Anti-interference optimization method, device and equipment for FPC high-performance computing chip interface

Abstract

The present application relates to the technical field of chip interface anti-interference, and discloses an anti-interference optimization method, device and equipment for an FPC high-performance computing chip interface. The method comprises: performing total variation regularization decomposition processing on the interface signal of the FPC to obtain effective data signal components; performing three-dimensional bending deformation processing on the structure of the FPC to obtain three-dimensional molding structure parameters; performing multi-layer electromagnetic shielding processing on the FPC to obtain a three-layer composite shielding structure; inputting the electromagnetic shielding effectiveness data of the three-layer composite shielding structure into a residual neural network model for feature extraction to obtain multi-scale feature data; performing covariance matrix tapering on the multi-scale feature data to obtain a reconstructed covariance matrix; performing functional domain allocation and signal optimization to obtain the anti-interference transmission data stream of each functional domain, thereby improving the accuracy of signal processing and realizing the optimized allocation of data streams for each functional domain.

Description

Technical Field

[0001] The present application relates to the technical field of chip interface anti-interference, and in particular to an anti-interference optimization method, device and equipment for an FPC high-performance computing chip interface. Background Art

[0002] With the rapid development of electronic information technology, high-performance computing chips are increasingly used in industrial control, autonomous driving and other fields. FPC (flexible printed circuit board) has become an important interface carrier for connecting high-performance computing chips with various sensors due to its thinness and bendability. However, in actual applications, the FPC interface is easily affected by various factors such as electromagnetic interference and mechanical stress, resulting in a decrease in signal transmission quality.

[0003] Traditional FPC interface anti-interference methods mainly rely on physical methods such as adding shielding layers and improving grounding design, but these methods often have problems such as limited shielding effect and complex structure. At the same time, as the integration of vehicle electronic systems continues to increase, signal interference between functional domains is becoming increasingly serious, and existing signal processing methods are difficult to meet the requirements of high-reliability transmission. Summary of the invention

[0004] The present application provides an anti-interference optimization method, device and equipment for an FPC high-performance computing chip interface. The present invention improves the accuracy of signal processing and realizes optimized allocation of data flows to each functional domain.

[0005] The first aspect of the present application provides an anti-interference optimization method for an FPC high-performance computing chip interface, the anti-interference optimization method for an FPC high-performance computing chip interface comprising:

[0006] Perform total variation regularization decomposition processing on the interface signal of FPC to obtain effective data signal components;

[0007] According to the effective data signal component, the structure of the FPC is subjected to three-dimensional bending deformation processing to obtain three-dimensional forming structure parameters;

[0008] Based on the three-dimensional molding structure parameters, the FPC is subjected to a multi-layer electromagnetic shielding treatment to obtain a three-layer composite shielding structure, wherein the first layer is a copper foil layer, the second layer is a high magnetic permeability alloy layer, and the third layer is a conductive polymer layer;

[0009] Inputting the electromagnetic shielding effectiveness data of the three-layer composite shielding structure into a preset residual neural network model for feature extraction to obtain multi-scale feature data;

[0010] Performing covariance matrix tapering on the multi-scale feature data to obtain a reconstructed covariance matrix;

[0011] Functional domain allocation and signal optimization processing are performed based on the reconstructed covariance matrix data to obtain the interference-resistant transmission data stream of each functional domain, and the functional domains include the automatic driving domain, the power domain, the chassis domain, the cockpit domain and the body domain.

[0012] The second aspect of the present application provides an anti-interference optimization device for an FPC high-performance computing chip interface, the anti-interference optimization device for an FPC high-performance computing chip interface comprising:

[0013] A decomposition module is used to perform total variation regularization decomposition processing on the interface signal of the FPC to obtain effective data signal components;

[0014] A deformation processing module, used for performing three-dimensional bending deformation processing on the structure of the FPC according to the effective data signal component to obtain three-dimensional forming structure parameters;

[0015] A shielding processing module, used for performing multi-layer electromagnetic shielding processing on the FPC based on the three-dimensional molding structure parameters to obtain a three-layer composite shielding structure, wherein the first layer is a copper foil layer, the second layer is a high magnetic permeability alloy layer, and the third layer is a conductive polymer layer;

[0016] A feature extraction module, used for inputting the electromagnetic shielding effectiveness data of the three-layer composite shielding structure into a preset residual neural network model for feature extraction to obtain multi-scale feature data;

[0017] A tapering module, used for performing covariance matrix tapering on the multi-scale feature data to obtain a reconstructed covariance matrix;

[0018] The optimization module is used to perform functional domain allocation and signal optimization processing based on the reconstructed covariance matrix data to obtain the anti-interference transmission data stream of each functional domain, and the functional domains include the automatic driving domain, the power domain, the chassis domain, the cockpit domain and the body domain.

[0019] The third aspect of the present application provides an electronic device, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory so that the electronic device executes the above-mentioned anti-interference optimization method of the FPC high-performance computing chip interface.

[0020] The fourth aspect of the present application provides a computer-readable storage medium, which stores instructions. When the computer-readable storage medium is run on a computer, it enables the computer to execute the above-mentioned anti-interference optimization method of the FPC high-performance computing chip interface.

[0021] Compared with the prior art, the present application has the following beneficial effects: through the combined application of total variation regularization decomposition and SCORE algorithm, the effective decomposition of FPC interface signals is achieved, and the effective data signal components are accurately extracted. A three-dimensional bending deformation processing method is adopted to control deformation in three dimensions of X-axis (0-180 degrees), Y-axis (0-90 degrees) and Z-axis (0-45 degrees), effectively optimizing the spatial structure of FPC. A three-layer composite shielding structure consisting of a copper foil layer, a high magnetic permeability alloy layer, and a conductive polymer layer is designed to form a complete electromagnetic shielding system. Based on the improved residual neural network model, combined with the feature pyramid structure and the context guidance module, multi-scale feature extraction of electromagnetic shielding effectiveness data is realized. The covariance matrix tapering processing technology is adopted, and the signal reconstruction effect is improved through the signal subspace reconstruction method. A complete functional domain allocation mechanism including autonomous driving domain, power domain, chassis domain, cockpit domain and body domain is established to optimize the data transmission between functional domains. Data caching and self-recovery mechanisms are designed to enhance the reliability of the system in practical applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0023] The structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification so as to facilitate understanding and reading by persons familiar with this technology. They are not used to limit the conditions under which the present invention can be implemented, and therefore have no substantive technical significance. Any structural modification, change in proportion or adjustment of size, without affecting the effects and purposes that can be achieved by the present invention, should still fall within the scope of the technical contents disclosed by the present invention.

[0024] Figure 1 It is a flow chart of an anti-interference optimization method for an FPC high-performance computing chip interface provided by an embodiment of the present invention;

[0025] Figure 2 It is a schematic block diagram of the structure of an anti-interference optimization device for an FPC high-performance computing chip interface provided by an embodiment of the present invention;

[0026] Figure 3 It is a schematic block diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0027] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0028] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0029] It should also be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in this application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0030] It should be further understood that the term "and / or" used in the specification and appended claims of this application refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations. Figure 1 In the embodiment of the present application, an embodiment of the anti-interference optimization method of the FPC high-performance computing chip interface includes:

[0031] Step 100, performing total variation regularization decomposition processing on the interface signal of the FPC to obtain an effective data signal component;

[0032] It is understandable that the execution subject of the present application can be an anti-interference optimization device for an FPC high-performance computing chip interface, or a terminal or a server, which is not specifically limited here. The present application embodiment is described by taking a server as the execution subject as an example.

[0033] Specifically, the initial signal of the FPC high-performance computing chip interface is collected. The signal data comes from the data transmission between the computing chip and the sensor, and contains the original signal information that may be interfered with. The original signal data is input into the total variation regularization algorithm with a predetermined regularization parameter for iterative decomposition processing. The total variation regularization algorithm controls the balance between the noise and the signal in the signal by adjusting the parameters, so that the signal can retain important information as much as possible after decomposition, while removing unnecessary noise and interference. In this process, the algorithm will iterate repeatedly and gradually optimize the signal decomposition effect until a certain convergence condition is met. When the decomposition algorithm is completed and the decomposition data is obtained, the Euclidean distance calculation is performed. The similarity between different signal components is evaluated to identify which signal components meet specific convergence conditions. By calculating the distance between each component, the signal data with high correlation and meeting the optimization criteria are screened out. The signal components that meet the convergence conditions are input into the SCORE algorithm for signal enhancement preprocessing. The SCORE algorithm enhances the clarity and strength of the signal by enhancing the signal to obtain the preprocessed signal component data. The preprocessed signal component data is timestamped. Assign an accurate time stamp to each signal component so that the signals are sorted and stored according to the time series. After the timestamp is completed, the signal component data with the time stamp is stored in the corresponding data buffer for storage and processing. Each buffer stores signal data of a specific type or a specific period of time. The signal-to-noise ratio is calculated for the classified and stored signal component data. The signal-to-noise ratio is an important indicator to measure the quality of the signal. By calculating the ratio of the signal strength to the noise, the effectiveness and clarity of the signal are judged. If the signal-to-noise ratio of the signal reaches the predetermined threshold requirement, the signal is considered to be a valid data signal component and is used for subsequent processing and analysis. For signals whose signal-to-noise ratio does not meet the requirements, they are automatically discarded or further processed to avoid unqualified signals interfering with subsequent work.

