CNN-based USRP modulation signal identification method

Abstract

The invention discloses a CNN-based USRP modulation signal identification method, and the method comprises the steps: carrying out the LabVIEW graphical programming on an upper computer, and achieving the switching of a receiving and transmitting system of a plurality of modulation signals; a USRP receiver collects a signal data set, and a convolutional neural network (CNN) is adopted to carry out modulation signal identification. The convolutional neural network model takes an acquired signal data set as the input of a recognition network and takes modulation signal judgment as the output of the recognition network, and can realize extraction and classification processing of autonomous modulation signal features without supervision. Experimental verification shows that the scheme has excellent performance close to that of a classical likelihood ratio scheme, and the recognition rate can reach 95% or above when the signal-to-noise ratio is larger than 5 dB.

Description

Technical Field

[0001] The invention relates to the field of software radio transmission, in particular to a USRP modulation signal recognition method based on CNN. Background Art

[0002] Modulation recognition technology has important application value in both commercial and military fields. Modulation recognition technology can be divided into modulation recognition based on likelihood ratio and modulation recognition based on signal features. Modulation recognition based on likelihood ratio uses the principle of statistical inference. By constructing a likelihood ratio model of the modulation type, comparing the likelihood ratios of known modulation types, and comparing the signal to be identified with the likelihood ratios of various modulation types, the modulation type of the unknown signal is determined; modulation recognition based on signal features extracts the characteristic parameters or characteristic sequences of the signal (including spectral characteristics, power spectral density, phase change, instantaneous amplitude, modulation index, etc.), and then determines the modulation type of the signal based on these characteristics.

[0003] Because the likelihood ratio-based method needs to accurately establish the likelihood ratio model of the modulation type, it may face the challenge of model construction for complex modulation types; the signal feature-based method is more flexible and can adapt to different modulation types according to different feature parameters, but it requires appropriate feature extraction and decision rule design, so the modulation recognition technology based on signal features is usually more widely used. The traditional modulation recognition technology based on signal features first designs and extracts signal features, and then designs classification rules to classify the modulation format of signal features. Such manually designed signal features and classification rules usually have limited recognition accuracy in complex channel environments. In recent years, the development of deep learning has promoted its widespread application in computer vision, natural language processing and other fields, and has achieved remarkable results. Therefore, deep learning is introduced in modulation recognition technology. Automatically learn classification rules from feature data to improve classification accuracy. How to use deep learning algorithms to realize automatic learning of signal features and classification rules in the field of modulation recognition, so as to further improve the performance of modulation recognition, is a technical problem that needs to be solved at this stage. Summary of the invention

[0004] The purpose of the present invention is to solve the above problems and design a USRP modulation signal recognition method based on CNN.

[0005] To achieve the above object, the technical solution of the present invention is that, further, in the above-mentioned USRP modulation signal recognition method based on CNN, the modulation signal recognition method comprises the following steps:

[0006] Use USRP platform and host computer with LabVIEW graphical programming to build wireless communication system;

[0007] There are 8 types of modulation signals for transmission signals between transceivers, namely: 2ASK, 4ASK, BPSK, 4PSF, 4QAM, 16QAM, 64QAM;

[0008] The wireless signal is transmitted through the transmitter. After receiving the signal, the USRP receiver receives and stores the signal through the signal acquisition program to obtain the signal training set and signal test set.

[0009] Establishing a CNN model for modulated signal recognition, and inputting the signal training set into the CNN model for training;

[0010] The trained CNN model is deployed to the test set data to verify the modulation signal recognition results.

[0011] Furthermore, in the above-mentioned USRP modulation signal recognition method based on CNN, the acquisition of device operation data in the sensor and the construction of a wireless communication system using a USRP platform and a host computer using LabVIEW graphical programming include:

[0012] The physical parameters of the USRP platform are as follows: The USRO RIO device has a 2×2 multi-input / multi-output RF transceiver for dual transmission and dual reception. The RF bandwidth transceiver has a maximum transmit signal power of 30dBm and is used to receive or transmit signals from 1.2GHz to 6GHz. The maximum instantaneous bandwidth reaches 40MHz. The maximum bidirectional point-to-point bandwidth of the USRP RIO device is up to 830MB / s, and its digital backend interface is a PCle×4 bus with a speed of 800MB / s.

