A communication signal detection and recognition method based on improved YOLOv5

Abstract

The present invention discloses a communication signal detection and recognition method based on improved YOLOv5. The method uses a lightweight improved YOLOv5, that is, a lightweight structure is used in the backbone network to reduce the amount of calculation, and the recognition ability is enhanced by combining a targeted scale feature fusion module and a feature attention module. The real-time recognition and extraction of features such as the existence, bandwidth, carrier frequency, duration and modulation mode of the communication signal can be achieved through the signal time-frequency diagram. The method proposes a signal detection and recognition scheme that can realize a series of functions such as signal acquisition, processing, training, real-time detection and recognition, and is easy to migrate and deploy as a whole. The present invention has high recognition performance in real-time detection and recognition of multiple parameters of communication signals, and has low difficulty in application and deployment in various terminal devices.

Description

Technical Field

[0001] The present invention relates to the technical field of signal detection and recognition, and more specifically to a communication signal detection and recognition method based on improved YOLOv5. Background Art

[0002] Over the past few decades, with the rapid development of mobile communications technology, especially the widespread rollout of 5G in recent years, the Internet of Things has taken off, and a variety of communication devices have emerged. Wireless spectrum resources are becoming increasingly scarce, and the wireless electromagnetic space is becoming increasingly complex and volatile. To ensure the legal and reasonable use of wireless spectrum resources and maintain the wireless communication environment, spectrum monitoring has become a key research topic in the field of wireless communications.

[0003] Spectrum monitoring is primarily responsible for monitoring, measuring, analyzing, and assessing radio communication signals. This allows for understanding the usage of the radio electromagnetic spectrum, the presence of signal sources, and other aspects of the spectrum, enabling radio management and scheduling. Spectrum monitoring often involves non-cooperative communication. This means that the monitoring party, acting as the receiving end, cannot confirm prior information about the communication signal with the transmitting end, resulting in blind signals. Detecting and identifying blind signals is the beginning of blind signal processing and a crucial step in spectrum monitoring under non-cooperative communication. Therefore, this paper focuses on the detection and identification of communication signals under non-cooperative communication.

[0004] Detecting and identifying signals under non-cooperative communication is an important part of spectrum monitoring. Its main goal is to estimate and identify parameters such as the existence, carrier frequency, bandwidth, duration, and modulation method of communication signals, laying the foundation for the subsequent identification and decryption of parameters such as channel coding, source coding, source identity, and communication information.

[0005] Communication signal detection and recognition algorithms have been a hot research topic with the popularization of mobile communication technology, but traditional detection and recognition algorithms often have the following defects: traditional algorithms are extremely dependent on the extraction of key features of communication signals, including instantaneous amplitude, power spectrum and high-order statistics, and the processing process is complex and computationally intensive.

[0006] With the rapid development of artificial intelligence (AI) technology, especially deep learning, the detection and identification of communication signals using deep learning has become a new research direction. Communication signal detection and identification is essentially still the extraction and induction of signal features, which are then used to complete classification. Deep learning, by simulating the cognitive structure of the human brain to establish learning models, acquires powerful feature extraction capabilities, and interprets various input data, showing significant advantages in classification problems. Leveraging the powerful feature learning capabilities of deep learning to design efficient signal detection and identification network models and realize communication signal detection and identification systems under non-cooperative communication has extremely high practical significance.

[0007] Traditional signal detection and recognition algorithms based on deep learning generally only detect a single parameter characteristic of a communication signal. When detecting multiple signal parameters, multiple algorithms must be combined, making them impractical for widespread application. Deep learning models for target classification tasks struggle to achieve classification capabilities when faced with parameters such as signal duration, carrier frequency, and bandwidth, as the number of parameter categories is nearly inexhaustible. While deep learning offers significant advantages in feature learning and fitting classification, its performance often relies on complex network models. Application reasoning for complex network models ultimately relies on local hardware resources, making them inconvenient for deployment. Summary of the Invention

[0008] Technical Problem: The purpose of this invention is to provide a communication signal detection and recognition method based on an improved YOLOv5. This method proposes a signal detection and recognition solution that can implement a series of functions such as signal acquisition, processing, training, and real-time detection and recognition, and is easy to migrate and deploy on various platforms. The present invention has high recognition performance in the real-time detection and recognition of multiple parameters of communication signals, and uses only a single model, which has a low computational load and is easy to deploy on various terminal devices.

[0009] Technical solution: In order to achieve the above-mentioned purpose of the invention, the present invention provides a communication signal detection and recognition method based on improved YOLOv5, which can realize real-time recognition and extraction of communication signal existence, bandwidth, carrier frequency, duration, modulation mode and other characteristics through signal time-frequency diagram.

