Vehicle-mounted display PCBA board anti-interference method, device and equipment

Abstract

The invention relates to the technical field of PCBA board anti-interference, and discloses a vehicle-mounted display PCBA board anti-interference method, device and equipment. The method comprises the following steps: carrying out electromagnetic interference data acquisition on a PCBA (Printed Circuit Board Assembly) of the vehicle-mounted display to obtain an electromagnetic interference data set, and constructing a PCBA structure matrix; performing feature extraction to obtain interference feature data; constructing a probability distribution function, generating a feature transformation matrix and performing joint operation to obtain an interference suppression matrix; constructing a multivariable constraint condition and solving an anti-interference optimization parameter; inputting the anti-interference optimization parameters and the interference signal matrix into a particle swarm-radial basis function neural network for training to obtain an anti-interference neural network model; and performing filtering processing and compensation correction on the original input signal of the PCBA board through the anti-interference neural network model to obtain a target output signal. Compensation correction can be dynamically carried out, the compensation coefficient is adjusted in real time according to the signal error, and the stability and accuracy of the output signal are ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of PCBA board anti-interference, and in particular to a vehicle-mounted display PCBA board anti-interference method, device and equipment. Background Art

[0002] As an important interface for human-computer interaction, the reliability and stability of the in-vehicle display directly affect driving safety and user experience. However, the PCBA board of the in-vehicle display faces serious electromagnetic interference problems in the complex in-vehicle environment. These interference sources include various electromagnetic radiations generated by in-vehicle electrical equipment, wireless communication equipment, and the external environment.

[0003] Traditional PCBA board anti-interference methods mainly rely on hardware shielding and filtering circuits. These methods not only increase product cost and weight, but also have difficulty coping with dynamically changing complex interference environments. Especially in the scenarios of high-speed data transmission and multi-signal parallel processing, the limitations of traditional anti-interference methods are becoming increasingly prominent, and they cannot meet the high requirements of vehicle-mounted displays for signal integrity and display quality. Although existing digital signal processing methods can suppress electromagnetic interference to a certain extent, most of them use a single filtering algorithm or a fixed compensation strategy, lacking in-depth analysis of interference characteristics and adaptive processing capabilities. At the same time, these methods often ignore the physical structural characteristics of PCBA boards and the impact of device layout on interference propagation, resulting in less than ideal anti-interference effects and difficulty in maintaining stable performance under different working conditions. Summary of the invention

[0004] The present invention provides a vehicle-mounted display PCBA board anti-interference method, device and equipment, which can dynamically perform compensation correction, adjust the compensation coefficient in real time according to the signal error, and ensure the stability and accuracy of the output signal.

[0005] In a first aspect, the present invention provides a vehicle-mounted display PCBA board anti-interference method, the vehicle-mounted display PCBA board anti-interference method comprising: Collect electromagnetic interference data of the PCBA board of the vehicle display to obtain an electromagnetic interference data set, and construct a PCBA board structure matrix containing device layout data; Performing fast Fourier transform and interference feature extraction on the electromagnetic interference data set to obtain interference feature data; Constructing a probability distribution function based on the interference feature data, generating a feature transformation matrix, and performing a joint operation on the feature transformation matrix and an interference signal matrix to obtain an interference suppression matrix; Constructing multivariable constraints according to the PCBA board structure matrix and the interference suppression matrix, and solving anti-interference optimization parameters based on the multivariable constraints; Inputting the anti-interference optimization parameters and the interference signal matrix into a particle swarm-radial basis function neural network for training to obtain an anti-interference neural network model; The original input signal of the PCBA board is filtered and compensated by the anti-interference neural network model to obtain the target output signal of the PCBA board.