[0013] Furthermore, in the above-mentioned USRP modulation signal recognition method based on CNN, there are 8 types of modulation signals for the transmission signals between the transceivers, namely: 2ASK, 4ASK, BPSK, 4PSF, 4QAM, 16QAM, 64QAM, including:

[0014] The transmitter and receiver are responsible for data communication. Multiple antennas can be configured on a USRP device, and SISO self-transmission and self-reception can be realized. Two single antennas are configured for transmission and reception. The two single antennas are located on the same device, and the data communication between the two is completed through a wireless channel.

[0015] Furthermore, in the above-mentioned USRP modulation signal recognition method based on CNN, there are 8 types of modulation signals for the transmission signals between the transceivers, namely: 2ASK, 4ASK, BPSK, 4PSF, 4QAM, 16QAM, 64QAM, and also include:

[0016] After source coding, channel coding, signal modulation, USRP up-conversion, and digital-to-analog conversion DAC, it is sent to the wireless channel through the RF transmission module, and an external expansion interface is added to the RF transmission module.

[0017] Furthermore, in the above-mentioned USRP modulation signal recognition method based on CNN, there are 8 types of modulation signals for the transmission signals between the transceivers, namely: 2ASK, 4ASK, BPSK, 4PSF, 4QAM, 16QAM, 64QAM, and also include:

[0018] At the receiving end, wireless information is received from the receiving antenna, passes through the RF receiving module, and then undergoes analog-to-digital conversion CAD and down-conversion processing to receive the modulated signal. The RF receiving module adds an external expansion interface.

[0019] Furthermore, in the above-mentioned USRP modulation signal recognition method based on CNN, the wireless signal is transmitted by the transmitting end, and after receiving the signal, the USRP receiving end receives and stores the signal through a signal acquisition program to obtain a signal training set and a signal test set, including:

[0020] The line signal is transmitted through the transmitter. After receiving the signal, the USRP receiver receives and stores the signal through the signal acquisition program, providing the required signal training set and signal test set for the CNN model of modulation signal recognition.

[0021] Furthermore, a system for a USRP modulation signal recognition method based on CNN is implemented, characterized in that the CNN model for modulation signal recognition is established, and the signal training set is input into the CNN model for training, including:

[0022] The CNN model uses two convolutional layers with a convolution step of 1. The first layer of the two convolutional layers has 64 kernels. The size of the first convolution kernel is 2×4, the activation function is ReLU, the bias term is used and initialized to zero, the weight initialization method is Glorot uniform distribution initialization, and the convolution kernel does not consider the completion method, regularization term and weight constraint. The number of convolution kernels in the second convolutional layer is 16, the size of the convolution kernel is 1×4, and the rest of the settings are the same as the first convolutional layer.

[0023] Furthermore, a system for a USRP modulation signal recognition method based on CNN is implemented, characterized in that the CNN model for modulation signal recognition is established, and the signal training set is input into the CNN model for training, including:

[0024] The fully connected neural network in the CNN model includes three fully connected layers. The first two layers are set to 64 and 16 neurons respectively, and the activation function is ReLU. The weights are initialized using Glorot uniform distribution, and regularization terms and weight constraints are used. The output layer activation function uses Softmax.

[0025] Furthermore, a system for a USRP modulation signal recognition method based on CNN is implemented, characterized in that the CNN model for modulation signal recognition is established, and the signal training set is input into the CNN model for training, including:

[0026] The training parameters of the CNN model are set to: use the Adam optimization method and the multi-category cross entropy loss function. Set the learning rate to 10 -4 , learning rate update related parameters β1 = 0.9, β2 = 0.999; fuzzy factor is set to 10 -8 , the decay value of the learning rate after update is 0.