[0034] Step 200, performing three-dimensional bending deformation processing on the structure of the FPC according to the effective data signal component to obtain three-dimensional forming structure parameters;

[0035] Specifically, based on the effective data signal component, the signal feature analysis is performed, and the key structural features are extracted from the data transmission process between the computing chip and the sensor as the basic data for the three-dimensional deformation processing of the FPC. Through analysis, the electrical characteristics, signal transmission paths and stress areas of the FPC under different working environments are identified. X-axis bending analysis is performed on the FPC. The bending angle change of the X-axis is determined, and the bending analysis is usually performed in the range of 0 to 180 degrees. On this basis, the stress concentration area that occurs during the bending process is evaluated by calculating the stress distribution at the X-axis bending point. By calculating these stress areas, the X-axis bending optimization data is obtained, which reflects the stress change trend of the FPC structure in the X-axis direction at a certain bending angle, thereby helping to avoid excessive stress concentration in the design and ensure the reliability of the circuit. Based on the X-axis bending optimization data, the Y-axis bending analysis is performed. Unlike the X-axis bending analysis, the bending angle change of the Y-axis is limited to the range of 0 to 90 degrees. In this process, the stress distribution at the Y-axis bending point is focused on, and the stress state of the FPC material at different bending angles is taken into account during the calculation process. In this way, the Y-axis bending optimization data is obtained, and the structural characteristics of the FPC are further improved to ensure its stress uniformity and anti-interference ability in the Y-axis direction. Based on the Y-axis bending optimization data, the Z-axis torsion analysis is performed. The torsion angle change of the Z-axis is generally set to a range of 0 to 45 degrees. Unlike bending analysis, torsion analysis needs to consider the shear force and stress concentration of the FPC material during the torsion process. By calculating the stress distribution of the Z-axis torsion point, the Z-axis torsion optimization data is obtained, which can help identify structural weaknesses or stress concentration areas that appear during the torsion process. Model the material stress distribution and establish a three-dimensional stress distribution model containing the X-axis, Y-axis and Z-axis stress distribution. This model comprehensively considers the bending and torsion of the FPC structure in different directions, and can provide a global perspective for subsequent material selection and structural optimization. Through the analysis of the three-dimensional stress distribution model, the various mechanical stresses to which the FPC is subjected in actual applications can be more accurately predicted to ensure that it can operate stably in various complex working environments. Material performance matching is performed on the three-dimensional stress distribution model data. According to the stress distribution in different directions, the appropriate material is selected to match the design requirements of the FPC. By matching the material properties, the mechanical and electrical properties of the FPC are optimized, making it more reliable in terms of anti-interference and signal transmission. At the same time, the choice of materials will also affect the flexibility and durability of the FPC, thereby providing the best physical support for the final sensor module. After the material performance matching is completed, the material structure parameter data is obtained. According to the material structure parameter data, the spatial layout of the sensor module is optimized to ensure that the connection method and arrangement between the modules can minimize signal interference and improve the stability and efficiency of signal transmission.Through this process, the optimal connection method between modules is determined and the module space layout data is obtained. Based on the module space layout data, the transmission path is calculated. By optimizing the transmission path, it is ensured that the signal transmission can achieve the best effect in the shortest time, while reducing the loss and interference during the transmission process. Through the precise calculation of the transmission path, the three-dimensional molding structure parameters are finally obtained.

[0036] Step 300: Based on the three-dimensional molding structure parameters, the FPC is subjected to multi-layer electromagnetic shielding treatment to obtain a three-layer composite shielding structure, wherein the first layer is a copper foil layer, the second layer is a high magnetic permeability alloy layer, and the third layer is a conductive polymer layer;

[0037] It should be noted that the electromagnetic shielding analysis is performed on the parameters of the three-dimensional molding structure to determine the coverage and thickness distribution of the shielding layer. By simulating the electromagnetic characteristics of the three-dimensional structure, the electromagnetic interference source in the signal transmission path is evaluated, and the layout of the shielding layer is reasonably planned to ensure effective shielding of external electromagnetic interference and maintain signal integrity. After completing the electromagnetic shielding analysis, the initial parameter data of the shielding structure is obtained. Based on the initial parameter data of the shielding structure, the structure design of the first copper foil layer is carried out. As the first shielding structure, the copper foil layer mainly blocks electromagnetic interference in the low-frequency and medium-frequency ranges. By designing the thickness, conductivity and contact method of the copper foil layer with the FPC, it is ensured that the copper foil layer has sufficient shielding effectiveness within the effective frequency band. The flexibility and strength of the copper foil layer are considered in the design process to ensure that it can adapt to the bending and deformation requirements of the FPC and avoid the problem of breakage or detachment during long-term use. In the design process of the copper foil layer, factors such as material cost, processing technology and environmental adaptability are comprehensively considered to achieve an ideal anti-interference effect. After the copper foil layer structure design is completed, the second layer of high magnetic permeability alloy layer is designed according to the structural parameter data of the copper foil layer. The high magnetic permeability alloy layer is used to shield high-frequency electromagnetic interference, especially for areas with higher electromagnetic wave frequencies. By selecting suitable alloy materials, optimizing their magnetic permeability and conductivity, it is ensured that the high magnetic permeability alloy layer can effectively shield the propagation of electromagnetic waves in actual work. The thickness design of the high magnetic permeability alloy layer is also very important, and it is precisely adjusted according to actual needs and shielding effectiveness requirements to maximize its electromagnetic shielding performance. At the same time, considering the flexibility requirements of FPC, the material selection and structural design of the high magnetic permeability alloy layer should ensure that it has sufficient elasticity to adapt to the bending and deformation of FPC without affecting its shielding effect. After completing the design of the high magnetic permeability alloy layer, the third layer of conductive polymer layer is designed based on the structural parameter data of the high magnetic permeability alloy layer. The conductive polymer layer mainly plays the role of enhancing shielding performance and improving structural flexibility. Compared with traditional metal shielding layers, conductive polymer materials have higher flexibility and can adapt to the bending and deformation requirements of FPC while maintaining good conductivity. When designing the conductive polymer layer, the conductivity, mechanical strength and environmental adaptability of the material are comprehensively considered. By properly selecting the type and thickness of the material, the conductive polymer layer is ensured to have good electromagnetic interference shielding capabilities in the high-frequency range, thus achieving comprehensive protection against complex electromagnetic environments. After completing the design of the three-layer shielding layer, the grounding position of the three-layer composite shielding layer is designed based on the structural parameter data of the conductive polymer layer. Grounding design is a key link in electromagnetic shielding. Reasonable grounding can significantly improve shielding effectiveness and reduce the impact of electromagnetic radiation on the system. Setting a grounding point at the connection of the three-layer shielding structure helps to effectively connect the multi-layer shielding structure with the system's grounding system to achieve good electromagnetic compatibility.In the process of grounding design, in addition to considering the location of the grounding point, it is also necessary to evaluate the grounding resistance, grounding method and the selection of grounding materials to ensure that the grounding system of the entire shielding structure can work stably and avoid electromagnetic interference problems caused by poor grounding. Based on the grounding structure parameter data, signal suppression is designed. Electromagnetic interference is suppressed by setting a common mode choke at the target position of the signal transmission channel. Common mode chokes can effectively reduce the noise caused by common mode interference in the transmission path and ensure the stability and clarity of signal transmission. By reasonably designing the parameters of the choke, such as the material, size, impedance, etc. of the coil, common mode interference can be suppressed to the greatest extent and the anti-interference ability of the entire system can be improved. The EMI (electromagnetic interference) suppression parameters are carefully adjusted to ensure that excellent anti-interference effects can be achieved in different frequency bands and working environments. According to the EMI suppression parameter data, the signal line protection structure is designed to ensure that the signal line is not affected by external interference during transmission, while reducing the electromagnetic interference generated by the signal line itself. The signal line protection structure needs to select appropriate shielding materials and structural forms according to specific usage requirements to achieve the best signal protection effect. At this stage, the layout of the signal line, the selection of shielding materials, and the spacing between the signal line and other circuits all need to be precisely designed to ensure optimal anti-interference capabilities. The signal line protection parameter data is integrated into a three-layer composite shielding structure to obtain a complete electromagnetic shielding system.

[0038] Step 400: Input the electromagnetic shielding effectiveness data of the three-layer composite shielding structure into a preset residual neural network model for feature extraction to obtain multi-scale feature data;

[0039] Specifically, the electromagnetic shielding effectiveness data of the three-layer composite shielding structure is input into the preset residual neural network model for processing. The electromagnetic shielding effectiveness data is initially extracted by the first convolutional layer with a ReLU activation function. The convolution kernel size of the convolutional layer is 3×3, the step size is 1, and the number of output channels is 64. Through these parameter settings, the convolutional layer can extract the local features of the input data and convert them into more abstract high-dimensional features. This process can effectively capture the preliminary information of the electromagnetic shielding effectiveness. Through the processing of the first convolutional layer, the first layer of feature data obtained contains a preliminary representation of the electromagnetic shielding effectiveness. The first layer of feature data is processed by the first residual block. The first residual block consists of two convolutional layers and a short-circuit connection, the number of output channels of each convolutional layer is 128, and a batch normalization layer is added after each convolutional layer. The role of the batch normalization layer is to accelerate the training process, avoid gradient disappearance, and help the model maintain a more stable performance during training. Through this processing, the network can effectively extract more complex features while maintaining the stability of the deep network. The addition of residual connections enables information to be transmitted more effectively in the deep layers of the network, thereby avoiding the degradation problem that occurs when the network depth increases. The first residual feature data is obtained through the processing of the first residual block. The first residual feature data is input into the context guidance module for feature enhancement to generate enhanced feature data. The module sets the attention weight threshold to 0.6 and performs feature selection through a dynamic weight adjustment mechanism to strengthen the important features in the electromagnetic shielding effectiveness data. The context guidance module can focus on the most critical feature areas according to the contextual relationship of the input data to improve the expressiveness and effectiveness of the features. Based on the data after feature enhancement, feature pyramid processing is performed. The feature pyramid structure can effectively capture the multi-scale information in the data and fuse features from different scales by constructing multiple layers of feature maps at different scales. By constructing a five-layer feature pyramid structure, the number of channels of each layer of feature maps is 64, 128, 256, 512, and 1024 respectively. The network can extract rich features at multiple scales and fuse them to obtain multi-scale fused feature data. The multi-scale fused feature data is input into the second residual block for deep feature extraction to obtain deep feature data. In the second residual block, two convolutional layers are included with an output channel number of 256, and a 1×1 convolutional layer is added at the residual connection for dimensionality reduction. Through these operations, the network can extract deeper feature information and reduce the amount of computation and memory consumption while maintaining the network depth. The role of the 1×1 convolutional layer is to reduce the number of channels of the feature map, thereby optimizing the computing performance and avoiding redundant information caused by the complexity of the feature map. Cross-layer feature connection processing is performed on the deep feature data. Through cross-layer feature connection, feature maps of different scales are unified to the same resolution through deconvolution operations and spliced ​​in the channel dimension.The deconvolution operation adjusts the resolution of the feature map so that information from different levels can be fused at the same scale to obtain multi-scale feature connection data. Based on the multi-scale feature connection data, the feature fusion module performs weighted fusion processing. The channel attention mechanism is used to adaptively assign weights to features of different scales. The channel attention mechanism can automatically adjust the weight of each channel in the fusion process according to its importance, strengthen the features that contribute most to the electromagnetic shielding effectiveness, and obtain multi-scale feature data.

