CN113726711A - OFDM receiving method and device, and channel estimation model training method and device - Google Patents
OFDM receiving method and device, and channel estimation model training method and device Download PDFInfo
- Publication number
- CN113726711A CN113726711A CN202111020964.9A CN202111020964A CN113726711A CN 113726711 A CN113726711 A CN 113726711A CN 202111020964 A CN202111020964 A CN 202111020964A CN 113726711 A CN113726711 A CN 113726711A
- Authority
- CN
- China
- Prior art keywords
- channel estimation
- frequency domain
- information
- resolution
- state information
- 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
- 238000012549 training Methods 0.000 title claims abstract description 81
- 238000000034 method Methods 0.000 title claims abstract description 45
- 238000007781 pre-processing Methods 0.000 claims abstract description 25
- 238000001514 detection method Methods 0.000 claims description 51
- 238000005070 sampling Methods 0.000 claims description 12
- 238000004590 computer program Methods 0.000 claims description 4
- 230000005540 biological transmission Effects 0.000 description 13
- 238000004891 communication Methods 0.000 description 10
- 230000006870 function Effects 0.000 description 9
- 238000010586 diagram Methods 0.000 description 8
- 230000008569 process Effects 0.000 description 8
- 230000004913 activation Effects 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 3
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 238000013507 mapping Methods 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 3
- 238000007493 shaping process Methods 0.000 description 3
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 2
- 230000002776 aggregation Effects 0.000 description 2
- 238000004220 aggregation Methods 0.000 description 2
- 239000000969 carrier Substances 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000003491 array Methods 0.000 description 1
- 230000001427 coherent effect Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000000644 propagated effect Effects 0.000 description 1
- 230000011664 signaling Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L27/00—Modulated-carrier systems
- H04L27/26—Systems using multi-frequency codes
- H04L27/2601—Multicarrier modulation systems
- H04L27/2647—Arrangements specific to the receiver only
- H04L27/2655—Synchronisation arrangements
- H04L27/2689—Link with other circuits, i.e. special connections between synchronisation arrangements and other circuits for achieving synchronisation
- H04L27/2695—Link with other circuits, i.e. special connections between synchronisation arrangements and other circuits for achieving synchronisation with channel estimation, e.g. determination of delay spread, derivative or peak tracking
-
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L27/00—Modulated-carrier systems
- H04L27/26—Systems using multi-frequency codes
- H04L27/2601—Multicarrier modulation systems
- H04L27/2647—Arrangements specific to the receiver only
- H04L27/2655—Synchronisation arrangements
- H04L27/2668—Details of algorithms
- H04L27/2669—Details of algorithms characterised by the domain of operation
- H04L27/2672—Frequency domain
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L27/00—Modulated-carrier systems
- H04L27/26—Systems using multi-frequency codes
- H04L27/2601—Multicarrier modulation systems
- H04L27/2647—Arrangements specific to the receiver only
- H04L27/2655—Synchronisation arrangements
- H04L27/2668—Details of algorithms
- H04L27/2681—Details of algorithms characterised by constraints
- H04L27/2688—Resistance to perturbation, e.g. noise, interference or fading
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L27/00—Modulated-carrier systems
- H04L27/26—Systems using multi-frequency codes
- H04L27/2601—Multicarrier modulation systems
- H04L27/2647—Arrangements specific to the receiver only
- H04L27/2655—Synchronisation arrangements
- H04L27/2689—Link with other circuits, i.e. special connections between synchronisation arrangements and other circuits for achieving synchronisation
- H04L27/2691—Link with other circuits, i.e. special connections between synchronisation arrangements and other circuits for achieving synchronisation involving interference determination or cancellation
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Power Engineering (AREA)
- Cable Transmission Systems, Equalization Of Radio And Reduction Of Echo (AREA)
Abstract
The application provides an OFDM receiving method and device, a channel estimation model training method and device, an electronic device and a storage medium, wherein the OFDM receiving method comprises the following steps: acquiring a frequency domain signal; preprocessing the frequency domain signal to obtain channel estimation information; determining a plurality of low-resolution two-dimensional images according to the channel estimation information; inputting a plurality of low-resolution two-dimensional images into a pre-trained channel estimation model to obtain channel state information; and determining bit information sent by the sending end based on the frequency domain signal and the channel state information. The channel estimation information is decomposed into a plurality of low-resolution two-dimensional images, the low-resolution two-dimensional images are input into a channel estimation model, the pre-trained channel estimation model is used for extracting characteristic information in the low-resolution two-dimensional images, and channel state information with high precision is determined, so that bit information sent by a sending end is determined based on the channel state information, and the receiving performance of an OFDM receiver is improved.
Description
Technical Field
The present application relates to the field of communications technologies, and in particular, to an OFDM receiving method and apparatus, a channel estimation model training method and apparatus, an electronic device, and a computer storage medium.
Background
Orthogonal Frequency Division Multiplexing (OFDM) is widely used in wireless communication and is a key modulation scheme of a wireless communication system. However, the peak-to-average power ratio of the OFDM system is relatively high, and the linear range of the amplifier is limited, which causes non-linear distortion of the signal, thereby seriously affecting the estimation accuracy and the channel capacity. In order to ensure high data rate and high reliability, the receiver of the OFDM system usually employs coherent demodulation, which requires estimation of channel state information.