[0027] Its beneficial effect is that, by targeting the signal modulation recognition problem in wireless communication and studying the existing wireless communication modulation signal recognition methods, a deep learning modulation signal recognition method based on the USRP platform is proposed. The USRP receiver collects signal data sets; a convolutional neural network is used for modulation signal recognition, and the deep learning technology is used to provide the data set to the convolutional neural network for training and learning, and offline modulation signal recognition processing is performed. It is effectively verified that this scheme can effectively realize modulation signal recognition in the actual environment, laying the foundation for subsequent signal processing and ensuring reliable information delivery. At the same time, it verifies the practicality and effectiveness of the CNN deep learning algorithm in wireless communication systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the following detailed description of the preferred embodiment.The drawings are only for the purpose of illustrating the preferred embodiments and are not to be construed as limiting the invention.

[0029] Figure 1 It is a schematic diagram of a first embodiment of a USRP modulation signal recognition method based on CNN in an embodiment of the present invention;

[0030] Figure 2 It is a system flow chart of a USRP modulation signal recognition method based on CNN in an embodiment of the present invention;

[0031] Figure 3 It is a functional block diagram of a transmitter of a USRP modulation signal recognition method based on CNN in an embodiment of the present invention;

[0032] Figure 41 is a functional block diagram of a receiver of a USRP modulation signal recognition method based on CNN in an embodiment of the present invention;

[0033] Figure 5 A CNN deep learning neural network structure of a CNN-based USRP modulation signal recognition method in an embodiment of the present invention;

[0034] Figure 6 A transmitter setting panel for a USRP modulation signal recognition method based on CNN in an embodiment of the present invention;

[0035] Figure 7 A receiver setting panel for a USRP modulation signal recognition method based on CNN in an embodiment of the present invention;

[0036] Figure 8 FIG. 1 is a USRP signal acquisition program diagram of a USRP modulation signal recognition method based on CNN in an embodiment of the present invention;

[0037] Fig. 9 The figure is a comparison of the recognition rates of four modulation signals under different signal-to-noise ratios in a USRP modulation signal recognition method based on CNN in an embodiment of the present invention;

[0038] Fig.10 The figure shows a comparison of the recognition rates of eight modulation signals under different signal-to-noise ratios in a CNN-based USRP modulation signal recognition method in an embodiment of the present invention.

[0039] Fig.11 This is a comparison chart of the recognition rates of a USRP modulation signal recognition method based on CNN in an embodiment of the present invention. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0041] Those skilled in the art will appreciate that, unless otherwise stated, the singular forms "a", "an", "said" and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present invention refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof.

[0042] Embodiment 1:

[0043] The present invention will be described in detail below in conjunction with the accompanying drawings. Figure 1As shown, a USRP modulation signal recognition method based on CNN, the modulation signal recognition method includes the following steps:

[0044] Step 101: Use the USRP platform and a host computer using LabVIEW graphical programming to build a wireless communication system;

[0045] Specifically, in this embodiment

[0046] In software radio, except for the basic frequency conversion, A / D, D / A conversion and RF driver implemented by the hardware platform USRP RIO, the rest of the functions are designed in the form of software. In addition to the basic transceiver function, other extended functions of the entire communication process need to be designed and programmed by yourself.

[0047] Some of the main physical parameters of the USRP RIO are as follows: The USRO RIO device has a 2×2 multi-input / multi-output RF transceiver, which can achieve dual transmission and dual reception. The RF bandwidth transceiver has a maximum transmit signal power of 30dBm, can receive or transmit signals from 1.2GHz to 6GHz, and the maximum instantaneous bandwidth can reach 40MHz. The maximum bidirectional point-to-point bandwidth of the USRP RIO device is up to 830MB / s, and its digital backend interface is a PCle×4 bus, which can reach a speed of 800MB / s.

[0048] The software part of the experiment uses Labview2015 and Python to design and debug the program. Based on the RF transceiver driver provided by the software itself, a series of functions required by the present invention are expanded to realize the entire invention.