[0006] In a second aspect, the present invention provides an anti-interference device for a PCBA board of a vehicle-mounted display, and the anti-interference device for a PCBA board of a vehicle-mounted display comprises: The acquisition module is used to collect electromagnetic interference data of the PCBA board of the vehicle display, obtain an electromagnetic interference data set, and construct a PCBA board structure matrix containing device layout data; An extraction module, used for performing fast Fourier transform and interference feature extraction on the electromagnetic interference data set to obtain interference feature data; A calculation module, used for constructing a probability distribution function based on the interference feature data, generating a feature transformation matrix, and performing a joint operation on the feature transformation matrix and an interference signal matrix to obtain an interference suppression matrix; A construction module, used to construct multivariable constraints according to the PCBA board structure matrix and the interference suppression matrix, and solve the anti-interference optimization parameters based on the multivariable constraints; A training module, used for inputting the anti-interference optimization parameters and the interference signal matrix into a particle swarm-radial basis function neural network for training to obtain an anti-interference neural network model; The correction module is used to filter and compensate the original input signal of the PCBA board through the anti-interference neural network model to obtain the target output signal of the PCBA board.

[0007] The third aspect of the present invention provides an anti-interference device for a vehicle-mounted display PCBA board, comprising: 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 vehicle-mounted display PCBA board anti-interference device executes the above-mentioned vehicle-mounted display PCBA board anti-interference method.

[0008] A fourth aspect of the present invention provides a computer-readable storage medium, in which instructions are stored. When the computer-readable storage medium is run on a computer, the computer executes the above-mentioned vehicle-mounted display PCBA board anti-interference method.

[0009] In the technical solution provided by the present invention, electromagnetic probes are arranged at the four vertices of the PCBA board, and the Maxwell equations and space coordinate mapping technology are combined to accurately obtain electromagnetic field distribution data, comprehensively grasp the interference propagation characteristics, and adopt a method combining fast Fourier transform and wavelet decomposition to realize multi-scale analysis of interference signals, accurately extract frequency domain features and time-varying features, and effectively identify the main interference components; Laplace probability distribution function and tapered matrix transformation are introduced to construct feature space mapping relationships, thereby improving the modeling ability and suppression effect of nonlinear interference; based on a multivariable constrained optimization method, the structural characteristics of the PCBA board are organically combined with the interference suppression strategy, and the particle swarm algorithm is combined with the radial basis neural network, and the learning ability and generalization performance of the network are improved through adaptive parameter optimization; a dynamic compensation correction mechanism is designed to adjust the compensation coefficient in real time according to the signal error to ensure the stability and accuracy of the output signal; through a multi-level and multi-dimensional signal processing strategy, the compensation coefficient is adjusted in real time to ensure the stability and accuracy of the output signal, and the display quality is significantly improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying any creative work.

[0011] Figure 1 A schematic diagram of the process of the vehicle display PCBA board anti-interference method provided in the embodiment of the present application; Figure 2 A schematic block diagram of the structure of the vehicle display PCBA board anti-interference device provided in an embodiment of the present application; Figure 3 A schematic block diagram of the structure of the vehicle display PCBA board anti-interference device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0012] 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.

[0013] 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 based on actual conditions.

[0014] 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.

[0015] It should be further understood that the term “and / or” used in the specification and appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0016] In conjunction with the accompanying drawings, some embodiments of the present application are described in detail below. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0017] See also Figure 1 , Figure 1 A schematic diagram of the process of the vehicle display PCBA board anti-interference method provided in the embodiment of the present application, such as Figure 1 As shown, the vehicle display PCBA board anti-interference method provided in the embodiment of the present application includes steps S100 to S600.

[0018] Step S100, collecting electromagnetic interference data of the PCBA board of the vehicle display to obtain an electromagnetic interference data set, and constructing a PCBA board structure matrix containing device layout data; It is understandable that the execution subject of the present invention can be a vehicle-mounted display PCBA board anti-interference device, or a terminal or a server, which is not limited here. The embodiment of the present invention is described by taking a server as an execution subject as an example.