In the prior art, the least square method is usually adopted for channel state information estimation. However, the least square method is adopted to estimate the channel state information, the accuracy of the obtained estimation result is low, and the accurate estimation of the channel state information cannot be realized, so that the receiving performance of the receiver is affected.
Disclosure of Invention
An object of the embodiments of the present application is to provide an OFDM receiving method and apparatus, a channel estimation model training method and apparatus, an electronic device, and a computer storage medium, which are used to improve the estimation accuracy of channel state information, thereby improving the receiving performance of a receiver.
In a first aspect, the present invention provides an OFDM receiving method, including: acquiring a frequency domain signal; preprocessing the frequency domain signal to obtain channel estimation information; determining a plurality of low-resolution two-dimensional images according to the channel estimation information; inputting the two-dimensional images with the low resolution into a pre-trained channel estimation model to obtain channel state information; and determining bit information sent by a sending end based on the frequency domain signal and the channel state information.
In the embodiment of the application, channel estimation information is determined according to a frequency domain signal, the channel estimation information (channel state information with low precision) is decomposed into a plurality of low-resolution two-dimensional images, the plurality of low-resolution two-dimensional images are input into a channel estimation model, and feature information in the plurality of low-resolution two-dimensional images is extracted by using a pre-trained channel estimation model, so that the channel state information with high precision is determined, and further bit information sent by a sending end is determined based on the channel state information, so that the receiving performance of an OFDM receiver is improved.
In an optional embodiment, the channel estimation model includes an image super-resolution network and an up-sampling network, and the inputting the plurality of low-resolution two-dimensional images into a pre-trained channel estimation model to obtain channel state information includes: extracting the features of the two-dimensional images with low resolution by using the super-resolution network to obtain the features corresponding to the two-dimensional images; and utilizing the up-sampling network to up-sample the characteristics corresponding to the two-dimensional image to obtain the channel state information.
In the embodiment of the application, the image super-resolution technology is used for extracting the features of a plurality of low-resolution two-dimensional images to obtain corresponding features, and then the up-sampling is carried out based on the features to determine the channel state information with higher precision, so that more accurate channel estimation is realized.
In an optional implementation manner, the determining, based on the frequency domain signal and the channel state information, bit information sent by a sending end includes: zero forcing equalization is carried out on the frequency domain signal and the channel state information, and transmitted symbol estimation information is determined; and inputting the estimation information of the transmitting symbols into a pre-trained signal detection model to obtain bit information sent by a sending end.
In an optional implementation manner, the signal detection model includes a denoising network and a parallel detection network, and the inputting the estimation information of the transmission symbol into a pre-trained signal detection model to obtain bit information sent by a sending end includes: inputting the emission symbol estimation information into the denoising network for denoising to obtain denoised emission symbol estimation information; and inputting the denoised emission symbol estimation information into the parallel detection network to obtain the bit information.
In the embodiment of the application, the transmitted symbol estimation information is denoised by using the denoising network, noise in the transmitted symbol estimation information is removed, and then the denoised transmitted symbol estimation information is input into the parallel detection network, so that the error rate of the determined bit information is reduced, and the receiving performance of the OFDM receiver is improved.
In an optional embodiment, the preprocessing the frequency domain signal to obtain channel estimation information includes: determining a pilot frequency signal according to the frequency domain signal and the pilot frequency position; and performing least square channel estimation on the pilot signal and a local pilot signal to determine the channel estimation information.
In a second aspect, the present invention provides a channel estimation model training method, including: acquiring a plurality of frequency domain signals and complete channel state information corresponding to the frequency domain signals; preprocessing each frequency domain signal to obtain corresponding channel estimation information; determining a plurality of low-resolution two-dimensional images according to the corresponding channel estimation information; and taking the plurality of low-resolution two-dimensional images as training samples, and inputting complete channel state information corresponding to each frequency domain signal as a training label into a preset channel estimation model for training until the model is trained to be converged to obtain a trained channel estimation model.
In the embodiment of the application, channel estimation information corresponding to a plurality of frequency domain signals is used as a training sample, complete channel state information corresponding to each frequency domain signal is used as a training label to train a channel estimation model, and the pre-trained channel estimation model is used for extracting characteristic information in a plurality of low-resolution two-dimensional images, so that channel state information close to real channel state information is determined, and the accuracy of channel state information estimation is improved.
In a third aspect, the present invention provides an OFDM receiving apparatus, comprising: the acquisition module is used for acquiring a frequency domain signal; the preprocessing module is used for preprocessing the frequency domain signal to obtain channel estimation information; a determining module for determining a plurality of low-resolution two-dimensional images according to the channel estimation information; the channel prediction module is used for inputting the low-resolution two-dimensional images into a pre-trained channel estimation model to obtain channel state information; and the bit signal determining module is used for determining bit information sent by a sending end based on the frequency domain signal and the channel state information.
In an optional embodiment, the channel estimation model includes an image super-resolution network and an up-sampling network, and the channel prediction module is specifically configured to perform feature extraction on the plurality of low-resolution two-dimensional images by using the super-resolution network to obtain features corresponding to the two-dimensional images; and utilizing the up-sampling network to up-sample the characteristics corresponding to the two-dimensional image to obtain the channel state information.
