Electromagnetic compatibility prediction system and method based on deep learning

Abstract

The invention relates to the field of electromagnetic compatibility prediction, and particularly discloses an electromagnetic compatibility prediction system and method based on deep learning, and the method comprises the steps: firstly obtaining to-be-judged electromagnetic data corresponding to to-be-judged electronic equipment and electromagnetic signal radiation intensity values of the to-be-judged electronic equipment at a plurality of preset time points, and then employing a deep learning technology to obtain electromagnetic signal radiation intensity values of the to-be-judged electronic equipment; and feature extraction and correlation analysis are carried out on the two data, and finally whether the electromagnetic compatibility of the electronic equipment is qualified or not is judged through a classifier, so that a more scientific and efficient method is provided for the electromagnetic compatibility evaluation of the electronic equipment, and the safety, reliability and market competitiveness of the equipment are ensured.

Description

Technical Field

[0001] The present application relates to the field of electromagnetic compatibility prediction, and more specifically, to an electromagnetic compatibility prediction system and method based on deep learning. Background Art

[0002] Electromagnetic compatibility (EMC) refers to the ability of electronic equipment to work properly in an electromagnetic environment without generating excessive electromagnetic interference. It ensures that the equipment can operate effectively under a certain level of electromagnetic interference without causing interference to other equipment. EMC research involves electromagnetic interference (EMI) and electromagnetic immunity (EMS), and is widely used in electronic product design, testing and certification to meet regulatory and standard requirements.

[0003] In the prior art, electromagnetic compatibility is usually evaluated through experimental testing methods. For example, a specific application environment is built in the laboratory, and indicators such as the degree of electromagnetic interference and failure rate between electronic equipment and surrounding equipment and devices are recorded to determine the electromagnetic compatibility results of electronic equipment. In this way, a specific experimental environment needs to be built in advance. The accuracy of the experimental test depends on the accuracy of the built experimental environment, and the accuracy of the simulation results depends on the accuracy of the model and the accuracy of the input parameters. For example, material parameters with large errors, electromagnetic radiation sources, etc., will affect the reliability of the simulation results.

[0004] Therefore, an electromagnetic compatibility prediction system and method based on deep learning is desired. Summary of the invention

[0005] In order to solve the above technical problems, this application is proposed. The embodiment of this application provides an electromagnetic compatibility prediction system and method based on deep learning, which first obtains the electromagnetic data to be judged corresponding to the electronic device to be judged and the electromagnetic signal radiation intensity values ​​of the electronic device to be judged at multiple predetermined time points, and then uses deep learning technology to perform feature extraction and correlation analysis on the two, and finally uses a classifier to judge whether the electromagnetic compatibility of the electronic device is qualified, thereby providing a more scientific and efficient method for the electromagnetic compatibility assessment of the electronic device, thereby ensuring the safety, reliability and market competitiveness of the equipment.

[0006] According to one aspect of the present application, there is provided an electromagnetic compatibility prediction system based on deep learning, which includes:

[0007] An electronic device electromagnetic compatibility data acquisition module is used to acquire electromagnetic data corresponding to the electronic device to be judged and electromagnetic signal radiation intensity values ​​of the electronic device to be judged at multiple predetermined time points;

[0008] An electronic equipment electromagnetic compatibility data processing module is used to extract a semantic association feature vector of the electromagnetic data to be judged and a feature vector of the electromagnetic signal radiation intensity of the electronic equipment to be judged from the electromagnetic data to be judged corresponding to the electronic equipment to be judged and the electromagnetic signal radiation intensity values ​​of the electronic equipment to be judged at the plurality of predetermined time points;

[0009] The electromagnetic compatibility qualification judgment module is used to judge whether the electronic equipment is electromagnetic compatible based on the semantic association feature vector of the electromagnetic data to be judged and the electromagnetic signal radiation intensity feature vector of the electronic equipment to be judged.

[0010] According to another aspect of the present application, a deep learning-based electromagnetic compatibility prediction method is provided, which includes:

[0011] Acquire electromagnetic data corresponding to the electronic device to be judged and electromagnetic signal radiation intensity values ​​of the electronic device to be judged at multiple predetermined time points;

[0012] Extracting a semantic association feature vector of the electromagnetic data to be judged and a feature vector of the electromagnetic signal radiation intensity of the electronic device to be judged from the electromagnetic data to be judged corresponding to the electronic device to be judged and the electromagnetic signal radiation intensity values ​​of the electronic device to be judged at the plurality of predetermined time points;

[0013] Based on the semantic association feature vector of the electromagnetic data to be judged and the electromagnetic signal radiation intensity feature vector of the electronic device to be judged, it is judged whether the electromagnetic compatibility of the electronic device is qualified.

[0014] Compared with the prior art, the present application provides an electromagnetic compatibility prediction system and method based on deep learning, which first obtains the electromagnetic data to be judged corresponding to the electronic device to be judged and the electromagnetic signal radiation intensity values ​​of the electronic device to be judged at multiple predetermined time points, and then uses deep learning technology to perform feature extraction and correlation analysis on the two, and finally uses a classifier to judge whether the electromagnetic compatibility of the electronic device is qualified, thereby providing a more scientific and efficient method for the electromagnetic compatibility evaluation of electronic equipment, thereby ensuring the safety, reliability and market competitiveness of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0016] Figure 1 is a block diagram of an electromagnetic compatibility prediction system based on deep learning according to an embodiment of the present application.

[0017] Figure 2 This is a block diagram of an electronic device electromagnetic compatibility data processing module in an electromagnetic compatibility prediction system based on deep learning according to an embodiment of the present application.

[0018] Figure 3 This is a block diagram of an electronic device electromagnetic data feature extraction unit in an electromagnetic compatibility prediction system based on deep learning according to an embodiment of the present application.

[0019] Figure 4 It is a block diagram of an electromagnetic compatibility qualification judgment module in an electromagnetic compatibility prediction system based on deep learning according to an embodiment of the present application.

