A Method and System for Radio Frequency Division Multiplexing Channel Estimation and Symbol Detection Based on Deep Learning

Abstract

The present invention discloses a method and system for analog radio frequency division multiplexing channel estimation and symbol detection based on deep learning. The method includes: Step S1, modulating the transmission symbols in the communication process using analog radio frequency division multiplexing technology, transmitting them through the channel, and obtaining the analog radio frequency division multiplexing received signal at the receiving end; Step S2, constructing a deep neural network model, using the I-channel and Q-channel in the analog radio frequency division multiplexing received signal as the input of the neural network, and the transmission symbol as the expected output of the neural network, training the neural network to enable it to have the ability to predict the transmission data symbols; Step S3, after the deep neural network model is trained, inputting the analog radio frequency division multiplexing signal received in real time during the communication process into the trained neural network model to obtain the predicted value of the transmission data symbols.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technologies, and particularly to a method and system for analog radio frequency division multiplexing channel estimation and symbol detection based on deep learning. Background Art

[0002] With the development of technology, people's demand for wireless communication under high-speed movement is gradually increasing, such as for high-speed railways, vehicle-to-everything, etc. Therefore, the next-generation wireless communication has put forward higher requirements for wireless communication under high-speed movement. In 4G and 5G wireless communication technologies, orthogonal frequency division multiplexing has been widely applied due to its good anti-multipath fading ability, high spectral utilization rate, and excellent performance in a linear time-invariant channel. However, in a linear time-varying channel, the orthogonality of orthogonal frequency division multiplexing will sharply decline, resulting in inter-carrier interference, and thus greatly reducing the performance of orthogonal frequency division multiplexing. In order to obtain good performance in a linear time-varying channel to meet the requirements of wireless communication under high-speed movement, many new signal modulation technologies have been proposed, and analog radio frequency division multiplexing was born in such a background.

[0003] The analog radio frequency division multiplexing technology is based on the generalization of the discrete Fourier transform - the discrete affine Fourier transform, which can achieve full diversity because its linear frequency modulation pulse parameters can adapt to the channel characteristics, thus realizing the complete delay-Doppler representation of the channel in the discrete affine Fourier transform domain. Experiments show that analog radio frequency division multiplexing has excellent performance in a linear time-varying channel and can meet the requirements of wireless communication under high-speed movement.

[0004] With the vigorous development of computer computing power in recent years, deep learning technology has also developed rapidly. Deep learning technology is widely applied in many fields, such as computer vision, natural language processing, etc. At the same time, deep learning technology also has a wide range of applications in the field of wireless communication, and its potential in the field of wireless communication is constantly being explored. However, there is still a lack of research on applying deep learning to analog radio frequency division multiplexing channel estimation and symbol detection. Summary of the Invention

[0005] To solve the above technical problems, the present invention provides a method for analog radio frequency division multiplexing channel estimation and symbol detection based on deep learning, and the method includes:

[0006] Step S1, modulating the transmission symbols in the communication process using the analog radio frequency division multiplexing technology, transmitting through the channel, and the receiving end obtains the analog radio frequency division multiplexing received signal;

[0007] Step S2: Construct a deep neural network model. Take the I-channel and Q-channel of the analog radio frequency division multiplexing received signal as the input of the deep neural network, and the transmitted symbol as the expected output of the deep neural network. Train the deep neural network to obtain a trained deep neural network model;

[0008] Step S3: Input the analog radio frequency division multiplexing signal received in real time during the communication process into the trained neural network model to obtain the predicted value of the transmitted data symbol.

[0009] Optionally, in step S1, the transmitted symbol consists of three parts: pilot, guard interval, and data. Among them, the pilot is known to both the transmitter and the receiver and is used for channel estimation when receiving the signal; the data carries the information to be sent; the guard interval is between the pilot and the data and is used to prevent interference between the pilot and the data.

