CN111865863B - RNN neural network-based OFDM signal detection method - Google Patents
RNN neural network-based OFDM signal detection method Download PDFInfo
- Publication number
- CN111865863B CN111865863B CN202010700419.3A CN202010700419A CN111865863B CN 111865863 B CN111865863 B CN 111865863B CN 202010700419 A CN202010700419 A CN 202010700419A CN 111865863 B CN111865863 B CN 111865863B
- Authority
- CN
- China
- Prior art keywords
- neural network
- model
- training
- signal detection
- ofdm signal
- 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.)
- Active
Links
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
- H04L27/2626—Arrangements specific to the transmitter only
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- 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
-
- 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
- H04L25/03—Shaping networks in transmitter or receiver, e.g. adaptive shaping networks
- H04L25/03006—Arrangements for removing intersymbol interference
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Power Engineering (AREA)
- Monitoring And Testing Of Transmission In General (AREA)
Abstract
The invention discloses an OFDM signal detection method based on an RNN neural network, which can be applied to a cognitive radio spectrum sensing technology of wireless mobile communication, and can effectively improve the recovery capability of an OFDM system in the aspect of receiver signal demodulation, thereby improving the signal detection performance of cognitive radio spectrum sensing. On the basis of analyzing the OFDM system, firstly, an OFDM system framework is built, and a time series data set for model learning is generated by utilizing the OFDM system framework. Then, a time series signal data set is learned through an LSTM model of an RNN neural network, and end-to-end spectrum signal detection is realized by adopting an Adam training algorithm to quickly optimize parameters and a small-batch training mode, so that the problem of receiving OFDM system signals with NBI and ICI is effectively solved, and the spectrum detection performance of the system is improved.
Description
Technical Field
The invention relates to the technical field of wireless mobile communication, in particular to an OFDM signal detection method based on an RNN neural network.
Background
The continued evolution of wireless communication technology has made information transfer faster and more reliable. For a recent time, OFDM technology has played an important role in facilitating fast flow of data streams as a key technology in wireless communication. The OFDM technology converts serial transmission of data stream into parallel transmission through orthogonal sub-carrier, thereby well coping with selective fading and narrow-band interference (NBI); the Cyclic Prefix (CP) is added at the same time, so that the method has better performance in resisting intersymbol interference (ISI) and intercarrier interference (ICI). In Cognitive Radio (CR), OFDM technology is also the preferred technology for CR transmission.
Although the OFDM technology has many advantages, in the OFDM system, information is affected by doppler shift and phase noise during high-speed transmission, which results in poor performance of the OFDM system. Although the CP is inserted into the OFDM system, the subcarrier interference resistance of the OFDM system is effectively improved, the orthogonality of the OFDM signal is still affected by the interference of other factors, so that a good signal detection effect cannot be obtained when the OFDM transmission technology is used for spectrum sensing in cognitive radio. The current techniques for reducing inter-subcarrier interference include: time domain windowing, frequency domain equalization, introduction of other techniques.
Since Hinton carries out layer-by-layer training to obtain high-dimensional features by modeling the hidden layer of the neural network through a limited Boltzmann machine, and a deep belief network is proposed. With the increasing number of hidden layers, its team redefines artificial neural networks as deep learning. Deep learning can actively learn data, high-dimensional features of the data are extracted, labor time is saved, and efficiency is improved. In recent years, deep learning has a great application value in image processing, automatic driving, robotics, and medical health systems. Some outstanding performances of deep learning meet the requirements of high signal transmission efficiency and high accuracy of future wireless communication systems. The optimization method adopts various advanced optimization algorithms such as Adagarad, Momentum, RMSprop, Adam and the like, improves the parameter updating performance of the neural network, prevents the local optimal problem and improves the convergence capability of the model.
In summary, in order to improve the signal detection capability of the OFDM system in wireless communication and increase the spectrum utilization rate, a neural network model based on deep learning is urgently needed to be found to improve the signal detection performance of the OFDM system by using the strong learning capability of deep learning.