[0049] Step 102: There are 8 types of modulation signals for transmission signals between transceivers, namely: 2ASK, 4ASK, BPSK, 4PSF, 4QAM, 16QAM, and 64QAM;

[0050] Specifically, in this embodiment:

[0051] The transmitter and receiver are responsible for data communication. Multiple antennas can be configured on a USRP device, and SISO self-transmission and self-reception can be realized. Two single antennas are configured for transmission and reception. The two single antennas are located on the same device, and the data communication between the two is completed through a wireless channel.

[0052] After source coding, channel coding, signal modulation, USRP up-conversion, and digital-to-analog conversion DAC, it is sent to the wireless channel through the RF transmission module, and an external expansion interface is added to the RF transmission module.

[0053] At the receiving end, wireless information is received from the receiving antenna, passes through the RF receiving module, and then undergoes analog-to-digital conversion CAD and down-conversion processing to receive the modulated signal. The RF receiving module adds an external expansion interface.

[0054] Step 103, the wireless signal is transmitted through the transmitting end, and after receiving the signal, the USRP receiving end receives and stores the signal through the signal acquisition program to obtain a signal training set and a signal test set;

[0055] Specifically, in this embodiment:

[0056] The line signal is transmitted through the transmitter. After receiving the signal, the USRP receiver receives and stores the signal through the signal acquisition program, providing the required signal training set and signal test set for the CNN model of modulation signal recognition.

[0057] Step 104: Establish a CNN model for modulated signal recognition, and input the signal training set into the CNN model for training;

[0058] Specifically, in this embodiment:

[0059] The CNN model uses two convolutional layers with a convolution step of 1. The first layer of the two convolutional layers has 64 kernels. The size of the first convolution kernel is 2×4, the activation function is ReLU, the bias term is used and initialized to zero, the weight initialization method is Glorot uniform distribution initialization, and the convolution kernel does not consider the completion method, regularization term and weight constraint. The number of convolution kernels in the second convolutional layer is 16, the size of the convolution kernel is 1×4, and the rest of the settings are the same as the first convolutional layer.

[0060] The fully connected neural network in the CNN model includes three fully connected layers. The first two layers are set to 64 and 16 neurons respectively, and the activation function is ReLU. The weights are initialized using Glorot uniform distribution, and regularization terms and weight constraints are used. The output layer activation function uses Softmax.

[0061] The training parameters of the CNN model are set to: use the Adam optimization method and the multi-category cross entropy loss function. Set the learning rate to 10 -4 , learning rate update related parameters β1 = 0.9, β2 = 0.999; fuzzy factor is set to 10 -8 , the decay value of the learning rate after update is 0.

[0062] Step 105: deploy the trained CNN model to the test set data to verify the modulation signal recognition result.

[0063] Its beneficial effect is that, by targeting the signal modulation recognition problem in wireless communication and studying the existing wireless communication modulation signal recognition methods, a deep learning modulation signal recognition method based on the USRP platform is proposed. The USRP receiver collects signal data sets; a convolutional neural network is used for modulation signal recognition, and the deep learning technology is used to provide the data set to the convolutional neural network for training and learning, and offline modulation signal recognition processing is performed. It is effectively verified that this scheme can effectively realize modulation signal recognition in the actual environment, laying the foundation for subsequent signal processing and ensuring reliable information delivery. At the same time, it verifies the practicality and effectiveness of the CNN deep learning algorithm in wireless communication systems.

[0064] Embodiment 2:

[0065] Modulation signal recognition refers to the process of identifying the modulation mode of the received signal in wireless communication. Modulation is the process of converting a digital or analog signal into a signal form suitable for transmission on a wireless channel, while modulation signal recognition is to determine its modulation mode by analyzing the characteristics of the signal after receiving the modulated signal. In wireless communication, modulation signal recognition is very important to ensure correct signal demodulation and high-quality communication, and is an important technical link in the wireless communication system. Therefore, this scheme discloses a USRP modulation signal recognition method based on CNN. The implementation steps of the USRP modulation signal recognition technology based on CNN are: LabVIEW graphical programming is performed on the host computer to realize a transceiver system that can switch multiple modulation signals; after receiving the signal, the USRP receiving end receives and stores the signal through the signal acquisition program, and provides the required signal training set and signal test set for the CNN model of modulation signal recognition; deep learning is used to perform modulation signal recognition and CNN model training; the trained CNN model is deployed to the test set data, and after the training is completed, the trained deep learning model is tested and verified. The trained modulation signal recognition module is deployed to the actual application scenario for modulation recognition of the receiver. Through the communication test environment built based on USRP software radio, it is verified that the signal modulation recognition technology based on deep learning can well solve the problem of modulation signal recognition in wireless communications.