[0010] Optionally, in step S1, the process of using the analog radio frequency division multiplexing technology to modulate the transmitted symbol during the communication process, transmitting it through the channel, and the receiving end obtaining the analog radio frequency division multiplexing received signal specifically includes:

[0011] Under integer Doppler frequency shift, the analog radio frequency division multiplexing received signal is expressed as:

[0012]

[0013] 0≤p≤N - 1, q=(p + loc i ) N·

[0014] where N is the number of chirp subcarriers; x[q] is the transmitted symbol modulated by modulation methods such as phase shift keying (PSK) and quadrature amplitude modulation (QAM); is the additive white Gaussian noise generated during the signal transmission process; P is the number of transmission channels; h i is the complex gain of the i-th channel; l i is the normalized delay of the i-th channel; loc i =(-α i +2Nc1l i ) N ; ɑ i is the normalized integer Doppler frequency shift of the i-th channel; to achieve full diversity of analog radio frequency division multiplexing, in the values of c1 and c2, c1=(2ɑ max +1) / 2N, and c2 is any irrational number or a rational number much smaller than 1 / 2N,

[0015] Optionally, in step S2, the constructed deep neural network model includes three layers: an input layer, a hidden layer, and an output layer. The simulation parameters are set as follows: the neural network has 8 layers, and the number of neuron nodes in each layer is 64, 128, 64, 32, 16, 8, 4, 1 respectively; the activation function of the hidden layer of the deep neural network is ReLu, the activation function of the output layer is Sigmoid, and Adam is used as the neural network optimizer.

[0016] The present invention also discloses a radio frequency division multiplexing channel estimation and symbol detection system based on deep learning. The system includes: a signal modulation module, a model training module, and a symbol detection module;

[0017] The signal modulation module is used to modulate the transmission symbols in the communication process using the radio frequency division multiplexing technology. After transmission through the channel, the radio frequency division multiplexing received signal is obtained at the receiving end;

[0018] The model training module is used to construct a deep neural network model, take the I channel and Q channel in the radio frequency division multiplexing received signal as the input of the deep neural network, and the transmission symbol as the expected output of the deep neural network, and train the deep neural network to obtain a trained deep neural network model;

[0019] The symbol detection module is used to input the radio frequency division multiplexing signal received in real time during the communication process into the trained neural network model to obtain the predicted value of the transmission data symbol.

[0020] Optionally, the transmission symbol in the signal modulation module consists of three parts: a pilot, a guard interval, and data. Among them, the pilot is data known to both the transmitter and the receiver, and is used for channel estimation when receiving the signal; the data carries the information to be sent; the guard interval is between the pilot and the data, and is used to prevent interference between the pilot and the data.

[0021] Optionally, the working process of the signal modulation module specifically includes:

[0022] Under integer Doppler frequency shift, the radio frequency division multiplexing received signal is expressed as:

[0023]

[0024] Among them, N is the number of chirp subcarriers; x[q] is the transmission symbol modulated by modulation methods such as phase shift keying (PSK) and quadrature amplitude modulation (QAM); is the additive white Gaussian noise generated during signal transmission; P is the number of transmission channels; h i is the complex gain of the i-th channel; l i is the normalized delay of the i-th channel; loc i= (-ɑ i + 2Nc1l i ) N ; ɑ i is the integer Doppler frequency shift after normalization of the i-th channel; to achieve full diversity in analog radio frequency division multiplexing, in the values of c1 and c2, c1 = (2α max + 1) / 2N, c2 is any irrational number or a rational number much smaller than 1 / 2N,

[0025] Optionally, in the model training module, the constructed deep neural network model includes three layers, an input layer, a hidden layer, and an output layer. The simulation parameters are set as follows: the neural network has 8 layers, and the number of neuron nodes in each layer is 64, 128, 64, 32, 16, 8, 4, 1 respectively; the activation function of the hidden layer of the deep neural network is ReLu, the activation function of the output layer is Sigmoid, and Adam is used as the neural network optimizer.

[0026] Compared with the prior art, the beneficial effects of the present invention are:

[0027] (1) The present invention is designed based on the application of deep learning in analog radio frequency division multiplexing technology, filling the gap in channel estimation and symbol detection of deep learning in analog radio frequency division multiplexing.

[0028] (2) The present invention combines the two steps of channel estimation and symbol detection into one, and only one link is required to obtain the predicted value of the input data, simplifying the system process.

[0029] (3) The present invention gives full play to the advantages of the deep neural network and shows excellent performance in analog radio frequency division multiplexing symbol detection; at the same time, the system has high robustness and can maintain good performance in scenarios with large interference. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions of the present invention, the drawings required for use in the embodiments are briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings without creative efforts.

