Electromagnetic signal prediction method and device of multi-domain transform joint optimization transformer

Abstract

The invention relates to the technical field of electromagnetic method signal processing, and particularly discloses an electromagnetic signal prediction method and device for multi-domain transformation combined optimization transformer, and the method comprises the following steps: carrying out the multi-domain analysis through electromagnetic data, and extracting feature parameters under multi-domain transformation; carrying out model hyper-parameter optimization by using an improved goose optimization algorithm; the multi-domain characteristic parameters and the signal samples are input into an optimization transformer for training, and an optimization model is obtained; electromagnetic data are tested and verified through the optimization model, and the model is applied to electromagnetic noise identification and signal prediction; the signals are reserved, identified noise is removed, the signals are predicted, and high-quality electromagnetic method data are obtained through integration and reconstruction according to the original sampling sequence. According to the method, the defects of a deep learning method in electromagnetic multi-domain data processing are overcome, the electromagnetic method signal processing effect is improved, and the low-frequency-band electromagnetic data quality is improved.

Description

Technical Field

[0001] The present application relates to the field of information technology and intelligent optimization deep learning, and specifically discloses an electromagnetic signal prediction method and device for a multi-domain transformation joint optimization transformer. Background Art

[0002] With the development of new-generation information technology, intelligent optimization and deep learning technologies have moved towards practical engineering applications.

[0003] Among them, the optimization method is an applied technology based on mathematics and used to solve various optimization problems; the intelligent optimization algorithm is a computational method that simulates biological behavior or physical phenomena in nature to solve complex optimization problems. It has achieved remarkable results in practical applications such as mechanical design, path optimization problems, electrical system optimization, and parameter optimization of machine learning models.

[0004] As a practical intelligent optimization method, deep learning is gradually developing towards large models. Transformer, as a deep learning model architecture for natural language processing and other sequence-to-sequence tasks, introduces a self-attention mechanism, which enables it to perform well when processing sequence data.

[0005] In the field of electromagnetic detection, noise limits the effect of electromagnetic detection and reduces the quality of electromagnetic data. Deep learning is applied to the denoising process in the field of electromagnetic detection. Although deep learning can analyze the noise type and influencing factors from different angles, the denoising effect is still difficult to evaluate and its application is limited, and it is impossible to finely improve the data quality.

[0006] Therefore, since the Transformer model was proposed, it has become an important cornerstone of natural language processing and other fields. In view of the Transformer model, the inventor provides an electromagnetic signal prediction method and device for jointly optimizing the Transformer by multi-domain transformation. Summary of the invention

[0007] The purpose of the present invention is to improve the goose optimization algorithm by using multiple strategies, and then use the improved goose optimization algorithm to search for the hyperparameters of the optimal transformer model, and then use the intelligent optimization model to suppress the noise of electromagnetic data, and realize the fine processing of electromagnetic data by intelligent optimization deep learning model, in order to solve the shortcomings of machine learning methods in electromagnetic denoising.

[0008] In order to achieve the above object, the basic scheme of the present invention provides an electromagnetic signal prediction method of a multi-domain transformation joint optimization transformer, comprising the following steps:

[0009] Step S1: Perform multi-domain analysis using electromagnetic data to extract characteristic parameters under multi-domain transformation;

[0010] Step S2: using chaotic mapping, adaptive weight adjustment and memory strategy to improve the goose optimization algorithm, and using the improved goose optimization algorithm to optimize the hyperparameters of the transformer model;

[0011] Step S3: inputting the feature parameters and signal samples under multi-domain transformation into the optimized transformer for training to obtain the optimized transformer model;

[0012] Step S4: Use the optimized transformer model to test and verify the electromagnetic data, and apply the model to electromagnetic noise identification and signal prediction;

[0013] Step S5: retain the signal, remove the identified noise and predict the signal, and integrate and reconstruct it in the original sampling order to obtain high-quality electromagnetic data.

[0014] Further, in step S1, the multi-domain transform uses Fourier transform, short-time Fourier transform and continuous wavelet transform to perform multi-domain analysis and extract multi-domain feature parameters. The multi-domain features are as follows:

[0015] y t =[Y 时域特征 ,Y 频域特征 ,Y 时频域特征 ] t

[0016] Where Y 时域特征 ,Y 频域特征 ,Y 时频域特征 Respectively represent the input multi-domain feature vector set, y t Represents the total set of multi-domain features.