[0020] Figure 5 Flow chart of an electromagnetic compatibility prediction method based on deep learning according to an embodiment of the present application. DETAILED DESCRIPTION

[0021] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0022] Figure 1 : is a block diagram of an electromagnetic compatibility prediction system based on deep learning according to an embodiment of the present application. Figure 1 As shown, according to the electromagnetic compatibility prediction system 100 based on deep learning according to the embodiment of the present application, it includes: an electronic device electromagnetic compatibility data acquisition module 110, which is used to obtain the electromagnetic data to be judged corresponding to the electronic device to be judged and the electromagnetic signal radiation intensity values ​​of the electronic device to be judged at multiple predetermined time points; an electronic device electromagnetic compatibility data processing module 120, which is used to extract the semantic association feature vector of the electromagnetic data to be judged and the electromagnetic signal radiation intensity feature vector of the electronic device to be judged from the electromagnetic data to be judged corresponding to the electronic device to be judged and the electromagnetic signal radiation intensity values ​​of the electronic device to be judged at the multiple predetermined time points; an electromagnetic compatibility qualified judgment module 130, which is used to judge whether the electromagnetic compatibility of the electronic device is qualified based on the semantic association feature vector of the electromagnetic data to be judged and the electromagnetic signal radiation intensity feature vector of the electronic device to be judged.

[0023] In the above-mentioned electromagnetic compatibility prediction system 100 based on deep learning, the electronic equipment electromagnetic compatibility data acquisition module 110 is used to obtain the electromagnetic data to be judged corresponding to the electronic equipment to be judged and the electromagnetic signal radiation intensity values ​​of the electronic equipment to be judged at multiple predetermined time points. It should be understood that electromagnetic compatibility (EMC) means that electronic equipment can operate normally in an electromagnetic environment without generating excessive electromagnetic interference. This means that the device can work effectively under a certain level of electromagnetic interference and will not affect the normal operation of other devices. The study of electromagnetic compatibility includes electromagnetic interference (EMI) and electromagnetic immunity (EMS), and is widely used in the design, testing and certification of electronic products to ensure that they comply with relevant regulations and standards. In the prior art, the evaluation of electromagnetic compatibility usually relies on experimental testing. The specific method is to build a specific test environment in the laboratory, and record key indicators such as the degree of electromagnetic interference and failure rate between the electronic equipment and other surrounding equipment, so as to obtain the evaluation result of electromagnetic compatibility. However, this method requires the establishment of an accurate experimental environment before testing, and its accuracy depends on the rationality of the experimental setting. At the same time, the reliability of the simulation results is also affected by the accuracy of the model and the input parameters. For example, errors in material parameters or electromagnetic radiation sources may have a negative impact on the simulation results, thereby affecting the reliability of the final evaluation. Therefore, in the technical solution of the present application, the electromagnetic data to be judged corresponding to the electronic device to be judged and the electromagnetic signal radiation intensity values ​​of the electronic device to be judged at multiple predetermined time points are obtained, and combined with deep learning technology, the electromagnetic compatibility of the electronic device is evaluated to determine whether it meets the standards, thereby providing a more scientific and efficient method. This will help ensure the safety, reliability and competitiveness of the equipment in the market.

[0024] Specifically, electronic devices generate electromagnetic radiation during operation, which may interfere with the surrounding environment and other electronic devices and affect their normal operation. Therefore, by real-time monitoring of the electromagnetic signals of these devices, it is possible to ensure that they are within the prescribed radiation limit range and avoid potential harm to users and other devices. On the other hand, by obtaining the electromagnetic signal radiation intensity values ​​of the device at multiple time points, the electromagnetic radiation characteristics of the device can be comprehensively analyzed. Different working states, different use environments, and different time points may affect the radiation intensity of the device. In the process of data acquisition, appropriate measurement equipment and technology must be used to ensure the accuracy and reliability of the data. Common measurement methods include using professional instruments such as electromagnetic field intensity meters and spectrum analyzers to detect and record the radiation intensity of the device at a specific frequency. These data can provide a basis for subsequent evaluation and analysis. In this way, a report on the electromagnetic radiation characteristics of the device can be generated to provide a reference for the design improvement of the device, provide consumers with necessary safety information, and help manufacturers conduct compliance testing before the product is launched to ensure that the product meets the requirements of market regulations and standards. This not only protects the health and safety of users, but also enhances the market competitiveness of products.

[0025] In the above-mentioned electromagnetic compatibility prediction system 100 based on deep learning, the electronic device electromagnetic compatibility data processing module 120 is used to extract the semantic association feature vector of the electromagnetic data to be judged and the electromagnetic signal radiation intensity feature vector of the electronic device to be judged from the electromagnetic data to be judged corresponding to the electronic device to be judged and the electromagnetic signal radiation intensity values ​​of the electronic device to be judged at the plurality of predetermined time points. In this way, a clearer basis for electromagnetic safety assessment can be provided for regulatory agencies and consumers, thereby enhancing the market competitiveness of products. By systematically understanding the electromagnetic behavior of equipment, manufacturers can carry out targeted optimization in the product development stage to ensure that the product meets both safety standards and market demand. The comprehensive application of this method provides a scientific and effective solution for improving electromagnetic compatibility.

[0026] Figure 2 FIG. 1 is a block diagram of an electronic device electromagnetic compatibility data processing module in an electromagnetic compatibility prediction system based on deep learning according to an embodiment of the present application. Figure 2 As shown, in a specific embodiment of the present application, the electronic device electromagnetic compatibility data processing module 120 includes: an electronic device electromagnetic data feature extraction unit 121, used to perform feature extraction on the electromagnetic data to be judged corresponding to the electronic device to be judged to obtain a semantic association feature vector of the electromagnetic data to be judged; an electromagnetic signal radiation intensity feature extraction unit 122, used to perform feature extraction on the electromagnetic signal radiation intensity values ​​of the electronic device to be judged at the multiple predetermined time points to obtain the electromagnetic signal radiation intensity feature vector of the electronic device to be judged.

[0027] It should be understood that raw electromagnetic data is usually high-dimensional and complex, containing a lot of noise and redundant information. Through feature extraction, the most representative data features can be extracted to reduce the complexity of data processing. This is very important for subsequent data analysis and modeling, because reducing data dimensions can improve computing efficiency, reduce the complexity of model training, and reduce the risk of overfitting. By analyzing these feature vectors, researchers can discover specific behavioral patterns of electromagnetic radiation and understand how the performance of the device changes in different environments or working conditions, which is of great significance for the optimization design and improvement of the equipment.