[0019] Specifically, electromagnetic probes are installed at the four vertices of the PCBA board to achieve coverage of the entire board area and ensure that the collected electromagnetic data can fully reflect the electromagnetic interference characteristics of each area on the PCBA board. The sampling frequency of the electromagnetic probe is set to 1MHz, so that it can collect electromagnetic interference signals at a higher frequency and capture rapidly changing electromagnetic field components. Under this setting, the PCBA board is sampled to obtain a complete electromagnetic interference data set. The electromagnetic interference data set is digitized, and the sampling time is set to 10 seconds to ensure that the collected data has a sufficient time span to analyze transient and steady-state interference signals. The electromagnetic interference data set is analog-to-digital converted to convert the original analog signal into an electromagnetic interference digital signal. The electromagnetic interference digital signal is spatially mapped, and a two-dimensional plane grid containing x-coordinates and y-coordinates is established on the PCBA board to describe the spatial distribution of electromagnetic interference on the board surface, and the distribution data of the electromagnetic field of the PCBA board is obtained. The electromagnetic field distribution data is analyzed for intensity to determine the electric field strength and magnetic field strength at different positions. The electric field strength and magnetic field strength of each two-dimensional plane grid point are calculated to obtain the distribution parameters of the electromagnetic field strength. The intensity distribution parameters describe the strength of electromagnetic interference in different areas of the PCBA board. Based on the electromagnetic field intensity distribution parameters, the electromagnetic wave conduction parameters are constructed. The propagation of electromagnetic waves is modeled by the Maxwell equations, and the propagation direction and attenuation coefficient of electromagnetic waves on the PCBA board are calculated to obtain the electromagnetic wave conduction model. The Maxwell equations are the basic equations that describe electromagnetic phenomena. By solving these equations, the propagation path of electromagnetic waves between different media and their attenuation characteristics with distance are determined, and a detailed description of how electromagnetic waves propagate on the PCBA board is obtained. Scan the physical layout of the PCBA board to understand the layout of the device. Scan the devices on the PCBA board, record the center coordinates, occupied area and device type of each device, and obtain the device layout data. According to the device layout data, the connection relationship between the devices is established, and the length, width and layer distribution of the connection are calculated to obtain the device interconnection data. The electromagnetic field intensity distribution parameters, electromagnetic wave conduction model, device layout data and device interconnection data are matrix synthesized to obtain the structural matrix of the PCBA board. The structural matrix integrates the distribution characteristics of the electromagnetic field, the physical layout of the device and the interconnection relationship between the devices, and can fully describe the electromagnetic interference situation on the PCBA board and the path of interference propagation.

[0020] Step S200, performing fast Fourier transform and interference feature extraction on the electromagnetic interference data set to obtain interference feature data; Specifically, the electromagnetic interference data set is segmented and the data is segmented according to a window length of 1024 points. Each segment of data is windowed using the Hanning window function. The spectrum leakage problem caused by the data boundary is reduced by the windowing operation, the accuracy of the spectrum analysis is improved, and a windowed data sequence is obtained to facilitate frequency domain transformation. The fast Fourier transform operation is performed on the windowed data sequence to calculate the corresponding spectrum distribution diagram. The original time domain signal is converted to the frequency domain through the fast Fourier transform to analyze the frequency components of the signal. The obtained spectrum distribution diagram shows the energy distribution of the signal at different frequencies. The spectrum distribution diagram is subjected to energy density analysis to obtain the characteristic data of the signal in the frequency domain, that is, the frequency domain characteristic data. Energy density analysis can quantify the energy of the signal at different frequencies, thereby clarifying the main frequency components and energy distribution of the interference. A frequency threshold is set for the frequency domain characteristic data to detect the strong interference frequency points therein. The main interference frequency band data is identified by detecting the frequency points in the spectrum where the energy density exceeds -60dBm. The main interference frequency band data is input into the db4 wavelet function for wavelet transform, and the signal is decomposed into 5 layers to extract the characteristics of the signal at different time scales. Wavelet transform is a time-frequency analysis method. By decomposing the signal into multiple wavelet coefficients of different scales, the signal changes at different frequencies and times are captured to obtain more fine-grained interference information. The wavelet coefficients of different scales are reconstructed to extract the frequency components, amplitude components, and phase components of each layer of the signal to obtain hierarchical feature data. Based on the hierarchical feature data, the feature vector matrix is ​​constructed, and the correlation coefficients between the feature components are calculated to obtain feature association data, which describes the relationship between different feature components and helps to reveal the intrinsic connection between different interference components in the signal. Time-varying analysis is performed on the feature association data to capture the changes in the features over time. The time domain change rate and change period of the feature components are calculated to obtain time-varying feature data. The time-varying feature data reflects the dynamic change law of the interference features over time, which helps to identify the interference components that change significantly in a specific time period. The frequency domain feature data, hierarchical feature data, feature association data, and time-varying feature data are feature fused to obtain interference feature data.