In an optional implementation manner, the bit signal determining module is specifically configured to perform zero-forcing equalization on the frequency domain signal and the channel state information, and determine transmit symbol estimation information; and inputting the estimation information of the transmitting symbols into a pre-trained signal detection model to obtain bit information sent by a sending end.
In an optional embodiment, the signal detection model includes a denoising network and a parallel detection network, and the bit signal determination module is specifically configured to input the transmission symbol estimation information into the denoising network for denoising, so as to obtain denoised transmission symbol estimation information; and inputting the denoised emission symbol estimation information into the parallel detection network to obtain the bit information.
In an optional embodiment, the preprocessing module is specifically configured to determine a pilot signal according to the frequency domain signal and a pilot position; and performing least square channel estimation on the pilot signal and a local pilot signal to determine the channel estimation information.
In a fourth aspect, the present invention provides a channel estimation model training apparatus, including: an obtaining module, configured to obtain a plurality of frequency domain signals and complete channel state information corresponding to the plurality of frequency domain signals; the preprocessing module is used for preprocessing each frequency domain signal to obtain corresponding channel estimation information; a determining module for determining a plurality of low-resolution two-dimensional images according to the corresponding channel estimation information; and the training module is used for taking the plurality of low-resolution two-dimensional images as training samples, inputting complete channel state information corresponding to each frequency domain signal as a training label into a preset channel estimation model for training until the model is trained to be converged, and obtaining a trained channel estimation model.
In a fifth aspect, the present invention provides an electronic device, comprising: a processor, a memory, and a bus; the processor and the memory are communicated with each other through the bus; the memory stores program instructions executable by the processor, the processor being capable of executing the method of any one of the preceding embodiments when invoked by the processor.
In a sixth aspect, the present invention provides a computer storage medium having stored thereon computer program instructions which, when read and executed by a computer, perform the method according to any of the preceding embodiments.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are required to be used in the embodiments of the present application will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present application and therefore should not be considered as limiting the scope, and that those skilled in the art can also obtain other related drawings based on the drawings without inventive efforts.
Fig. 1 is a flowchart of an OFDM receiving method according to an embodiment of the present application;
fig. 2 is a flowchart of a channel estimation model training method according to an embodiment of the present disclosure;
fig. 3 is a structural diagram of a channel estimation model according to an embodiment of the present application;
fig. 4 is a structural diagram of a signal detection model according to an embodiment of the present application;
fig. 5 is a block diagram of an OFDM receiving apparatus according to an embodiment of the present application;
fig. 6 is a block diagram of a channel estimation model training apparatus according to an embodiment of the present disclosure;
fig. 7 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
Icon: 500-OFDM receiving means; 501-an obtaining module; 502-preprocessing module; 503-a determination module; 504-a channel prediction module; 505-a bit signal determination module; 600-a channel estimation model training means; 601-an obtaining module; 602-a pre-processing module; 603-a determination module; 604-a training module; 700-an electronic device; 701-a processor; 702 — a communication interface; 703-a memory; 704-bus.
Detailed Description
The technical solutions in the embodiments of the present application will be described below with reference to the drawings in the embodiments of the present application.
Referring to fig. 1, fig. 1 is a flowchart of an OFDM receiving method according to an embodiment of the present disclosure, where the OFDM receiving method includes the following steps:
step 101: and acquiring a frequency domain signal.
Step 102: and preprocessing the frequency domain signal to obtain channel estimation information.
Step 103: a plurality of low resolution two-dimensional images is determined from the channel estimation information.
Step 104: and inputting a plurality of low-resolution two-dimensional images into a pre-trained channel estimation model to obtain channel state information.
Step 105: and determining bit information sent by the sending end based on the frequency domain signal and the channel state information.
The above-described flow will be described in detail with reference to examples.
According to the above steps, in the present application, a channel estimation model trained in advance is used for channel estimation, so as to obtain channel state information, and further, bit information sent by a sending end is determined according to the channel state information and a frequency domain signal. To facilitate understanding of the present solution, before describing step S101, a training process of the channel estimation model is described.
Referring to fig. 2, fig. 2 is a flowchart of a channel estimation model training method according to an embodiment of the present disclosure, where the channel estimation model training method includes the following steps:
step 201: and acquiring a plurality of frequency domain signals and complete channel state information corresponding to the frequency domain signals.
Step 202: and preprocessing each frequency domain signal to obtain corresponding channel estimation information.
Step 203: a plurality of low resolution two-dimensional images is determined from the corresponding channel estimation information.
Step 204: and taking the plurality of low-resolution two-dimensional images as training samples, and inputting complete channel state information corresponding to each frequency domain signal as a training label into a preset channel estimation model for training to obtain a trained channel estimation model.
The above steps 201-203 will be described in detail below.
Step 201: and acquiring a plurality of frequency domain signals and complete channel state information corresponding to the frequency domain signals.
In the embodiment of the application, multiple groups of sending signals are generated locally and randomly, after the multiple groups of sending signals pass through an OFDM (orthogonal frequency division multiplexing) sender and a nonlinear amplifier, the multiple groups of sending signals are placed in channels with different signal-to-noise ratios for transmission, and the signals are received at an OFDM receiver. In an OFDM transmission signal, N is containedcSub-carriers and NsA time slot, wherein the pilot frequency adopts a comb pattern and has a size of Ncp×Nsp。
The signal received by the OFDM receiver is a time domain signal, and the received time domain signal is subjected to fast Fourier transform to obtain a frequency domain signal Y.