Disclosure of Invention
Aiming at the problem that the OFDM signal detection accuracy is reduced due to the influence of narrowband interference and noise in the OFDM transmission technology applied to CR spectrum sensing, the invention provides an OFDM signal detection method based on an RNN neural network for the OFDM transmission technology based on deep learning, and the strong learning and memory capacity of the RNN neural network is utilized to complete signal detection.
The invention adopts the following technical scheme for solving the technical problems:
the invention provides an OFDM signal detection method based on an RNN neural network, which finishes OFDM signal detection through the RNN neural network, and the realization mode comprises the following steps:
step 1: establishing a spectrum signal detection model and initializing network parameters, and setting iteration times and other hyper-parameters;
step 2: inputting a data set comprising a training set and a verification set, carrying out forward propagation training, and calculating loss;
and step 3: finishing a time back propagation algorithm through a gradient descent algorithm, and updating parameters until training is finished;
and 4, step 4: counting a loss curve and an accuracy fitting curve, simultaneously paying attention to training time, if the training time is not converged, adjusting a hyper-parameter, returning to the step 2, and if the training time is not converged, continuing to carry out the next step;
and 5: inputting test set sequence data to complete model test.
As a further technical scheme of the invention, the RNN neural network is built by using a Matlab2019 deep learning toolbox.
As a further technical scheme of the invention, the RNN depth detection network uses an LSTM model in a CPU mode, and model training is completed by using an Adam optimization algorithm and a small-batch data processing mode.
As a further technical scheme of the invention, the learning rate of model training is 0.01, the small batch size is 1000, the number of input nodes is 256, the number of hidden layer nodes is 16, the number of hidden layer layers is 1, the number of full-connection layers is 4, and the learning rate attenuation factor is 0.1, and verification is performed every 50 times of training.
As a further technical scheme of the invention, pilot frequency is added in a signal transmitter for better channel estimation, thus being beneficial to the detection of signals; in the last OFDM signal demodulation and recovery process of the receiver, a deep learning model is added, the memory capacity of the RNN neural network to time sequence signals can be well exerted, and the memory of OFDM signal channel information is completed through continuous cyclic training in a small batch mode until the loss function output value is optimal.
As a further technical scheme of the invention, the CP number inserted into the model is ensured to be unchanged, and the training and the testing of the detection model are carried out under the condition of different pilot numbers.
As a further technical solution of the present invention, a loss function of the RNN deep inspection network model is defined as:
in the formula (I), the compound is shown in the specification,theoretical value of the model, y(m)Representing the model predicted values.
As a further technical solution of the present invention, when a QPSK modulation scheme or other signal modulation schemes are adopted, the OFDM signal detection model can be equivalently migrated and learned, so as to implement OFDM signal detection under different modulation schemes.
Due to the adoption of the technical scheme, compared with the prior art, the invention has the following technical effects: the optimization of OFDM signal detection is achieved through the strong memory capacity of an LSTM model of an RNN neural network, the problem that the detection accuracy is reduced due to the influence of noise and other interference factors in the OFDM transmission technology in CR spectrum sensing is solved, the problem of receiving signals of an OFDM system with NBI and ICI is effectively solved, and the spectrum detection performance of the system is improved.
Drawings
FIG. 1 is a schematic diagram of an overall structure of an RNN neural network-based OFDM signal detection model designed by the present invention;
FIG. 2 is a comparison graph of RNN neural network model detection performance for an insertion pilot number of 32, a subcarrier number of 64, and a QPSK modulation mode according to the present invention;
FIG. 3 is a comparison graph of RNN neural network model detection performance for an insertion pilot number of 48, a subcarrier number of 64, and a QPSK modulation mode according to the present invention;
fig. 4 is a comparison graph of RNN neural network model detection performance for an inserted pilot number of 64, a subcarrier number of 64, and a modulation scheme of QPSK in accordance with the present invention.
Detailed Description
The technical solution in the embodiment of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiment of the present invention:
as can be seen from the accompanying drawings, in the RNN neural network-based OFDM signal detection method according to this embodiment, an overall model structure of an RNN neural network detection model is shown in fig. 1, and the steps are as follows:
step 1: and (5) building an OFDM system framework.