[0028] Furthermore, the raw electromagnetic signal radiation intensity data is usually high-dimensional and complex, and may be affected by multiple factors. Through feature extraction, the most representative features in the data can be extracted to remove noise and redundant information. By inputting these vectors into the model, electromagnetic radiation can be predicted and evaluated, and potential risks or failure modes can be identified. This method can not only improve the accuracy of detection, but also realize automated monitoring and real-time evaluation, reducing the need for manual intervention.

[0029] Figure 3 FIG. 1 is a block diagram of an electronic device electromagnetic data feature extraction unit in an electromagnetic compatibility prediction system based on deep learning according to an embodiment of the present application. Figure 3 As shown, in a specific embodiment of the present application, the electronic device electromagnetic data feature extraction unit 121 includes: an electromagnetic data to be judged word segmentation subunit 1211, used to perform word segmentation on the electromagnetic data to be judged corresponding to the electronic device to be judged to obtain multiple electromagnetic data items to be judged; and an electromagnetic data to be judged feature encoding subunit 1212, used to perform feature encoding on the multiple electromagnetic data items to be judged to obtain a semantic association feature vector of the electromagnetic data to be judged.

[0030] It should be understood that the electromagnetic data to be judged usually exists in an unstructured or semi-structured form and contains a variety of information. This information may involve equipment type, radiation intensity, working status, timestamp, etc., and contains a lot of redundant information. Through word segmentation, this information can be structured to make it easier to analyze and understand. The structured data form facilitates subsequent classification, clustering and other data analysis operations. The word segmentation process can help identify and extract key information items in the data. For example, in electromagnetic data, certain words or phrases may represent specific electromagnetic characteristics, standards or measurement conditions. Through word segmentation, these key information can be extracted separately to form meaningful electromagnetic data items to be judged, thereby providing a basis for subsequent analysis. Specifically, word segmentation can help capture the semantic associations between different data items, which is very important for identifying patterns and trends of electromagnetic signals. By analyzing the relationship between these data items to be judged, the electromagnetic radiation characteristics of the equipment under different working conditions can be revealed, thereby optimizing equipment design and operation.

[0031] Furthermore, the electromagnetic data items to be judged usually appear in different formats and units, and directly using the original data for analysis may lead to inconsistency and misleading. When performing feature encoding, different data items can be converted into a unified format to ensure the consistency and comparability of the data. After normalization, it is easier for analysts to understand and apply the data. The feature encoding process can extract the most representative information from multiple electromagnetic data items to be judged, and then convert it into feature vectors. These feature vectors can simplify the dimensions of the original data, reduce the complexity of the data, and thus improve the efficiency of subsequent analysis. Through dimensionality reduction, features that have an important impact on model training and prediction can be selected to reduce noise interference.

[0032] In a specific embodiment of the present application, the electromagnetic data feature encoding subunit 1212 includes: semantically encoding multiple electromagnetic data items to be judged based on deep learning technology to obtain multiple electromagnetic data feature vectors to be judged; and performing mixed convolution encoding on the multiple electromagnetic data feature vectors to be judged to obtain the semantically associated feature vector of the electromagnetic data to be judged.

[0033] It should be understood that electromagnetic data is usually nonlinear and complex, containing multi-dimensional features and complex relationships between samples. The multi-layer network structure in deep learning can effectively capture these complex relationships and generate feature vectors that reflect the intrinsic semantics of the data. By semantically encoding multiple electromagnetic data items to be judged, the deep learning model can learn a common feature representation through a large number of training samples. Among them, deep learning can map high-dimensional data to a low-dimensional feature space for effective dimensionality reduction and information compression. This feature not only improves data processing efficiency, but also preserves important information structures. In the context of electromagnetic data, the feature vector obtained by semantic encoding can focus on key electromagnetic radiation characteristics and reduce the impact of irrelevant information.

[0034] Furthermore, in electromagnetic data, local changes in signals often contain important physical meanings. Hybrid convolutional coding can integrate features at different levels to enhance the ability of feature representation. Hybrid convolutional coding can capture the complex semantic relationships between electromagnetic data to be judged through a combination of multiple convolution kernels. Since electromagnetic signals usually involve various influencing factors, such as device status, environmental interference, etc., there may be nonlinear and complex interactions between various factors. Through hybrid convolutional coding, these dynamic relationships can be analyzed more comprehensively, thereby generating feature vectors with more semantic relevance.

[0035] In a specific embodiment of the present application, multiple electromagnetic data items to be judged are semantically encoded based on deep learning technology to obtain multiple electromagnetic data feature vectors to be judged, including: passing the multiple electromagnetic data items to be judged through an electromagnetic data context encoder to be judged including an embedding layer to obtain the multiple electromagnetic data feature vectors to be judged.

[0036] It should be understood that electromagnetic data is often complex, time-varying, multidimensional, and contains a lot of noise and irrelevant information. Through the embedding layer, the model can learn which part of the data is the most important, rather than simply relying on the original data form for processing. The embedding layer helps the model to show better generalization ability when facing unseen data by learning contextual relationships. The feature vector generated by the context encoder can make it easier to understand the relationship between data. By examining the output of the embedding layer, it is possible to intuitively see how different electromagnetic data items interact in the feature space, thereby providing a certain degree of interpretability, which is very important in many practical applications. Specifically, the embedding layer of the electromagnetic data context encoder to be judged containing the embedding layer is used to map each of the multiple electromagnetic data items to be judged into an electromagnetic word embedding vector to obtain a sequence of electromagnetic word embedding vectors; the converter-based Bert model of the electromagnetic data context encoder to be judged containing the embedding layer is used to perform global context semantic encoding on the sequence of electromagnetic word embedding vectors to obtain multiple electromagnetic data feature vectors to be judged.

[0037] In a specific embodiment of the present application, the multiple electromagnetic data feature vectors to be judged are subjected to mixed convolution encoding to obtain the semantically associated feature vector of the electromagnetic data to be judged, including: splicing the multiple electromagnetic data feature vectors to be judged to obtain an electromagnetic data splicing feature vector; passing the electromagnetic data splicing feature vector through an electromagnetic convolution neural network model based on a mixed convolution layer to obtain the semantically associated feature vector of the electromagnetic data to be judged.

[0038] It should be understood that multiple electromagnetic data feature vectors to be judged usually come from different data dimensions or processing stages, and each of them captures different characteristics and related information of the electromagnetic signal. By splicing these feature vectors, a lot of information can be integrated into a new high-dimensional feature vector, ensuring that the model can take into account all relevant information when performing analysis, thereby avoiding information loss.