[0066] In the software radio of this embodiment, except for the basic frequency conversion, A / D, D / A conversion and RF driver implemented by the hardware platform USRP RIO, the rest of the functions are designed and completed in the form of software. Except for the basic transceiver function, the other extended functions of the entire communication process need to be designed and programmed by yourself.

[0067] Some of the main physical parameters of the USRP RIO are as follows: The USRO RIO device has a 2×2 multi-input / multi-output RF transceiver, which can achieve dual transmission and dual reception. The RF bandwidth transceiver has a maximum transmit signal power of 30dBm, can receive or transmit signals from 1.2GHz to 6GHz, and the maximum instantaneous bandwidth can reach 40MHz. The maximum bidirectional point-to-point bandwidth of the USRP RIO device is up to 830MB / s, and its digital backend interface is a PCle×4 bus, which can reach a speed of 800MB / s.

[0068] The software part of the experiment uses Labview2015 and Python to design and debug the program. Based on the RF transceiver driver provided by the software itself, a series of functions required by the present invention are expanded to realize the entire invention.

[0069] System flow see Figure 2 In the experimental environment, a transmitter and a receiver are set up. After the signal is transmitted and received by the USRP through wireless communication, LabVIEW collects data on the received modulated signal to form a data set. Using deep learning technology, the data set is provided to the convolutional neural network for training and learning, and offline modulation signal recognition and processing is performed.

[0070] The experimental program is mainly divided into four parts: transmitter and receiver, data set extraction module, and CNN recognition module.

[0071] Transmitter and receiver experimental setup:

[0072] The transmitter and receiver are responsible for data communication in the experiment. Multiple antennas can be configured on a USRP device, and SISO self-transmission and self-reception can be realized. Two single antennas are configured as transmission and reception. Although they are physically located on the same device, the data communication between the two is completed through the wireless channel, which can effectively simulate the function of transmitting data in a wireless environment. Taking the transmitter and receiver as an example, Figure 3 This is the functional block diagram of the transmitter. Figure 4 This is the functional block diagram of the receiver.

[0073] Depend on Figure 3 It can be seen that at the transmitter, starting from the signal source, after source coding, channel coding, signal modulation, USRP up-conversion, digital-to-analog conversion (DAC), it is sent to the wireless channel through the RF transmission module. In order to modify the RF parameters of the transmitter in real time, such as center frequency, gain, etc., an external expansion interface is added to the RF transmission module to realize the call of the transmission parameter modification module.

[0074] Depend on Figure 4It can be seen that at the receiving end, wireless information is received from the receiving antenna, and after passing through the RF receiving module, it undergoes analog-to-digital conversion (CAD) and down-conversion processing to receive the modulated signal. The RF receiving module adds an external expansion interface to achieve real-time modification of the RF parameters of the receiving end.

[0075] Dataset extraction module settings:

[0076] The wireless signal is transmitted through the transmitter. After receiving the signal, the USRP receiver receives and stores the signal through the signal acquisition program, providing the required signal training set and signal test set for the CNN model of modulated signal recognition.

[0077] CNN recognition module settings:

[0078] Figure 5 Finalize the convolutional neural network for the deep learning neural network structure of CNN, and select appropriate network parameters and training parameters for the deep learning framework. The specific settings are:

[0079] 1. Feature extraction network design: Two convolution layers are used, and the convolution step size is 1. A two-layer convolution layer scheme with 64 kernels in the first layer is adopted. The size of the first convolution kernel is 2×4, the excitation function is ReLU, the bias term is used and initialized to zero, the weight initialization method is Glorot uniform distribution initialization, and the convolution kernel does not consider the completion method, regularization term and weight constraint; the number of convolution kernels in the second convolution layer is 16, the size of the convolution kernel is 1×4, and the rest of the settings are the same as the first convolution layer.