The OFDM symbol may be represented as:
in the formula, ts≤t≤ts+ T, N being the number of subchannels, T being the OFDM symbol width, di(i 1.., N-1) is a data symbol of a subchannel, fcIs a carrier frequency, t being t at time tsAnd starting.
In the published literature, the mathematical expression for the equivalent baseband signal describing the OFDM output signal is:
in the formula, ts≤t≤ts+T。
As can be seen from equation (2), the OFDM signal is different from other transmission signals in that it is a complex baseband signal and has orthogonality. Then let t s0, t ═ kT/N (k ═ 0,1,2,. N-1) yields:
wherein k is not less than 0 and not more than N-1, s (k) represents the k-th path diAn Inverse Discrete Fourier Transform (IDFT) operation.
In order to obtain the original information-bearing symbol diThe receiver computes a Discrete Fourier Transform (DFT) transform on s (k), the mathematical expression being:
theoretical studies have shown that in OFDM systems, the modulation and demodulation are replaced by mathematical IDFT and DFT transforms. When used in the field, OFDM systems utilize a more optimal IFFT and FFT transformation.
In step 1, the transmitter of the OFDM system specifically implements the following steps:
step 1.1: and (5) encoding.
Step 1.2: and (4) interleaving. Interleaving is the data channel used to reduce burst errors. IQ completes signal mapping through a series-parallel converter. Where I denotes the in-phase signal and Q denotes the quadrature signal.
Step 1.3: and (4) digital modulation. And completing constellation mapping in the OFDM system mode. The IQ values produced in this process are filtered and sent to the IFFT for conversion.
Step 1.4: and inserting a pilot frequency.
Step 1.5: and (4) performing series-parallel conversion. The serial input signal is output in parallel onto the M lines.
Step 1.6: an IFFT. The frequency domain discrete signal is converted into a time domain discrete signal.
Step 1.7: and (4) parallel conversion.
Step 1.8: inserting a CP and windowing.
Step 2: the method comprises the steps of modeling a transmitter and a receiver of the OFDM system, transmitting OFDM signals through the steps of modulation of the transmitter, pilot frequency insertion, serial-parallel conversion, IFFT, CP insertion and the like, and receiving the OFDM signals through the steps of CP removal, parallel-serial conversion, FFT and the like of the receiver. In order to ensure the correlation of time domain signals, OFDM signal time sequence data is received according to the time sequence, the total number of data is 10000 groups, one OFDM signal comprises OFDM data symbols and OFDM pilot symbols, and each training sample comprises a received OFDM packet. The ratio of the training set, the validation set and the test set is 4:1:1, and the OFDM signal sequence data is defined as x ═ x [ -1: [ ]1,x2,x3,x4,L,xn]Adding a corresponding label y to the data ═ y1,y2,y3,y4,L,ym](y i0, 1). The data set is expressed by matrix form as:
random noise is added in the OFDM signal generation process to ensure that the data is more authoritative.
And step 3: and (5) training and testing the model. The most common training method of the recurrent neural network LSTM model is training by using a time back propagation algorithm, which can be divided into two parts, namely forward propagation of a time series data stream and back propagation of an error. The forward propagation is that time sequence data is input from an input layer, then the time sequence data enters a hidden layer to extract high-dimensional characteristics, and learning of weight and bias in the recurrent neural network is completed; the time back propagation is to perform supervised learning, calculate the error of the previous layer by using the error of the output layer in a back-stepping mode, advance the previous layer by layer, and correct the parameters of the model by using the errors among the layers.
Specifically, a recurrent neural network model with an N-layer structure is established, wherein the 1 st layer and the N-th layer are an input layer and an output layer respectively. The forward training process is a layer-by-layer derivation process from an input layer to an output layer, and the existing m pairs of training data { (x)(1),y(1)),(x(2),y(2)),L,(x(m),y(m)) In which x(i)For time series of OFDM signals, y(i)Is the label to which it is attached. The training cyclic neural network adopts a gradient descent algorithm, and the output mapping mathematical expression from the input layer to the output layer is as follows:
in the formula, W and b are respectively the weight and bias value of the LSTM model to be trained.