[0031] Figure 1 is a schematic flow chart of the method for analog radio frequency division multiplexing channel estimation and symbol detection based on deep learning of the present invention;

[0032] Figure 2 is a schematic diagram of the input symbols of the method for analog radio frequency division multiplexing channel estimation and symbol detection based on deep learning of the present invention;

[0033] Figure 3It is the system block diagram of the method for analog radio frequency division multiplexing channel estimation and symbol detection based on deep learning of the present invention;

[0034] Figure 4 It is the schematic architecture diagram of the method for analog radio frequency division multiplexing channel estimation and symbol detection based on deep learning of the present invention;

[0035] Figure 5 It is the schematic diagram of the bit error rate performance under ζ = 0 in the embodiment of the present invention;

[0036] Figure 6 It is the schematic diagram of the bit error rate performance under ζ = 2 in the embodiment of the present invention. Detailed implementation manners

[0037] To make the above objects, features, and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific implementation manners.

[0038] Embodiment 1

[0039] A method for analog radio frequency division multiplexing channel estimation and symbol detection based on deep learning, as Figure 1 shown, the method includes:

[0040] Step S1: Modulate the transmission symbols in the communication process using analog radio frequency division multiplexing technology, and transmit them through the channel to obtain the analog radio frequency division multiplexing received signal at the receiving end.

[0041] Modulate a large number of transmission symbols including pilots, data, and guard intervals using analog radio frequency division multiplexing, and transmit them through a specific channel to obtain the analog radio frequency division multiplexing output signal. A large number of input and output signals form a data set for training the deep neural network.

[0042] In this embodiment, under integer Doppler frequency shift, the analog radio frequency division multiplexing received signal is expressed as:

[0043]

[0044] where N is the number of chirp subcarriers; x[q] is the input signal modulated by modulation methods such as phase shift keying (PSK) and quadrature amplitude modulation (QAM); is the additive white Gaussian noise generated during signal transmission; P is the number of transmission channels; h i is the complex gain of the i-th channel; l i is the delay of the i-th channel after normalization; loc i = (-α i + 2Nc1l i ) N ; α iis the normalized integer Doppler frequency shift of the i-th channel; To achieve full diversity in analog radio frequency division multiplexing, in terms of the values of c1 and c2,

[0045] c1 = (2α max + 1) / 2N, c2 is any irrational number or a rational number much smaller than 1 / 2N,

[0046] Under fractional Doppler frequency shift, the received signal will be expressed as:

[0047]

[0048] where, a i is the fractional part of the Doppler frequency shift. For the case of fractional Doppler frequency shift, c1 = (2(α max + ζ) + 1) / 2N, ζ is a positive integer, used to reduce the interference between different channels in the case of fractional Doppler frequency shift, but at the same time it will increase the size of the guard interval and reduce the spectrum utilization rate.

[0049] As Figure 2 shown, this embodiment provides a symbol structure diagram of analog radio frequency division multiplexing transmission to illustrate the symbol distribution of the system's transmitted signal. It is composed of three types of symbols: pilot, data, and guard interval. Among them, the pilot is known to both the transmitter and the receiver and is used for channel estimation when receiving the signal; the data carries the information to be transmitted; the guard interval is between the pilot and the data and is used to prevent interference between the pilot and the data.

[0050] Step S2, construct a deep neural network model. Take the I-channel and Q-channel in the analog radio frequency division multiplexing received signal as the input of the neural network, and the transmitted symbol as the expected output of the neural network, and train the neural network to make it have the ability to predict the transmitted data symbol.

[0051] In symbol detection at the receiving end: Deploy the deep neural network trained in the offline training stage to the receiving end. The user can input the I-channel and Q-channel of the received analog radio frequency division multiplexing signal into the trained neural network model, and without pre-estimating the channel, the predicted value of the transmitted symbol can be obtained.

[0052] Construct the training data set of the deep neural network. The data set is composed of the analog radio frequency division multiplexing received signal and the transmitted symbol. The input of the deep neural network is the I-channel and Q-channel of the analog radio frequency division multiplexing received signal, and the transmitted symbol is used as the expected output of the deep neural network. To improve the accuracy of symbol detection, multiple deep neural networks are used, and each deep neural network is responsible for detecting one data symbol. After determining the structure of the deep neural network, the data set can be used for training. As Figure 3As shown, the input of this embodiment is the real and imaginary parts of the symbol of the analog radio frequency division multiplexing reception. By deploying multiple deep neural networks, each network is responsible for outputting a data symbol, thereby increasing the prediction accuracy of the system.

[0053] Step S3: After the deep neural network model is trained, input the analog radio frequency division multiplexing signal received in real time during the communication process into the trained neural network model to obtain the predicted value of the transmitted symbol.