[0017] Further, in step S2, the improved goose optimization algorithm uses chaotic mapping, adaptive weight adjustment and memory strategy to improve the algorithm, and uses the initialization particle x n The population chaos map calculation is as follows:

[0018] x n+1 =r×(sin(Π×x n )+cos(Π×x n ))

[0019] In the formula, the control parameter r adjusts the degree of chaos of the system;

[0020] The adaptive weight adjustment is calculated as follows:

[0021] a=2×exp[-6×(t÷Max_IT)]

[0022] In the formula, the adaptive parameter a is dynamically adjusted according to the current number of iterations t and the total number of iterations Max_IT;

[0023] The memory strategy is calculated as follows:

[0024] M is the array storing the best historical solutions, M_scores is used to store the fitness scores of the corresponding solutions, and for the newly discovered solution Best_pos and its fitness score Best_score, the update process is as follows:

[0025] Find the maximum value in M_scores and its index idx:

[0026] max_score,idx=max(M_scores)

[0027] If the best range Best_score is less than the maximum range max_score, then update:

[0028] M_[idx]=Best_pos

[0029] M_scores[idx]=Best_score

[0030] Checking memory: When the performance of the algorithm stagnates, that is, no better solution is found after multiple consecutive iterations, you can try to jump out of the local optimum by checking the solution in memory. The process of checking memory and selecting the optimal solution is expressed as: Find the minimum value and its index in M_scores:

[0031] Best_score,idx=min(M_scores)

[0032] Return the corresponding solution:

[0033] Best_pos=M_[idx].

[0034] Further, in step S2, in step S2, the improved goose optimizes the transformer model hyperparameters, namely, the number of attention heads, learning rate and batch size of the transformer model.

[0035] Further, in step S3, the transformer model consists of an encoder, a decoder, and an output layer;

[0036] The encoder consists of a multi-head attention mechanism and a fully connected neural network, the decoder consists of two multi-head attention mechanisms and a fully connected neural network, and the output layer undergoes a linear transformation and Softmax to obtain the output probability distribution.

[0037] Further, in step S4 and step S5, the electromagnetic data testing and prediction using the optimized transformer model are as follows:

[0038]

[0039] In the formula, x, y, xy represent the multi-domain feature data segments, j represents the number of segments, represents the prediction process, Y j+1 Represents the predicted value.

[0040] The basic scheme of the present invention also provides an electromagnetic signal prediction device for a multi-domain transform jointly optimized transformer, comprising a memory, a control processor, and a computer program stored in the memory and executable on the control processor, wherein the control processor executes the program to implement the electromagnetic signal prediction method for a multi-domain transform jointly optimized transformer as described above.

[0041] The basic scheme of the present invention also provides a control system, including the electromagnetic signal prediction device of the aforementioned multi-domain transformation joint optimization transformer.

[0042] The basic scheme of the present invention also provides a computer-readable storage medium, which stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute the electromagnetic signal prediction method of the multi-domain transformation joint optimization transformer as mentioned above.

[0043] The principle and effect of this scheme are:

[0044] The present invention is implemented in sequence through steps to achieve the improvement of intelligent optimization and the hyperparameter optimization of the transformer model, improve the difficulties in electromagnetic denoising, solve the problem of poor denoising effect caused by inappropriate parameter selection in machine learning and deep learning algorithms in processing electromagnetic data, and enhance the application prospect of the improved intelligent optimization transformer model in electromagnetic signal processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] 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 those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0046] Figure 1 A flowchart of an electromagnetic signal prediction method of a multi-domain transformation joint optimization transformer proposed in an embodiment of the present application is shown;

[0047] Figure 2 A schematic diagram of the structure of a transformer model proposed in an embodiment of the present application is shown;

[0048] Figure 3 A schematic diagram showing the feature input and optimization process proposed in the embodiment of the present application is shown;

[0049] Figure 4 The figure shows the effect of processing the analog signal proposed in the embodiment of the present application by the method;

[0050] Figure 5 The figure shows the recognition effect of the measured data proposed in the embodiment of the present application through the method;

[0051] Figure 6 A comparison diagram of the measured point curve before and after processing proposed in the embodiment of the present application is shown. DETAILED DESCRIPTION

[0052] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation mode, structure, characteristics and effects of the present invention are described in detail below in combination with the accompanying drawings and preferred embodiments.