In the training of the channel estimation model, the condition of the channel is determined by artificial simulation. Therefore, the complete channel state information corresponding to the frequency domain signal is a known quantity, and can be directly determined according to the condition of the artificially simulated channel.
Step 202: and preprocessing each frequency domain signal to obtain corresponding channel estimation information.
In the embodiment of the application, after the frequency domain signal Y is obtained, the pilot frequency position is referred to, and the pilot frequency signal Y is extractedp. Will be the local pilot signal XpAnd a pilot signal YpPerforming least square channel estimation to determine channel estimation information corresponding to the frequency domain signal Y
Wherein, Xp,i，Yp,i，Hp,i，
Respectively representing a local pilot signal, a pilot signal, frequency domain channel information and channel estimation information at the time of the ith training. Since the channel is an artificially simulated channel, the frequency domain channel information is a known quantity. diag (X)p,i) A square diagonal matrix representing pilots, vector Xp,iIs located on the main diagonal of the square diagonal matrix.
Step 203: a plurality of low resolution two-dimensional images is determined from the corresponding channel estimation information.
In the embodiment of the application, the real part of the channel estimation information in the ith training is used
And imaginary part
And splitting and splicing the two-dimensional tensor into a two-dimensional tensor. The two-dimensional tensor can be considered as a two-dimensional low resolution image. Since the pilot size is Ncp×NspThus, the two-dimensional tensor can be thought of as an Ncp×NspOf the matrix of (a). Decomposing the two-dimensional tensor into N according to the time slot dimension of the pilot frequencyspN iscpA column vector of x 1. Then according to pilot frequency NspSize of (2), N isspN iscpX 1 column vector shaping to NspA two-dimensional image of low resolution.
In the shaping, the N may be the same as the NcpWhether or not the squaring can determine the form of the low-resolution two-dimensional image. For example, if NcpIs 9, NcpCan be squared to form NcpThe x 1 column vector is shaped into a 3 x 3 low resolution two-dimensional image; if N is presentcpIs 10, NcpCan not be squared, thenWill NcpThe x 1 column vector is shaped into a 2 x 5 low resolution two-dimensional image. It should be understood that the above shaping manner is only one specific implementation manner provided in the embodiments of the present application, and those skilled in the art may make corresponding adjustments according to actual needs, and the present application is not limited thereto.
Step 204: and taking a plurality of low-resolution two-dimensional images as training samples, and inputting complete channel state information corresponding to each frequency domain signal as a training label into a preset channel estimation model for training until the model is trained to be converged to obtain a trained channel estimation model.
In this embodiment, as an optional implementation manner, as shown in fig. 3, the preset channel estimation model may include an image super-resolution network and an upsampling network.
The input layer of the image super-resolution network is a convolutional layer. Wherein the convolutional layer comprises a convolutional kernel and a filter. Specifically, a convolution kernel of 3 × 3 is adopted, and in order to ensure that the number of pictures is not changed, N is adoptedspA filter. The hidden layer is composed of a plurality of Residual Feature Aggregation (RFA) modules and a convolutional layer. By adding a jump connection mode between residual error feature aggregation modules, features are directly propagated on each local residual error branch, and residual error features in pilot frequency information are more effectively extracted. Each RFA module consists of four residual modules, each consisting of two convolutional layers (step size 1) and one leakage ReLU activation function. The jumps of the first three residual modules are connected to the end of the RFA module and combined with the last residual module. Finally, all features are integrated by one convolutional layer.
Image super-resolution network pair NspAnd extracting features of the two-dimensional image with low resolution, and extracting pilot frequency information and correlation.
The up-sampling network replaces the traditional interpolation process, consists of two layers of neural networks, has the size twice of the state information of the complete channel, and does not adopt an activation function. The first layer network is used for mapping the low-resolution image into a high-resolution image, the second layer network finely adjusts the image through the weight and the offset of the neuron, and finally the high-resolution image is outputEstimation result of rate
Will be provided with
Real part of
And imaginary part
Combining to obtain channel state information
In actual use, the number of network layers and the network parameters of each layer can be modified according to actual conditions so as to improve the generalization capability and robustness of the network.
The penalty function used to train the preset channel estimation model is:
wherein, thetas,θmRespectively super-resolution network and up-sampling network parameters, fs,fmThe function is activated for both networks.
As an alternative implementation, 50 batches of channel estimation model training are adopted, the training times are 4000, and each training comprises 300 time slot samples; the learning rate is initialized to 0.01 and decreases as the number of training times increases; the optimizer adopts an Adam optimizer.
It can be understood that the values of the training batch, the training times, the time slot samples, and the learning rate are all specific implementations provided in the embodiments of the present application, and those skilled in the art may perform corresponding adjustments according to actual needs, which is not limited to this application.
As an optional implementation manner, after the channel state information is determined, zero-forcing equalization is performed on the frequency domain signal and the channel state information to determine transmission symbol estimation information, and then the transmission symbol estimation information is input into a pre-trained signal detection model to obtain bit information sent by a sending end. For ease of understanding, the training process of the signal detection model will be described first.
It should be noted that, the training step of the signal detection model may be performed after the training step of the channel estimation model, and the frequency domain signal determined by the simulated channel condition during the training of the channel estimation model and the channel state information corresponding to the frequency domain signal are used.