[0039] Furthermore, the concatenated feature vector of electromagnetic data is processed through an electromagnetic convolutional neural network model based on a hybrid convolutional layer to obtain the semantically associated feature vector of the electromagnetic data to be judged, which is an efficient feature learning and information extraction method. The hybrid convolutional layer combines different types of convolution operations and can extract multi-level feature information from the concatenated feature vector. By applying convolution kernels of different sizes and types, the model can capture local and global features in electromagnetic data, which is crucial for identifying important patterns in signals. The hybrid convolutional layer can effectively handle complex spatiotemporal dependencies, thereby improving the ability of feature extraction. By applying hybrid convolutional layers in the electromagnetic convolutional neural network, the model can learn the combination patterns of multiple features during the training process, which improves its generalization ability to new data. Specifically, each hybrid convolutional layer of the electromagnetic convolutional neural network model based on the hybrid convolutional layer includes a parallel first convolution branch structure, a second convolution branch structure, a third convolution branch structure and a fourth convolution branch structure, and a multi-scale fusion structure connected to the first to fourth convolution branch structures, wherein the first convolution branch uses a first convolution kernel having a first size, the second convolution branch uses a second convolution kernel having a first size and a first void ratio, the third convolution branch uses a third convolution kernel having a first size and a second void ratio, and the fourth convolution branch uses a fourth convolution kernel having a first size and a third void ratio. More specifically, each hybrid convolutional layer of the electromagnetic convolutional neural network model based on the hybrid convolutional layer is further used to perform the following operations on the input data in the forward pass of the layer: using a first convolutional kernel with a first size to perform convolutional encoding on the electromagnetic data splicing feature vector to obtain a first scale feature map; using a second convolutional kernel with a first void rate to perform convolutional encoding on the electromagnetic data splicing feature vector to obtain a second scale feature map; using a third convolutional kernel with a second void rate to perform convolutional encoding on the electromagnetic data splicing feature vector to obtain a third scale feature map; using a fourth convolutional kernel with a third void rate to perform convolutional encoding on the electromagnetic data splicing feature vector to obtain a fourth scale feature map; The invention relates to a method for obtaining a feature map, wherein the first convolution kernel, the second convolution kernel, the third convolution kernel and the fourth convolution kernel have the same size, and the second convolution kernel, the third convolution kernel and the fourth convolution kernel have different hole rates; the first scale feature map, the second scale feature map, the third scale feature map and the fourth scale feature map are aggregated along the channel dimension to obtain an aggregate feature map; the aggregate feature map is pooled to generate a pooled feature map; and the pooled feature map is activated to generate an activated feature map; wherein the output of the last layer of the convolutional neural network model including multiple hybrid convolution layers is the semantic association feature vector of the electromagnetic data to be judged.

[0040] In a specific embodiment of the present application, the electromagnetic signal radiation intensity feature extraction unit 122 includes: arranging the electromagnetic signal radiation intensity values ​​of the electronic device to be judged at the multiple predetermined time points into an electromagnetic signal radiation intensity input vector of the electronic device to be judged; passing the electromagnetic signal radiation intensity input vector of the electronic device to be judged through an electronic device electromagnetic signal radiation intensity timing encoder to obtain the electromagnetic signal radiation intensity feature vector of the electronic device to be judged.

[0041] It should be understood that the radiation intensity of electromagnetic signals usually changes over time, forming a time series characteristic. Arranging the signal intensity values ​​into vectors in chronological order helps to preserve the continuity and order relationship in time. This time series feature can provide key information for subsequent analysis, especially in tasks such as electromagnetic interference pattern recognition and anomaly detection. By uniformly arranging the radiation intensity values ​​into vectors, the randomness of data at different time points can be eliminated, making the input more systematic and consistent, thereby improving the accuracy of subsequent analysis. For example, the vectorized form facilitates batch processing and matrix operations, providing a suitable input format for machine learning and deep learning models. In machine learning and deep learning, the model is very sensitive to the format and structure of the input data. Arranging the radiation intensity values ​​into vectors can effectively input data into various algorithms and enhance the performance of the model in the learning and prediction process. The model will be able to improve its understanding of the electromagnetic radiation characteristics of the device by learning the patterns and relationships in this structured input. Arranging the electromagnetic signal radiation intensity values ​​into input vectors makes it easier for the model to identify the changing trends and periodic characteristics of the signal. For example, by analyzing the changes in the vector, the performance of the device under different working conditions or environments can be revealed. This is very helpful for curve fitting and trend prediction.

[0042] Furthermore, the electromagnetic signal radiation intensity of electronic devices varies with time and contains rich time-dependent information. The time series encoder can process time series data and identify potential patterns and regularities by capturing the changes in the signal at different time points. This ability is crucial for analyzing the normal working state and abnormal behavior of the equipment. The original radiation intensity value in the input vector may contain redundancy and noise. The time series encoder can compress the input data into a more representative feature vector by learning the effective features in the time series. This feature vector not only reduces the amount of data, but also increases the density of information, making subsequent analysis more efficient. The time series encoder usually adopts architectures such as recurrent neural network (RNN), long short-term memory network (LSTM) or gated recurrent unit (GRU), which can effectively process long sequence data. These models can capture long-term dependencies, thereby identifying complex change patterns in the input data and supporting more accurate device status judgment. By converting the input vector into a feature vector, the time series encoder can optimize the input format of machine learning and deep learning models. This feature vector contains the key features in the time series, making it easier for the model to learn effective mapping relationships during training, thereby improving the accuracy of prediction. There may be a nonlinear relationship between the radiation intensity of the electromagnetic signal and time. The design of the time series encoder enables it to handle such complex relationships. Through nonlinear activation functions and multi-layer structures, it captures the implicit patterns in the input data, thereby enhancing the expressive power of the model. In real-time monitoring systems, timely analysis of the radiation intensity of electromagnetic signals is crucial. The time series encoder can quickly process the input vector in a continuously updated data stream, generate a real-time feature vector, and provide instant feedback for the evaluation and early warning of the device status. This flexibility improves the response speed and accuracy of the system. Specifically, the fully connected layer of the electromagnetic signal radiation intensity time series encoder of the electronic device is used to fully connect the input vector of the electromagnetic signal radiation intensity of the electronic device to be judged to extract the high-dimensional implicit features of the eigenvalues ​​at each position in the electromagnetic signal radiation intensity input vector of the electronic device to be judged; and the one-dimensional convolutional layer of the electromagnetic signal radiation intensity time series encoder of the electronic device is used to one-dimensionally encode the electromagnetic signal radiation intensity input vector of the electronic device to be judged to extract the high-dimensional implicit correlation features of the correlation between the eigenvalues ​​at each position in the electromagnetic signal radiation intensity input vector of the electronic device to be judged.

[0043] In the above-mentioned electromagnetic compatibility prediction system 100 based on deep learning, the electromagnetic compatibility qualified judgment module 130 is used to judge whether the electromagnetic compatibility of the electronic equipment is qualified based on the semantic association feature vector of the electromagnetic data to be judged and the electromagnetic signal radiation intensity feature vector of the electronic equipment to be judged. It should be understood that electromagnetic compatibility (EMC) refers to the ability of electronic equipment to work normally in its working environment without interfering with other equipment. In order to ensure that the equipment meets the EMC standard, it is necessary to conduct a comprehensive assessment of its electromagnetic radiation intensity and possible electromagnetic interference. On the other hand, the electromagnetic signal radiation intensity feature vector directly describes the electromagnetic radiation level generated by the equipment when it is working. By establishing an association model, the relationship between the electromagnetic signal radiation intensity and the functional status of the equipment can be analyzed, and potential abnormal situations can be effectively identified.

[0044] Figure 4 FIG. 1 is a block diagram of an electromagnetic compatibility qualification judgment module in an electromagnetic compatibility prediction system based on deep learning according to an embodiment of the present application. Figure 4 As shown, in a specific embodiment of the present application, the electromagnetic compatibility qualified judgment module 130 includes: an electromagnetic compatibility feature fusion unit 131, which is used to perform target domain adaptation fusion based on kernel association fusion on the semantic association feature vector of the electromagnetic data to be judged and the electromagnetic signal radiation intensity feature vector of the electronic device to be judged to obtain an electromagnetic compatibility classification judgment feature vector; an electronic device electromagnetic compatibility qualified judgment classification unit 132, which is used to pass the electromagnetic compatibility classification judgment feature vector through a classifier to obtain a classification result, and the classification result is used to judge whether the electronic device is electromagnetic compatible.

[0045] It should be understood that in electromagnetic compatibility analysis, there may be subtle changes between the normal working state of the equipment and the potential electromagnetic interference. By analyzing the fused feature vectors, these changes can be more effectively identified, thereby timely discovering electromagnetic compatibility problems and reducing the risk of equipment failure.

[0046] In particular, considering that the semantic association feature vector of the electromagnetic data to be judged is obtained from the electromagnetic data to be judged after word segmentation, embedding layer encoding and electromagnetic convolutional neural network model based on hybrid convolution layer. This process mainly captures the structured information and contextual relationship in the electromagnetic data, emphasizing the pattern recognition and semantic understanding of electromagnetic signals in the time or frequency dimension. These feature vectors reflect the complex patterns inside the electromagnetic data and their potential semantic meanings, and are suitable for describing the behavioral characteristics of electromagnetic signals. In contrast, the feature vector of the electromagnetic signal radiation intensity of the electronic device to be judged is obtained by arranging the electromagnetic signal radiation intensity values ​​at multiple predetermined time points and processing them through a time series encoder. This feature vector focuses more on describing the time series characteristics of the electromagnetic signal, such as the changing trend and periodic fluctuation of the radiation intensity. They provide important information about the state of the electromagnetic environment, but pay more attention to the changes in the measured values ​​at the physical level rather than semantic interpretation. Since the above two feature vectors focus on the information at the semantic level and the physical measurement level respectively, even if they are both used to describe the electromagnetic compatibility of electronic equipment, they will be different in numerical representation and carry different types of data distribution and statistical properties. Therefore, when trying to directly transfer the knowledge of one field (for example, a model trained only for semantic association features) to another field (i.e., a new environment that integrates semantic and physical measurements), it is difficult for the model to directly and effectively transfer its learned knowledge. Based on this, in the technical solution of the present application, the semantic association feature vector of the electromagnetic data to be judged and the electromagnetic signal radiation intensity feature vector of the electronic device to be judged are subjected to target domain adaptation fusion based on kernel association fusion to obtain an electromagnetic compatibility classification judgment feature vector.

[0047] Specifically, in an embodiment of the present application, the semantic association feature vector of the electromagnetic data to be judged and the feature vector of the electromagnetic signal radiation intensity of the electronic device to be judged are subjected to target domain adaptation fusion based on kernel association fusion to obtain an electromagnetic compatibility classification judgment feature vector, including: mapping the semantic association feature vector of the electromagnetic data to be judged and the feature vector of the electromagnetic signal radiation intensity of the electronic device to be judged to an intrinsic decomposition space to obtain the semantic association intrinsic feature vector of the electromagnetic data to be judged and the intrinsic feature vector of the electromagnetic signal radiation intensity of the electronic device to be judged; domain mapping the semantic association intrinsic feature vector of the electromagnetic data to be judged and the intrinsic feature vector of the electromagnetic signal radiation intensity of the electronic device to be judged. The electromagnetic data semantics-electromagnetic signal radiation intensity fine-grained semantic information field is constructed between the electromagnetic data semantics-electromagnetic signal radiation intensity fine-grained semantic information field; based on the electromagnetic data semantics-electromagnetic signal radiation intensity fine-grained semantic information field, the electromagnetic data semantics-electromagnetic signal radiation intensity fine-grained semantic information field is used to fuse the electromagnetic data semantics-electromagnetic signal radiation intensity modulated intrinsic feature vector to obtain the electromagnetic compatibility classification judgment feature vector.

[0048] More specifically, in an embodiment of the present application, an electromagnetic data semantics-electromagnetic signal radiation intensity fine-grained semantic information field is constructed between the intrinsic feature vector after modulation of the semantic association of the electromagnetic data to be judged and the intrinsic feature vector after modulation of the electromagnetic signal radiation intensity of the electronic device to be judged, including: multiplying the transposed vector of the intrinsic feature vector after modulation of the semantic association of the electromagnetic data to be judged and the intrinsic feature vector after modulation of the electromagnetic signal radiation intensity of the electronic device to be judged, and then dividing by the scale of the intrinsic feature vector after modulation of the electromagnetic signal radiation intensity of the electronic device to be judged to obtain an electromagnetic data semantics-electromagnetic signal radiation intensity fine-grained interaction matrix; and performing convolution encoding on the electromagnetic data semantics-electromagnetic signal radiation intensity fine-grained matrix to obtain the electromagnetic data semantics-electromagnetic signal radiation intensity fine-grained semantic information field.

[0049] More specifically, in an embodiment of the present application, the electromagnetic data semantics-electromagnetic signal radiation intensity fine-grained semantic information field is used to fuse the modulated intrinsic feature vector of the electromagnetic data semantic association to be judged and the modulated intrinsic feature vector of the electromagnetic signal radiation intensity of the electronic device to be judged to obtain an electromagnetic compatibility classification judgment feature vector, including: multiplying the electromagnetic data semantics-electromagnetic signal radiation intensity fine-grained semantic information field by the modulated intrinsic feature vector of the electromagnetic data semantic association to be judged to obtain the information field mapping electromagnetic data semantics association feature vector; multiplying the electromagnetic data semantics-electromagnetic signal radiation intensity fine-grained semantic information field by the modulated intrinsic feature vector of the electromagnetic signal radiation intensity of the electronic device to be judged to obtain the information field mapping electromagnetic signal radiation intensity feature vector of the electronic device to be judged; weighted fusion of the electromagnetic data semantics association feature vector after the information field mapping and the electromagnetic signal radiation intensity feature vector of the electronic device to be judged after the information field mapping to obtain the electromagnetic compatibility classification judgment feature vector.

[0050] In the embodiment of the present application, specifically, the semantic association feature vector of the electromagnetic data to be judged and the feature vector of the electromagnetic signal radiation intensity of the electronic device to be judged are subjected to target domain adaptation fusion based on kernel association fusion to obtain the electromagnetic compatibility classification judgment feature vector, including: processing the semantic association feature vector of the electromagnetic data to be judged and the feature vector of the electromagnetic signal radiation intensity of the electronic device to be judged according to the following formula to obtain the electromagnetic compatibility classification judgment feature vector; wherein the formula is:

[0051]

[0052]

[0053] v c =av kt +βv st

[0054] Wherein, V1 represents the semantic association feature vector of the electromagnetic data to be judged, V2 represents the feature vector of the electromagnetic signal radiation intensity of the electronic device to be judged, PCA(·) represents the principal component analysis, U1 represents the sequence of the eigendecomposition vectors of the semantic association of the electromagnetic data to be judged, Λ1 represents the diagonal matrix of the semantic association of the electromagnetic data to be judged, λ 11 and λ 1m They represent the eigenvalues ​​of the first and mth positions of the semantic association diagonal matrix of the electromagnetic data to be judged, T represents the transpose of the vector, concat{·;·} represents the vector concatenation, and v 11 、v 12 、v 1m They represent the first, second and mth eigenvectors of the sequence of semantic association eigendecomposition vectors of electromagnetic data to be judged, Vk represents the semantic association eigenvector of the electromagnetic data to be judged, U2 represents the sequence of eigendecomposition vectors of the electromagnetic signal radiation intensity of the electronic equipment to be judged, Λ2 represents the diagonal matrix of the semantic association of the electromagnetic data to be judged, λ 21 and λ 2m They represent the eigenvalues ​​of the first and mth positions of the diagonal matrix of the electromagnetic signal radiation intensity of the electronic device to be judged, respectively. 21 、v 22 、v 2m They represent the first, second and mth eigenvectors of the sequence of eigendecomposition vectors of the electromagnetic signal radiation intensity of the electronic device to be judged, V s represents the intrinsic characteristic vector of the electromagnetic signal radiation intensity of the electronic device to be judged, v i represents the ith eigenvalue of the semantically associated eigenvector of the electromagnetic data to be judged, v j represents the jth eigenvalue of the intrinsic eigenvector of the electromagnetic signal radiation intensity of the electronic device to be judged, represents the square of the vector norm, exp represents the natural exponential function, V' k represents the intrinsic feature vector after semantic association modulation of the electromagnetic data to be judged, V' s It represents the intrinsic eigenvector after the electromagnetic signal radiation intensity of the electronic device to be judged is modulated. represents matrix multiplication, S represents the scale of the vector, M x Represents the electromagnetic data semantics-electromagnetic signal radiation intensity fine-grained interaction matrix, Conv 3×3 (M x ) indicates that M x A convolutional encoding operation based on a convolution kernel of 3×3 is performed, Ω represents the semantics of electromagnetic data - the fine-grained semantic information field of electromagnetic signal radiation intensity, v kt represents the semantic association feature vector of electromagnetic data to be judged after information field mapping, v st represents the characteristic vector of the electromagnetic signal radiation intensity of the electronic device to be judged after the information field is mapped, α and β represent weighted hyperparameters, and v c Represents the electromagnetic compatibility classification judgment feature vector.

[0055] That is, in the technical solution of the present application, the semantic association feature vector of the electromagnetic data to be judged and the feature vector of the electromagnetic signal radiation intensity of the electronic device to be judged are subjected to target domain adaptation fusion based on kernel association fusion. The process first maps the semantic association feature vector of the electromagnetic data to be judged and the feature vector of the electromagnetic signal radiation intensity of the electronic device to be judged to the intrinsic decomposition space to obtain the semantic association intrinsic feature vector of the electromagnetic data to be judged and the intrinsic feature vector of the electromagnetic signal radiation intensity of the electronic device to be judged. A person of ordinary skill in the art should know that the process of mapping to the intrinsic decomposition space is a combination of dimensionality reduction and feature extraction. In a specific example, the eigenvalue decomposition technology of principal component analysis (PCA) is used to find the intrinsic decomposition space, and the intrinsic decomposition space mapping is performed. The semantic association intrinsic feature vector of the electromagnetic data to be judged and the intrinsic feature vector of the electromagnetic signal radiation intensity of the electronic device to be judged after information field mapping not only represent the projection of the original feature vector in the new coordinate system, but also reflect the main direction of change of the data, which helps to avoid the "dimensionality disaster", that is, the problem that the sample density drops sharply as the number of features increases.

[0056] Next, after mapping the feature vector to the eigendecomposition space, it is necessary to perform inter-domain feature calibration on the semantically associated eigenfeature vector of the electromagnetic data to be judged and the eigenfeature vector of the electromagnetic signal radiation intensity of the electronic device to be judged to adjust the source domain feature vector to better match the data distribution of the target domain. This operation aims to capture and compensate for the systematic deviation between the two domains, so as to reduce the hidden disturbance of the class expression of the local feature distribution of the semantically associated eigenfeature vector of the electromagnetic data to be judged and the eigenfeature vector of the electromagnetic signal radiation intensity of the electronic device to be judged to the overall class expression. After modulation, the modulated eigenfeature vector of the semantically associated eigenfeature vector of the electromagnetic data to be judged and the modulated eigenfeature vector of the electromagnetic signal radiation intensity of the electronic device to be judged can retain the source domain characteristics while being more in line with the actual needs of the target domain, thereby enhancing the model's understanding and application of the target domain data.

[0057] After completing the scale modulation of the feature vector, the electromagnetic data semantics-electromagnetic signal radiation intensity fine-grained semantic information field is constructed between the intrinsic feature vector after the semantic association modulation of the electromagnetic data to be judged and the intrinsic feature vector after the electromagnetic signal radiation intensity modulation of the electronic device to be judged. The electromagnetic data semantics-electromagnetic signal radiation intensity fine-grained semantic information field is used to describe a more detailed relationship between feature vectors. The introduction of the electromagnetic data semantics-electromagnetic signal radiation intensity fine-grained semantic information field enables the model to go beyond simple point-to-point comparison and consider more complex patterns and contextual information, thereby enhancing the model's sensitivity to subtle differences and improving the quality of feature representation. Finally, based on the electromagnetic data semantics-electromagnetic signal radiation intensity fine-grained semantic information field, the intrinsic feature vector after the semantic association modulation of the electromagnetic data to be judged and the intrinsic feature vector after the electromagnetic signal radiation intensity modulation of the electronic device to be judged are fused to obtain the electromagnetic compatibility classification judgment feature vector.

[0058] Furthermore, the information extracted from the electromagnetic compatibility classification judgment feature vector includes the functional status of the device and the electromagnetic radiation level, which can be used by the classifier to identify whether the device has electromagnetic interference problems. By processing the electromagnetic compatibility classification judgment feature vector, the classifier can use the previous training data to build a model. During the training process, the classifier learns how to extract features related to electromagnetic compatibility from historical data and associate these features with qualified and unqualified labels. This data-driven approach enhances the objectivity of decision-making and reduces the bias that may be caused by human judgment.

[0059] In summary, the embodiment of the present application first obtains the electromagnetic data to be judged corresponding to the electronic device to be judged and the electromagnetic signal radiation intensity values ​​of the electronic device to be judged at multiple predetermined time points, and then uses deep learning technology to perform feature extraction and correlation analysis on the two. Finally, a classifier is used to determine whether the electromagnetic compatibility of the electronic device is qualified, thereby providing a more scientific and efficient method for the electromagnetic compatibility assessment of electronic equipment, thereby ensuring the safety, reliability and market competitiveness of the equipment.

[0060] As described above, the electromagnetic compatibility prediction system 100 based on deep learning according to the embodiment of the present application can be implemented in various terminal devices. In one example, the electromagnetic compatibility prediction system 100 based on deep learning can be integrated into the terminal device as a software module and / or a hardware module. For example, the electromagnetic compatibility prediction system 100 based on deep learning can be a software module in the operating system of the terminal device, or can be an application developed for the terminal device; of course, the electromagnetic compatibility prediction system 100 based on deep learning can also be one of the many hardware modules of the terminal device.

[0061] Alternatively, in another example, the deep learning-based electromagnetic compatibility prediction system 100 and the terminal device may also be separate devices, and the deep learning-based electromagnetic compatibility prediction system 100 may be connected to the terminal device via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0062] Figure 5 FIG. 1 is a flow chart of an electromagnetic compatibility prediction method based on deep learning according to an embodiment of the present application. Figure 5 As shown, according to the electromagnetic compatibility prediction method based on deep learning according to the embodiment of the present application, it includes: S110, obtaining the electromagnetic data to be judged corresponding to the electronic device to be judged and the electromagnetic signal radiation intensity values ​​of the electronic device to be judged at multiple predetermined time points; S120, extracting the semantic association feature vector of the electromagnetic data to be judged and the electromagnetic signal radiation intensity feature vector of the electronic device to be judged from the electromagnetic data to be judged corresponding to the electronic device to be judged and the electromagnetic signal radiation intensity values ​​of the electronic device to be judged at the multiple predetermined time points; S130, judging whether the electromagnetic compatibility of the electronic device is qualified based on the semantic association feature vector of the electromagnetic data to be judged and the electromagnetic signal radiation intensity feature vector of the electronic device to be judged.

[0063] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned electromagnetic compatibility prediction method based on deep learning have been referred to above. Figures 1 to 4 It has been introduced in detail in the description of the deep learning-based electromagnetic compatibility prediction system, and therefore, its repeated description will be omitted.

[0064] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only illustrative, for example, the module division is only a logical function division, and there may be other division methods in actual implementation.

[0065] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0066] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0067] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0068] Therefore, no matter from which point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the appended claims rather than the above description, so it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any attached figure mark in the claims should not be regarded as limiting the claims involved.

[0069] In addition, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices stated in the system claim can also be implemented by one unit or device through software or hardware. The second and other words are used to indicate names, but not to indicate any particular order.

[0070] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. An electromagnetic compatibility prediction system based on deep learning, characterized in that: include: An electronic device electromagnetic compatibility data acquisition module is used to acquire electromagnetic data corresponding to the electronic device to be judged and electromagnetic signal radiation intensity values ​​of the electronic device to be judged at multiple predetermined time points; An electronic equipment electromagnetic compatibility data processing module is used to extract a semantic association feature vector of the electromagnetic data to be judged and a feature vector of the electromagnetic signal radiation intensity of the electronic equipment to be judged from the electromagnetic data to be judged corresponding to the electronic equipment to be judged and the electromagnetic signal radiation intensity values ​​of the electronic equipment to be judged at the plurality of predetermined time points; The electromagnetic compatibility qualification judgment module is used to judge whether the electronic equipment is electromagnetic compatible based on the semantic association feature vector of the electromagnetic data to be judged and the electromagnetic signal radiation intensity feature vector of the electronic equipment to be judged.

2. The electromagnetic compatibility prediction system based on deep learning according to claim 1, characterized in that: The electronic equipment electromagnetic compatibility data processing module comprises: An electronic device electromagnetic data feature extraction unit, used for extracting features from the electromagnetic data to be judged corresponding to the electronic device to be judged to obtain a semantically associated feature vector of the electromagnetic data to be judged; The electromagnetic signal radiation intensity feature extraction unit is used to extract the features of the electromagnetic signal radiation intensity values ​​of the electronic device to be judged at the plurality of predetermined time points to obtain the electromagnetic signal radiation intensity feature vector of the electronic device to be judged.

3. The electromagnetic compatibility prediction system based on deep learning according to claim 2, characterized in that: The electronic device electromagnetic data feature extraction unit comprises: The electromagnetic data to be judged word segmentation subunit is used to segment the electromagnetic data to be judged corresponding to the electronic device to be judged to obtain a plurality of electromagnetic data items to be judged; The feature encoding subunit of the electromagnetic data to be judged is used to perform feature encoding on the multiple electromagnetic data items to be judged to obtain a semantically associated feature vector of the electromagnetic data to be judged.

4. The electromagnetic compatibility prediction system based on deep learning according to claim 3 is characterized in that: The electromagnetic data feature encoding subunit to be determined includes: Based on deep learning technology, semantic encoding is performed on multiple electromagnetic data items to be judged to obtain multiple electromagnetic data feature vectors to be judged; The multiple electromagnetic data feature vectors to be judged are subjected to mixed convolution coding to obtain the semantically associated feature vector of the electromagnetic data to be judged.

5. The electromagnetic compatibility prediction system based on deep learning according to claim 4 is characterized in that: Based on deep learning technology, multiple electromagnetic data items to be judged are semantically encoded to obtain multiple electromagnetic data feature vectors to be judged, including: The multiple electromagnetic data items to be judged are passed through an electromagnetic data context encoder to be judged including an embedding layer to obtain the multiple electromagnetic data feature vectors to be judged.

6. The electromagnetic compatibility prediction system based on deep learning according to claim 5, characterized in that: The plurality of electromagnetic data feature vectors to be determined are subjected to mixed convolution coding to obtain the semantic association feature vector of the electromagnetic data to be determined, including: splicing the plurality of electromagnetic data feature vectors to be determined to obtain an electromagnetic data splicing feature vector; The electromagnetic data splicing feature vector is passed through an electromagnetic convolutional neural network model based on a hybrid convolutional layer to obtain the semantic association feature vector of the electromagnetic data to be judged.

7. The electromagnetic compatibility prediction system based on deep learning according to claim 6, characterized in that: The electromagnetic signal radiation intensity feature extraction unit comprises: Arranging the electromagnetic signal radiation intensity values ​​of the electronic device to be determined at the plurality of predetermined time points into an electromagnetic signal radiation intensity input vector of the electronic device to be determined; The electromagnetic signal radiation intensity input vector of the electronic device to be determined is passed through an electronic device electromagnetic signal radiation intensity timing encoder to obtain the electromagnetic signal radiation intensity feature vector of the electronic device to be determined.

8. The electromagnetic compatibility prediction system based on deep learning according to claim 7, characterized in that: The electromagnetic compatibility qualification judgment module includes: An electromagnetic compatibility feature fusion unit, used for performing target domain adaptation fusion based on kernel association fusion on the semantic association feature vector of the electromagnetic data to be judged and the electromagnetic signal radiation intensity feature vector of the electronic device to be judged to obtain an electromagnetic compatibility classification judgment feature vector; The electronic equipment electromagnetic compatibility qualified judgment classification unit is used to pass the electromagnetic compatibility classification judgment feature vector through a classifier to obtain a classification result, and the classification result is used to judge whether the electronic equipment is electromagnetic compatible or not.

9. The electromagnetic compatibility prediction system based on deep learning according to claim 8, characterized in that: The electromagnetic compatibility feature fusion unit is used to: Mapping the semantic association feature vector of the electromagnetic data to be determined and the feature vector of the electromagnetic signal radiation intensity of the electronic device to be determined to an eigendecomposition space to obtain the semantic association eigenfeature vector of the electromagnetic data to be determined and the eigenfeature vector of the electromagnetic signal radiation intensity of the electronic device to be determined; Performing inter-domain feature calibration on the semantic association intrinsic feature vector of the electromagnetic data to be determined and the intrinsic feature vector of the electromagnetic signal radiation intensity of the electronic device to be determined to obtain the modulated intrinsic feature vector of the semantic association of the electromagnetic data to be determined and the modulated intrinsic feature vector of the electromagnetic signal radiation intensity of the electronic device to be determined; Constructing an electromagnetic data semantics-electromagnetic signal radiation intensity fine-grained semantic information field between the intrinsic feature vector after semantic association modulation of the electromagnetic data to be determined and the intrinsic feature vector after electromagnetic signal radiation intensity modulation of the electronic device to be determined; Based on the electromagnetic data semantics-electromagnetic signal radiation intensity fine-grained semantic information field, the semantic association modulated intrinsic feature vector of the electromagnetic data to be judged and the electromagnetic signal radiation intensity modulated intrinsic feature vector of the electronic device to be judged are fused to obtain the electromagnetic compatibility classification judgment feature vector.

10. An electromagnetic compatibility prediction method based on deep learning, characterized in that: include: Acquire electromagnetic data corresponding to the electronic device to be judged and electromagnetic signal radiation intensity values ​​of the electronic device to be judged at multiple predetermined time points; Extracting a semantic association feature vector of the electromagnetic data to be judged and a feature vector of the electromagnetic signal radiation intensity of the electronic device to be judged from the electromagnetic data to be judged corresponding to the electronic device to be judged and the electromagnetic signal radiation intensity values ​​of the electronic device to be judged at the plurality of predetermined time points; Based on the semantic association feature vector of the electromagnetic data to be judged and the electromagnetic signal radiation intensity feature vector of the electronic device to be judged, it is judged whether the electromagnetic compatibility of the electronic device is qualified.

Images