[0053] A method for predicting electromagnetic signals by jointly optimizing transformers using multi-domain transformations, as shown in Example 1 Figure 1 As shown:

[0054] The steps include:

[0055] Step S1: Perform multi-domain analysis using electromagnetic data to extract characteristic parameters under multi-domain transformation;

[0056] In this step, the electromagnetic data is subjected to multi-domain analysis including the time domain, frequency domain and time-frequency domain, and the characteristic parameters extracted under the multi-domain transformation include the morphology, energy, amplitude, spectrum, time-frequency, polarization direction, wavelet scale and other characteristics;

[0057] Step S2: using chaotic mapping, adaptive weight adjustment and memory strategy to improve the goose optimization algorithm, and using the improved goose optimization algorithm to optimize the hyperparameters of the transformer model;

[0058] like Figure 3 As shown in Figure 1, in this step, the goose optimization algorithm is improved by using chaotic mapping, adaptive weight adjustment and memory strategy. The improved goose optimizes the transformer model hyperparameters, namely, the number of attention heads, learning rate and batch size of the transformer model.

[0059] Using the initialized particle x n The population chaos map calculation is as follows:

[0060] x n+1 =r×(sin(Π×x n )+cos(Π×x n ))

[0061] Wherein, the control parameter r adjusts the degree of chaos of the system.

[0062] The adaptive weight adjustment is calculated as follows:

[0063] a=2×exp[-6×(t÷Max_IT)]

[0064] Wherein, the adaptive parameter a is dynamically adjusted according to the current number of iterations t and the total number of iterations Max_IT to ensure that the search behavior gradually shifts from extensive exploration to concentrated development.

[0065] The memory strategy is calculated as follows:

[0066] M is the array storing the best historical solutions, M_scores is used to store the fitness scores of the corresponding solutions, and for the newly discovered solution Best_pos and its fitness score Best_score, the update process is as follows:

[0067] Find the maximum value in M_scores and its index idx:

[0068] max_score,idx=max(M_scores)

[0069] If the best range Best_score is less than the maximum range max_score, then update:

[0070] M_[idx]=Best_pos

[0071] M_scores[idx]=Best_score

[0072] Checking memory: When the performance of the algorithm stagnates, that is, no better solution is found after multiple consecutive iterations, you can try to jump out of the local optimum by checking the solution in memory. The process of checking memory and selecting the optimal solution is expressed as: Find the minimum value and its index in M_scores:

[0073] Best_score,idx=min(M_scores)

[0074] Return the corresponding solution:

[0075] Best_pos=M_[idx]

[0076] Step S3: inputting the feature parameters and signal samples under multi-domain transformation into the optimized transformer for training to obtain the optimized transformer model;

[0077] In this step, the multi-domain feature parameters and signal samples are input into the optimized transformer for training to obtain the optimized transformer model. The transformer model consists of an encoder, a decoder, and an output layer; the encoder consists of a multi-head attention mechanism and a fully connected neural network; the decoder consists of two multi-head attention mechanisms and a fully connected neural network; the output layer undergoes a linear transformation and Softmax to obtain the output probability distribution.

[0078] Step S4: Use the optimized transformer model to test and verify the electromagnetic data, and apply the model to electromagnetic noise identification and signal prediction;

[0079] In this step, the optimized transformer model is used to test and predict electromagnetic data as follows:

[0080]

[0081] In the formula, x, y, xy represent the multi-domain feature data segments, j represents the number of segments, represents the prediction process, Y j+1 Represents the predicted value.

[0082] Step S5: retain the signal, remove the identified noise and predict the signal, and integrate and reconstruct it in the original sampling order to obtain high-quality electromagnetic data.

[0083] Through the above steps, the noise identification and elimination of electromagnetic data of the improved intelligent optimization transformer model is realized, which solves the effect and efficiency of deep learning algorithm in processing electromagnetic data. The improved intelligent optimization transformer model can reduce the artificial setting of parameters, effectively improve the denoising effect of the electromagnetic method, and enhance data quality.