[0021] Step S300: construct a probability distribution function based on the interference feature data, generate a feature transformation matrix, and perform a joint operation on the feature transformation matrix and the interference signal matrix to obtain an interference suppression matrix; Specifically, the interference feature data is statistically analyzed, the mean and standard deviation of the interference feature are calculated, and the feature statistical parameters are obtained to reflect the central tendency and dispersion of the interference feature distribution in the data set. Based on the feature statistical parameters, the Laplace probability density function is constructed to obtain the feature distribution function. The Laplace distribution is a double exponential distribution, which is suitable for describing the characteristics of signals with spikes and long tails, and can better capture the significant features and sudden changes in electromagnetic interference signals. By constructing the Laplace probability density function, a mathematical description model of the interference feature is established. The feature distribution function is matrix expanded, and a parameter matrix containing position parameters and scale parameters is constructed to obtain a distribution parameter matrix. The distribution parameter matrix describes the distribution characteristics of the interference feature at different positions and scales. According to the distribution parameter matrix, a tapering function is constructed. The tapering function is a function used to measure the similarity between different points in the feature space, and the correlation between any two points in the feature space is calculated by the function. Correlation is an important indicator for measuring the similarity between features. By calculating the correlation between any two points, a matrix describing the feature transformation relationship, namely the feature transformation matrix, is constructed. The feature transformation matrix is ​​used to describe the mapping relationship between different features of the interference signal. The interference feature data is rearranged according to the frequency component, amplitude component and phase component to construct the interference signal matrix. The interference signal matrix is ​​a structured representation of the interference feature data, so that these data can participate in matrix operations and thus perform joint operations with the feature transformation matrix. The feature transformation matrix and the interference signal matrix are subjected to tensor product operations to obtain a fusion matrix. The tensor product operation is a high-dimensional matrix operation method that maps the feature relationship in the feature transformation matrix to the interference signal matrix to achieve multi-dimensional analysis and fusion of the interference signal. Based on the fusion matrix, the minimum variance distortion-free response is calculated to solve the weight vector. The minimum variance distortion-free response is an optimization method for suppressing interference, and its goal is to reduce the variance of the interference signal as much as possible without distorting the signal. By solving the optimization problem of the minimum variance distortion-free response, a weight vector is obtained, which is used to describe the relative importance and influence of each feature in interference suppression. The solved weight vector is subjected to matrix multiplication operation with the fusion matrix to obtain the interference suppression matrix.

[0022] Step S400: construct multivariable constraints according to the PCBA board structure matrix and the interference suppression matrix, and solve the anti-interference optimization parameters based on the multivariable constraints; Specifically, matrix multiplication is performed on the PCBA board structure matrix and the interference suppression matrix to obtain a feature fusion matrix. The feature fusion matrix contains comprehensive information between the physical structure of the PCBA board and the electromagnetic interference characteristics, and can characterize how the interference signal interacts with the structural characteristics. Geometric constraints are constructed based on the feature fusion matrix. Geometric constraints are used to describe the geometric characteristics and spatial layout restrictions in the interference suppression process, which helps to ensure that the optimization process can meet the geometric structure characteristics of the PCBA board and ensure the feasibility of the solution in the actual physical structure. In order to effectively control the interference energy, the interference suppression matrix is ​​traced and energy constraints are constructed. The trace operation is to sum the sum of the diagonal elements of the matrix, which is used to quantify the overall energy level of the interference suppression matrix. By constructing energy constraints, it is ensured that the final suppression scheme meets expectations in terms of energy, thereby effectively suppressing the interference signal while avoiding excessive impact on system performance. At the same time, in order to ensure the rationality of the matrix characteristics and the stability of the system solution, the interference suppression matrix is ​​singular value decomposition to construct rank constraints. Singular value decomposition is a method of decomposing a matrix into multiple rank components. This process quantifies the rank characteristics of the interference suppression matrix, so that the rank characteristics are taken into consideration when constructing the constraints, ensuring that the optimization solution has appropriate sparsity and structured characteristics while maintaining sufficient degrees of freedom. Rank constraints help reduce redundant information, improve the efficiency of optimization calculations, and make the final solution more compact and stable in structure. The geometric constraints, energy constraints, and rank constraints are combined in series to construct the optimization objective function of multivariable constraints. The gradient of the optimization objective function is calculated to obtain the sensitivity information of each parameter. By calculating the first-order and second-order derivatives of the optimization objective function with respect to each parameter, the gradient matrix is ​​obtained to describe the change trend of the optimization objective function in the parameter space. Based on the gradient matrix, the iteration step size and termination condition are set, and the gradient descent method is used for optimization iteration. The gradient descent method gradually approaches the minimum value of the function by iterating along the negative gradient direction of the objective function. In order to ensure that the optimization process can converge effectively, an appropriate iteration step size and termination condition are set. The iteration step size determines the magnitude of each iteration adjustment, while the termination condition is used to determine whether the iteration has reached the expected optimal solution. By setting these parameters, the optimization process is ensured to be carried out within the effective range, and the convergence speed and the accuracy of the results are improved. During the iteration of the gradient descent method, an optimization solution sequence is generated to describe the gradual evolution from the initial solution to the final optimal solution. In order to judge the effectiveness of the optimization solution, the convergence of the optimization solution sequence is judged, that is, whether the solution gradually approaches a stable value during the iteration process. If the optimization solution sequence meets the termination condition, it means that the expected optimal solution has been found, and this solution is selected as the final anti-interference optimization parameter.