[0040] Step 500, performing covariance matrix tapering on the multi-scale feature data to obtain a reconstructed covariance matrix;

[0041] Specifically, the distribution characteristics of multi-scale feature data are analyzed, the statistical distribution pattern of multi-scale feature data is identified, and a tapered matrix is ​​constructed based on the Laplace distribution. The characteristics of the Laplace distribution enable it to provide more accurate modeling in the signal processing process, especially in a noisy environment. When constructing the tapered matrix, the dimension of the matrix is ​​set to n×n, where n is the number of signal channels, which means that each element of the matrix represents the relationship between different signal channels. The covariance of the tapered matrix is ​​calculated to obtain the signal covariance matrix R, where the size of R is n×n. In the covariance matrix, the diagonal elements represent the variance of the signal, while the non-diagonal elements represent the covariance between the signal channels. The covariance matrix provides a quantitative description of the relationship between different signal channels, which can effectively reflect the distribution characteristics of the signal in space and the correlation between different signals. The covariance matrix R is tapered. The tapered parameter α is set to 0.8, and this setting value affects the degree of compression during the matrix tapering process. By setting the tapering parameter α, the sparsity of the matrix is ​​effectively controlled, the redundant information in the data is reduced, the computational efficiency is improved, and the tapered matrix parameter data is obtained. The tapering process compresses and optimizes the covariance matrix to make the structure of the matrix more concise while retaining the main information of the signal. The tapered matrix parameter data is reconstructed by matrix multiplication. The matrix multiplication operation can effectively transform the covariance matrix to obtain a new matrix representation. In the reconstruction process, the matrix is ​​adjusted so that it can remove redundancy and noise while maintaining the signal characteristics. After the matrix reconstruction is completed, the minimum variance distortion-free response optimization is performed. The minimum variance distortion-free response optimization improves the signal quality by minimizing the signal variance while ensuring that no distortion is introduced, and obtains the optimized matrix data. The optimized matrix data is subjected to eigenvalue decomposition. By decomposing the matrix data into the form of eigenvalues ​​and eigenvectors, the key information in the signal is extracted. For the decomposed eigenvalues, they are sorted according to size, and the dimension of the signal subspace is determined by the minimum description length criterion. The minimum description length criterion is an information criterion that reduces the complexity of the model and improves the accuracy and efficiency of signal processing by selecting the dimension that can most effectively represent the data. Through this criterion, the most important eigenvectors are selected based on the eigenvalues ​​to determine the main components of the signal and obtain the eigendecomposition data. The signal subspace is reconstructed for the eigendecomposition data, and the eigenvectors corresponding to the largest k eigenvalues ​​are selected. These eigenvectors contain the main features of the signal and can effectively represent the core information of the signal. By combining these eigenvectors, the subspace representation of the signal is reconstructed. In the reconstruction process, by selecting and combining the eigenvectors, the representation of the signal can be made more accurate, and unimportant components can be removed to obtain subspace reconstructed data.The matrix tapering optimization is performed on the subspace reconstruction data to improve the interpretability and computational efficiency of the matrix while ensuring that the main characteristics of the signal are not affected, and the reconstructed covariance matrix is ​​obtained.

[0042] Step 600: Perform functional domain allocation and signal optimization processing based on the reconstructed covariance matrix data to obtain the interference-resistant transmission data stream of each functional domain, where the functional domains include the autonomous driving domain, the power domain, the chassis domain, the cockpit domain, and the body domain.

[0043] Specifically, the reconstructed covariance matrix data is distributed to the controllers of each functional domain for regional centralized processing. The matrix data is distributed according to the requirements of different functional domains, and the controller of each functional domain is responsible for processing the data in its specific area. The functional domains include the autonomous driving domain, the power domain, the chassis domain, the cockpit domain, and the body domain. The data flow requirements of these domains are different. The signal area of ​​each functional domain is centrally controlled to ensure the efficiency and stability of data transmission, and the initial allocation data of the functional domain is obtained. The data cache structure is processed for the initial allocation data of the functional domain. In order to improve the efficiency of data processing and avoid data delays during transmission, a 256KB cache area is set using a circular queue structure. The setting of the cache area can effectively store and manage data, and dynamically read and write management is performed according to the data transmission priority of each functional domain to obtain the data cache configuration parameters. Based on the data cache configuration parameters, the functional domain data channel is established. The bandwidth of the data exchange channel between each functional domain is set to ensure that the data transmission channel of each functional domain can meet its data flow requirements. The data transmission requirements of different functional domains are different, and the data channel bandwidth of each functional domain is adjusted according to the actual situation. Through reasonable bandwidth allocation, data transmission conflicts between different functional domains can be effectively avoided to ensure efficient data transmission. The data transmission routes between functional domains are optimized and configured. By optimizing the transmission path, the loss and interference in the data transmission process are reduced, the signal transmission quality is improved, and the channel allocation data is obtained. The channel allocation data is tested for real-time signal quality to ensure that the signal quality during data transmission is always within an acceptable range. The signal quality detection data is compared with the preset threshold. If the signal quality is detected to be lower than the set standard, corresponding processing is performed. The detection process can timely discover problems in the signal and provide data basis for subsequent signal optimization. By detecting the signal quality, it is ensured that the data transmission between the functional domains is carried out under good signal quality assurance, thereby minimizing the impact of interference on data transmission. After the signal quality detection is completed, the signal transmission status evaluation processing is performed. When the signal-to-noise ratio of the signal is lower than the target value, the automatic retransmission mechanism is started to ensure that the lost data can be retransmitted in time. This mechanism effectively ensures the integrity and accuracy of the data and avoids data loss or transmission errors in harsh signal environments. During the retransmission process, the number and results of each retransmission are recorded. By statistically analyzing the number of retransmissions and retransmission results, the stability and reliability of signal transmission are evaluated, and optimization suggestions are provided. Fault location analysis is performed based on transmission status evaluation data. By analyzing data transmission anomalies in each functional domain, potential fault points can be identified, and fault location data can be generated to help locate problems that occur during data transmission, such as signal interference, hardware failures, and other factors. After completing fault location, self-recovery processing is performed. By analyzing the fault location data, automatic repair or adjustment measures are taken to restore the normal operation of the system.The self-recovery execution data can ensure that the system can quickly recover to a normal state when encountering a fault, thereby minimizing system downtime and performance degradation. Through self-recovery processing, the system can maintain efficient and stable operation in a real-time environment, thereby ensuring that the data transmission of each functional domain is not interrupted. After the self-recovery execution data is processed, the data is input into the functional domain optimization module for data flow optimization. At this stage, the data transmission of each functional domain is controlled in real time under the interference-to-signal ratio condition within the preset range. By dynamically adjusting the transmission rate and priority of the data flow, the anti-interference ability of each functional domain is improved, ensuring that data transmission can still be carried out efficiently and stably in a complex and unstable signal environment, and obtaining the anti-interference transmission data flow of each functional domain.

[0044] In the embodiment of the present application, through the combined application of total variation regularization decomposition and SCORE algorithm, the effective decomposition of the FPC interface signal is achieved, and the effective data signal component is accurately extracted. A three-dimensional bending deformation processing method is adopted to carry out deformation control in three dimensions of the X-axis (0-180 degrees), the Y-axis (0-90 degrees) and the Z-axis (0-45 degrees), and the spatial structure of the FPC is effectively optimized. A three-layer composite shielding structure consisting of a copper foil layer, a high magnetic permeability alloy layer, and a conductive polymer layer is designed to form a complete electromagnetic shielding system. Based on the improved residual neural network model, combined with the feature pyramid structure and the context guidance module, the multi-scale feature extraction of electromagnetic shielding effectiveness data is realized. The covariance matrix tapering processing technology is adopted, and the signal reconstruction effect is improved by the signal subspace reconstruction method. A complete functional domain allocation mechanism including the autonomous driving domain, the power domain, the chassis domain, the cockpit domain and the body domain is established to optimize the data transmission between the functional domains. Data caching and self-recovery mechanisms are designed to enhance the reliability of the system in practical applications.

[0045] In a specific embodiment, the process of executing step 100 may specifically include the following steps:

[0046] The initial signal of the FPC high-performance computing chip interface is collected to obtain the original signal data containing the data transmission information between the computing chip and the sensor;

[0047] Inputting the original signal data into a total variation regularization algorithm with a predetermined regularization parameter for iterative decomposition to obtain decomposed data;

[0048] Perform Euclidean distance calculation on the decomposed data to obtain component data that meets the convergence conditions;

[0049] The component data that meets the convergence condition is input into the SCORE algorithm for signal enhancement preprocessing to obtain the preprocessed signal component data;

[0050] Performing time stamp marking on the preprocessed signal component data to obtain signal component data with time stamp, and storing the signal component data with time stamp into corresponding data buffers for storage processing to obtain classified stored signal component data;

[0051] The signal-to-noise ratio is calculated for the classified and stored signal component data to obtain signal quality characteristic data, and a threshold value is judged for the signal quality characteristic data to output a valid data signal component that meets the signal-to-noise ratio requirement.