[0010] The method comprises the following steps:

[0011] Step 1. Obtain a communication signal time-frequency diagram dataset: obtain communication signal simulation data through signal simulation, and then convert the signal simulation data into a signal time-frequency diagram to obtain a communication signal time-frequency diagram target simulation dataset and a communication signal time-frequency diagram dataset obtained by actual measurement using instruments and receiving equipment. Preprocess the communication signal time-frequency diagram, annotate it according to the parameters of the communication signal, such as duration, bandwidth, carrier frequency, presence, and modulation mode, and divide it into a training dataset and a test dataset.

[0012] Step 2: Build an improved YOLOv5 model. This involves using a lightweight structure in the YOLOv5 backbone network and improving the YOLOv5 model by combining a targeted scale feature fusion module and a feature attention module. The targeted scale feature fusion module is improved based on the bidirectional feature pyramid BiFPN, and the feature attention module is based on the simple parameter-free attention module SimAM attention mechanism.

[0013] Step 3: Input the training data set into the improved YOLOv5 model for training to obtain a trained improved YOLOv5 model;

[0014] Step 4: Input the test data set into the trained improved YOLOv5 model to obtain the detection and recognition results.

[0015] This method specifically includes the following four parts:

[0016] In the first part, a receiving device receives communication signals in real time through cables and antennas. After analog-to-digital conversion, signal processing, and signal acquisition, the communication signal data is obtained. The communication signal data is converted into a signal time-frequency diagram and transmitted in real time.

[0017] The second part is to establish, train and deploy the improved YOLOv5 model, including: establishing the improved YOLOv5 model, that is, using a lightweight structure in the YOLOv5 backbone network, and improving it by combining the targeted scale feature fusion module and the feature attention module. The targeted scale feature fusion module is improved based on BiFPN, and the feature attention module is based on the SimAM attention mechanism. The activation function uses ReLU, Leaky ReLU, Swish, Mish, PReLU and other activation functions, and the loss function uses IOU, CIOU, Wise IOU and other loss functions; training the improved YOLOv5 model, that is, inputting the training data set into the improved YOLOv5 model for training to obtain the trained improved YOLOv5 model; deploying the improved YOLOv5 model, that is, converting the trained improved YOLOv5 model into a format suitable for deployment using TensorFlow Lite, ONNX, and optimize the trained improved YOLOv5 model, using multiple quantization methods and branch reduction operations such as int8, FP16, and FP32. Finally, the model is deployed on the target platform, using multiple platforms such as servers, edge devices, and mobile devices;

[0018] In the third part, the real-time signal frequency graph is input into the deployed improved YOLOv5 model for real-time detection and recognition, using image or video input;

[0019] In the fourth part, the recognition results of the improved YOLOv5 model are converted into multiple real-time parameters of the communication signal, combining the characteristics of the system operation time and the signal time-frequency diagram. This includes converting the positioning box of the communication signal in the time-frequency diagram into parameters such as duration, bandwidth, carrier frequency, and existence, and converting the target type of the signal into the type of modulation parameter. At the same time, multiple parameters are sorted to form a detection and recognition report that is more in line with the public's reading habits.

[0020] The lightweight structure used in the YOLOv5 backbone network makes lightweight improvements to the standard Conv unit, CBL unit, Res unit unit, and CSP unit in YOLOv5, specifically including:

[0021] The standard Conv unit is split into a lightweight Conv unit. First, the input of the standard Conv unit is evenly split into two parts. One part is convolved according to the standard convolution process. The convolution kernel size remains unchanged, but the number of output channels is halved, to obtain half of the normal output feature map. The other part first passes through the standard convolution with a convolution kernel size of 1×1, and then the convolution output is grouped and convolved. The convolution kernel size remains unchanged, the number of output channels is halved, and the number of groups is 1 / 4 of the number of input channels, to obtain the other half of the output feature map. The output feature maps of the two parts are spliced ​​together by the splicing layer and the channels are rearranged to form a complete output feature map.

[0022] The standard Conv unit in the CBL unit is replaced by a lightweight Conv unit. At the same time, the BN layer and ReLU layer superimposed after the standard Conv unit are placed before the output feature map splicing. The outputs of the two parts of the convolution process are executed separately to form a lightweight CBL unit.