[0084] Example 2 is a specific implementation process of this method. Specifically,

[0085] like Figure 1 As shown, this embodiment discloses an electromagnetic signal prediction method of a multi-domain transformation joint optimization transformer, comprising the following steps:

[0086] Step S1: Perform multi-domain analysis using electromagnetic data to extract characteristic parameters under multi-domain transformation; perform time domain, frequency domain and time-frequency domain analysis on electromagnetic data to extract characteristic parameters such as morphology, energy, amplitude, spectrum, time-frequency, polarization direction, wavelet scale, etc. under multi-domain transformation;

[0087] Step S2: Using chaotic mapping, adaptive weight adjustment and memory strategy to improve the goose optimization algorithm, and using the improved goose optimization algorithm to adjust the hyperparameters of the transformer model such as the number of attention heads, learning rate and batch size; Figure 2 The figure shows the structure of the transformer model, which includes the input end, encoder, decoder and output end. The encoder is composed of a multi-head attention mechanism and a feedforward neural network, and the decoder is composed of two multi-head attention mechanisms, two feedforward neural networks, and two residual connections and normalizations.

[0088] like Figure 3 The figure shows the feature input and optimization process, including the input multi-domain data and its features, the process of the optimization algorithm and its optimization parameter process, specifically: the chaotic map population is initialized, and then the exploration stage and the development stage are entered in sequence, and then the memory strategy is updated to obtain the optimal solution;

[0089] Step S3: Input the multi-domain feature parameters and signal samples into the optimized transformer for training to obtain the optimized transformer model; use a multi-head attention mechanism and a fully connected neural network to form an encoder for data output, use two multi-head attention mechanisms and a fully connected neural network to form a decoder for data transformation, and use the output layer to obtain the output probability distribution after a linear transformation and Softmax to output the result; Figure 2 and Figure 3 As shown;

[0090] Step S4: Use the optimized transformer model to test and verify the electromagnetic data, and apply the model to electromagnetic noise identification and signal prediction;

[0091]

[0092] In the formula, x,y,xy They represent multi-domain feature data segments, j represents the number of segments, represents the prediction process, Y j+1 Represents the predicted value.

[0093] Step S5: retain the signal, remove the identified noise and predict the signal, and integrate and reconstruct it in the original sampling order to obtain high-quality electromagnetic data.

[0094] In order to verify the effectiveness of the effect of this embodiment, the method of the present invention is processed in simulated noisy signals and measured data. Figure 4 The above is a diagram showing the processing effect of the simulated noisy signal according to the embodiment of the present application by the method of the present invention; Figure 5It can be seen that the original data contains typical noise types. After being processed by the method of the present invention, different types of noise can be effectively identified, and it also shows that the effective signal is retained with high precision. Comprehensive analysis Figure 4 and Figure 5 It can be shown that the method of this embodiment can effectively realize intelligent signal-noise recognition and denoising processing of electromagnetic data, thereby improving the quality of electromagnetic data.

[0095] By comparing the electric field curves before and after the actual measurement point processing. Figure 6 As shown, since the original electromagnetic data is affected by noise or abnormal waveforms, the curve shows different degrees of jumps in the low frequency band, the amplitude is unstable, and the curve quality is low; it is not difficult to find that after being processed by the method of this embodiment, the electric field curve of the measured point shows a stable trend and there is no sudden change in abnormal frequency points, indicating that the electromagnetic method data after processing is effectively improved.

[0096] This embodiment also provides an electromagnetic signal prediction device for a multi-domain transformation jointly optimized transformer, comprising at least one module, wherein at least one module is used to execute an electromagnetic signal prediction method for a multi-domain transformation jointly optimized transformer in this embodiment, and at least one module used to implement this method is an improved goose optimized transformer module.

[0097] In the above embodiment, an improved goose optimized transformer network model is set, and electromagnetic data is used for multi-domain analysis to extract characteristic parameters under multi-domain transformation; the goose optimization algorithm is improved by using chaotic mapping, adaptive weight adjustment and memory strategy to optimize transformer model hyperparameters; multi-domain characteristic parameters and signal samples are input into the optimized transformer for training to obtain the optimized transformer model; the optimized transformer model is used to test and verify electromagnetic data, and the model is applied to electromagnetic noise identification and signal prediction; the signal is retained, the identified noise is eliminated and the signal is predicted, and the original sampling order is integrated and reconstructed to obtain high-quality electromagnetic data. After the experimental results are analyzed, the present invention improves the existing machine learning algorithm in the electromagnetic signal-noise identification and processing effect, and has certain practical value and innovation.

[0098] The present invention also provides an electromagnetic signal prediction device for a multi-domain transform jointly optimized transformer, comprising a memory, a control processor, and a computer program stored in the memory and executable on the control processor, wherein the control processor executes the program to implement the aforementioned electromagnetic signal prediction method for a multi-domain transform jointly optimized transformer.