[0080] 2. Recognition network design: The subsequent fully connected neural network includes three fully connected layers. The first two layers are set to 64 and 16 neurons respectively, and the activation function is ReLU; the weights are initialized using Glorot uniform distribution, and regularization terms and weight constraints are used. The output layer activation function uses Softmax.

[0081] The specific training parameters for CNN are: using Adam optimization method and multi-category cross entropy loss function. The learning rate is set to 10 -4 , learning rate update related parameters β1 = 0.9, β2 = 0.999; fuzzy factor is set to 10 -8 , the decay value of the learning rate after update is 0.

[0082] The experimental process of this embodiment is as follows:

[0083] Step 1: Set the signal transmission and reception parameters and collect signal data, see Figure 6 and Figure 7, are the adjustable RF parameters of the transmitter and receiver respectively. The RF parameters of the transmitter are set as follows: RF antenna, center frequency, transmit gain, modulation mode, etc. The RF parameters of the receiver are set as follows: RF antenna, center frequency, receive gain, etc. Wireless communication transmission of different modulation signals under different signal-to-noise ratios can be achieved through configuration settings.

[0084] Step 2: The user starts preparing for communication. There are 8 types of modulation signals for transmission between transceivers, namely: 2ASK, 4ASK, BPSK, 4PSF, 4QAM, 16QAM, 64QAM;

[0085] Step 3: Collect signal data. Figure 8 It is a USRP signal acquisition program. The wireless signal is transmitted through the transmitter. After receiving the signal, the USRP receiver receives and stores the signal through the signal acquisition program, providing the required signal training set and signal test set for the CNN model of modulation signal recognition.

[0086] Step 4: Train the CNN model. The generated data set is used to train the CNN model for modulated signal recognition proposed in this paper.

[0087] Step 5: Deploy the trained CNN model to the test set data to verify the modulation signal recognition results.

[0088] Fig. 9 and Fig.10 The recognition rate of each signal under different signal-to-noise ratios is compared. The experiment shows that in the modulation signal recognition effect experiment under different signal sets, relatively ideal recognition effects are achieved. According to the signal recognition rate, it can be analyzed that low-order modulation signals can achieve good modulation signal recognition performance under different signal-to-noise ratios, and high-order modulation signals need a better channel environment, that is, a higher signal-to-noise ratio, to achieve better modulation signal recognition performance.

[0089] Fig.11 The modulation signal recognition performance of the CNN-based modulation signal recognition scheme is compared with that of the traditional machine learning-based scheme, namely the digital feature plus fully connected network scheme (TF-NN), and the average likelihood ratio scheme (ALRT). The experimental results show that the modulation signal recognition scheme based on CNN has a superior classification and recognition performance close to the traditional likelihood ratio-based scheme, and can achieve a recognition rate of more than 95% when the signal-to-noise ratio is greater than 5dB. Compared with the machine learning scheme, the recognition rate is also better, showing that the CNN signal extraction scheme through the convolution layer is better than the manually set high-order data features. This shows the superiority of the CNN convolution layer feature extraction.

[0090] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A USRP modulation signal recognition method based on CNN, characterized in that: The modulation signal recognition method comprises the following steps: Use USRP platform and host computer with LabVIEW graphical programming to build wireless communication system; There are 8 types of modulation signals for transmission signals between transceivers, namely: 2ASK, 4ASK, BPSK, 4PSF, 4QAM, 16QAM, 64QAM; The wireless signal is transmitted through the transmitter. After receiving the signal, the USRP receiver receives and stores the signal through the signal acquisition program to obtain the signal training set and signal test set. Establishing a CNN model for modulated signal recognition, and inputting the signal training set into the CNN model for training; The trained CNN model is deployed to the test set data to verify the modulation signal recognition results.