[0023] The convolution process of the standard Conv unit in the CBL unit is replaced by the grouped convolution process to form a grouped CBL unit;

[0024] The CBL unit in the Res unit is replaced by a lightweight CBL unit and a grouped CBL unit to form a lightweight Res unit: first, the two-way process of the Res unit is changed to an average input split, and then split into two parts; one part replaces the CBL units with convolution kernel sizes of 1×1 and 3×3 in the Res unit with improved CBL units with a convolution kernel of 1×1 and a grouped CBL unit with a convolution kernel of 3×3 to extract features; the other part uses a separate 1×1 CBL unit, which is responsible for the change in the output feature size dimension but not for feature extraction; finally, the addition layer in the Res unit is replaced by a splicing layer to complete the combination of the output feature maps of the split part, and the channel rearrangement is used to output the complete feature map;

[0025] The CBL unit and Res unit in the CSP unit are replaced by lightweight CBL unit and lightweight Res unit, and the structure is simplified to form a lightweight CSP unit: first, the CBL unit with a 1×1 convolution kernel and the standard Conv layer with a 1×1 convolution kernel in the CSP unit are merged into a lightweight CBL unit with a 1×1 convolution kernel. The subsequent two operations share the output of the lightweight CBL unit; at the same time, the convolution unit structure consisting of the convolution layer, splicing layer, BN layer, and ReLU layer at the output end of the CSP unit is deleted, retaining only the combined output features of the splicing layer, and subsequent processing is completed by channel rearrangement.

[0026] The targeted scale feature fusion module uses a BiFPN-based targeted scale feature fusion module. The BiFPN-based targeted scale feature fusion module blocks the path for fusing small-scale features to large-scale features, and large-scale features are only fused with moderate-scale features. The targeted scale feature fusion module focuses on the moderate scale and uses the bidirectional structure in BiFPN to fuse features upward and downward from the moderate scale respectively, thereby realizing large-scale and small-scale feature extraction. At the same time, an additional path is added to the moderate scale to fuse the three scale features again, thereby extracting the final moderate-scale feature.

[0027] The feature attention module is constructed based on the SimAM attention mechanism; the feature attention module based on the SimAM attention mechanism realizes the weight allocation of each neuron through separability and activation function sigmoid, completes the spatial and channel attention allocation, and adds the fusion of features of different scales in YOLOv5 and the feature extraction of different scales in the targeted scale feature fusion module based on BiFPN.

[0028] The preprocessing of the communication signal time-frequency diagram includes: adding noise, changing brightness, cropping, translating, rotating and scaling the data set; and adjusting the duration range, bandwidth range, sampling frequency and power display range of the time-frequency diagram display.

[0029] The labeling according to the parameters of the communication signal, such as duration, bandwidth, carrier frequency, existence, and modulation mode, includes: labeling each image in the training set using a graphic labeling tool, and automatically generating a labeling text file according to rules using a code.

[0030] Beneficial effects: The present invention discloses a communication signal detection and recognition method based on improved YOLOv5, which can realize the real-time recognition and extraction of multiple parameter features such as the existence, bandwidth, carrier frequency, duration and modulation mode of the communication signal through the signal time-frequency diagram. The method uses a lightweight improved YOLOv5, that is, a lightweight structure is improved in the backbone network to reduce the amount of calculation, and the recognition ability is enhanced by combining a targeted scale feature fusion module and a feature attention module. The method proposes a signal detection and recognition scheme, which can realize a series of functions such as signal acquisition, processing, training, real-time detection and recognition, and is easy to migrate and deploy on various platforms as a whole. The present invention has high recognition performance in the real-time detection and recognition of multiple parameters of communication signals, and only uses one model with a small amount of calculation and low difficulty in application and deployment in various terminal devices. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 This is a diagram of a communication signal detection and identification scheme of the present invention;

[0032] Figure 2This is a diagram showing the lightweight improvement of the standard Conv unit of the present invention;

[0033] Figure 3 This is a diagram showing the lightweight improvement of the CBL unit of the present invention;

[0034] Figure 4 This is a diagram showing the lightweight improvement of the Res unit of the present invention;

[0035] Figure 5 This is a diagram showing the lightweight improvement of the CSP unit of the present invention;

[0036] Figure 6 Schematic diagram of the targeted scale feature fusion module of the present invention;

[0037] Figure 7 Schematic diagram of the feature attention module based on the SimAM attention mechanism of the present invention;

[0038] Figure 8 This is a comparison chart of the computational complexity of the improved YOLOv5 and the standard YOLOv5. DETAILED DESCRIPTION

[0039] To further clarify the objectives, technical solutions, and advantages of the embodiments of the present invention, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. It should be noted that the described embodiments are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without inventive effort are within the scope of protection of the present invention. It should be noted that the terms used herein are intended only to describe specific embodiments and are not intended to limit the exemplary embodiments according to this application. As used herein, unless the context clearly indicates otherwise, the singular is intended to include the plural. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they specify the presence of features, steps, operations, devices, components, and / or combinations thereof. Unless otherwise specified, the relative arrangement of components and steps, numerical expressions, and numerical values ​​described in these embodiments do not limit the scope of the present invention. It should also be understood that for ease of description, the dimensions of the various parts shown in the drawings are not drawn to scale. Technologies, methods, and devices known to persons of ordinary skill in the relevant art may not be discussed in detail, but where appropriate, such technologies, methods, and devices should be considered part of the specification. In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not limiting. Therefore, other examples of the exemplary embodiments may have different values. It should be noted that similar reference numerals and letters represent similar items in the following figures, and therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures.