[0099] The present invention also provides a control system, including the electromagnetic signal prediction device of the aforementioned multi-domain transformation joint optimization transformer.

[0100] The electromagnetic signal prediction method of the multi-domain transformation joint optimization transformer according to the embodiment of the present disclosure can be written as a computer program and stored on a computer-readable storage medium. When the computer program is executed by a processor, the electromagnetic signal prediction method of the multi-domain transformation joint optimization transformer as described above can be implemented.

[0101] Examples of computer-readable storage media include read-only memory (ROM), random-access programmable read-only memory (PROM), electrically erasable programmable read-only memory (EEPROM), random-access memory (RAM), dynamic random-access memory (DRAM), static random-access memory (SRAM), flash memory, nonvolatile memory, CD-ROM, CD-R, CD+R, CD-RW, CD+RW, DVD-ROM, DVD-R, DVD+R, DVD-RW, DVD+RW, DVD-RAM, BD-ROM, BD-R, BD-R LTH, BD-RE, Blu-ray or optical disk storage, hard disk drive (HDD), solid state drive (SSD), card storage (such as, multimedia card, secure digital (SD) card or extreme digital (XD) card), magnetic tape, floppy disk, magneto-optical data storage device, optical data storage device, hard disk, solid state disk and any other device, any other device is configured to store computer program and any associated data, data files and data structures in a non-transitory manner and provide the computer program and any associated data, data files and data structures to a processor or computer so that the processor or computer can execute the computer program. In one example, the computer program and any associated data, data files and data structures are distributed on a networked computer system so that the computer program and any associated data, data files and data structures are stored, accessed and executed in a distributed manner by one or more processors or computers.

[0102] Although the above methods are illustrated and described as a series of actions for simplicity of explanation, it should be understood and appreciated that these methods are not limited by the order of the actions, because according to one or more embodiments, some actions may occur in different orders and / or concurrently with other actions from the illustrations and descriptions herein or not illustrated and described herein but understandable to those skilled in the art. It will be further appreciated by those skilled in the art that the various illustrative logic blocks, modules, circuits, and algorithm steps described in conjunction with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or a combination of the two. To clearly explain this interchangeability of hardware and software, various illustrative components, frames, modules, circuits, and steps are generally described in the form of their functionality above. Whether such functionality is implemented as hardware or software depends on the specific application and the design constraints imposed on the overall system. The technical staff can implement the described functionality in different ways for each specific application, but such implementation decisions should not be interpreted as resulting in a departure from the scope of the present invention. The various illustrative logic blocks, modules, and circuits described in conjunction with the embodiments disclosed herein may be implemented or executed with a general purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. A general purpose processor may be a microprocessor, but in an alternative, the processor may be any conventional processor, a battery compartment control board, a micro-battery compartment control board, or a state machine. The processor may also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, multiple microprocessors, one or more microprocessors in cooperation with a DSP core, or any other such configuration. The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. The software module may reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to a processor so that the processor can read information from the storage medium and can write information to the storage medium. In an alternative, the storage medium can be integrated into the processor. The processor and the storage medium can reside in an ASIC. The ASIC can reside in a user terminal. In an alternative, the processor and the storage medium can reside in a user terminal as discrete components. In one or more exemplary embodiments, the functions described can be implemented in hardware, software, firmware, or any combination thereof. If implemented as a computer program product in software, each function can be stored on a computer-readable medium or transmitted therefrom as one or more instructions or codes.Computer-readable media include both computer storage media and communication media, including any media that facilitates a computer program to be transferred from one place to another. Storage media can be any available media that can be accessed by a computer. As an example and not limitation, such computer-readable media may include RAM, ROM, EEPROM, CD-ROM or other optical disk storage, disk storage or other magnetic storage devices, or any other media that can be used to carry or store the agreed program code in the form of instructions or data structures and can be accessed by a computer. Any connection is also properly referred to as computer-readable media. For example, if software is transmitted from a web site, a central control computer, or other remote sources using coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwaves, then the coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwaves are included in the definition of medium. Disk and disc as used herein include compact disc (CD), laser disc, optical disc, digital versatile disc (DVD), floppy disk and Blu-ray disc, wherein disk often reproduces data magnetically, while disc reproduces data optically with lasers. Combinations of the above should also be included within the scope of computer-readable media.