[0054] As Figure 4 shown, in this embodiment, by deploying a pre-trained deep neural network at the signal receiving end, the analog radio frequency division multiplexing transmitted symbol is predicted. This embodiment includes two stages, namely the offline training stage and the online deployment stage. In the offline training stage, a deep neural network is constructed, and a training data set composed of the transmitted and received analog radio frequency division multiplexing signals is used to train the neural network so that the output of the neural network is closest to the actual transmitted data symbol. In the online deployment stage, the trained deep neural network model is deployed at the receiving end, and the user inputs the analog radio frequency division multiplexing signal received in real time during the communication process into the pre-trained neural network, and the predicted value of the transmitted symbol can be output.

[0055] Embodiment 2

[0056] The present invention discloses an analog radio frequency division multiplexing channel estimation and symbol detection system based on deep learning. The system includes: a signal modulation module, a model training module, and a symbol detection module;

[0057] The signal modulation module is used to modulate the transmitted symbol in the communication process using the analog radio frequency division multiplexing technology. After transmission through the channel, the receiving end obtains the analog radio frequency division multiplexing received signal.

[0058] A large number of transmitted symbols including pilots, data, and guard intervals are modulated by analog radio frequency division multiplexing and transmitted through a specific channel to obtain an analog radio frequency division multiplexing output signal. A large number of input and output signals form a data set for training the deep neural network.

[0059] In this embodiment, under integer Doppler frequency shift, the analog radio frequency division multiplexing received signal is expressed as:

[0060]

[0061] where N is the number of chirp subcarriers; x[q] is the transmitted symbol modulated by modulation methods such as phase shift keying (PSK) and quadrature amplitude modulation (QAM); is the additive white Gaussian noise generated during signal transmission; P is the number of transmission channels; h i is the complex gain of the i-th channel; l iis the delay of the i-th channel after normalization; loc i = (-α i + 2Nc1l i ) N ; α i is the integer Doppler frequency shift after normalization of the i-th channel; To achieve full diversity of analog RF division multiplexing, in terms of the values of c1 and c2,

[0062] c1 = (2α max + 1) / 2N, c2 is any irrational number or a rational number much less than 1 / 2N,

[0063] Under fractional Doppler frequency shift, the received signal will be expressed as:

[0064]

[0065] where, a i is the fractional part of the Doppler frequency shift. For the case of fractional Doppler frequency shift, c1 = (2(α max + ζ) + 1) / 2N, ζ is a positive integer used to reduce the interference between different channels in the case of fractional Doppler frequency shift, but at the same time it will increase the size of the guard interval and reduce the spectrum utilization rate.

[0066] The model training module is used to construct a deep neural network model, take the I-channel and Q-channel in the analog RF division multiplexing received signal as the input of the deep neural network, take the transmission symbol as the expected output of the deep neural network, and train the deep neural network to obtain a trained deep neural network model.

[0067] To improve the accuracy of symbol detection, we use multiple deep neural networks with the same structure. The constructed deep neural network model includes three layers, an input layer, a hidden layer and an output layer. The simulation parameters are set as follows: the neural network has 8 layers, and the number of neuron nodes in each layer is 64, 128, 64, 32, 16, 8, 4, 1 respectively; the activation function of the hidden layer of the deep neural network is ReLu, the activation function of the output layer is Sigmoid, and Adam is used as the neural network optimizer. Each neural network is responsible for detecting a transmission symbol. After determining the structure of the deep neural network, the dataset can be used for training.

[0068] The symbol detection module is used to input the analog RF division multiplexing signal received in real time during the communication process into the trained neural network model to obtain the predicted value of the transmission symbol.

[0069] After the neural network training is completed, the trained neural network can be deployed to the receiving end. When the user receives the analog radio frequency division multiplexing signal, input it into the neural network, and the transmitted symbol can be directly predicted without prior channel estimation.

[0070] Embodiment 3

[0071] To illustrate the technological advancement of the method in this embodiment, the bit error rate performance of the analog radio frequency division multiplexing channel estimation and symbol detection method based on deep learning proposed in this embodiment is simulated on the Python platform.