[0040] The application scenario of this invention mainly refers to the detection and identification of blind signals in third-party non-cooperative communication scenarios. The specific implementation plan for unknown signals is as follows:

[0041] In the first part, a receiving device is used to receive communication signals in real time through cables and antennas, and the communication signal data is obtained through analog-to-digital conversion, signal processing, signal acquisition and other steps. The communication signal data is converted into a signal time-frequency diagram and transmitted in real time.

[0042] The second part is to establish, train and deploy the improved YOLOv5 model, including: establishing the improved YOLOv5 model, that is, using a lightweight structure in the YOLOv5 backbone network, and improving it by combining the targeted scale feature fusion module and the feature attention module. The targeted scale feature fusion module is improved based on BiFPN, and the feature attention module is based on the SimAM attention mechanism. The activation function can use ReLU, Leaky ReLU, Swish, Mish, PReLU and other activation functions, and the loss function can use IOU, CIOU, Wise IOU and other loss functions; training the improved YOLOv5 model, that is, inputting the training data set into the improved YOLOv5 model for training to obtain the trained improved YOLOv5 model; deploying the improved YOLOv5 model, that is, converting the trained improved YOLOv5 model into a format suitable for deployment, which can be done using TensorFlow Lite, ONNX, and other formats, and optimize the trained improved YOLOv5 model, using various quantization methods and branch reduction operations such as int8, FP16, and FP32, and finally deploy the model to the target platform, which can be a variety of platforms such as servers, edge devices, and mobile devices;

[0043] In the third part, the real-time signal frequency graph is input into the deployed improved YOLOv5 model for real-time detection and recognition. This can be done using a variety of input methods, such as images and videos.

[0044] In the fourth part, the recognition results of the improved YOLOv5 model are converted into multiple real-time parameters of the communication signal, combining the characteristics of the system operation time and the signal time-frequency diagram. This includes converting the positioning box of the communication signal in the time-frequency diagram into parameters such as duration, bandwidth, carrier frequency, and existence, and converting the target type of the signal into the type of modulation parameter. At the same time, multiple parameters are sorted out to form a detection and recognition report that is more in line with the public's reading habits.

[0045] Furthermore, the steps for building and training the improved YOLOv5 model in the second part are as follows:

[0046] Step 1. Obtain a communication signal frequency spectrum data set: obtain communication signal simulation data through signal simulation, then convert the signal simulation data into a signal frequency spectrum to obtain a communication signal frequency spectrum target simulation data set and a communication signal frequency spectrum data set measured by instruments and receiving equipment. Preprocess the communication signal frequency spectrum, annotate it according to parameters such as the duration, bandwidth, carrier frequency, presence, and modulation mode of the communication signal, and divide it into a training data set and a test data set.

[0047] Step 2. Establish an improved YOLOv5 model. That is, use a lightweight structure in the YOLOv5 backbone network and improve it by combining the targeted scale feature fusion module and the feature attention module. The targeted scale feature fusion module is improved based on BiFPN, and the feature attention module is based on the simple parameter-free attention module SimAM attention mechanism. The activation function can use ReLU, Leaky ReLU, Swish, Mish, PReLU and other activation functions, and the loss function can use IOU, CIOU, Wise IOU and other loss functions.

[0048] Step 3: Input the training data set into the improved YOLOv5 model for training to obtain a trained improved YOLOv5 model;

[0049] Step 4: Input the test data set into the trained improved lightweight YOLOv5 model to obtain the detection and recognition results.

[0050] Furthermore, the key points of the improved YOLOv5 model in step 2 are as follows:

[0051] The lightweight structure is used in the backbone network of YOLOv5 to make lightweight improvements to the standard Conv unit, CBL unit, Res unit unit, and CSP unit in YOLOv5.

[0052] The standard Conv unit is improved by lightweight, such as Figure 2 As shown in the figure, the standard Conv unit is split into two parts. The input of the standard Conv unit is first split evenly into two parts. One part is convolved according to the standard convolution process, with the kernel size unchanged but the number of output channels halved, resulting in half of the normal output feature map. The other part first undergoes a standard convolution with a kernel size of 1×1. The convolution output is then grouped and convolved, with the kernel size unchanged but the number of output channels halved to 1 / 4 of the input channels, resulting in the other half of the output feature map. The two output feature maps are then concatenated by the concatenation layer, followed by channel rearrangement to form the complete output feature map.