[0103] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

[0104] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technical personnel in this field can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A method for electromagnetic signal prediction by jointly optimizing transformers using multi-domain transformations, characterized in that: The steps include: Step S1: Perform multi-domain analysis using electromagnetic data to extract characteristic parameters under multi-domain transformation; Step S2: using chaotic mapping, adaptive weight adjustment and memory strategy to improve the goose optimization algorithm, and using the improved goose optimization algorithm to optimize the hyperparameters of the transformer model; Step S3: inputting the feature parameters and signal samples under multi-domain transformation into the optimized transformer for training to obtain the optimized transformer model; Step S4: Use the optimized transformer model to test and verify the electromagnetic data, and apply the model to electromagnetic noise identification and signal prediction; Step S5: retain the signal, remove the identified noise and predict the signal, and integrate and reconstruct it in the original sampling order to obtain high-quality electromagnetic data.

2. According to claim 1, the electromagnetic signal prediction method of multi-domain transformation joint optimization transformer is characterized in that: In step S1, the multi-domain transform uses Fourier transform, short-time Fourier transform and continuous wavelet transform to perform multi-domain analysis and extract multi-domain feature parameters. The multi-domain features are as follows: and t =[And 时域特征 ,AND 频域特征 ,AND 时频域特征 ] t Where Y 时域特征 ,Y 频域特征 ,Y 时频域特征 Respectively represent the input multi-domain feature vector set, y t Represents the total set of multi-domain features.

3. The electromagnetic signal prediction method of a multi-domain transformation joint optimization transformer according to claim 1 is characterized in that: In step S2, the improved goose optimization algorithm uses chaotic mapping, adaptive weight adjustment and memory strategy to improve the algorithm, and uses the initialization particle x n The population chaos map calculation is as follows: x n+1 =r×(sin(Π×x n )+cos(Π×x n )) In the formula, the control parameter r adjusts the degree of chaos of the system; The adaptive weight adjustment is calculated as follows: a=2×exp[-6×(t÷Max_IT)] In the formula, the adaptive parameter a is dynamically adjusted according to the current number of iterations t and the total number of iterations Max_IT; The memory strategy is calculated as follows: M is the array storing the best historical solutions, M_scores is used to store the fitness scores of the corresponding solutions, and for the newly discovered solution Best_pos and its fitness score Best_score, the update process is as follows: Find the maximum value in M_scores and its index idx: max_score,idx=max(M_scores) If the best range Best_score is less than the maximum range max_score, then update: M_[idx]=Best_pos M_scores[idx]=Best_score Checking memory: When the performance of the algorithm stagnates, that is, no better solution is found after multiple consecutive iterations, you can try to jump out of the local optimum by checking the solution in memory. The process of checking memory and selecting the optimal solution is expressed as: Find the minimum value and its index in M_scores: Best_score,idx=min(M_scores) Return the corresponding solution: Best_pos=M_[idx].

4. The electromagnetic signal prediction method of a multi-domain transformation joint optimization transformer according to claim 3 is characterized in that: In step S2, the improved goose optimizes the transformer model hyperparameters, namely, the number of attention heads, learning rate and batch size of the transformer model.

5. The electromagnetic signal prediction method of a multi-domain transformation joint optimization transformer according to claim 1 is characterized in that: In step S3, the transformer model consists of an encoder, a decoder and an output layer; the encoder in the transformer model consists of a multi-head attention mechanism and a fully connected neural network, the decoder consists of two multi-head attention mechanisms and a fully connected neural network, and the output layer undergoes a linear transformation and Softmax to obtain the output probability distribution.

6. The electromagnetic signal prediction method of multi-domain transformation joint optimization transformer according to claim 5 is characterized in that: In step S4 and step S5, the electromagnetic data testing and prediction using the optimized transformer model are as follows: In the formula, x, y, xy represent the multi-domain feature data segments, j represents the number of segments, represents the prediction process, Y j+1 Represents the predicted value.

7. An electromagnetic signal prediction device for multi-domain transformation joint optimization transformer, characterized in that: It includes a memory, a control processor, and a computer program stored in the memory and executable on the control processor, wherein the control processor executes the computer program to implement the electromagnetic signal prediction method of the multi-domain transformation joint optimization transformer as described in any one of claims 1 to 6.

8. A control system, characterized in that: An electromagnetic signal prediction device comprising the multi-domain transformation joint optimization transformer as described in claim 7.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute the electromagnetic signal prediction method of the multi-domain transformation joint optimization transformer as described in any one of claims 1 to 6.

Images