Due to theoretical values in the forward propagation processAnd predicted value y(i)Certain errors exist, so the invention adopts a more suitable cross entropy function as a loss function, and the mathematical expression can be expressed as:
in order to reduce cross entropy loss to the minimum, the parameters of each layer of the neural network are continuously optimized and adjusted by using a back propagation algorithm.
After the recurrent neural network is trained through the time back propagation algorithm, in order to verify the effectiveness of the OFDM spectrum sensing based on the recurrent neural network, the recurrent neural network model trained above needs to be tested. Similarly, let us assume that the time-series data of n pairs of test samples of the recurrent neural network model is { (x)(m+1),y(m+1)),(x(m+2),y(m +2)),L,(x(m+n),y(m+n)) And updating the weight and the bias to obtain an error between an output value and a theoretical value as follows:
μ=||fW,b(x(i))-y(i)|| (8)
then, find the coincidence mu ≦ mutThe mathematical expression of the final accuracy obtained by the number l of samples is as follows:
therefore, the testing process is completed, the accuracy of signal detection can be judged by calculating the error rate, and the lower the error rate is, the higher the accuracy is and the better the spectrum sensing performance is. In order to verify the excellent signal detection performance of the trained neural network model, the trained neural network model is compared with other methods through simulation experiments under different wireless communication parameter conditions, and the superiority of the method is highlighted.
The simulation parameters are as follows:
modulation system | QPSK | Number of deep RNN network layers | 15 |
Number of subcarriers | 64 | Small batch size | 1000 |
Number of FFT points | 128 | Number of data sets | 100000 |
Interpolated pilot number | 32 | Number of rounds of training | 800 |
|
16 | ||
Baseband sampling frequency | 4Mbps |
Under the simulation conditions, the detection performance of the OFDM signal is compared with that shown in FIG. 2, and it can be seen that the RNN neural network detection model is obviously superior to the method of the least square error, and the error rate is averagely reduced by about 1.9 dB.
The simulation parameters are as follows:
under the simulation conditions, the detection performance of the OFDM signal is compared with that shown in FIG. 3, and it can be seen that the RNN neural network detection model is obviously superior to the method of the least square error, and the error rate is averagely reduced by about 2.1 dB.
Simulation experiment 3
The simulation parameters are as follows:
modulation system | QPSK | Number of deep RNN network layers | 15 |
Number of subcarriers | 64 | Small batch size | 1000 |
Number of FFT points | 128 | Number of data sets | 100000 |
Interpolated pilot number | 64 | Number of rounds of training | 800 |
|
16 | ||
Baseband sampling frequency | 4Mbps |
Under the simulation conditions, the detection performance of the OFDM signal is compared with that shown in FIG. 4, and it can be seen that the RNN neural network detection model is obviously superior to the method of the least square error, and the error rate is averagely reduced by about 2.3 dB. As can be seen from comparison among fig. 2, fig. 3, and fig. 4, as the number of pilots increases, the bit error rate reduction amount of the RNN neural network detection model gradually increases, which can fully illustrate that the RNN neural network has stronger memory capacity while the information amount increases.
In summary, the method for detecting the OFDM signal of the RNN neural network of the present invention can be applied to the cognitive radio spectrum sensing technology of wireless mobile communication, and effectively improve the capability of the OFDM system in the aspect of receiver signal demodulation, thereby improving the signal detection performance of cognitive radio spectrum sensing. The method adopts innovative logic design, firstly, an OFDM system framework is built on the basis of analyzing an OFDM system, and a time series data set for model learning is generated by utilizing the OFDM system framework; then, a time series signal data set is learned through an LSTM model of an RNN neural network, parameters are quickly optimized through an Adam training algorithm, a small-batch training mode is adopted, end-to-end spectrum signal detection is achieved, the problem that detection accuracy is reduced due to the fact that an OFDM transmission technology is influenced by noise and other interference factors in cognitive radio spectrum sensing is solved, the problem of receiving of OFDM system signals with NBI and ICI is effectively solved, and the accuracy of OFDM signal detection is improved. Simulation experiments show that the method has stronger detection advantages compared with a least square error detection algorithm.