In the embodiment of the application, a plurality of groups of frequency domain signals and channel state information corresponding to the plurality of groups of frequency domain signals are obtained first. And performing zero-forcing equalization on the channel state information and the frequency domain signal corresponding to the channel state information to determine a transmitting symbol estimation value, and taking the transmitting symbol estimation value as a training sample of a signal detection model.
Specifically, zero-forcing equalization of the frequency domain signal and the channel state information may be expressed as:
wherein, YiAnd
respectively an ith training time-frequency domain signal and channel state information,
is the estimated value of the transmitted symbol determined in the ith training.
Bit information B to be transmitted by a transmitting endiAs a training label for the signal detection model. It should be noted that, in the embodiments of the present application, a frequency domain signal determined by a simulated channel condition during channel estimation model training and channel state information corresponding to the frequency domain signal are used. Therefore, the bit information transmitted by the transmitting end is a known quantity.
Will transmit a symbol estimate
Real part of
And imaginary part
Splitting and splicing the two-dimensional tensor into a two-dimensional tensor as a training sample of a signal detection model, and transmitting bit information B sent by a transmitting endiAnd inputting the training labels serving as the signal detection models into a preset signal detection model for training until the model is trained to be converged to obtain the trained signal detection model.
As shown in fig. 4, the preset signal detection model may include a denoising network and a parallel detection network.
As an alternative embodiment, the denoising network includes an input layer, an intermediate hidden layer, and an output layer. Specifically, the input layer of the denoised network uses 16 filters of size 3 × 3 × 1 for generating 16 feature maps, while adding bulk normalization and a nonlinear activation function ReLU. The intermediate hidden layer consists of a plurality of convolutional layers, each convolutional layer having a filter of 16 sizes, and a batch normalization and nonlinear activation function ReLU is introduced at each layer. The output layer uses 1 filter of size 3 × 3 × 16 to reconstruct the signal and output the denoised transmit symbol estimates.
The parallel detection network can be composed of three same parallel networks, each detection network is composed of a multilayer neural network, and the transmission symbol estimation value after denoising is mapped into bit information of the transmitting end by adopting a full connection mode between every two layers of neural networks. The number of neurons of the output layer of each parallel detection network is the mapping bit number of the current network, and the bit numbers of the multiple parallel networks are connected in series to form bit information sent by the sending end.
In the training process, the average mean square error is used as a loss function of a training signal detection model:
wherein, thetaDAs a network parameter, fDIs a non-linear mapping function.
As an alternative embodiment, the signal detection model training uses 50 batches, the training frequency is 2000 times, and each training comprises 300 time slot samples; the learning rate is initialized to 0.01 and decreases as the number of training times increases; the optimizer adopts an Adam optimizer.
It can be understood that the values of the training batch, the training times, the time slot samples, and the learning rate are all specific implementations provided in the embodiments of the present application, and those skilled in the art may perform corresponding adjustments according to actual needs, which is not limited to this application.
In the embodiment of the application, the channel is a time-varying channel and has white gaussian noise when OFDM is transmitted. And adding a denoising network in the preset signal detection model for removing the Gaussian white noise in the emission symbol estimation value, so that the trained signal detection model can remove the Gaussian white noise introduced by the time-varying channel in use, and the accuracy of signal detection is improved.
The embodiment of the application introduces deep learning into the traditional OFDM receiver for OFDM channel estimation and signal detection, improves the bit error rate performance of the receiver and can overcome the nonlinear distortion of a nonlinear amplifier to the signal. The receiver trains a channel estimation model and a signal detection model in an offline training mode, and can perform transfer training on the channel estimation model and the signal detection model in actual use to obtain better performance. The two models both adopt a multi-signal-to-noise ratio training mode, and can complete the estimation and signal detection of the nonlinear channel under the condition of no prior information such as channel, noise and the like.
After the training procedures of the channel estimation model and the signal detection model are introduced, the above steps S101 to S105 will be described in detail with reference to the examples.
Step 101: and acquiring a frequency domain signal.
In the embodiment of the present application, the OFDM system may adopt a single-input single-output mode, and one OFDM transmission signal includes NcSub-carriers and NsA time slot, wherein the pilot frequency adopts a comb pattern and has a size of Ncp×Nsp. An OFDM transmission signal passes through an OFDM transmitter and a nonlinear amplifier, and then a time domain signal is received by an OFDM receiver through transmission of a channel. And carrying out fast Fourier transform on the received time domain signal to obtain a frequency domain signal.
Step 102: and preprocessing the frequency domain signal to obtain channel estimation information.
Step 103: a plurality of low resolution two-dimensional images is determined from the channel estimation information.
In the embodiment of the present application, step 102 and step 103 correspond to step 202 and step 203, and for brevity of the description, the same or similar parts may be referred to each other, and are not described herein again.
Step 104: and inputting a plurality of low-resolution two-dimensional images into a pre-trained channel estimation model to obtain channel state information.
In the embodiment of the present application, as can be known from the foregoing description of the training process of the channel estimation model, the channel estimation model includes an image super-resolution network and an upsampling network. The super-resolution network performs feature extraction on a plurality of low-resolution two-dimensional images to obtain features corresponding to the two-dimensional images; and the up-sampling network up-samples the characteristics corresponding to the two-dimensional image to obtain the channel state information. The specific implementation process of step 104 corresponds to step 204, and for the sake of brevity, the same or similar parts may be referred to each other, and are not described herein again.