[0052] Specifically, the data transmission between the computing chip and the sensor is signal collected to obtain the original signal data containing the data transmission information. The original signal data is input into the total variation regularization algorithm with a predetermined regularization parameter set for iterative decomposition. The total variation regularization algorithm is a technology used for signal denoising and decomposition, which can effectively separate the useful information and noise components in the signal. In this process, through continuous iteration, the algorithm gradually optimizes the result of signal decomposition according to the predetermined regularization parameter. The setting of the regularization parameter depends on the characteristics of the signal and the denoising requirements. λ is used as the adjustment factor. When the λ value is large, the denoising effect will be stronger, but more detail information will be lost. Conversely, more detail information will be retained, but the denoising effect will be weaker. The Euclidean distance of the decomposed data is calculated. The signal is further optimized so that the obtained component data meets certain convergence conditions. Euclidean distance is a method to measure the similarity between two signal data. Its calculation formula is:

[0053] ;

[0054] in, and There are two signal data in the first The values ​​in the dimensions, is the number of dimensions of the signal data, is the Euclidean distance between the two signals. By calculating the Euclidean distance, the similarity between the two signal data is evaluated. When the distance is less than a certain threshold, it means that the data is close to convergence. The convergence condition is set as the distance value is lower than a predetermined threshold. When, for example, When , it means that the signal has converged and meets the processing requirements. The component data that meets the convergence conditions are input into the SCORE algorithm for signal enhancement preprocessing. The SCORE algorithm is a technology that improves signal quality by optimizing and enhancing signal components. According to the characteristic information of the signal, it is adaptively enhanced. The SCORE algorithm can remove noise components and strengthen the effective information in the signal through multiple processing and optimization of component data to obtain preprocessed signal component data. The signal component data is timestamped to provide time series information for the signal data. The signal component data with time stamps are stored in the corresponding data buffers for storage. The buffer uses a first-in-first-out method to manage data and sets a reasonable cache strategy according to the size of the buffer area. For example, a 256KB buffer area is used to cache data through a circular queue structure. When the buffer area reaches capacity, the new data will overwrite the oldest data to ensure that the latest data is stored in time. The signal-to-noise ratio is calculated for the classified stored signal component data. The signal-to-noise ratio is an important indicator for measuring signal quality, which represents the ratio of signal power to noise power. The higher the signal-to-noise ratio, the better the signal quality and the less interference. The signal-to-noise ratio is calculated as:

[0055] ;

[0056] in, is the power of the signal, is the power of the noise. The signal power is estimated by the mean of the square values ​​of the signal, while the noise power is the mean of the square values ​​of the noise part of the signal. The calculated signal-to-noise ratio helps to judge the quality of the signal. If the signal-to-noise ratio is too low, it means that the signal is subject to greater interference and requires further optimization or enhancement. Perform threshold judgment on the signal quality feature data, and output the valid data signal component that meets the signal-to-noise ratio requirement according to the preset signal-to-noise ratio requirement. Set a reasonable signal-to-noise ratio threshold. When the signal-to-noise ratio is greater than a certain set threshold, the signal quality is considered to be good and is used for subsequent data transmission and processing; when the signal-to-noise ratio is lower than the threshold, it means that the signal quality is poor and further denoising or retransmission is required.

[0057] In a specific embodiment, the process of executing step 200 may specifically include the following steps:

[0058] Based on the effective data signal components, the signal characteristics are analyzed to extract the data transmission structure characteristics between the computing chip and the sensor, and the basic data of the three-dimensional deformation of the FPC is obtained;

[0059] Perform X-axis bending analysis on the basic data of FPC 3D deformation, determine the change in X-axis bending angle within the range of 0-180 degrees, and calculate the stress distribution of the X-axis bending point to obtain X-axis bending optimization data;

[0060] Based on the X-axis bending optimization data, the Y-axis bending analysis is performed to determine the bending angle variation of the Y-axis within the range of 0-90 degrees, and the stress distribution of the Y-axis bending point is calculated to obtain the Y-axis bending optimization data;

[0061] Based on the Y-axis bending optimization data, the Z-axis torsion analysis is performed to determine the Z-axis torsion angle variation within the range of 0-45 degrees, and the stress distribution of the Z-axis torsion point is calculated to obtain the Z-axis torsion optimization data;

[0062] Model the material stress distribution of the Z-axis torsion optimization data and establish a three-dimensional stress distribution model data including the X-axis, Y-axis and Z-axis stress distribution;

[0063] The material properties of the three-dimensional stress distribution model data are matched to obtain the material structure parameter data, and the spatial layout of the sensor module is optimized according to the material structure parameter data to determine the optimal connection method between the modules and obtain the module spatial layout data;

[0064] The transmission path is calculated based on the module space layout data to obtain the three-dimensional molding structure parameters.

[0065] Specifically, the structural characteristics of data transmission between the computing chip and the sensor are extracted from the effective data signal components. By performing frequency domain analysis and time domain feature extraction on the effective data signal components, the structural characteristics of data transmission, such as signal transmission delay, frequency characteristics, and signal strength, are obtained to reveal the physical path of signal transmission in the FPC. Write it down and perform a three-dimensional deformation analysis of the FPC on the structural characteristics. As a flexible circuit board, FPC has good flexibility, but during signal transmission, the deformation of FPC will affect the quality of the signal. Perform X-axis bending analysis on the basic data of the three-dimensional deformation of FPC. The change in the bending angle of FPC on the X-axis needs to be calculated within the range of 0 to 180 degrees. According to the change in the bending angle, the stress distribution at different bending angles is derived. The bending stress is determined by the elastic modulus and bending angle of the FPC material. The formula for the bending stress is:

[0066] ;

[0067] in, is the bending stress on the X-axis, is the bending moment caused by the bending angle, is the second moment of area of ​​the FPC, It is the distance from any point in the material to the neutral axis. By calculating the stress distribution of the X-axis bending point, the X-axis bending optimization data is obtained. This data helps evaluate the mechanical properties of the FPC during the bending process and ensures that the signal transmission quality is not affected during the deformation process. Y-axis bending analysis is performed based on the X-axis bending optimization data. Due to the three-dimensional structure of the FPC, the bending characteristics of the Y-axis will be affected by the X-axis bending results, and the change in the bending angle on the Y-axis needs to be determined within the range of 0 to 90 degrees. The stress distribution at the Y-axis bending point is also affected by the material properties, bending angle and external force. In this process, a bending stress formula similar to the X-axis is used for calculation to obtain the Y-axis bending optimization data. For FPC, the Y-axis bending optimization data helps to determine the stress concentration of the FPC after bending, and then optimize the design to ensure that the FPC can maintain good signal transmission performance during the bending process. Based on the Y-axis bending optimization data, the Z-axis torsion analysis is processed. The Z-axis torsion analysis mainly involves the change in the torsion angle of the FPC on the Z-axis, which is usually between 0 and 45 degrees. In the torsion analysis of the Z axis, torsional stress will also affect the stability of signal transmission. The torsional stress distribution on the Z axis is determined by the torque and the polar moment of inertia of the material, and the calculation formula is:

[0068] ;

[0069] in, is the torsional stress in the Z axis, is the torque, is the polar moment of inertia of the FPC cross section, is the distance from the torsion axis. The Z-axis torsion optimization data is obtained by calculating the stress distribution of the Z-axis torsion point. The material stress distribution model is performed on the Z-axis torsion optimization data, and the three-dimensional stress distribution model data including the stress distribution of the X-axis, Y-axis and Z-axis is established. Based on the three-dimensional stress distribution model, the material properties of the FPC are matched to determine the best material selection. The material selection achieves the best balance in terms of strength, durability and conductivity to ensure that the FPC is not interfered with during the signal transmission between the high-performance computing chip and the sensor. The material structure parameter data obtained by material performance matching provides a basis for the subsequent optimization of the sensor module space layout. The spatial layout optimization aims to achieve the optimal performance of signal transmission by reasonably arranging each sensor module and signal transmission path. Based on the material structure parameter data, the connection method between modules is optimized by finite element analysis and other methods. The module space layout data is obtained by calculation to determine the optimal position of each sensor module to avoid signal distortion due to excessive bending or stress concentration. Based on the module space layout data, the transmission path is calculated to determine the optimal transmission path of the signal. The physical shape of the FPC, the spatial relationship between modules and various electromagnetic interference factors in the signal transmission process are comprehensively considered. The goal of transmission path calculation is to reduce signal loss and interference during transmission, thereby ensuring that the signal can be stably transmitted to the target location. The calculation formula of the transmission path usually involves factors such as electromagnetic wave propagation and transmission medium loss, and is expressed by the following simplified formula:

[0070] ;

[0071] in, is the signal output power, is the signal input power, is the attenuation coefficient of the transmission medium, It is the distance of signal transmission. By optimizing the transmission path, signal loss is minimized while ensuring signal quality, and finally the three-dimensional molding structure parameters are obtained.

[0072] In a specific embodiment, the process of executing step 300 may specifically include the following steps:

[0073] Conduct electromagnetic shielding analysis on the parameters of the three-dimensional molding structure, determine the coverage and thickness distribution of the shielding layer, and obtain the initial parameter data of the shielding structure;

[0074] Performing a first copper foil layer structure design on the initial parameter data of the shielding structure to obtain copper foil layer structure parameter data;

[0075] Design the structure of the second high magnetic permeability alloy layer according to the copper foil layer structure parameter data to obtain the high magnetic permeability alloy layer structure parameter data;

[0076] Design the structure of the third conductive polymer layer according to the structural parameter data of the high magnetic permeability alloy layer, and obtain the structural parameter data of the conductive polymer layer;

[0077] The grounding position of the three-layer composite shielding layer is designed based on the structural parameter data of the conductive polymer layer, and the grounding point is set at the connection of the three-layer shielding structure to obtain the grounding structure parameter data;

[0078] Signal suppression is performed based on grounding structure parameter data, and a common mode choke is set at a target position of the signal transmission channel to obtain EMI suppression parameter data;

[0079] A signal line protection structure is designed according to the EMI suppression parameter data to obtain the signal line protection parameter data, and a three-layer composite shielding structure is integrated on the signal line protection parameter data to obtain the three-layer composite shielding structure.