2. A USRP modulation signal recognition method based on CNN as claimed in claim 1, characterized in that: The construction of the wireless communication system using the USRP platform and the host computer using LabVIEW graphical programming includes: The physical parameters of the USRP platform are as follows: The USRO RIO device has a 2×2 multi-input / multi-output RF transceiver for dual transmission and dual reception. The RF bandwidth transceiver has a maximum transmit signal power of 30dBm and is used to receive or transmit signals from 1.2GHz to 6GHz. The maximum instantaneous bandwidth reaches 40MHz. The maximum bidirectional point-to-point bandwidth of the USRP RIO device is up to 830MB / s, and its digital backend interface is a PCle×4 bus with a speed of 800MB / s.

3. A USRP modulation signal recognition method based on CNN as claimed in claim 1, characterized in that: There are 8 types of modulation signals for the transmission signals between the transceivers, namely: 2ASK, 4ASK, BPSK, 4PSF, 4QAM, 16QAM, 64QAM, including: The transmitter and receiver are responsible for data communication. Multiple antennas can be configured on a USRP device, and SISO self-transmission and self-reception can be realized. Two single antennas are configured for transmission and reception. The two single antennas are located on the same device, and the data communication between the two is completed through a wireless channel.

4. A USRP modulation signal recognition method based on CNN as claimed in claim 1, characterized in that: There are 8 types of modulation signals for the transmission signals between the transceivers, namely: 2ASK, 4ASK, BPSK, 4PSF, 4QAM, 16QAM, 64QAM, and also include: After source coding, channel coding, signal modulation, USRP up-conversion, and digital-to-analog conversion DAC, it is sent to the wireless channel through the RF transmission module, and an external expansion interface is added to the RF transmission module.

5. A USRP modulation signal recognition method based on CNN as claimed in claim 1, characterized in that: There are 8 types of modulation signals for the transmission signals between the transceivers, namely: 2ASK, 4ASK, BPSK, 4PSF, 4QAM, 16QAM, 64QAM, and also include: At the receiving end, wireless information is received from the receiving antenna, passes through the RF receiving module, and then undergoes analog-to-digital conversion CAD and down-conversion processing to receive the modulated signal. The RF receiving module adds an external expansion interface.

6. A USRP modulation signal recognition method based on CNN as claimed in claim 1, characterized in that: The wireless signal is transmitted through the transmitting end. After receiving the signal, the USRP receiving end receives and stores the signal through the signal acquisition program to obtain a signal training set and a signal test set, including: The line signal is transmitted through the transmitter. After receiving the signal, the USRP receiver receives and stores the signal through the signal acquisition program, providing the required signal training set and signal test set for the CNN model of modulation signal recognition.

7. A USRP modulation signal recognition method based on CNN as claimed in claim 1, characterized in that: The step of establishing a CNN model for modulated signal recognition and inputting the signal training set into the CNN model for training includes: The CNN model uses two convolutional layers with a convolution step of 1. The first layer of the two convolutional layers has 64 kernels. The size of the first convolution kernel is 2×4, the activation function is ReLU, the bias term is used and initialized to zero, the weight initialization method is Glorot uniform distribution initialization, and the convolution kernel does not consider the completion method, regularization term and weight constraint. The number of convolution kernels in the second convolutional layer is 16, the size of the convolution kernel is 1×4, and the rest of the settings are the same as the first convolutional layer.

8. A USRP modulation signal recognition method based on CNN as claimed in claim 1, characterized in that: The step of establishing a CNN model for modulated signal recognition and inputting the signal training set into the CNN model for training further includes: The fully connected neural network in the CNN model includes three fully connected layers. The first two layers are set to 64 and 16 neurons respectively, and the activation function is ReLU. The weights are initialized using Glorot uniform distribution, and regularization terms and weight constraints are used. The output layer activation function uses Softmax.

9. A method for identifying USRP modulation signals based on CNN as claimed in claim 1, characterized in that: The step of establishing a CNN model for modulated signal recognition and inputting the signal training set into the CNN model for training further includes: The training parameters of the CNN model are set to: use the Adam optimization method and the multi-category cross entropy loss function. Set the learning rate to 10 -4 , learning rate update related parameters β1=0.9,β2=0.999; fuzzy factor is set to 10 -8 , the decay value of the learning rate after update is 0.

Images