Step 105: and determining bit information sent by the sending end based on the frequency domain signal and the channel state information.
As an alternative implementation, step 105 may include the following steps:
firstly, performing zero forcing equalization on a frequency domain signal and channel state information to determine transmitted symbol estimation information;
and secondly, inputting the estimation information of the transmitted symbols into a pre-trained signal detection model to obtain bit information sent by a sending end.
Specifically, zero-forcing equalization of the frequency domain signal and the channel state information may be expressed as:
wherein,
is the channel state information of the k-th sub-carrier outputted by the channel estimation module, y (k) is the k-th sub-carrier of the frequency domain signal,
and transmitting symbol estimation information of the k subcarrier.
Will be provided with
Real part of
And imaginary part
Splitting and splicing the signals into a two-dimensional tensor, inputting the two-dimensional tensor into a pre-trained signal detection model, and obtaining bit information sent by a sending end.
According to the foregoing description of the training process of the signal detection model, the signal detection model includes a denoising network and a parallel detection network. The working principle of the denoising network and the parallel detection network corresponds to the introduction of the signal detection model in the training process, and the same or similar parts can be referred to each other for the sake of concise specification, and are not repeated herein.
To sum up, in the embodiment of the present application, channel estimation information is determined according to a frequency domain signal, the channel estimation information (channel state information with lower accuracy) is decomposed into a plurality of low-resolution two-dimensional images, the plurality of low-resolution two-dimensional images are input into a channel estimation model, and feature information in the plurality of low-resolution two-dimensional images is extracted by using a pre-trained channel estimation model, so that channel state information with higher accuracy is determined, and further, bit information sent by a sending end is determined based on the channel state information, so that the receiving performance of an OFDM receiver is improved.
Based on the same inventive concept, the embodiment of the application also provides an OFDM receiving device. Referring to fig. 5, fig. 5 is a block diagram of an OFDM receiving apparatus according to an embodiment of the present disclosure, where the OFDM receiving apparatus 500 may include:
an obtaining module 501, configured to obtain a frequency domain signal;
a preprocessing module 502, configured to preprocess the frequency domain signal to obtain channel estimation information;
a determining module 503, configured to determine a plurality of low-resolution two-dimensional images according to the channel estimation information;
a channel prediction module 504, configured to input the multiple low-resolution two-dimensional images into a pre-trained channel estimation model to obtain channel state information;
a bit signal determining module 505, configured to determine bit information sent by a sending end based on the frequency domain signal and the channel state information.
In an optional embodiment, the channel estimation model includes an image super-resolution network and an up-sampling network, and the channel prediction module 504 is specifically configured to perform feature extraction on the multiple low-resolution two-dimensional images by using the super-resolution network to obtain features corresponding to the two-dimensional images; and utilizing the up-sampling network to up-sample the characteristics corresponding to the two-dimensional image to obtain the channel state information.
In an optional implementation manner, the bit signal determining module 505 is specifically configured to perform zero-forcing equalization on the frequency domain signal and the channel state information, and determine transmit symbol estimation information; and inputting the estimation information of the transmitting symbols into a pre-trained signal detection model to obtain bit information sent by a sending end.
In an optional implementation manner, the signal detection model includes a denoising network and a parallel detection network, and the bit signal determination module 505 is specifically configured to input the transmission symbol estimation information into the denoising network for denoising, so as to obtain denoised transmission symbol estimation information; and inputting the denoised emission symbol estimation information into the parallel detection network to obtain the bit information.
In an optional embodiment, the preprocessing module 502 is specifically configured to determine a pilot signal according to the frequency domain signal and a pilot position; and performing least square channel estimation on the pilot signal and a local pilot signal to determine the channel estimation information.
In addition, the embodiment of the application also provides a device for training the channel estimation model. Referring to fig. 6, fig. 6 is a block diagram illustrating a channel estimation model training apparatus according to an embodiment of the present disclosure, where the channel estimation model training apparatus 600 may include:
an obtaining module 601, configured to obtain a plurality of frequency domain signals and complete channel state information corresponding to the plurality of frequency domain signals;
a preprocessing module 602, configured to preprocess each frequency domain signal to obtain corresponding channel estimation information;
a determining module 603, configured to determine a plurality of low-resolution two-dimensional images according to corresponding channel estimation information;
the training module 604 is configured to use the multiple low-resolution two-dimensional images as training samples, and input complete channel state information corresponding to each frequency domain signal as a training tag into a preset channel estimation model for training until the model is trained to converge, so as to obtain a trained channel estimation model.
Referring to fig. 7, fig. 7 is a schematic structural diagram of an electronic device 700 according to an embodiment of the present application, where the electronic device 700 includes: at least one processor 701, at least one communication interface 702, at least one memory 703 and at least one bus 704. Wherein the bus 704 is used for implementing direct connection communication of these components, the communication interface 702 is used for communicating signaling or data with other node devices, and the memory 703 stores machine-readable instructions executable by the processor 701. When the electronic device 700 is in operation, the processor 701 communicates with the memory 703 via the bus 704, and the machine readable instructions, when invoked by the processor 701, perform the OFDM reception method as described above.