[0080] Specifically, by performing electromagnetic shielding analysis on the parameters of the three-dimensional molding structure, the coverage and thickness distribution of the shielding layer are determined, and the initial parameter data of the shielding structure are obtained. The goal of electromagnetic shielding is to reduce electromagnetic interference and ensure the stability and clarity of the signal during transmission. The effect of electromagnetic shielding is closely related to factors such as the conductivity, magnetic permeability and thickness of the shielding material. When performing shielding analysis, a model of the shielding layer material and geometry is established, and the propagation path of the electromagnetic wave is calculated. According to the electromagnetic characteristics of the shielding layer, preliminary shielding parameters such as the thickness, surface conductivity, and magnetic permeability of the shielding layer are obtained, and the electromagnetic shielding effect of each layer of material in practical applications is further analyzed. According to the initial parameter data of the shielding structure, the structure design of the first copper foil layer is carried out. As the base layer of electromagnetic shielding, the copper foil layer mainly reflects and absorbs electromagnetic waves. The conductivity of copper enables it to effectively isolate external electromagnetic interference and provide electromagnetic shielding between chip signal transmission channels. When designing the copper foil layer, it is necessary to consider the thickness of copper, the laying method, and the contact resistance of the copper layer. The thickness of the copper foil layer should be designed according to the signal frequency and transmission distance to ensure that it has sufficient shielding effect and low resistance. The resistance of the copper foil layer is expressed by the following formula:

[0081] ;

[0082] in, is the resistance of the copper foil layer, is the resistivity of copper, is the length of the copper foil, is the cross-sectional area of ​​the copper foil. By adjusting the thickness and design of the copper foil layer, the copper foil layer can achieve the expected electromagnetic shielding effect, and the structural parameter data of the copper foil layer is obtained. The structural design of the second layer of high magnetic permeability alloy layer is carried out. The main function of the high magnetic permeability alloy layer is to absorb electromagnetic waves and effectively convert electromagnetic waves into heat to reduce reflected signals. High magnetic permeability alloys use materials with higher magnetic permeability, such as iron-nickel alloys, for electromagnetic shielding in high frequency ranges. At this stage, the structural parameters of the high magnetic permeability alloy layer are calculated according to the design parameters of the copper foil layer, including the thickness, magnetic permeability, and magnetic permeability of the alloy layer. The magnetic permeability of the alloy layer affects its ability to absorb electromagnetic waves. Assume that the magnetic permeability of the material is , according to the magnetic properties of the material, its absorption degree of electromagnetic waves is calculated and described by the following formula:

[0083] ;

[0084] in, is the electromagnetic wave power absorbed by the material, is the angular frequency of the electromagnetic wave, is the magnetic field strength, It is the magnetic permeability of the material. By designing the magnetic permeability and thickness of the alloy layer, it is ensured that the alloy layer absorbs electromagnetic waves within the effective range, thereby optimizing the shielding effect and obtaining the structural parameter data of the high magnetic permeability alloy layer. According to the structural parameter data of the high magnetic permeability alloy layer, the structure of the third conductive polymer layer is designed. As the last layer in the shielding structure, the conductive polymer layer mainly reflects and absorbs electromagnetic waves through its conductivity. At the same time, it has good flexibility and is suitable for the design of flexible circuit boards for high-performance computing chip interfaces. Conductive polymer materials have high electrical conductivity and good anti-interference ability, and are suitable for use in signal transmission systems requiring high frequency and high density. When designing the conductive polymer layer, it is mainly necessary to determine its thickness, conductive properties and contact resistance with the first two layers. The electrical conductivity of the conductive polymer material is expressed by the following formula:

[0085] ;

[0086] in, is the conductivity of the conductive polymer material, is the resistivity of the material. By calculating the conductivity and thickness of the conductive polymer layer, the structural parameter data of the conductive polymer layer is obtained, and its role in the shielding structure is ensured to be realized. Based on the structural parameter data of the conductive polymer layer, the grounding position of the three-layer composite shielding layer is designed. Grounding is an important part of electromagnetic shielding design. By grounding the shielding structure, the shielding effect is improved and the reflection of electromagnetic waves is reduced. The location and method of grounding directly affect the propagation path and attenuation of electromagnetic waves. The grounding point is set at the connection of the shielding structure, that is, the junction of the copper foil layer, the high magnetic permeability alloy layer and the conductive polymer layer, forming an effective current loop, thereby maximally suppressing electromagnetic interference. The parameter data of the grounding structure is determined by analyzing the current distribution of the shielding structure. Based on the grounding structure parameter data, EMI (electromagnetic interference) is suppressed for the signal transmission channel. In order to reduce electromagnetic interference during signal transmission, a common mode choke is set at the key position of the signal transmission channel. The common mode choke effectively suppresses the common mode noise in the transmission signal and improves the quality of the signal. The effect of EMI suppression is determined by the frequency characteristics of the choke, and the frequency characteristics are calculated by the following formula:

[0087] ;

[0088] in, is the impedance of the common mode choke, is the frequency of the signal, is the inductance of the choke. By adjusting the design parameters of the common mode choke, electromagnetic interference can be effectively suppressed, and EMI suppression parameter data can be obtained. Based on the EMI suppression parameter data, the signal line protection structure is designed. The design purpose of signal line protection is to protect the signal integrity during signal transmission by reducing electromagnetic interference. Signal line protection uses highly conductive materials to ensure stable signal transmission. When designing the signal line protection structure, a comprehensive analysis is performed based on the shielding structure, EMI suppression, and the transmission characteristics of the signal line, and finally the design parameter data of the signal line protection is obtained, and the three-layer composite shielding structure is integrated to ensure that the electromagnetic shielding effect of the FPC high-performance computing chip interface is optimal, and to ensure the stability and reliability of the signal in a complex electromagnetic environment.

[0089] In a specific embodiment, the process of executing step 400 may specifically include the following steps:

[0090] The electromagnetic shielding effectiveness data of the three-layer composite shielding structure is input into the preset residual neural network model, and feature extraction is performed through the first convolution layer with a ReLU activation function. The convolution kernel size of the first convolution layer is 3×3, the step size is 1, and the number of output channels is 64, so as to obtain the first layer feature data;

[0091] Performing a first residual block processing on the first layer feature data, wherein the first residual block includes two convolutional layers and a short-circuit connection, the number of output channels of each convolutional layer is 128, and a batch normalization layer is added after the convolutional layer to obtain the first residual feature data;

[0092] The first residual feature data is input into the context guidance module for feature enhancement, wherein the attention weight threshold is set to 0.6, and feature selection is performed through dynamic weight adjustment to obtain enhanced feature data;

[0093] Perform feature pyramid processing on the enhanced feature data to construct a 5-layer feature pyramid structure. The number of feature map channels in each layer is 64, 128, 256, 512, and 1024, respectively, to obtain multi-scale fusion feature data;

[0094] The multi-scale fusion feature data is input into the second residual block for deep feature extraction, where the second residual block contains two convolutional layers with an output channel number of 256, and a 1×1 convolutional layer is added at the residual connection for dimensionality reduction to obtain deep feature data;

[0095] Perform cross-layer feature connection processing on deep feature data, unify feature maps of different scales to the same resolution through deconvolution operation, and perform channel dimension splicing to obtain multi-scale feature connection data;

[0096] The multi-scale feature connection data is input into the feature fusion module for weighted fusion processing, and the channel attention mechanism is used to adaptively allocate weights to features of different scales to obtain multi-scale feature data.

[0097] Specifically, the electromagnetic shielding effectiveness data of the three-layer composite shielding structure is input into the preset residual neural network model, and feature extraction is performed through the convolution layer. The input data is processed through a convolution layer with a RelU activation function set to enhance the features of the data in a nonlinear activation manner. The convolution kernel size of the first convolution layer is set to , the step size is 1, and the number of output channels is set to 64, so as to extract preliminary features from the input data. The function of the convolution layer is to filter the input signal through the convolution operation, so as to highlight the features of different frequency ranges and spatial positions. The calculation formula of this convolution layer is expressed as:

[0098] ;

[0099] in, is the output feature map, is the input data, is the convolution kernel, is the bias term, is the ReLU activation function, defined as:

[0100] ;

[0101] In this way, preliminary feature information is extracted from the electromagnetic shielding effectiveness data and passed to the next layer of the network for deeper analysis. The first layer of feature data is processed by the first residual block, which alleviates the gradient vanishing problem in the deep neural network training process by introducing short-circuit connections. The first residual block contains two convolutional layers, each with an output channel number of 128, and a batch normalization layer is added after the convolutional layer to help accelerate the network training process and improve the stability of the training. The role of the batch normalization layer is to reduce internal covariate shift by normalizing the output data of each layer, thereby accelerating the convergence speed of the neural network. The formula for the batch normalization operation is as follows:

[0102] ;

[0103] in, is the mean of the mini-batch data, is the variance, is a small constant used to prevent division by zero errors. is the standardized output. Through this processing, the first residual block can effectively enhance the feature representation, extract deeper electromagnetic shielding effectiveness features, and obtain the first residual feature data. The first residual feature data is input into the context guidance module for feature enhancement. The context guidance module selects the most important features by setting the attention weight threshold to 0.6 and adopts dynamic weight adjustment. The model adaptively selects the features most relevant to the electromagnetic shielding effectiveness and suppresses those redundant or irrelevant features. Through the feature selection mechanism, the model can improve the accuracy of electromagnetic shielding optimization. The feature enhancement process of the context guidance module is similar to the application of the attention mechanism, in which each feature is assigned a weight, and features with high weights are more important. Assume that the feature set is , and its corresponding weight is , then the enhanced feature set is calculated by weighted average:

[0104] ;

[0105] The enhanced feature data can reflect more accurate electromagnetic shielding effectiveness information, and make the model more focused on important signal features, thereby improving the final optimization effect. The enhanced feature data is processed using a feature pyramid. The feature pyramid network can effectively process multi-scale features and construct multi-level feature maps to enhance the network's expressiveness at different resolutions. The feature pyramid captures information of different scales by gradually enlarging the size of the feature map, and uses upsampling and downsampling to construct the pyramid structure. A 5-layer feature pyramid structure is constructed, with the number of feature map channels of each layer being 64, 128, 256, 512, and 1024, respectively. Through the pyramid structure, the network can extract multi-level feature information of electromagnetic shielding effectiveness at different scales. The upsampling process of the feature pyramid is expressed as:

[0106] ;

[0107] in, is a low-resolution feature map, It is a high-resolution feature map obtained by upsampling. In this way, multi-scale fused feature data is obtained. The multi-scale fused feature data is input to the second residual block for deep feature extraction. The second residual block contains two convolutional layers, the number of output channels is set to 256, and a The convolutional layer is used for dimensionality reduction. Through this design, the network can better compress and optimize deep features, making the feature representation more refined. In this process, the dimensionality reduction operation is expressed by the following formula:

[0108] ;

[0109] in, yes Convolutional layers, is the input feature map, is the output dimensionality reduction feature map. In this way, the second residual block improves the quality of the feature and passes it to the next layer for deeper processing. When performing cross-layer feature connection processing on deep feature data, the network unifies the feature maps of different scales to the same resolution through deconvolution operation and splices them in the channel dimension to obtain multi-scale feature connection data. The purpose of cross-layer feature connection is to fuse features from different levels to enhance the expression ability of the network. The deconvolution operation can restore the low-resolution feature map to a higher resolution and enhance the spatial detail performance of the feature map. The splicing operation splices the feature maps from different scales by channel to form a richer feature map. The multi-scale feature connection data is input into the feature fusion module for weighted fusion processing. The feature fusion module uses the channel attention mechanism to adaptively assign weights to features of different scales. The channel attention mechanism determines the importance of features of different scales by learning the weights of each feature map and adjusts the contribution of the feature map according to these weights. Through weighted fusion, multi-scale feature data is finally obtained.

[0110] In a specific embodiment, the process of executing step 500 may specifically include the following steps:

[0111] Perform distribution feature analysis on multi-scale feature data and construct a tapered matrix based on Laplace distribution. The dimension of the tapered matrix is ​​n×n, where n is the number of signal channels.

[0112] Perform covariance calculation on the tapered matrix to calculate the signal covariance matrix R. The size of the signal covariance matrix R is n×n, the diagonal elements are the signal variance, and the non-diagonal elements are the covariance between signal channels.

[0113] The signal covariance matrix is ​​tapered and the tapered parameter α is set to 0.8 to obtain tapered matrix parameter data;

[0114] Reconstructing the tapered matrix parameter data by matrix multiplication, reconstructing the covariance matrix by matrix multiplication to obtain reconstructed matrix data, and performing minimum variance distortion-free response optimization on the reconstructed matrix data to obtain optimized matrix data;

[0115] Perform eigenvalue decomposition on the optimized matrix data, sort the decomposed eigenvalues, and determine the signal subspace dimension by the minimum description length criterion to obtain eigendecomposition data;

[0116] The signal subspace is reconstructed for the eigendecomposition data, the eigenvectors corresponding to the largest k eigenvalues ​​are selected to obtain the subspace reconstructed data, and the subspace reconstructed data is optimized by matrix taper to obtain the reconstructed covariance matrix.

[0117] Specifically, the distribution characteristics of the multi-scale feature data are analyzed. Statistical methods are used to describe the distribution characteristics of the data in order to extract useful signal information from it. The Laplace distribution is used as the basis to construct the tapered matrix. The Laplace distribution has a heavier tail, which makes it more robust in modeling signal noise and abnormal data. Assume that the multi-scale feature data is ,in Represents the number of signal channels. By assuming Laplace distribution, a tapered matrix is ​​constructed with the dimension The function of the tapered matrix is ​​to process the covariance matrix of the signal to enhance the important features of the signal and suppress the influence of noise. In the process of constructing the tapered matrix, the covariance of the signal is calculated. Signal covariance matrix Describes the linear correlation between different signal channels, and its dimension is also , where the diagonal elements represent the variance of the signal channel and the off-diagonal elements represent the covariance between the signal channels. Signal covariance matrix Calculated by the following formula:

[0118] ;

[0119] in, Indicates The characteristic data of the signal channels, is the mean of the signal data. By calculating the covariance matrix , understand the relationship between signal channels. Set the tapered parameters of the signal covariance matrix and set the tapered parameters Set to 0.8. The role of the tapering parameter is to adjust the noise level of the signal data, thereby improving the quality of the signal. Taper matrix Combined with the covariance matrix in the following way:

[0120] ;

[0121] in, is the identity matrix, is the tapering parameter, and its value range is [0,1]. The value of , balances the weight of signal and noise, so that the reconstructed matrix can more accurately reflect the signal characteristics. The tapered matrix is ​​reconstructed by matrix multiplication. Using the structural characteristics of the signal data, the noise part is minimized to obtain a more accurate covariance matrix. Assume that the reconstructed matrix is , and its calculation process is expressed as:

[0122] ;

[0123] The matrix multiplication operation readjusts the covariance structure of the signal by combining the tapered matrix and the original covariance matrix, so that the influence of noise is effectively suppressed. The reconstructed covariance matrix is ​​optimized for minimum variance distortion-free response to improve signal quality. By minimizing the variance of the signal, distortion-free transmission of the signal is ensured at the same time. The optimization process is expressed by the following formula:

[0124] ;

[0125] in, is the optimized weight vector, is the desired direction vector, indicating the direction of the signal. Through this optimization method, the signal-to-noise ratio of the signal is effectively improved while suppressing the influence of noise. The optimized matrix is ​​subjected to eigenvalue decomposition to obtain the principal components of the signal data. The eigenvalue decomposition of the signal covariance matrix is ​​expressed as:

[0126] ;

[0127] in, is a diagonal matrix containing the eigenvalues ​​of the covariance matrix, It is an eigenvector matrix, which represents the basis of the signal space. The size of the eigenvalue reflects the strength of the signal feature. The component with a larger eigenvalue represents the most important feature in the signal, while the component with a smaller eigenvalue is noise or an unimportant signal. By sorting the eigenvalues, the eigenvector corresponding to the largest eigenvalue is selected. These eigenvectors represent the main directions in the signal data. The dimension of the signal subspace is determined by the minimum description length criterion, and a suitable signal subspace dimension is selected. , so that the reconstruction error is minimized while avoiding overfitting. The calculation of the minimum description length criterion is expressed by the following formula:

[0128] ;

[0129] in, It is before The largest eigenvalue, is the total number of samples, is the characteristic dimension. By minimizing the minimum description length criterion, the optimal dimension of the signal subspace is determined, thereby effectively reconstructing the signal subspace. The signal subspace is reconstructed for the characteristic decomposition data. By selecting the largest The eigenvector corresponding to the eigenvalue is used to obtain the subspace reconstruction data of the signal. The subspace reconstruction is expressed by the following formula:

[0130] ;

[0131] in, Is the selected feature vectors, is the square root matrix of the eigenvalues. Through the reconstructed signal subspace, the main components of the signal are more accurately represented and further processed. The matrix taper optimization is performed on the subspace reconstruction data to optimize the covariance matrix so that the reconstructed covariance matrix more accurately reflects the characteristics of the signal. The matrix taper optimization is performed similar to the tapering process:

[0132] ;

[0133] The final reconstructed covariance matrix can effectively improve the signal quality.

[0134] In a specific embodiment, the process of executing step 600 may specifically include the following steps:

[0135] The reconstructed covariance matrix data is distributed to the controllers of each functional domain for regional centralized processing, and the signal regional centralized control is performed on the autonomous driving domain, power domain, chassis domain, cockpit domain and body domain to obtain the initial distribution data of the functional domain;

[0136] The data cache structure is processed for the initial allocation data of the functional domain, a 256KB cache area is set using a circular queue structure, and data read and write management is performed according to the data transmission priority of each functional domain to obtain data cache configuration parameters;

[0137] Establishing functional domain data channels based on data cache configuration parameters, setting the bandwidth of data exchange channels between functional domains, and optimizing the data transmission routes between functional domains to obtain channel allocation data;

[0138] Perform real-time signal quality detection on the channel allocation data to obtain detection data, and compare the detection data with a preset threshold to obtain signal quality detection data;

[0139] The signal transmission status is evaluated based on the signal quality detection data. When the signal-to-noise ratio is lower than the target value, the automatic retransmission mechanism is started, and the number of retransmissions and the retransmission results are recorded to obtain the transmission status evaluation data.

[0140] Based on the transmission status evaluation data, fault location is performed, data transmission anomalies of each functional domain are analyzed to obtain fault location data, and self-recovery processing is performed on the fault location data to obtain self-recovery execution data;

[0141] The self-recovery execution data is input into the functional domain optimization module for data flow optimization. Under the interference-signal ratio condition within the preset range, the data transmission of each functional domain is regulated in real time to obtain the interference-resistant transmission data flow of each functional domain.

[0142] Specifically, the reconstructed covariance matrix data is preprocessed and assigned to the controllers of each functional domain. The signal area is centrally controlled for the autonomous driving domain, power domain, chassis domain, cockpit domain and body domain, and the preliminary functional domain data allocation results are obtained. The data cache structure is processed for the initial allocation data of the functional domain. By establishing an appropriate cache mechanism, it is ensured that each functional domain can meet the bandwidth and delay requirements during signal transmission. In order to optimize the data transmission efficiency, a circular queue structure is used to set the cache area, and the cache area size is set to 256KB. This structure can effectively avoid the cache overflow problem caused by frequent data read and write operations, thereby ensuring the smoothness of data transmission between functional domains. At the same time, data read and write management is performed according to the data transmission priority of each functional domain to ensure that the high-priority functional domain is not blocked by low-priority tasks during data transmission, thereby improving the efficiency and reliability of the system. Based on the results of the data cache configuration, the data channel between the functional domains is established. The data exchange bandwidth between each functional domain is determined, and the bandwidth is reasonably configured to meet the data transmission requirements of different functional domains. By finely configuring the bandwidth of the data channel, data congestion in some functional domains due to insufficient bandwidth can be avoided. Optimize the data transmission routes between functional domains. Optimizing the routes can prevent data from going through too many jumps during transmission, thereby reducing latency, improving overall transmission efficiency, and obtaining channel allocation data to ensure that communication between each functional domain is optimal and reduce latency and packet loss. Perform real-time signal quality detection on channel allocation data to monitor the transmission quality of signals between each functional domain to ensure data integrity and accuracy. Signal quality detection data is evaluated by comparing the current signal quality with a preset threshold. If the signal quality of a functional domain is lower than the set standard, the system will issue a warning and analyze and adjust according to the preset standard. Based on the signal quality detection results, the signal transmission status is evaluated. Dynamically adjust the signal transmission strategy based on changes in the signal-to-noise ratio (SNR) during transmission. The formula for calculating the signal-to-noise ratio is:

[0143] ;

[0144] in, is the power of the signal, is the power of the noise. When it is detected that the signal-to-noise ratio is lower than the target value, the system automatically starts the retransmission mechanism to ensure the integrity of the data. The retransmission mechanism helps the system evaluate the reliability and stability of the current transmission by recording the number and results of each retransmission. By analyzing the number of retransmissions and their results, its transmission strategy is gradually improved to reduce data loss and improve transmission efficiency. This process generates transmission status evaluation data. In the fault location phase, the data transmission anomalies of each functional domain are analyzed to identify the areas where the problem occurs. By analyzing the signal quality, the number of retransmissions and the transmission route, it is accurately identified which functional domains or data channels have problems. These fault location data are used for subsequent self-recovery processing to ensure that the system can automatically take measures to repair when a fault occurs. For example, when the signal quality of a functional domain is lower than the target threshold for a long time, it automatically switches to a backup channel or reconfigures the bandwidth to reduce the impact of the fault on the overall system performance. After the self-recovery processing is completed, it enters the data flow optimization phase. Through the self-recovery execution data, the system can adjust the data flow of each functional domain in real time according to the preset interference signal ratio (C / 1) condition. The interference signal ratio (C / 1) is a measure of the ratio between the system signal and the interference, and the formula is:

[0145] ;

[0146] According to the specific needs of each functional domain, the data flow and signal transmission strategy are adjusted in real time. For example, when the signal interference of a functional domain is large, the signal quality is improved by adjusting its signal strength or selecting a more suitable transmission path. Through the above operations, the data transmission of each functional domain is guaranteed to be free from interference and run in the optimal state, and finally the anti-interference transmission data flow between each functional domain is realized.

[0147] The above describes the anti-interference optimization method of the FPC high-performance computing chip interface in the embodiment of the present application. The following describes the anti-interference optimization device 10 of the FPC high-performance computing chip interface in the embodiment of the present application. Figure 2 In the embodiment of the present application, an anti-interference optimization device 10 for the FPC high-performance computing chip interface includes:

[0148] The decomposition module 11 is used to perform total variation regularization decomposition processing on the interface signal of the FPC to obtain effective data signal components;

[0149] The deformation processing module 12 is used to perform three-dimensional bending deformation processing on the structure of the FPC according to the effective data signal component to obtain three-dimensional forming structure parameters;

[0150] The shielding processing module 13 is used to perform multi-layer electromagnetic shielding processing on the FPC based on the three-dimensional molding structure parameters to obtain a three-layer composite shielding structure, wherein the first layer is a copper foil layer, the second layer is a high magnetic permeability alloy layer, and the third layer is a conductive polymer layer;

[0151] The feature extraction module 14 is used to input the electromagnetic shielding effectiveness data of the three-layer composite shielding structure into a preset residual neural network model for feature extraction to obtain multi-scale feature data;

[0152] A tapering module 15, used for performing covariance matrix tapering on multi-scale feature data to obtain a reconstructed covariance matrix;

[0153] The optimization module 16 is used to perform functional domain allocation and signal optimization processing based on the reconstructed covariance matrix data to obtain the anti-interference transmission data stream of each functional domain, and the functional domains include the autonomous driving domain, the power domain, the chassis domain, the cockpit domain and the body domain.

[0154] Through the synergy of the above components, the combined application of total variation regularization decomposition and SCORE algorithm, the effective decomposition of FPC interface signals is achieved, and the effective data signal components are accurately extracted. The three-dimensional bending deformation processing method is adopted to control the deformation in three dimensions of X axis (0-180 degrees), Y axis (0-90 degrees) and Z axis (0-45 degrees), effectively optimizing the spatial structure of FPC. A three-layer composite shielding structure consisting of a copper foil layer, a high magnetic permeability alloy layer and a conductive polymer layer is designed to form a complete electromagnetic shielding system. Based on the improved residual neural network model, combined with the feature pyramid structure and context guidance module, the multi-scale feature extraction of electromagnetic shielding effectiveness data is realized. The covariance matrix tapering processing technology is adopted, and the signal reconstruction effect is improved through the signal subspace reconstruction method. A complete functional domain allocation mechanism including autonomous driving domain, power domain, chassis domain, cockpit domain and body domain is established to optimize the data transmission between functional domains. Data caching and self-recovery mechanisms are designed to enhance the reliability of the system in practical applications.

[0155] See also Figure 3 , Figure 3 This is a schematic block diagram of the structure of an electronic device 300 provided in an embodiment of the present application. The electronic device 300 includes a processor 301 and a memory 302. The processor 301 and the memory 302 are connected via a device bus 303, wherein the memory 302 may include a non-volatile storage medium and an internal memory.

[0156] The non-volatile storage medium can store a computer program. The computer program includes program instructions, and when the program instructions are executed by the processor 301, the processor 301 can execute any of the above-mentioned anti-interference optimization methods for the FPC high-performance computing chip interface.

[0157] The processor 301 is used to provide computing and control capabilities to support the operation of the entire electronic device 300 .

[0158] The internal memory provides an environment for the operation of the computer program in the non-volatile storage medium. When the computer program is executed by the processor 301, the processor 301 can execute any of the above-mentioned anti-interference optimization methods for the FPC high-performance computing chip interface.

[0159] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a partial structure related to the scheme of the present application, and does not constitute a limitation on the electronic device 300 involved in the scheme of the present application. The specific electronic device 300 may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0160] It should be understood that the processor 301 may be a central processing unit (CPU), and the processor 301 may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0161] It should be noted that technical personnel in the relevant field can clearly understand that for the convenience and simplicity of description, the specific working process of the electronic device 300 described above can refer to the corresponding process of the anti-interference optimization method of the aforementioned FPC high-performance computing chip interface, and will not be repeated here.

[0162] An embodiment of the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by one or more processors, the one or more processors implement the anti-interference optimization method of the FPC high-performance computing chip interface provided in the embodiment of the present application.

[0163] The computer-readable storage medium may be an internal storage unit of the electronic device 300 in the aforementioned embodiment, such as a hard disk or memory of the electronic device 300. The computer-readable storage medium may also be an external storage device of the electronic device 300, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped with the electronic device 300.

[0164] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0165] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable an electronic device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., various media that can store program codes.

[0166] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An anti-interference optimization method for an FPC high-performance computing chip interface, characterized in that: The method comprises: Perform total variation regularization decomposition processing on the interface signal of FPC to obtain effective data signal components; According to the effective data signal component, the structure of the FPC is subjected to three-dimensional bending deformation processing to obtain three-dimensional forming structure parameters; specifically comprising: performing signal feature analysis based on the effective data signal component, extracting the data transmission structure features between the computing chip and the sensor, and obtaining basic data of the three-dimensional deformation of the FPC; performing X-axis bending analysis on the basic data of the three-dimensional deformation of the FPC, determining the bending angle change of the X-axis within the range of 0-180 degrees, and calculating the stress distribution of the X-axis bending point to obtain X-axis bending optimization data; performing Y-axis bending analysis based on the X-axis bending optimization data, determining the bending angle change of the Y-axis within the range of 0-90 degrees, and calculating the stress distribution of the Y-axis bending point to obtain the Y-axis Bending optimization data; performing Z-axis torsion analysis based on the Y-axis bending optimization data, determining the torsion angle change of the Z-axis within the range of 0-45 degrees, and calculating the stress distribution of the Z-axis torsion point to obtain Z-axis torsion optimization data; performing material stress distribution modeling on the Z-axis torsion optimization data, and establishing three-dimensional stress distribution model data including X-axis, Y-axis and Z-axis stress distributions; performing material property matching on the three-dimensional stress distribution model data to obtain material structure parameter data, and optimizing the spatial layout of the sensor module based on the material structure parameter data, determining the optimal connection method between modules, and obtaining module spatial layout data; performing transmission path calculation based on the module spatial layout data to obtain three-dimensional molding structure parameters; Based on the three-dimensional molding structure parameters, the FPC is subjected to a multi-layer electromagnetic shielding treatment to obtain a three-layer composite shielding structure, wherein the first layer is a copper foil layer, the second layer is a high magnetic permeability alloy layer, and the third layer is a conductive polymer layer; Inputting the electromagnetic shielding effectiveness data of the three-layer composite shielding structure into a preset residual neural network model for feature extraction to obtain multi-scale feature data; Performing covariance matrix tapering on the multi-scale feature data to obtain a reconstructed covariance matrix; Functional domain allocation and signal optimization processing are performed based on the reconstructed covariance matrix data to obtain the interference-resistant transmission data stream of each functional domain, and the functional domains include the automatic driving domain, the power domain, the chassis domain, the cockpit domain and the body domain.

2. The anti-interference optimization method for the FPC high-performance computing chip interface according to claim 1 is characterized in that: The total variation regularization decomposition process is performed on the interface signal of the FPC to obtain the effective data signal component, including: The initial signal of the FPC high-performance computing chip interface is collected to obtain the original signal data containing the data transmission information between the computing chip and the sensor; Inputting the original signal data into a total variation regularization algorithm with a predetermined regularization parameter for iterative decomposition to obtain decomposed data; Performing Euclidean distance calculation on the decomposed data to obtain component data that meets a convergence condition; Inputting the component data satisfying the convergence condition into the SCORE algorithm for signal enhancement preprocessing to obtain preprocessed signal component data; Performing time stamp marking on the preprocessed signal component data to obtain signal component data with time stamp, and storing the signal component data with time stamp into corresponding data buffers for storage processing to obtain classified stored signal component data; The signal-to-noise ratio is calculated for the classified and stored signal component data to obtain signal quality characteristic data, and a threshold value is judged for the signal quality characteristic data to output a valid data signal component that meets the signal-to-noise ratio requirement.