[0072] The simulation parameter settings are as follows. The carrier frequency and bandwidth are 4 GHz and 32 kHz respectively; the number of chirp subcarriers N = 32; the number of guard intervals for preventing interference between pilots and data in the input symbols is 2Q = 2(2(α max +ζ)+1)(l max +1)-1), where there are Q guard intervals on both the left and right sides of the pilot. For the case of integer Doppler frequency shift, the value of ζ is always 0; for the case of fractional Doppler frequency shift, increasing ζ can reduce the interference between different channels, but at the same time increase the number of guard intervals, thereby reducing the spectral efficiency. The maximum delay of the channel is 31.25 μs, and after normalization, the maximum delay is 1; the maximum Doppler frequency shift of the channel is 1 kHz, and after normalization, the maximum Doppler frequency shift is 1; the number of channels P = 2; the complex gain of the channel is generated by independent complex random variables with zero mean and variance of 1 / P; the normalized delay of the channel 1 = [0, 1]; the normalized Doppler frequency shift of the channel in the i-th channel is v i =v max cosθ i , where, v max represents the maximum Doppler frequency shift, and θ i is uniformly distributed on [-π, π]; the signal-to-noise ratio SNR p of the pilot = 30 dB. In terms of the settings of the neural network parameters, including the input layer, hidden layer, and output layer, the neural network has 8 layers. From the input layer to the output layer, the number of neuron nodes in each layer is 64, 128, 64, 32, 16, 8, 4, 1 respectively; the activation function of the hidden layer of the neural network is ReLu, the activation function of the output layer is Sigmoid, and Adam is used as the optimizer of the neural network. Specifically, Figure 5 and Figure 6 show the performance of the method in the case of fractional Doppler frequency shift. Among them, Figure 5 shows the bit error rate performance under ζ = 0, Figure 6The bit error rate performance under ζ = 2 is shown. The bit error rate performance of the deep learning-based radio frequency division multiplexing channel estimation and symbol detection method is obtained from the simulation results and compared with the existing embedded channel estimation method and the performance under ideal channel state information. Among them, both of these two schemes use Linear Minimum Mean Square Error (LMMSE) for symbol detection.

[0073] It can be seen from the simulation that this embodiment has the following technical advantages:

[0074] Compared with the traditional embedded channel estimation and symbol detection technology, the present invention has higher robustness. In the case of less interference, such as ζ = 2, the present invention has comparable performance with this scheme. However, when the interference increases, such as ζ = 0, the present invention can still maintain better performance, while the performance of the traditional method has a significant decline.

[0075] This embodiment discloses a deep learning-based radio frequency division multiplexing channel estimation and symbol detection method, which includes the following steps: Step S1, modulating the transmission symbols in the communication process using the radio frequency division multiplexing technology and transmitting them through the channel to obtain the radio frequency division multiplexing received signal at the receiving end; Step S2, constructing a deep neural network model, using the I-channel and Q-channel in the radio frequency division multiplexing received signal as the input of the neural network, and the transmission symbol as the expected output of the neural network, and training the neural network to enable it to have the ability to predict the transmission data symbol. Among them, in order to improve the accuracy of symbol detection, we use multiple deep neural networks, and each neural network is responsible for detecting one data symbol. The method of using multiple neural networks greatly improves the system's ability to detect radio frequency division multiplexing symbols; Step S3, after the deep neural network model is trained, input the received radio frequency division multiplexing signal into the trained neural network model, and the predicted value of the transmission symbol can be obtained without pre-estimating the channel.

[0076] The above-described embodiments are only descriptions of the preferred embodiments of the present invention, and do not limit the scope of the present invention. Without departing from the design spirit of the present invention, various deformations and improvements made by those of ordinary skill in the art to the technical solutions of the present invention shall fall within the protection scope determined by the claims of the present invention.

Claims

1. A radio frequency division multiplexing based deep learning channel estimation and symbol detection method, characterized in that, The method includes: Step S1: Modulate the transmission symbols in the communication process using the analog radio frequency division multiplexing technology, transmit through the channel, and the receiving end obtains the analog radio frequency division multiplexing received signal; Step S2: Construct a deep neural network model, use the I-channel and Q-channel in the analog radio frequency division multiplexing received signal as the input of the deep neural network, and the transmission symbol as the expected output of the deep neural network, and train the deep neural network to obtain a trained deep neural network model; The constructed deep neural network model includes three layers, an input layer, a hidden layer, and an output layer. The simulation parameters are set as follows: The neural network has 8 layers, and the number of neuron nodes in each layer is 64, 128, 64, 32, 16, 8, 4, 1 respectively; The activation function of the deep neural network hidden layer is ReLu, the activation function of the output layer is Sigmoid, and Adam is used as the neural network optimizer; Step S3: Input the analog radio frequency division multiplexing signal received in real time during the communication process into the trained neural network model to obtain the predicted value of the transmission data symbol.