The processor 701 may be an integrated circuit chip having signal processing capabilities. The Processor 701 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 Gate or transistor logic devices, discrete hardware components. Which may implement or perform the various methods, steps, and logic blocks disclosed in the embodiments of the present application. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The Memory 703 may include, but is not limited to, Random Access Memory (RAM), Read Only Memory (ROM), Programmable Read Only Memory (PROM), Erasable Read Only Memory (EPROM), electrically Erasable Read Only Memory (EEPROM), and the like.
It will be appreciated that the configuration shown in fig. 7 is merely illustrative and that electronic device 700 may include more or fewer components than shown in fig. 7 or have a different configuration than shown in fig. 7. The components shown in fig. 7 may be implemented in hardware, software, or a combination thereof. In this embodiment, the electronic device 700 may be, but is not limited to, an entity device such as a desktop, a laptop, a smart phone, an intelligent wearable device, and a vehicle-mounted device, and may also be a virtual device such as a virtual machine. In addition, the electronic device 700 is not necessarily a single device, but may also be a combination of multiple devices, such as a server cluster, and the like.
In addition, an embodiment of the present application further provides a computer storage medium, where a computer program is stored on the computer storage medium, and when the computer program is executed by a computer, the steps of the OFDM receiving method in the foregoing embodiment are executed.
In the embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other ways. The above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is only one logical division, and there may be other divisions when actually implemented, and for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection of devices or units through some communication interfaces, and may be in an electrical, mechanical or other form.
In addition, units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
Furthermore, the functional modules in the embodiments of the present application may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.
It should be noted that the functions, if implemented in the form of software functional modules and sold or used as independent products, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application or portions thereof that substantially contribute to the prior art may be embodied in the form of a software product stored in a storage medium and including instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present application. And the aforementioned storage medium includes: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.
In this document, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions.
The above description is only an example of the present application and is not intended to limit the scope of the present application, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Claims (10)
1. An OFDM receiving method, comprising:
acquiring a frequency domain signal;
preprocessing the frequency domain signal to obtain channel estimation information;
determining a plurality of low-resolution two-dimensional images according to the channel estimation information;
inputting the two-dimensional images with the low resolution into a pre-trained channel estimation model to obtain channel state information;
and determining bit information sent by a sending end based on the frequency domain signal and the channel state information.
2. The method of claim 1, wherein the channel estimation model comprises an image super-resolution network and an up-sampling network, and the inputting the plurality of low-resolution two-dimensional images into a pre-trained channel estimation model to obtain channel state information comprises:
extracting the features of the two-dimensional images with low resolution by using the super-resolution network to obtain the features corresponding to the two-dimensional images;
and utilizing the up-sampling network to up-sample the characteristics corresponding to the two-dimensional image to obtain the channel state information.
3. The method of claim 1, wherein the determining bit information transmitted by a transmitting end based on the frequency domain signal and the channel state information comprises:
zero forcing equalization is carried out on the frequency domain signal and the channel state information, and transmitted symbol estimation information is determined;
and inputting the estimation information of the transmitting symbols into a pre-trained signal detection model to obtain bit information sent by a sending end.
4. The method of claim 3, wherein the signal detection model includes a de-noising network and a parallel detection network, and the inputting the estimation information of the transmitted symbol into a pre-trained signal detection model to obtain the bit information sent by the transmitting end includes:
inputting the emission symbol estimation information into the denoising network for denoising to obtain denoised emission symbol estimation information;
and inputting the denoised emission symbol estimation information into the parallel detection network to obtain the bit information.
5. The method of claim 1, wherein the pre-processing the frequency domain signal to obtain channel estimation information comprises:
determining a pilot frequency signal according to the frequency domain signal and the pilot frequency position;
and performing least square channel estimation on the pilot signal and a local pilot signal to determine the channel estimation information.
6. A method for training a channel estimation model, comprising:
acquiring a plurality of frequency domain signals and complete channel state information corresponding to the frequency domain signals;
preprocessing each frequency domain signal to obtain corresponding channel estimation information;
determining a plurality of low-resolution two-dimensional images according to the corresponding channel estimation information;
and taking the plurality of low-resolution two-dimensional images as training samples, and inputting complete channel state information corresponding to each frequency domain signal as a training label into a preset channel estimation model for training until the model is trained to be converged to obtain a trained channel estimation model.
7. An OFDM receiving apparatus, characterized in that the apparatus comprises:
the acquisition module is used for acquiring a frequency domain signal;
the preprocessing module is used for preprocessing the frequency domain signal to obtain channel estimation information;
a determining module for determining a plurality of low-resolution two-dimensional images according to the channel estimation information;
the channel prediction module is used for inputting the low-resolution two-dimensional images into a pre-trained channel estimation model to obtain channel state information;
and the bit signal determining module is used for determining bit information sent by a sending end based on the frequency domain signal and the channel state information.
8. An apparatus for training a channel estimation model, the apparatus comprising:
an obtaining module, configured to obtain a plurality of frequency domain signals and complete channel state information corresponding to the plurality of frequency domain signals;
the preprocessing module is used for preprocessing each frequency domain signal to obtain corresponding channel estimation information;
a determining module for determining a plurality of low-resolution two-dimensional images according to the corresponding channel estimation information;
and the training module is used for taking the plurality of low-resolution two-dimensional images as training samples, inputting complete channel state information corresponding to each frequency domain signal as a training label into a preset channel estimation model for training until the model is trained to be converged, and obtaining a trained channel estimation model.