The above description is only for illustrating one embodiment of the present invention, but the scope of the present invention is not limited thereto, and other modifications or equivalent substitutions made by the technical solution of the present invention by those skilled in the art should be covered within the scope of the present invention as long as they do not depart from the spirit and scope of the technical solution of the present invention.
Claims (5)
1. An OFDM signal detection method based on RNN neural network is characterized in that OFDM signal detection is completed through the RNN neural network, and the steps are as follows:
step 1: establishing a spectrum signal detection model and initialized network parameters based on an RNN neural network, and setting iteration times and other hyper-parameters; the RNN neural network is built by using a Matlab2019 deep learning tool box; the RNN depth detection network uses an LSTM model in a CPU mode, and model training is completed by using an Adam optimization algorithm and a small-batch data processing mode;
step 2: inputting a data set comprising a training set and a verification set, carrying out forward propagation training, and calculating loss, wherein a loss function of the RNN deep detection network model is defined as:
in the formula (I), the compound is shown in the specification,theoretical value of the model, y(m)Representing the model predicted value;
and step 3: finishing a time back propagation algorithm through a gradient descent algorithm, and updating parameters until training is finished;
and 4, step 4: counting a loss curve and an accuracy fitting curve, simultaneously checking the training time, if the training time is not converged, adjusting the hyperparameter, repeating the step 2, and if the training time is not converged, entering the step 5;
and 5: inputting test set sequence data to complete model test.
2. The RNN neural network-based OFDM signal detection method of claim 1, wherein:
the learning rate of model training is 0.01, the small batch size is 1000, the number of input nodes is 256, the number of hidden layer nodes is 16, the number of hidden layer layers is 1, the number of full-connection layers is 4, the learning rate attenuation factor is 0.1, and verification is performed every 50 times of training.
3. The RNN neural network-based OFDM signal detection method of claim 2, wherein:
pilot frequency for better channel estimation is added into a signal transmitter, so that stable detection of signals is facilitated; in the last OFDM signal demodulation and recovery process of the receiver, a deep learning model is added to exert the memory capacity of RNN to the time sequence; and continuously carrying out cyclic training in a small batch mode to complete the memory of the OFDM signal channel information until the loss function output value is optimal.
4. The RNN neural network-based OFDM signal detection method according to claim 1, wherein:
and ensuring that the number of Cyclic Prefixes (CP) inserted into the model is unchanged, and training and testing the detection model under different pilot frequency conditions.
5. The RNN neural network-based OFDM signal detection method according to claim 1, wherein:
when a QPSK modulation mode is adopted, the OFDM signal detection model is subjected to equivalent transfer learning, and OFDM signal detection under different modulation modes is realized.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010700419.3A CN111865863B (en) | 2020-07-20 | 2020-07-20 | RNN neural network-based OFDM signal detection method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010700419.3A CN111865863B (en) | 2020-07-20 | 2020-07-20 | RNN neural network-based OFDM signal detection method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111865863A CN111865863A (en) | 2020-10-30 |
CN111865863B true CN111865863B (en) | 2021-07-20 |
Family
ID=73000701
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010700419.3A Active CN111865863B (en) | 2020-07-20 | 2020-07-20 | RNN neural network-based OFDM signal detection method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111865863B (en) |
Families Citing this family (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113285902B (en) * | 2021-05-19 | 2023-03-14 | 南京航空航天大学 | Design method of OFDM system detector |