[0023] Step S500, inputting the anti-interference optimization parameters and the interference signal matrix into the particle swarm-radial basis function neural network for training to obtain an anti-interference neural network model; Specifically, the anti-interference optimization parameters and interference signal matrix are normalized, and the data is mapped to a range of mean 0 and variance 1 to reduce the impact of data scale differences on model training. After standardization, the data is divided into training data set, validation data set and test data set in a ratio of 8:1:1 to ensure the generalization performance of the model during training and the effectiveness of the final test. A three-layer particle swarm-radial basis neural network structure is constructed, which includes three parts: input layer, hidden layer and output layer. The input layer contains N input nodes, N is equal to the dimension of the feature vector; the hidden layer contains M Gaussian radial basis function nodes, which use radial basis functions to transform input features to enhance the nonlinear characteristics of the network; the output layer contains 1 output node to generate the final prediction value. The parameters of the neural network are particle encoded, and each particle contains three parts: center vector, width vector and weight vector. These parameters determine the specific performance of the radial basis neural network. By randomly initializing each particle, an initialized particle swarm is obtained. In order to optimize the parameters of the neural network, a fitness function of particle swarm optimization is constructed. The fitness function is used to measure the quality of each particle. The fitness value is determined by the cross-validation method to obtain the optimization evaluation function. The construction of the fitness function takes into account the error between the network output and the actual target, and can effectively evaluate the impact of the particle parameter configuration on the model performance. For each particle in the initialized particle swarm, the corresponding encoding parameters are decoded into neural network parameters, and then the training data is input into the neural network for calculation, and the fitness value is calculated by the fitness function. According to the fitness value, the individual optimal position of each particle and the global optimal position of the entire swarm are updated. In order to optimize the particle position, the speed update formula is used to calculate the new speed, and the position of each particle is updated by the position update formula to obtain a new generation of particle swarms. Through iterative optimization, the fitness of the particles is gradually improved, so as to find the optimal neural network parameter configuration. The parameters corresponding to the optimal position of the swarm are decoded into the optimal parameters of the neural network, including the optimal center vector, the optimal width vector and the optimal weight vector. Based on these parameters, the radial basis function is reconstructed to obtain the optimized neural network structure. The optimized neural network structure is trained by back propagation to refine the model performance. The training data is input into the network, forward propagation is performed to calculate the hidden layer output and the final network output, and then back propagation is performed to calculate the error gradient. The parameters are updated by using a learning rate of η = 0.01, and the network is iteratively trained until the error converges. In this way, the optimized radial basis neural network further converges to obtain a trained neural network model. After the neural network training is completed, it is validated and tested. The validation data set is input into the trained neural network, and the mean square error on the validation set is calculated.The mean square error is used to measure the prediction accuracy of the model on the validation data to ensure that the trained model can not only perform well on the training data, but also have good generalization ability on unseen data. If the mean square error of the validation set is less than the preset threshold, it is considered that the model has met the expected performance requirements, and the trained neural network is used as the final anti-interference neural network model.

[0024] Step S600: filtering and compensating the original input signal of the PCBA board through the anti-interference neural network model to obtain the target output signal of the PCBA board.

[0025] Specifically, the original input signal of the PCBA board is normalized, and the amplitude of the input signal is mapped to a standard range, so that all input signals are processed at the same scale to reduce the impact of amplitude differences on model training and reasoning, and obtain standardized input signals. The standardized input signal is segmented according to the time window to obtain a segmented signal sequence. The continuous signal is divided into multiple short-time segments so that the neural network model can process local features and perform subsequent filtering and compensation operations. The segmented signal sequence is input into the input layer of the anti-interference neural network model for signal feature extraction. In the input layer, the Gaussian radial basis function is used to calculate the activation value of each input node to obtain the input layer feature map. The Gaussian radial basis function is a kernel function with nonlinear mapping capability, which can convert the input signal to a high-dimensional space, so that some complex interference features can be better represented and captured. According to the input layer feature map, a nonlinear transformation is performed in the hidden layer of the anti-interference neural network model, and the output value of each hidden layer node is calculated to obtain the representation of the interference feature. The hidden layer can perform complex nonlinear mapping and feature extraction on the input signal through the combination of multiple Gaussian radial basis function nodes, so that the characteristics of the interference signal can be clearly represented. Based on the interference feature representation, the output layer weight vector of the anti-interference neural network model is linearly combined to obtain the filtered signal data. The weight vector of the output layer is obtained through the previous particle swarm optimization and back propagation training. The optimized weight vector can effectively suppress the interference characteristics and output the filtered signal data. The filtered signal data is compensated to ensure that the amplitude and phase of the output signal achieve the desired effect. The error between the filtered signal data and the expected signal is calculated, and the compensation coefficient is calculated based on the error. The amplitude and phase of the signal are accurately corrected through the dynamic compensation function to compensate for the phase distortion and amplitude change introduced by the filtering process. The compensated signal is denormalized and the amplitude of the signal is mapped back to the original amplitude range to obtain the initial output signal. The adjacent time windows of the initial output signal are smoothly connected, and the signals between adjacent time windows are seamlessly connected to obtain the final target output signal.

[0026] In the embodiment of the present invention, electromagnetic probes are arranged at four vertices of the PCBA board, and the Maxwell equations and space coordinate mapping technology are combined to accurately obtain electromagnetic field distribution data, comprehensively grasp the interference propagation characteristics, and adopt a method combining fast Fourier transform and wavelet decomposition to realize multi-scale analysis of interference signals, accurately extract frequency domain features and time-varying features, and effectively identify the main interference components; Laplace probability distribution function and tapered matrix transformation are introduced to construct feature space mapping relationships, thereby improving the modeling ability and suppression effect of nonlinear interference; based on a multivariable constrained optimization method, the structural characteristics of the PCBA board are organically combined with the interference suppression strategy, and the particle swarm algorithm is combined with the radial basis neural network, and the learning ability and generalization performance of the network are improved through adaptive parameter optimization; a dynamic compensation correction mechanism is designed to adjust the compensation coefficient in real time according to the signal error to ensure the stability and accuracy of the output signal; through a multi-level and multi-dimensional signal processing strategy, the compensation coefficient is adjusted in real time to ensure the stability and accuracy of the output signal, and the display quality is significantly improved.

[0027] In a specific embodiment, the process of executing step S100 may specifically include the following steps: Electromagnetic probes were installed at the four vertices of the PCBA board, and the sampling frequency of the electromagnetic probes was set to 1 MHz. The PCBA board was sampled to obtain an electromagnetic interference data set. The electromagnetic interference data set is digitized, the sampling time is set to 10s, the electromagnetic interference data set is converted into digital form, the electromagnetic interference digital signal is obtained, and the electromagnetic interference digital signal is mapped into spatial coordinates, a two-dimensional plane grid containing x-coordinates and y-coordinates is established, and the electromagnetic field distribution data of the PCBA board is obtained; Perform intensity analysis on the electromagnetic field distribution data of the PCBA board, calculate the electric field intensity and magnetic field intensity of each grid point, and obtain the electromagnetic field intensity distribution parameters; The electromagnetic wave conduction parameters are constructed according to the electromagnetic field intensity distribution parameters, and the propagation direction and attenuation coefficient of the electromagnetic wave are calculated by the Maxwell equations to obtain the electromagnetic wave conduction model; Scan the PCBA board for device layout, record the center coordinates, occupied area and device type of each device, obtain device layout data, and establish the connection relationship between devices based on the device layout data, calculate the length, width and layer distribution of the connection, and obtain device interconnection data; The electromagnetic field intensity distribution parameters, electromagnetic wave conduction model, device layout data and device interconnection data are matrix synthesized to obtain the PCBA board structure matrix.

[0028] Specifically, electromagnetic probes are installed at the four vertices of the PCBA board to achieve comprehensive collection of electromagnetic interference signals on the board. The sampling frequency of each probe is set to 1MHz, and the probe collects 1 million data points per second, effectively capturing the rapid change characteristics of the electromagnetic signal and obtaining an electromagnetic interference data set containing time series data. The electromagnetic interference data set is digitized. The sampling time is set to 10 seconds to ensure that a variety of interference conditions can be covered, including short-term transient interference and longer-lasting interference. In this process, the analog signal is converted to a digital signal, and the electromagnetic interference signal is converted from a continuous analog quantity to a discrete digital signal to obtain an electromagnetic interference digital signal. The analog-to-digital conversion process is expressed as: ; in, Indicates The digital signal of the sampling points, To quantize the operation, For analog signals, is the sampling interval is the total number of sampling points. The electromagnetic interference digital signal is mapped to a spatial coordinate system to establish a Coordinates and The two-dimensional plane grid of coordinates is used to describe the spatial distribution of the electromagnetic field on the PCBA board. By dividing the board surface according to a certain grid density, each grid point corresponds to a specific spatial position, and the intensity value of the electromagnetic signal is mapped to these positions to obtain the electromagnetic field distribution data of the PCBA board. The electromagnetic field distribution is represented by a matrix, in which each element corresponds to the electric field strength and magnetic field strength of a grid point. Perform intensity analysis on the electromagnetic field distribution data of the PCBA board. By calculating the electric field strength at each grid point and magnetic field strength , the distribution of the electromagnetic field on the entire PCBA board is described. The relationship between the electric field strength and the magnetic field strength is described by the Maxwell equations, where the electric field strength and magnetic field strength The calculation of is approximated by the following formula: ; in, represents the curl of the magnetic field strength, is the current density, is the time rate of change of electric flux density. By calculation, the direction of electromagnetic wave propagation on the PCBA board and the electromagnetic field strength in different areas are determined. Based on the intensity data, the electromagnetic wave conduction parameters are constructed, and the propagation direction and attenuation coefficient of the electromagnetic wave are calculated by solving the Maxwell equations to obtain the electromagnetic wave conduction model. Layout scan of the devices on the PCBA board. Scan all electronic components on the PCBA board, record the center coordinates, occupied area and device type of each device, and obtain a detailed record containing device layout data. According to the layout data of the device, the connection relationship between the devices is established, and the length, width and hierarchical distribution of each connection line are calculated to obtain the device interconnection data, which is used to describe the physical connection relationship between the components on the board and the influence of these connections on the electromagnetic characteristics. The electromagnetic field intensity distribution parameters, electromagnetic wave conduction model, device layout data and device interconnection data are matrix synthesized to obtain the structural matrix of the PCBA board, which is used to describe the electromagnetic characteristics of the entire PCBA board, including the electromagnetic field distribution on the board, device layout and interconnection between devices. The synthesis of the structural matrix is ​​expressed as: ; Among them, S represents the structural matrix of the PCBA board, and E is the electromagnetic field intensity distribution matrix, which contains the electric field and magnetic field information of each grid point; It is an electromagnetic wave conduction model, which describes the propagation direction and attenuation characteristics of electromagnetic waves; The device layout matrix includes the location, type and occupied area of ​​each device; It is the device interconnection matrix, which describes the connection relationship, length and hierarchical distribution between devices.