[0053] The CBL unit is improved to be lightweight, such as Figure 3 As shown in the figure, the standard Conv unit in the CBL unit is replaced by a lightweight Conv unit. At the same time, the BN layer and ReLU layer superimposed after the standard Conv unit are placed before the output feature map is spliced. The outputs of the two parts of the convolution process are executed separately to form a lightweight CBL unit.

[0054] The Res unit is improved to be lightweight, such as Figure 4As shown in the figure, the CBL unit in the Res unit is replaced with a lightweight CBL unit and a grouped CBL unit to form a lightweight Res unit. First, the two-way process of the Res unit is changed to an average input split, splitting it into two parts. In one part, the CBL units with convolution kernel sizes of 1×1 and 3×3 in the Res unit are replaced with improved CBL units with a convolution kernel of 1×1 and grouped CBL units with a convolution kernel of 3×3 to extract features. In the other part, a single 1×1 CBL unit is used to change the output feature size dimension but not for feature extraction. Finally, the addition layer in the Res unit is replaced with a splicing layer to complete the combination of the output feature maps of the split parts, and the channels are rearranged to output the complete feature map.

[0055] The grouped CBL unit is formed by replacing the convolution process of the standard Conv unit in the CBL unit with the grouped convolution process.

[0056] The CSP unit is improved to be lightweight, such as Figure 5 As shown in the figure, the CBL unit and Res unit in the CSP unit are replaced with lightweight CBL units and lightweight Res units, and the structure is streamlined to form a lightweight CSP unit. First, the CBL unit with a 1×1 convolution kernel and the standard Conv layer with a 1×1 convolution kernel in the CSP unit are merged into a lightweight CBL unit with a 1×1 convolution kernel. Subsequent operations on the two channels share the output of this lightweight CBL unit. At the same time, the convolution unit structure at the output of the CSP unit, consisting of the convolution layer, concatenation layer, BN layer, and ReLU layer, is deleted, retaining only the combined output features of the concatenation layer. Subsequent processing is completed by channel rearrangement.

[0057] The improvement of the joint targeted scale feature fusion module is to replace the FPN+PAN in YOLOv5 with a targeted scale feature fusion module based on BiFPN.

[0058] The BiFPN-based targeted scale feature fusion module, such as Figure 6 As shown in the figure, the path from small-scale features to large-scale features is blocked, and large-scale features are only fused with moderate-scale features. The targeted scale feature fusion module focuses on the moderate scale and uses the bidirectional structure of BiFPN to fuse features from the moderate scale upward and downward, respectively, to achieve large-scale and small-scale feature extraction. At the same time, an additional path is added at the moderate scale to fuse the three scale features again, and then extract the final moderate-scale feature.

[0059] The joint feature attention module is improved to be a feature attention module based on SimAM attention mechanism. The feature attention module based on SimAM attention mechanism is as follows: Figure 7 As shown in the figure, the weight distribution of each neuron is achieved by combining separability with the activation function sigmoid, completing the spatial and channel attention distribution, adding the fusion of features of different scales in YOLOv5, and the feature extraction of different scales in the targeted scale feature fusion module based on BiFPN.

[0060] The present invention presents two experimental examples to evaluate the recognition performance of the communication signal detection and recognition method based on the improved YOLOv5 and the application performance of the communication signal detection and recognition scheme based on the improved YOLOv5.

[0061] In Experimental Example 1, a signal simulation time-frequency graph dataset was used. It contained communication signals using 11 modulation schemes, including BPSK, QPSK, QAM8, 16QAM, QAM32, 64QAM, AM, AM-DSB, AM-SSB, FM, and FSK. During the signal presence test, all modulation scheme signal samples were classified into one category: the target communication signal samples.

[0062] The parameters of the signal simulation time-frequency graph dataset are shown in Table 1. Because YOLOv5 image object detection adaptively scales target images, the image size of the signal simulation time-frequency graph dataset is not 640×640, but all images remain the same size. The signals in the signal simulation time-frequency graph dataset are primarily low-frequency, narrowband signal graphs, but the results are still meaningful for high-frequency, wideband signals. When these high-frequency, wideband signals are processed into time-frequency graphs, they are relatively "low-frequency, narrowband signal graphs" compared to the overall time-frequency graph. The image features are essentially the same, and the time-frequency graphs store the relative frequency band information of the signals. The noise environment of the signal simulation time-frequency graph dataset is divided into -20dB, -10dB, 0dB, 10dB, and 20dB, based on the ratio of the average power of the entire signal sequence to the channel noise. All signals in a single time-frequency graph are exposed to the same noise environment. The test set to training set ratio is 9:1.

[0063] The signal simulation time-frequency graph dataset is divided into six subsets based on the noise environment: a mixed noise environment dataset, containing 500 time-frequency graphs each for noise environments of -20dB, -10dB, 0dB, 10dB, and 20dB, for a total of 2,500 graphs; a -20dB noise environment dataset, containing 1,000 -20dB time-frequency graphs; a -10dB noise environment dataset, containing 1,000 -10dB time-frequency graphs; a 0dB noise environment dataset, containing 1,000 0dB time-frequency graphs; a 10dB noise environment dataset, containing 1,000 10dB time-frequency graphs; and a 20dB noise environment dataset, containing 1,000 20dB time-frequency graphs. The total time-frequency graph dataset contains 7,500 time-frequency graphs, with over 45,000 signal samples.

[0064] System configuration: CPU+GPU is an edge terminal based on the Jetson Orin module, the operating system is Ubuntu 20.04, the software environment is Python 3.8, CUDA 11.4, and the software framework is PyTorch V1.12+torchvision v0.13.0.

[0065] like Figure 8 As shown in the figure, the computational amount comparison chart of the improved YOLOv5 of the present invention and the standard YOLOv5 shows the computational gap between different versions of YOLOv5. FLOPs is the number of floating point operations, which is the computational amount used to measure the complexity of the model. GFLOPs is 1 billion floating point operations. Figure 8 As shown, the improved YOLOv5 is compatible with different YOLOv5 versions, exhibiting varying computational loads depending on the version. Compared to the standard YOLOv5 versions, the improved YOLOv5 versions have significantly reduced computational load, with the difference ranging from a minimum of 1.6 times to a maximum of 5.7 times. The improved YOLOv5s has a computational load of only 5.1 GFLOPs, similar to that of the standard YOLOv5n, while the lightweight YOLOv5n has a computational load of only 1.6 GFLOPs. This reduced computational load greatly facilitates the deployment and application of the improved YOLOv5 on various platforms.

[0066] Table 2 shows the simulated signal presence detection results of the improved YOLOv5 algorithm in this invention. In the case of third-party non-cooperative communication, detection of the communication signal presence focuses more on the omission of the target signal, specifically the R value. A higher R value indicates fewer target signal omissions. As shown in Table 2, in a -20dB noise environment, due to signal quality limitations, omissions are high, and the average precision is also low, with an mAP50-95 value of less than 30%, which conforms to general cognitive principles. The higher P value is because the energy in the signal sequence is not evenly distributed, but rather concentrated in the signal, while the noise follows a Gaussian distribution. At higher energy levels, there is a greater probability of the target signal, rather than noise, being present. As signal quality improves and the noise environment improves, the R value continues to increase, reaching 80% at -10dB. In noise environments above 0dB, the R value is always above 80%. In a noise environment of 20dB, the R value reaches a maximum of 97.4%, with almost no omission of the target signal. The average accuracy at an IOU threshold of 50% is close to 100%, proving that the improved YOLOv5 model can effectively detect the presence of target signals. In mixed noise environments, the R value also reaches 82.8%, proving that the improved YOLOv5 model still maintains excellent performance in complex noise environments and can effectively detect target signals.

[0067] Table 3 shows the simulated signal presence detection results of the improved YOLOv5 proposed in this invention. Due to the increased number of communication signal classification targets, the improved YOLOv5's modulation recognition performance declined compared to signal presence detection. As shown in Table 3, in a noise environment of -20dB, limited by signal quality, the P value was 57.4%, the R value was 46.3%, and the mAP50-95 value was less than 25%. However, at -10dB, the P value increased to approximately 75%, and the R value increased to approximately 70%. When the noise environment was better than 0dB, the P value was above 85%, the R value was above 880%, and the mAP50 value was no less than 90%. In a 10dB noise environment, the P value achieved an average precision of 93.1%. Furthermore, in a mixed noise environment, the P value reached 83.3%, the R value reached 72.5%, the mAP50 value was 79.9%, and the mAP50-95 value exceeded 55%, maintaining modulation recognition capabilities in complex noise environments.

[0068] In Experimental Example 2, a time-frequency graph of a measured signal was used for testing. The communication signal was generated in real time by the SMW 200A, received and processed by the LimeSDR into a time-frequency graph, and then submitted to the improved YOLOv5 model for real-time recognition. Table 4 shows the parameters of the measured signal time-frequency graph dataset of the present invention. The system configuration is consistent with that in Experimental Example 1.