2. The method for analog radio frequency division multiplexing channel estimation and symbol detection based on deep learning according to claim 1, wherein In step S1, the transmission symbol consists of three parts: a pilot, a guard interval, and data. Among them, the pilot is known to both the transmitter and the receiver and is used for channel estimation when receiving the signal; The data carries the information to be sent; The guard interval is between the pilot and the data and is used to prevent interference between the pilot and the data.

3. The method for radio frequency division multiplexing channel estimation and symbol detection based on deep learning according to claim 1, characterized in that In step S1, the process of modulating the transmission symbols in the communication process using the analog radio frequency division multiplexing technology, transmitting through the channel, and the receiving end obtaining the analog radio frequency division multiplexing received signal specifically includes: Under integer Doppler frequency shift, the analog radio frequency division multiplexing received signal is expressed as: 0 ≤ p ≤ N - 1, q = (p + loc i ) N where N is the number of chirp subcarriers; x[q] is the transmission symbol modulated by phase shift keying (PSK) or quadrature amplitude modulation (QAM); is the additive white Gaussian noise generated during signal transmission; P is the number of transmission channels; h i is the complex gain of the i-th channel; l i is the normalized delay of the i-th channel; loc i = (-α i + 2Nc1l i ) N ; α i is the normalized integer Doppler shift of the i-th channel; to achieve full diversity of analog radio frequency division multiplexing, in terms of the values of c1 and c2, c1 = (2α max + 1) / 2N, and c2 is any irrational number or a rational number much smaller than 1 / 2N, 4. A radio frequency division multiplexing-like channel estimation and symbol detection system based on deep learning, wherein the symbol detection system is used to implement the symbol detection method according to any one of claims 1-3, characterized in that The system includes: a signal modulation module, a model training module, and a symbol detection module; The signal modulation module is used to modulate the transmission symbols in the communication process using the analog radio frequency division multiplexing technology, transmit through the channel, and the receiving end obtains the analog radio frequency division multiplexing received signal; The model training module is used to construct a deep neural network model, use the I-channel and Q-channel in the analog radio frequency division multiplexing received signal as the input of the deep neural network, and the transmission symbol as the expected output of the deep neural network, and train the deep neural network to obtain a trained deep neural network model; The constructed deep neural network model includes three layers, an input layer, a hidden layer, and an output layer. The simulation parameters are set as follows: The neural network has 8 layers, and the number of neuron nodes in each layer is 64, 128, 64, 32, 16, 8, 4, 1 respectively; The activation function of the deep neural network hidden layer is ReLu, the activation function of the output layer is Sigmoid, and Adam is used as the neural network optimizer; The symbol detection module is used to input the analog radio frequency division multiplexing signal received in real time during the communication process into the trained neural network model to obtain the predicted value of the transmission data symbol.

5. The deep learning-based radio frequency division multiplexing channel estimation and symbol detection system according to claim 4, characterized in that The transmission symbols in the signal modulation module consist of three parts: pilots, guard intervals, and data. Among them, the pilots are known to both the transmitter and the receiver and are used for channel estimation when receiving signals; the data carries the information to be transmitted; the guard interval is between the pilots and the data and is used to prevent interference between the pilots and the data.

6. The deep learning-based radio frequency division multiplexing channel estimation and symbol detection system according to claim 4, wherein, The working process of the signal modulation module specifically includes: The analog radio frequency division multiplexing received signal under integer Doppler frequency shift is expressed as: 0 ≤ p ≤ N - 1, q = (p + loc i ) N . Where, N is the number of chirp subcarriers; x[q] is the transmission symbol modulated by phase shift keying (PSK) or quadrature amplitude modulation (QAM). is the additive white Gaussian noise generated during signal transmission; P is the number of transmission channels; h i is the complex gain of the i-th channel; l i is the delay of the i-th channel after normalization; loc i = (-α i + 2Nc1l i ) N ; α i is the normalized integer Doppler shift of the i-th channel; to achieve full diversity of analog radio frequency division multiplexing, for the values of c1 and c2, c1 = (2α max + 1) / 2N, and c2 is any irrational number or a rational number much smaller than 1 / 2N.

Images