CN114499605B (en) | Signal transmission method, signal transmission device, electronic equipment and storage medium | |
CN114448478B (en) | Signal transmission method, signal transmission device, electronic equipment and storage medium | |
CN116633399B (en) | Semantic communication waveform design method, device, equipment and storage medium | |
CN116761223B (en) | Method for realizing 4G radio frequency communication by using 5G baseband chip and vehicle-mounted radio frequency system | |
CN114268409B (en) | Method and device for constructing index sequence of polarization code | |
CN115882908A (en) | Feedback method for intelligent super-surface time correlation cascade channel | |
CN115242274B (en) | Beam searching method and device and electronic equipment | |
CN110417692B (en) | Uplink channel tracking method and device | |
Hajjaj | Joint dictionary and deep learning for cosparse recovery of the millimeter wave massive MIMO channels | |
Lin et al. | A Literature Review on Antenna Data Analysis in MIMO System | |
CN117220729A (en) | Codebook generation method and system in network optimization, electronic equipment and storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |