A data enhancement method, device and equipment in signal machine learning

Abstract

The present application discloses a data enhancement method, device and equipment in signal machine learning, the method comprising: automatically collecting original signals, performing data enhancement on the collected original signals to obtain enhanced signals, sending the enhanced signals to a neural network for training, and obtaining a signal recognition model. In data enhancement, the original signal is subjected to channelization transformation, signal-to-interference ratio transformation, signal-to-noise ratio transformation, and carrier frequency deviation transformation to obtain an enhanced signal. The present application improves the diversity of training signals through data enhancement technology to cover various possible scenarios, reduces the complexity of signal acquisition, and shortens the development time to support signal detection.

Description

Technical Field

[0001] The present application relates to the fields of artificial intelligence and signal processing, and specifically to a method, device and equipment for data enhancement in signal machine learning in signal recognition. Background Art

[0002] Signal detection, such as communication signal detection and wireless communication signal detection and identification, has always attracted much attention. The reason is that, on the one hand, wireless communication technology has developed rapidly, and there are more and more wireless communication products, such as mobile phones, remote control cars, remote control aircraft, engineering communication equipment, civil and even military drones, etc. Due to the needs of control and safety, the importance of signal detection is becoming increasingly obvious.

[0003] Signal detection means that when a target object appears within the search range, the appearance of the target object can be detected in various ways. Currently, the commonly used detection methods include physical detection based on radar reflection waves, visual detection based on cameras, auditory detection based on sound, and detection based on radio signals.

[0004] With the development of artificial intelligence, the role of machine learning in image recognition and signal recognition has gradually become more prominent in recent years. Machine recognition mainly extracts the features of input signals by training neural networks. Figure 1 This is a simple diagram of machine learning, where circles represent values ​​and arrows represent operations. The essence of the neural network training process is to find the optimal operation parameters (i.e. Figure 1 The present invention is based on machine learning technology, generates an optimal neural network by training the neural network, and then detects and identifies the target through the network.

[0005] In terms of detection based on communication signals, a variety of methods have been developed. For example, since communication signals need to be modulated before they can be sent, the common method of signal detection is automatic modulation recognition, such as frequency hopping, amplitude modulation, phase modulation, etc. In order to improve the anti-noise performance, a high-order statistical method has also been developed to detect based on the high-order cumulant or cyclic spectrum of the signal, but this method is limited by the selection of features and the determination of the decision threshold in the identification of complex multi-class modulated signals. In addition, a signal recognition method based on data enhancement has emerged. Patent CN201910930192.9 relates to a method for identifying a target remote control signal based on frequency domain characteristics, which performs mixing, phase detection, sampling, and channelization on the received signal to obtain the frequency measurement output of multiple channels, thereby realizing the detection and identification of communication signals.

[0006] However, in the process of training the neural network of signals, the collection scenarios of the collected signals are often limited. The model obtained after training with the signals collected in limited scenarios is difficult to accurately identify the signals in more other scenarios, that is, the generalization ability of the obtained training model is poor. For example, a model trained with signals collected in one scenario can correctly identify subsequent signals in the same scenario, but may not be able to identify the same signals in other scenarios. Most of the current solutions are to manually collect signals in as many scenarios as possible through scene selection and setting. This greatly prolongs the research and development cycle of signal recognition devices and equipment, and cannot solve the problem of low recognition success rate in new scenarios. Summary of the invention

[0007] The purpose of the present invention is to overcome the deficiencies of the prior art and provide a data enhancement method, device and equipment for signal machine learning in signal recognition, which does not require the actual provision of signals in many scenarios, but only requires the collection of signals in some conventional scenarios, and then obtains signals in a variety of possible scenarios through signal transformation enhancement, and sends them to a neural network for training to obtain a signal recognition model. The advantage of the present invention is that it avoids the artificial collection of signals of various scenarios, reduces the complexity of signal collection, and improves the generalization ability of the model.

[0008] Machine learning mainly extracts the features of input signals by training neural networks. The present invention uses data enhancement technology to obtain target signals in various scenarios based on target signals in conventional scenarios, generates an optimized neural network by training the neural network, and then uses the network to detect and identify targets.

[0009] The specific technical solutions of the present invention are as follows:

[0010] In a first aspect, the present application provides a signal detection method based on machine learning, the method comprising:

[0011] Step 1: Use signal acquisition equipment to automatically collect the original signal.

[0012] Step 2: Use the data enhancement module to transform and enhance the collected signals, and use the transformed and enhanced data as input for neural network training.

[0013] Step 3: Use a neural network module to train the transformed and enhanced data to obtain a signal recognition model, and send the trained signal recognition model to the signal recognition module.

[0014] Step 4: Use the signal recognition module to recognize the received signal to be recognized and output the target signal recognition result.

[0015] In step 1, the signals collected by the signal collection device include communication signals and environmental signals.

[0016] In the step 1, after the signal is collected, the collected signal is added to the signal library.

[0017] In the step 2, data enhancement is performed on the collected signal, including channelization transformation, signal-to-interference ratio transformation, signal-to-noise ratio transformation, and carrier frequency deviation transformation performed in sequence according to data enhancement parameters.

[0018] A wireless channel refers to a data signal transmission channel that uses wireless signals as the transmission medium. There are usually multiple wireless channels for communication between the target object and the control receiving end. Channelization conversion is to convert the wireless signal on a certain channel into a signal on any other channel.

[0019] The channelization transformation is specifically as follows: Assume that the original signal is {x[0], x[1], x[2], ..., x[N-1]}, where N is the length of the wireless signal. Assume that the channel frequency of the original signal is f0, and the set of all channel frequencies of the signal is {f0, f1, ..., f M}, where M is the total number of channels. Assume that the signal frequency point to be transformed in the data enhancement parameter is f m , the sampling frequency is f s , then the signal is transformed into:

[0020]

[0021] The signal-to-interference ratio conversion is to simulate the actual situation that the receiving end in the wireless communication system receives both the useful signal and the interference signal when receiving the signal. The signal-to-interference ratio conversion is specifically:

[0022] Assume that the required signal-to-interference ratio is α, and the interference signal is {I[0], I[1], I[2], ..., I[N-1]}, then

[0023]

[0024] in:

[0025] Then, the signal is transformed into:

[0026]

[0027] The signal-to-noise ratio conversion is to simulate the attenuation effect of different degrees of attenuation due to the transmission distance when the wireless signal reaches the receiving end. The signal-to-noise ratio conversion is specifically:

[0028] Assume the noise power is P noise, the required signal-to-noise ratio is β, then:

[0029]

[0030] in:

[0031] Then, the signal is transformed into:

[0032]

[0033] in, The power is P noise is a circularly symmetric zero-mean complex Gaussian random variable.

[0034] The carrier frequency deviation transformation is to simulate the carrier frequency deviation δf between the receiving end and the transmitting end due to the deviation of the local oscillator of the transmitting and receiving end and the Doppler frequency shift in the wireless communication system. The carrier frequency deviation has a range [δf min ,δf max ]. The carrier frequency deviation transformation is specifically as follows:

[0035] A deviation δf is randomly selected within the deviation range, and then the signal is transformed according to the variation as follows:

[0036]

[0037] In step 4, the signal recognition module stores a signal recognition model that can be used to identify target signals in multiple scenarios, that is, the signal recognition model trained in step 3. When a target communication signal is received, the signal recognition module will identify the received communication signal through the stored signal recognition model and output the communication signal recognition result.

[0038] In a second aspect, the present application provides a signal detection device based on machine learning, the device comprising:

[0039] The signal acquisition module automatically collects the original signal.

[0040] The signal library is connected to the signal acquisition module and the data enhancement module, and is used to store the collected original signals and send the stored signals to the data enhancement module.

[0041] The data enhancement module transforms and enhances the collected original signals, and uses the transformed and enhanced data as the input for neural network training.

[0042] The neural network module is connected with the data enhancement module and is used to train the transformed and enhanced signals to obtain a signal recognition model.

[0043] The signal recognition module is used to recognize the signal. The neural network module also sends the trained signal recognition model to the signal recognition module to enable the signal recognition module to have the ability to recognize the target signal. The signal recognition module stores the signal recognition model trained by the neural network module, and recognizes the collected signal to be distinguished through the signal recognition model.

[0044] The original signal of the target object collected by the signal collection module includes communication signals and environmental signals. The steps of the automatic collection method are the same as those described above and will not be repeated here.

[0045] The data enhancement module performs data enhancement on the collected signal, and includes a channelization transformation module, a signal-to-interference ratio transformation module, a signal-to-noise ratio transformation module, and a carrier frequency deviation transformation module which are arranged in sequence.

[0046] The channelization transformation module, signal-to-interference ratio transformation module, signal-to-noise ratio transformation module, and carrier frequency deviation transformation module respectively implement the channelization transformation, signal-to-interference ratio transformation, signal-to-noise ratio transformation, and carrier frequency deviation transformation as described above, which will not be repeated here.

[0047] The signal recognition module stores a signal recognition model that can be used to recognize target signals in various scenarios. When a communication signal is received, the signal recognition module will recognize the received communication signal through the stored signal recognition model and output the communication signal recognition result.

[0048] In a third aspect, the present application also provides a signal detection device including a data enhancement device in signal machine learning, wherein the device includes a memory and a processor.

[0049] The memory is used to store program codes and transmit the program codes to the processor.

[0050] The processor is used to execute the data enhancement method and signal detection method in signal machine learning described in the first aspect according to the instructions in the program code.

[0051] In a fourth aspect, the present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the data enhancement method in signal machine learning and the signal detection method described in the first aspect.

[0052] The beneficial effects of this application are:

[0053] 1. This application can identify target signals in various scenarios, avoids the manual collection of signals in various scenarios, reduces the complexity of signal collection, improves the generalization ability of the model, greatly improves the success rate of signal recognition in new scenarios, and shortens the development time of supporting signal detection.

[0054] 2. In the data enhancement link, the present application sequentially performs channelization transformation, signal-to-interference ratio transformation, signal-to-noise ratio transformation, and carrier frequency deviation transformation on the original signal. The signal recognition model obtained by training the data after the channelization transformation can accurately identify the communication signal of the constantly changing communication channel. The signal-to-interference ratio transformation is used for data enhancement, so that the signal recognition model can recognize the communication signals under various interferences. The signal-to-noise ratio transformation is used for data enhancement, so that the signal recognition model can recognize the communication signals after different degrees of attenuation. The carrier frequency deviation transformation is used for data enhancement, so that the signal recognition model can recognize the communication signals under different carrier frequency deviations. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0056] Figure 1 is a schematic diagram of the structure of a neural network model in a signal detection device based on machine learning;

[0057] Figure 2 A flowchart of a signal detection method based on machine learning provided in an embodiment of the present application;

[0058] Figure 3 A schematic diagram of the structure of a data enhancement device in signal machine learning provided in an embodiment of the present application;

[0059] Figure 4 A structural schematic diagram of a signal detection device based on machine learning provided in an embodiment of the present application. DETAILED DESCRIPTION

[0060] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0061] In the current training process of the neural network based on radio signals, the actual scenarios in the signal collection process are often limited. The model obtained after training with the target signals collected in limited scenarios is difficult to accurately identify the target signals in more scenarios, and the generalization ability of the obtained training model is poor. The existing solution is to manually collect target signals in as many scenarios as possible through scene selection and setting. This greatly prolongs the R&D cycle of recognition devices and equipment, and cannot solve the problem of low recognition success rate in new scenarios.

[0062] To this end, the present application provides a data enhancement method for signal machine learning in signal recognition, which only requires collecting target signals in some conventional scenarios, and then obtaining target signals in multiple possible scenarios through signal transformation enhancement, so that the signal recognition model obtained after training can recognize target signals in complex scenarios.

[0063] The following specifically introduces an embodiment of a data enhancement method in signal machine learning and a signal detection method provided by the present application, with reference to Figure 2 , is a flow chart of a method provided in an embodiment of the present application, the method specifically includes:

[0064] S101: Automatically collect signals using signal collection equipment.

[0065] In the embodiment of the present application, the signal collected by the signal collection device includes the communication signal used when the target object communicates with the ground and the control end, and also includes the environmental signal. In addition, after the signal is collected, the collected signal is added to the signal library.

[0066] S102: using the data enhancement module to transform and enhance the collected original signal according to the data enhancement parameters, and using the transformed and enhanced data as input for neural network training.

[0067] S103: Using a neural network module to train the transformed and enhanced signal to obtain a signal recognition model, and sending the trained signal recognition model to a signal recognition module.

[0068] S104: using the signal recognition module to recognize the collected signal to be identified, and outputting the signal recognition result.

[0069] In S102, data enhancement is performed on the collected signal, including channelization transformation, signal-to-interference ratio transformation, signal-to-noise ratio transformation, and carrier frequency deviation transformation performed in sequence.

[0070] A wireless channel refers to a data signal transmission channel that uses wireless signals as the transmission medium. There are usually multiple wireless channels for communication between the target object and the control receiving end. Channelization conversion is to convert the wireless signal on a certain channel into a signal on any other channel.

[0071] The channelization transformation is specifically as follows: Assume that the original signal is {x[0], x[1], x[2], ..., x[N-1]}, where N is the length of the wireless signal. Assume that the channel frequency of the original signal is f0, and the set of all channel frequencies of the signal is {f0, f1, ..., f M}, where M is the total number of channels. Assume that the signal frequency point to be transformed in the data enhancement parameter is f m , the sampling frequency is f s , then the signal is transformed into:

[0072]

[0073] The signal-to-interference ratio conversion is to simulate the actual situation that the receiving end in the wireless communication system receives both the useful signal and the interference signal when receiving the signal. The signal-to-interference ratio conversion is specifically:

[0074] Assume that the required signal-to-interference ratio is α, and the interference signal is {I[0], I[1], I[2], ..., I[N-1]}, then

[0075]

[0076] in:

[0077] Then, the signal is transformed into:

[0078]

[0079] The signal-to-noise ratio conversion is to simulate the attenuation effect of different degrees of attenuation due to the transmission distance when the wireless signal reaches the receiving end. The signal-to-noise ratio conversion is specifically:

[0080] Assume the noise power is P noise , the required signal-to-noise ratio is β, then:

[0081]

[0082] in:

[0083] Then, the signal is transformed into:

[0084]

[0085] in, The power is Pnoise is a circularly symmetric zero-mean complex Gaussian random variable.

[0086] The carrier frequency deviation transformation is to simulate the carrier frequency deviation δf between the receiving end and the transmitting end due to the deviation of the local oscillator of the transmitting and receiving end and the Doppler frequency shift in the wireless communication system. The carrier frequency deviation has a range [δf min ,δf max ]. The carrier frequency deviation transformation is specifically as follows:

[0087] A deviation δf is randomly selected within the deviation range, and then the signal is transformed according to the variation as follows:

[0088]

[0089] In an embodiment of the present application, the signal recognition module stores a signal recognition model that can be used to identify target signals in multiple scenarios, that is, the signal recognition model obtained by training in S103. When a communication signal to be identified is received, the signal recognition module will identify the received communication signal through the stored signal recognition model and output the target signal identification result.

[0090] Corresponding to the above method embodiments, the present application also provides a data enhancement device in signal machine learning, the device comprising:

[0091] The signal acquisition module automatically collects the original signal.

[0092] The signal library is connected to the signal acquisition module and the data enhancement module, and is used to store the collected original signals and send the stored signals to the data enhancement module.

[0093] The data enhancement module transforms and enhances the collected original signals, and uses the transformed and enhanced data as the input for neural network training.

[0094] The neural network module is connected with the data enhancement module and is used to train the transformed and enhanced signals to obtain a signal recognition model.

[0095] The signal recognition module is used to recognize the signal. The neural network module also sends the trained signal recognition model to the signal recognition module to enable the signal recognition module to have the ability to recognize the target signal.

[0096] The signal recognition module stores a signal recognition model obtained by training the neural network module, and the collected signal to be distinguished is recognized through the signal recognition model.

[0097] The original signal of the target object collected by the signal collection module includes communication signals and environmental signals. The steps of the automatic collection method are the same as those described above and will not be repeated here.

[0098] The data enhancement module performs data enhancement on the collected signal, including a channelization conversion module, a signal-to-interference ratio conversion module, a signal-to-noise ratio conversion module, and a carrier frequency deviation conversion module which are arranged in sequence. Figure 3 described.

[0099] The channelization transformation module, signal-to-interference ratio transformation module, signal-to-noise ratio transformation module, and carrier frequency deviation transformation module respectively implement the channelization transformation, signal-to-interference ratio transformation, signal-to-noise ratio transformation, and carrier frequency deviation transformation as described above based on the data enhancement parameters, which will not be repeated here.

[0100] The signal recognition module stores a signal recognition model that can be used to recognize target signals in various scenarios. When a communication signal to be recognized is received, the signal recognition module will recognize the received communication signal through the stored signal recognition model and output the target signal recognition result.

[0101] The embodiment of the present invention also provides a signal detection device based on machine learning, see Figure 4 As shown, it may include:

[0102] Processor 401, memory 402, input device 403 and output device 404. The number of processors 401 in the signal detection device based on machine learning can be one or more. Figure 4 In some embodiments of the present invention, the processor 401, the memory 402, the input device 403 and the output device 404 may be connected via a bus or other means, wherein: Figure 4 The example of connecting through bus is taken in the following.

[0103] The memory 402 can be used to store software programs and modules. The processor 401 executes various functional applications and data processing of the signal detection device based on machine learning by running the software programs and modules stored in the memory 402. The memory 402 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, at least one application required for a function, etc. In addition, the memory 402 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage device. The input device 403 may be used to receive input digital or character information, and to generate signal input related to user settings and function control of the signal detection device based on machine learning.

[0104] Specifically in this embodiment, the processor 401 will load the executable files corresponding to the processes of one or more applications into the memory 402 according to the following instructions, and the processor 401 will run the applications stored in the memory 402, thereby realizing the various functions of the above-mentioned signal detection method based on machine learning.

[0105] The detection method, detection device, signal enhancement device, etc. disclosed in this application can all be used for drone detection.

[0106] For the device embodiment, since it basically corresponds to the method embodiment, the relevant parts can refer to the partial description of the method embodiment. The device embodiment described above is only schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative work.

[0107] It should be noted that, in this article, the terms "comprises", "includes" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or equipment that includes a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or equipment.

[0108] The above is a detailed introduction to the signal detection and recognition method, device and equipment based on machine learning provided in the embodiments of the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for general technical personnel in this field, according to the ideas of the present application, there will be changes in the specific implementation methods and application scopes. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A data enhancement method in signal machine learning, the method is used for signal detection, characterized in that: The method comprises: Automatically collect the original signal; Perform data enhancement on the collected original signal and use the enhanced data as input for neural network training; The data enhancement specifically includes: performing channelization transformation, signal-to-interference ratio transformation, signal-to-noise ratio transformation, and carrier frequency deviation transformation in sequence; Sending the enhanced data to a neural network for training to obtain a signal recognition model, and sending the trained signal recognition model to a signal recognition module; and The collected signal to be distinguished is sent to the signal recognition module to determine whether there is a target signal in the signal and output the recognition result; in, The channelization transformation comprises: The original signal collected is {x[0], x[1], x[2], ..., x[N-1]}, where N is the signal length, the channel frequency of the original signal is f0, and the set of all channel frequencies of the signal is {f0, f1, ..., f M }, where M is the total number of channels, and the signal frequency point that needs to be transformed in the data enhancement parameter is f m , the sampling frequency is f s , the channelization transform transforms the original signal x[n] into y1[n]: The signal-to-interference ratio transformation includes: The channelization transformation obtains y1[n], the target signal-to-interference ratio of the signal-to-interference ratio transformation is α, and the interference signal is {I[0], I[1], I[2], ..., I[N-1]}. The signal-to-interference ratio transformation transforms the signal y1[n] into y2[n]: in, The signal-to-noise ratio transformation comprises: The signal-to-interference ratio is transformed to obtain the signal y2[n], and the power of the noise is P noise , the target signal-to-noise ratio is β, and the signal-to-noise ratio transformation transforms the signal y2[n] into y3[n]: in, The power is P noise A cyclically symmetric zero-mean complex Gaussian random variable; The carrier frequency deviation transformation includes: The signal-to-noise ratio transformation obtains the signal y3[n], and the carrier frequency deviation is δf. The carrier frequency deviation transformation transforms the signal y3[n] into y4[n]: Among them, f s is the sampling frequency.