3. The anti-interference optimization method for the FPC high-performance computing chip interface according to claim 1 is characterized in that: Based on the three-dimensional molding structure parameters, the FPC is subjected to multi-layer electromagnetic shielding treatment to obtain a three-layer composite shielding structure, wherein the first layer is a copper foil layer, the second layer is a high magnetic permeability alloy layer, and the third layer is a conductive polymer layer, including: Performing electromagnetic shielding analysis on the three-dimensional molding structure parameters to determine the coverage and thickness distribution of the shielding layer and obtain initial parameter data of the shielding structure; Performing a first copper foil layer structure design on the initial parameter data of the shielding structure to obtain copper foil layer structure parameter data; Designing the structure of the second high magnetic permeability alloy layer according to the copper foil layer structure parameter data to obtain the high magnetic permeability alloy layer structure parameter data; Design the structure of the third conductive polymer layer according to the structural parameter data of the high magnetic permeability alloy layer to obtain the structural parameter data of the conductive polymer layer; Design the grounding position of the three-layer composite shielding layer based on the conductive polymer layer structure parameter data, set a grounding point at the connection of the three-layer shielding structure, and obtain grounding structure parameter data; Perform signal suppression based on the grounding structure parameter data, set a common mode choke coil at a target position of the signal transmission channel, and obtain EMI suppression parameter data; A signal line protection structure is designed according to the EMI suppression parameter data to obtain signal line protection parameter data, and a three-layer composite shielding structure is integrated on the signal line protection parameter data to obtain a three-layer composite shielding structure.

4. The anti-interference optimization method for the FPC high-performance computing chip interface according to claim 3 is characterized in that: The electromagnetic shielding effectiveness data of the three-layer composite shielding structure is input into a preset residual neural network model for feature extraction to obtain multi-scale feature data, including: Inputting the electromagnetic shielding effectiveness data of the three-layer composite shielding structure into a preset residual neural network model, performing feature extraction through a first convolutional layer provided with a ReLU activation function, wherein the convolution kernel size of the first convolutional layer is 3×3, the step size is 1, and the number of output channels is 64, to obtain first-layer feature data; Performing a first residual block processing on the first layer feature data, wherein the first residual block includes two convolutional layers and a short-circuit connection, the number of output channels of each convolutional layer is 128, and a batch normalization layer is added after the convolutional layer to obtain first residual feature data; Inputting the first residual feature data into a context guidance module for feature enhancement, wherein the attention weight threshold is set to 0.6, and feature selection is performed through dynamic weight adjustment to obtain enhanced feature data; Performing feature pyramid processing on the enhanced feature data to construct a 5-layer feature pyramid structure, where the number of feature map channels in each layer is 64, 128, 256, 512, and 1024, respectively, to obtain multi-scale fused feature data; Inputting the multi-scale fusion feature data into the second residual block for deep feature extraction, wherein the second residual block includes two convolutional layers, the number of output channels is 256, and a 1×1 convolutional layer is added at the residual connection for dimensionality reduction to obtain deep feature data; Performing cross-layer feature connection processing on the deep feature data, unifying feature maps of different scales to the same resolution through deconvolution operations, and performing channel dimension splicing to obtain multi-scale feature connection data; The multi-scale feature connection data is input into the feature fusion module for weighted fusion processing, and the channel attention mechanism is used to adaptively allocate weights to features of different scales to obtain multi-scale feature data.

5. The anti-interference optimization method for the FPC high-performance computing chip interface according to claim 4 is characterized in that: The step of performing covariance matrix tapering on the multi-scale feature data to obtain a reconstructed covariance matrix includes: Performing distribution feature analysis on the multi-scale feature data, and constructing a tapered matrix based on Laplace distribution, wherein the dimension of the tapered matrix is ​​n×n, where n is the number of signal channels; Performing covariance calculation on the tapered matrix to calculate a signal covariance matrix R, wherein the size of the signal covariance matrix R is n×n, the diagonal elements are signal variances, and the non-diagonal elements are covariances between signal channels; Performing tapered parameter setting on the signal covariance matrix, setting the tapered parameter α to 0.8, and obtaining tapered matrix parameter data; Performing matrix multiplication reconstruction processing on the tapered matrix parameter data, reconstructing the covariance matrix through matrix multiplication to obtain reconstructed matrix data, and performing minimum variance distortion-free response optimization on the reconstructed matrix data to obtain optimized matrix data; Performing eigenvalue decomposition on the optimized matrix data, sorting the decomposed eigenvalues, and determining the signal subspace dimension by a minimum description length criterion to obtain eigendecomposition data; The signal subspace is reconstructed for the eigendecomposition data, eigenvectors corresponding to the largest k eigenvalues ​​are selected to obtain subspace reconstructed data, and matrix tapering optimization is performed on the subspace reconstructed data to obtain a reconstructed covariance matrix.

6. The anti-interference optimization method for the FPC high-performance computing chip interface according to claim 5 is characterized in that: The functional domain allocation and signal optimization processing are performed based on the reconstructed covariance matrix data to obtain the anti-interference transmission data stream of each functional domain, wherein the functional domain includes an automatic driving domain, a power domain, a chassis domain, a cockpit domain and a body domain, including: The reconstructed covariance matrix data is distributed to the controllers of each functional domain for regional centralized processing, and the signal regional centralized control is performed on the autonomous driving domain, the power domain, the chassis domain, the cockpit domain and the body domain to obtain the initial distribution data of the functional domain; Perform data cache structure processing on the functional domain initial allocation data, set a 256KB cache area using a circular queue structure, and perform data read and write management according to the data transmission priority of each functional domain to obtain data cache configuration parameters; Establishing a functional domain data channel based on the data cache configuration parameters, setting the bandwidth of the data exchange channel between the functional domains, and optimizing the data transmission route between the functional domains to obtain channel allocation data; Performing real-time signal quality detection on the channel allocation data to obtain detection data, and comparing the detection data with a preset threshold to obtain signal quality detection data; Performing signal transmission status evaluation processing according to the signal quality detection data, when the signal-to-noise ratio is lower than the target value, starting the automatic retransmission mechanism, and recording the number of retransmissions and the retransmission results to obtain transmission status evaluation data; Perform fault location based on the transmission status evaluation data, analyze data transmission anomalies in each functional domain to obtain fault location data, and perform self-recovery processing on the fault location data to obtain self-recovery execution data; The self-recovery execution data is input into the functional domain optimization module for data flow optimization. Under the interference-signal ratio condition in a preset range, the data transmission of each functional domain is regulated in real time to obtain the interference-resistant transmission data flow of each functional domain.

7. An anti-interference optimization device for an FPC high-performance computing chip interface, characterized in that: The anti-interference optimization device of the FPC high-performance computing chip interface includes: A decomposition module is used to perform total variation regularization decomposition processing on the interface signal of the FPC to obtain effective data signal components; The deformation processing module is used to perform three-dimensional bending deformation processing on the structure of the FPC according to the effective data signal component to obtain three-dimensional forming structure parameters; specifically comprising: performing signal feature analysis based on the effective data signal component, extracting the data transmission structure features between the computing chip and the sensor, and obtaining basic data of the three-dimensional deformation of the FPC; performing X-axis bending analysis on the basic data of the three-dimensional deformation of the FPC, determining the bending angle change of the X-axis within the range of 0-180 degrees, and calculating the stress distribution of the X-axis bending point to obtain X-axis bending optimization data; performing Y-axis bending analysis based on the X-axis bending optimization data, determining the bending angle change of the Y-axis within the range of 0-90 degrees, and calculating the stress distribution of the Y-axis bending point, Obtain Y-axis bending optimization data; perform Z-axis torsion analysis based on the Y-axis bending optimization data, determine the torsion angle change of the Z-axis within the range of 0-45 degrees, and calculate the stress distribution of the Z-axis torsion point to obtain Z-axis torsion optimization data; perform material stress distribution modeling on the Z-axis torsion optimization data, and establish three-dimensional stress distribution model data including X-axis, Y-axis and Z-axis stress distributions; perform material property matching on the three-dimensional stress distribution model data to obtain material structure parameter data, and perform spatial layout optimization on the sensor module according to the material structure parameter data, determine the optimal connection method between modules, and obtain module space layout data; perform transmission path calculation based on the module space layout data to obtain three-dimensional molding structure parameters; A shielding processing module, used for performing multi-layer electromagnetic shielding processing on the FPC based on the three-dimensional molding structure parameters to obtain a three-layer composite shielding structure, wherein the first layer is a copper foil layer, the second layer is a high magnetic permeability alloy layer, and the third layer is a conductive polymer layer; A feature extraction module, used for inputting the electromagnetic shielding effectiveness data of the three-layer composite shielding structure into a preset residual neural network model for feature extraction to obtain multi-scale feature data; A tapering module, used for performing covariance matrix tapering on the multi-scale feature data to obtain a reconstructed covariance matrix; The optimization module is used to perform functional domain allocation and signal optimization processing based on the reconstructed covariance matrix data to obtain the anti-interference transmission data stream of each functional domain, and the functional domains include the automatic driving domain, the power domain, the chassis domain, the cockpit domain and the body domain.

8. An electronic device, characterized in that: The electronic device comprises: a memory and at least one processor, wherein instructions are stored in the memory; The at least one processor calls the instructions in the memory so that the electronic device executes the anti-interference optimization method for the FPC high-performance computing chip interface as described in any one of claims 1-6.

9. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instruction is executed by the processor, the anti-interference optimization method of the FPC high-performance computing chip interface as described in any one of claims 1-6 is implemented.

Images