9. An electronic device, comprising: a processor, a memory, and a bus; the processor and the memory are communicated with each other through the bus; the memory stores program instructions executable by the processor, the processor invoking the program instructions to perform the method of any of claims 1-6.
10. A computer storage medium having computer program instructions stored thereon that, when read and executed by a computer, perform the method of any one of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111020964.9A CN113726711B (en) | 2021-09-01 | 2021-09-01 | OFDM receiving method and device, and channel estimation model training method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111020964.9A CN113726711B (en) | 2021-09-01 | 2021-09-01 | OFDM receiving method and device, and channel estimation model training method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113726711A true CN113726711A (en) | 2021-11-30 |
| CN113726711B CN113726711B (en) | 2023-09-15 |
Family
ID=78680529
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111020964.9A Active CN113726711B (en) | 2021-09-01 | 2021-09-01 | OFDM receiving method and device, and channel estimation model training method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113726711B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114301742A (en) * | 2021-12-23 | 2022-04-08 | 北京邮电大学 | Channel estimation method and device |
| US11979265B2 (en) | 2021-04-13 | 2024-05-07 | Samsung Electronics Co., Ltd. | Learning-based common phase error estimation |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110266620A (en) * | 2019-07-08 | 2019-09-20 | 电子科技大学 | 3D MIMO-OFDM system channel estimation method based on convolutional neural networks |
| WO2021112360A1 (en) * | 2019-12-01 | 2021-06-10 | 엘지전자 주식회사 | Method and device for estimating channel in wireless communication system |
| CN113285899A (en) * | 2021-05-20 | 2021-08-20 | 南京邮电大学 | Time-varying channel estimation method and system based on deep learning |
-
2021
- 2021-09-01 CN CN202111020964.9A patent/CN113726711B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110266620A (en) * | 2019-07-08 | 2019-09-20 | 电子科技大学 | 3D MIMO-OFDM system channel estimation method based on convolutional neural networks |
| WO2021112360A1 (en) * | 2019-12-01 | 2021-06-10 | 엘지전자 주식회사 | Method and device for estimating channel in wireless communication system |
| CN113285899A (en) * | 2021-05-20 | 2021-08-20 | 南京邮电大学 | Time-varying channel estimation method and system based on deep learning |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11979265B2 (en) | 2021-04-13 | 2024-05-07 | Samsung Electronics Co., Ltd. | Learning-based common phase error estimation |
| CN114301742A (en) * | 2021-12-23 | 2022-04-08 | 北京邮电大学 | Channel estimation method and device |
| CN114301742B (en) * | 2021-12-23 | 2024-01-09 | 北京邮电大学 | Channel estimation method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113726711B (en) | 2023-09-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113726711B (en) | OFDM receiving method and device, and channel estimation model training method and device | |
| CN113141325B (en) | Training method, identification method and device for optical OFDM signal subcarrier modulation format identification model | |
| CN110266620A (en) | 3D MIMO-OFDM system channel estimation method based on convolutional neural networks | |
| CN109995449A (en) | A kind of millimeter-wave signal detection method based on deep learning | |
| CN110417694A (en) | A kind of modulation mode of communication signal recognition methods | |
| CN113708855B (en) | OTFS data driving and receiving method, system and medium based on deep learning | |
| CN111431831A (en) | Multi-dimensional OFDM environment-based adaptive modulation method and system | |
| CN115664898A (en) | OFDM system channel estimation method and system based on complex convolution neural network | |
| An et al. | Robust automatic modulation classification in low signal to noise ratio | |
| CN114826832A (en) | Channel estimation method, neural network training method, device and equipment | |
| CN115250216A (en) | Underwater sound OFDM combined channel estimation and signal detection method based on deep learning | |
| CN114548201B (en) | Automatic modulation identification method and device for wireless signal, storage medium and equipment | |
| Zhang et al. | Deep learning with a self-adaptive threshold for OTFS channel estimation | |
| CN117640327A (en) | Anti-interference multi-carrier data transmission method and device | |
| CN114884780A (en) | Underwater acoustic communication signal modulation identification method and device based on passive time reversal mirror | |
| CN114244675B (en) | MIMO-OFDM system channel estimation method based on deep learning | |
| CN116070136A (en) | Multi-mode fusion wireless signal automatic modulation recognition method based on deep learning | |
| CN113869227B (en) | Signal modulation mode identification method, device, equipment and readable storage medium | |
| Zhang et al. | Deep learning‐based digital signal modulation identification under different multipath channels | |
| CN114826850A (en) | Modulation identification method, device and equipment based on time-frequency diagram and deep learning | |
| CN115086123A (en) | Modulation identification method and system based on fusion of time-frequency graph and constellation diagram | |
| CN113242201A (en) | Wireless signal enhanced demodulation method and system based on generation classification network | |
| Ivanov et al. | Hybrid noise-resilient deep learning architecture for modulation classification in cognitive radio networks | |
| Bai et al. | An efficient automatic modulation classification method based on the convolution adaptive noise reduction network | |
| CN111917674A (en) | Modulation identification method based on deep learning |
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 |