[0069] As shown in Table 5, the communication signal detection and recognition scheme proposed in this paper, based on the improved YOLOv5, can basically recognize third-party non-cooperative measured modulation signals. The detection values ​​for all eight modulation schemes are close to 100%. As the dataset continues to expand, this paper will further improve its ability to recognize third-party non-cooperative communication signals.

[0070] Any details not described in detail herein are generally known to those skilled in the art. The preferred embodiments of the present invention have been described in detail above. It should be understood that a person skilled in the art can make numerous modifications and variations based on the concepts of the present invention without inventive effort. Therefore, any technical solution that a person skilled in the art can arrive at through logical analysis, reasoning, or limited experimentation based on the concepts of the present invention and the prior art should be within the scope of protection defined by the claims.

[0071] Table 1 Parameters of the signal simulation time-frequency diagram dataset of the present invention

[0072]

[0073] Table 2 The results of the simulation signal existence detection of the improved YOLOv5 of the present invention

[0074]

[0075] Table 3. Results of simulation signal modulation recognition using improved YOLOv5 by the present invention

[0076]

[0077] Table 4 Parameters of the measured signal time-frequency diagram dataset of the present invention

[0078]

[0079] Table 5: Measured signal modulation recognition results of the improved YOLOv5 of the present invention

[0080] type P R mAP50 mAP50-95 all 99.7％ 99.8％ 99.4％ 97.9％ 128QAM 100％ 100％ 99.5％ 98.5％ 64QAM 98.3％ 98.7％ 98.7％ 97.8％ 32QAM 99.9％ 100％ 99.5％ 98.1％ 8PSK 100％ 100％ 99.5％ 97.7％ QPSK 99.9％ 100％ 99.5％ 97.9％ BPSK 99.9％ 100％ 99.5％ 97.2％ ASK 99.6％ 99.6％ 99.5％ 98.5％ MSK 100％ 100％ 99.5％ 97.4％

Claims

1. A communication signal detection and recognition method based on improved YOLOv5, characterized in that: The method comprises the following steps: Step 1. Obtain a communication signal time-frequency diagram dataset: obtain communication signal simulation data through signal simulation, and then convert the signal simulation data into a signal time-frequency diagram to obtain a communication signal time-frequency diagram target simulation dataset and a communication signal time-frequency diagram dataset obtained by actual measurement using instruments and receiving equipment. Preprocess the communication signal time-frequency diagram, annotate it according to the parameters of the communication signal, such as duration, bandwidth, carrier frequency, presence, and modulation mode, and divide it into a training dataset and a test dataset. Step 2: Build an improved YOLOv5 model. This involves using a lightweight structure in the YOLOv5 backbone network and improving the YOLOv5 model by combining a targeted scale feature fusion module and a feature attention module. The targeted scale feature fusion module is improved based on the bidirectional feature pyramid BiFPN, and the feature attention module is based on the simple parameter-free attention module SimAM attention mechanism. The lightweight structure used in the YOLOv5 backbone network makes lightweight improvements to the standard Conv unit, CBL unit, Res unit unit, and CSP unit in YOLOv5, specifically including: The standard Conv unit is split into a lightweight Conv unit. First, the input of the standard Conv unit is evenly split into two parts. One part is convolved according to the standard convolution process. The convolution kernel size remains unchanged, but the number of output channels is halved, to obtain half of the normal output feature map. The other part first passes through the standard convolution with a convolution kernel size of 1×1, and then the convolution output is grouped and convolved. The convolution kernel size remains unchanged, the number of output channels is halved, and the number of groups is 1 / 4 of the number of input channels, to obtain the other half of the output feature map. The output feature maps of the two parts are spliced ​​together by the splicing layer and the channels are rearranged to form a complete output feature map. The standard Conv unit in the CBL unit is replaced by a lightweight Conv unit. At the same time, the BN layer and ReLU layer superimposed after the standard Conv unit are placed before the output feature map splicing. The outputs of the two parts of the convolution process are executed separately to form a lightweight CBL unit. The convolution process of the standard Conv unit in the CBL unit is replaced by the grouped convolution process to form a grouped CBL unit; The CBL unit in the Res unit is replaced by a lightweight CBL unit and a grouped CBL unit to form a lightweight Resunit unit: first, the two-way process of the Res unit is changed to an average input split, split into two parts; one part replaces the CBL units with convolution kernel sizes of 1×1 and 3×3 in the Res unit unit with improved CBL units with a convolution kernel of 1×1 and grouped CBL units with a convolution kernel of 3×3 to extract features; the other part uses a separate 1×1 CBL unit, which is responsible for the change in the output feature size dimension but not for feature extraction; finally, the addition layer in the Res unit unit is replaced by a splicing layer to complete the combination of the output feature maps of the split part, and use channel rearrangement to output the complete feature map; The CBL unit and Res unit in the CSP unit are replaced by lightweight CBL unit and lightweight Res unit, and the structure is simplified to form a lightweight CSP unit: first, the CBL unit with a convolution kernel of 1×1 and the standard Conv layer with a convolution kernel of 1×1 in the CSP unit are merged into a lightweight CBL unit with a convolution kernel of 1×1. The subsequent two operations share the output of the lightweight CBL unit; at the same time, the convolution unit structure consisting of the convolution layer, splicing layer, BN layer, and ReLU layer at the output end of the CSP unit is deleted, retaining only the combined output features of the splicing layer, and the subsequent processing is completed by channel rearrangement; Step 3: Input the training data set into the improved YOLOv5 model for training to obtain a trained improved YOLOv5 model; Step 4: Input the test data set into the trained improved YOLOv5 model to obtain the detection and recognition results.

2. The communication signal detection and recognition method based on improved YOLOv5 according to claim 1, characterized in that This method specifically includes the following four parts: In the first part, a receiving device receives communication signals in real time through cables and antennas. After analog-to-digital conversion, signal processing, and signal acquisition, the communication signal data is obtained. The communication signal data is converted into a signal time-frequency diagram and transmitted in real time. The second part is to establish, train and deploy the improved YOLOv5 model, including: establishing the improved YOLOv5 model, that is, using a lightweight structure in the YOLOv5 backbone network, and improving it by combining the targeted scale feature fusion module and the feature attention module. The targeted scale feature fusion module is improved based on BiFPN, and the feature attention module is based on the SimAM attention mechanism. The activation function uses ReLU, Leaky ReLU, Swish, Mish, PReLU and other activation functions, and the loss function uses IOU, CIOU, Wise IOU and other loss functions; training the improved YOLOv5 model, that is, inputting the training data set into the improved YOLOv5 model for training to obtain the trained improved YOLOv5 model; deploying the improved YOLOv5 model, that is, converting the trained improved YOLOv5 model into a format suitable for deployment, using TensorFlow Lite, ONNX and other formats, and optimizing the trained improved YOLOv5 model, using int8 , FP16, FP32 multiple quantization methods and branch reduction operations, and finally deploy the model to the target platform, using multiple platforms such as servers, edge devices, and mobile devices; In the third part, the real-time signal frequency graph is input into the deployed improved YOLOv5 model for real-time detection and recognition, using image or video input; In the fourth part, the recognition results of the improved YOLOv5 model are converted into multiple real-time parameters of the communication signal, combining the characteristics of the system operation time and the signal time-frequency diagram. This includes converting the positioning box of the communication signal in the time-frequency diagram into duration, bandwidth, carrier frequency, and existence parameters, and converting the target type of the signal into the type of modulation parameter. At the same time, multiple parameters are sorted to form a detection and recognition report that is more in line with the public's reading habits.

3. The communication signal detection and recognition method based on improved YOLOv5 according to claim 2, characterized in that The targeted scale feature fusion module uses a targeted scale feature fusion module based on BiFPN; The targeted scale feature fusion module based on BiFPN blocks the path of fusion of small-scale features to large-scale features, and large-scale features are only fused with moderate-scale features; The targeted scale feature fusion module focuses on the moderate scale and adopts the bidirectional structure in BiFPN to fuse features from the moderate scale upward and downward respectively to realize large-scale and small-scale feature extraction; at the same time, an additional path is added on the moderate scale to fuse the three scale features again and extract the final moderate-scale feature.

4. The communication signal detection and recognition method based on improved YOLOv5 according to claim 2, characterized in that The feature attention module is constructed based on the SimAM attention mechanism; the feature attention module based on the SimAM attention mechanism realizes the weight allocation of each neuron through separability and activation function sigmoid, completes the spatial and channel attention allocation, and adds the fusion of features of different scales in YOLOv5 and the feature extraction of different scales in the targeted scale feature fusion module based on BiFPN.

5. The communication signal detection and recognition method based on improved YOLOv5 according to claim 1, characterized in that The preprocessing of the communication signal time-frequency diagram includes: adding noise, changing brightness, cropping, translating, rotating and scaling the data set; and adjusting the duration range, bandwidth range, sampling frequency and power display range of the time-frequency diagram display.

6. The communication signal detection and recognition method based on improved YOLOv5 according to claim 1, characterized in that The labeling according to the parameters of the communication signal, such as duration, bandwidth, carrier frequency, existence, and modulation mode, includes: labeling each image in the training set using a graphic labeling tool, and automatically generating a labeling text file according to rules using a code.

Images