CN113315593A (en) * | 2021-05-20 | 2021-08-27 | 南京工业大学 | Frequency spectrum sensing algorithm based on FLOM covariance matrix and LSTM neural network |
CN114629763B (en) * | 2021-09-27 | 2023-10-13 | 亚萨合莱国强(山东)五金科技有限公司 | OFDM system IQ signal demodulation method and device based on neural network |
CN113872911B (en) * | 2021-10-15 | 2023-10-24 | 齐鲁工业大学 | Model-driven method and system for suppressing peak-to-average ratio of orthogonal frequency division multiplexing system |
CN114679363A (en) * | 2022-04-11 | 2022-06-28 | 浙江工业大学 | Artificial intelligence assisted OFDM receiver offline learning method |
Family Cites Families (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108304917B (en) * | 2018-01-17 | 2020-11-24 | 华南理工大学 | P300 signal detection method based on LSTM network |
CN110858289B (en) * | 2018-08-24 | 2023-11-03 | 河南工业大学 | Grain moisture detection method and system based on deep learning |
CN109308522B (en) * | 2018-09-03 | 2022-02-22 | 河海大学常州校区 | GIS fault prediction method based on recurrent neural network |
CN109194595B (en) * | 2018-09-26 | 2020-12-01 | 东南大学 | Neural network-based channel environment self-adaptive OFDM receiving method |
CN109831801B (en) * | 2019-01-04 | 2021-09-28 | 东南大学 | Base station caching method for user behavior prediction based on deep learning neural network |
CN110113288B (en) * | 2019-05-23 | 2021-06-22 | 徐州中矿康普盛通信科技有限公司 | Design and demodulation method of OFDM demodulator based on machine learning |
CN110598859B (en) * | 2019-08-01 | 2022-12-13 | 北京光锁科技有限公司 | Nonlinear equalization method based on gated cyclic neural network |
CN110738138A (en) * | 2019-09-26 | 2020-01-31 | 哈尔滨工程大学 | Underwater acoustic communication signal modulation mode identification method based on cyclic neural network |
-
2020
- 2020-07-20 CN CN202010700419.3A patent/CN111865863B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN111865863A (en) | 2020-10-30 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111865863B (en) | RNN neural network-based OFDM signal detection method | |
Liao et al. | ChanEstNet: A deep learning based channel estimation for high-speed scenarios | |
Mao et al. | RoemNet: Robust meta learning based channel estimation in OFDM systems | |
CN111683024B (en) | Time-varying OFDM system channel estimation method based on deep learning | |
CN103873422B (en) | Multi-path jamming removing method in underwater sound ofdm system symbol | |
CN101945066B (en) | Channel estimation method of OFDM/OQAM system | |
CN101485125A (en) | A method and system for frequency division multiplexing | |
CN111669344B (en) | Time-varying OFDM system signal detection method based on deep learning | |
CN102111205A (en) | Channel estimation for communication system with multiple transmitting antennas | |
CN110311876A (en) | The implementation method of underwater sound OFDM receiver based on deep neural network | |
CN108881080B (en) | OFDM anti-ICI detection method based on sliding window and deep learning | |
CN112215335B (en) | System detection method based on deep learning | |
CN113852575A (en) | Iterative OTFS symbol detection method based on time domain channel equalization assistance | |
CN114338305A (en) | Symbol detection method for orthogonal time-frequency-space modulation system | |
CN115250216A (en) | Underwater sound OFDM combined channel estimation and signal detection method based on deep learning | |
CN113285902B (en) | Design method of OFDM system detector | |
CN101018219A (en) | Space frequency signal processing method | |
Zhang et al. | Deep learning based underwater acoustic OFDM receiver with joint channel estimation and signal detection | |
CN110958204A (en) | Non-orthogonal multi-carrier underwater communication system of asymmetric complex deep neural network | |
CN112564830B (en) | Deep learning-based dual-mode orthogonal frequency division multiplexing index modulation detection method and device | |
CN112636855A (en) | OFDM signal detection method | |
CN104301282B (en) | A kind of ICI Adaptive Suppression methods of ultrahigh speed OFDM in Mobile | |
Ali et al. | GFDM transceiver based on ann channel estimation | |
CN113709075B (en) | Method for realizing underwater acoustic communication receiver by using underwater acoustic channel multipath effect | |
Liu et al. | RecNet: Deep learning-based OFDM receiver with semi-blind channel estimation |
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 |