2. The method according to claim 1, characterized in that The original signal includes a communication signal.

3. A signal detection device based on machine learning, characterized in that: The device comprises: Signal acquisition module, automatically collects the original signal; The signal library is connected to the signal acquisition module and the data enhancement module, and is used to store the collected original signals and send the stored signals to the data enhancement module; The data enhancement module transforms and enhances the collected original signals, and uses the transformed and enhanced data as the input for neural network training; The data enhancement specifically includes: performing channelization transformation, signal-to-interference ratio transformation, signal-to-noise ratio transformation, and carrier frequency deviation transformation in sequence; A neural network module, connected to the data enhancement module, for training the transformed and enhanced signal to obtain a signal recognition model; and A signal recognition module stores a signal recognition model obtained by training the neural network module and is used to recognize the target signal; in, The channelization transformation comprises: The original signal collected is {x[0], x[1], x[2], ..., x[N-1]}, where N is the signal length, the channel frequency of the original signal is f0, and the set of all channel frequencies of the signal is {f0, f1, ..., f M }, where M is the total number of channels, and the signal frequency point that needs to be transformed in the data enhancement parameter is f m , the sampling frequency is f s , the channelization transform transforms the original signal x[n] into y1[n]: The signal-to-interference ratio transformation includes: The channelization transformation obtains y1[n], the target signal-to-interference ratio of the signal-to-interference ratio transformation is α, and the interference signal is {I[0], I[1], I[2], ..., I[N-1]}. The signal-to-interference ratio transformation transforms the signal y1[n] into y2[n]: in, The signal-to-noise ratio transformation comprises: The signal-to-interference ratio is transformed to obtain the signal y2[n], and the power of the noise is P noise , the target signal-to-noise ratio is β, and the signal-to-noise ratio transformation transforms the signal y2[n] into y3[n]: in, The power is P noise A cyclically symmetric zero-mean complex Gaussian random variable; The carrier frequency deviation transformation includes: The signal-to-noise ratio transformation obtains the signal y3[n], and the carrier frequency deviation is δf. The carrier frequency deviation transformation transforms the signal y3[n] into y4[n]: Among them, f s is the sampling frequency.

4. A data enhancement device in signal machine learning, characterized in that: The device comprises: The data enhancement module transforms and enhances the data, and uses the transformed and enhanced data as the input for neural network training; The data enhancement module includes a channelization transformation module, a signal-to-interference ratio transformation module, a signal-to-noise ratio transformation module, and a carrier frequency deviation transformation module which are arranged in sequence; in, The channelization transformation comprises: The original signal collected is {x[0], x[1], x[2], ..., x[N-1]}, where N is the signal length, the channel frequency of the original signal is f0, and the set of all channel frequencies of the signal is {f0, f1, ..., f M }, where M is the total number of channels, and the signal frequency point that needs to be transformed in the data enhancement parameter is f m , the sampling frequency is f s , the channelization transform transforms the original signal x[n] into y1[n]: The signal-to-interference ratio transformation includes: The channelization transformation obtains y1[n], the target signal-to-interference ratio of the signal-to-interference ratio transformation is α, and the interference signal is {I[0], I[1], I[2], ..., I[N-1]}. The signal-to-interference ratio transformation transforms the signal y1[n] into y2[n]: in, The signal-to-noise ratio transformation comprises: The signal-to-interference ratio is transformed to obtain the signal y2[n], and the power of the noise is P noise , the target signal-to-noise ratio is β, and the signal-to-noise ratio transformation transforms the signal y2[n] into y3[n]: in, The power is P noise A cyclically symmetric zero-mean complex Gaussian random variable; The carrier frequency deviation transformation includes: The signal-to-noise ratio transformation obtains the signal y3[n], and the carrier frequency deviation is δf. The carrier frequency deviation transformation transforms the signal y3[n] into y4[n]: Among them, f s is the sampling frequency.

5. A signal detection device based on machine learning, characterized in that: The device comprises a memory and a processor; The memory is used to store program code and transmit the program code to the processor; The processor is configured to execute the method according to any one of claims 1 to 2 according to the instructions in the program code.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 2 are implemented.

Images