[0029] In a specific embodiment, the process of executing step S200 may specifically include the following steps: The electromagnetic interference data set is segmented and the data is segmented according to a window length of 1024 points. Each segment of data is windowed using the Hanning window function to obtain a windowed data sequence. Performing a fast Fourier transform operation on the windowed data sequence to calculate a frequency spectrum distribution diagram, and performing energy density analysis on the frequency spectrum distribution diagram to obtain frequency domain feature data; A frequency threshold is set for the frequency domain feature data. The main interference frequency band data is obtained by detecting the frequency point where the spectrum energy density exceeds -60dBm. The main interference frequency band data is input into the db4 wavelet function, and the signal is decomposed into 5 layers to obtain wavelet coefficients of different scales. Reconstruct wavelet coefficients of different scales, extract the frequency component, amplitude component, and phase component of each layer of signals, obtain hierarchical feature data, and construct a feature vector matrix based on the hierarchical feature data, calculate the correlation coefficient between each feature component, and obtain feature correlation data; Perform time-varying analysis on the feature correlation data, calculate the time domain change rate and change period of the feature component, and obtain the time-varying feature data; The frequency domain feature data, hierarchical feature data, feature association data and time-varying feature data are fused to obtain interference feature data.

[0030] Specifically, the electromagnetic interference data set is segmented, and the entire data set is segmented according to a window length of 1024 data points. Decompose large-scale signal data into multiple smaller parts for local frequency domain analysis and interference feature extraction. The data is divided into windows of length 1024 points, and the signal in each window is regarded as a relatively stable part in a short time. In order to reduce the impact of edge effects on frequency domain analysis, the Hanning window function is applied to each signal segment for windowing. The Hanning window function is a gradually changing window function that can smooth the edges of the signal and reduce spectrum leakage. The Hanning window function is expressed as: ; in, is the weight of the window function, Indicates the length of the window (i.e. 1024 points), is the index of the current data point. By multiplying each segment of the signal by the Hanning window function , and obtain the windowed data sequence. Perform a fast Fourier transform operation on the windowed data sequence to convert the signal from the time domain to the frequency domain and obtain the signal's spectrum distribution diagram. The fast Fourier transform can reveal the signal's frequency structure and the amplitude of each frequency component by representing the time domain signal as a linear combination of frequency components. The mathematical formula for the fast Fourier transform operation is: ; in, Indicates The complex value of the frequency components, is the value of the time domain signal, is an imaginary unit, is the number of points of the fast Fourier transform. Perform energy density analysis on the spectrum distribution diagram and calculate the energy density at each frequency. The frequency domain feature data reflects the energy distribution of the signal at different frequencies, which helps to identify the main interference source. Based on the frequency domain feature data, set the frequency threshold and set the threshold to -60dBm. Determine the main interference frequency band data by detecting the frequency points in the spectrum where the energy density exceeds -60dBm. The main interference frequency band data contains frequency components that have a significant interference effect on the entire electromagnetic environment. Input the main interference frequency band into the db4 wavelet function for wavelet analysis. Wavelet analysis is a time-frequency analysis method that can describe the characteristics of the signal at different scales. Perform 5-layer wavelet decomposition on the signal to obtain wavelet coefficients of different scales, and each layer of coefficients represents the characteristics of the signal in different frequency ranges. Perform reconstruction operations on the wavelet coefficients of different scales to extract the frequency components, amplitude components and phase components of each layer of signals to obtain hierarchical feature data. For example, after 5-layer decomposition, the information of the signal in different frequency bands is extracted respectively, such as the trend information of the low frequency band and the detail information of the high frequency band. Based on the hierarchical feature data, a feature vector matrix is ​​constructed to describe the amplitude, phase and frequency characteristics of each frequency component in the signal. By calculating the correlation coefficient between each feature component, feature association data is obtained to describe the relationship between different frequency components and reveal the mutual influence and correlation of interference components in the signal. Time-varying analysis is performed on the feature association data to understand the temporal changes of the interference characteristics. The time domain change rate and change period of the feature component are calculated to obtain time-varying feature data. Time-varying feature data can reflect the dynamic change law of the interference characteristics in the signal and help identify interference components that change significantly in a specific time period. For example, if the energy density of a certain interference frequency rises sharply over time, it means that a certain device has generated significant electromagnetic radiation in this time period. The frequency domain feature data, hierarchical feature data, feature association data and time-varying feature data are feature fused to obtain the final interference feature data. Feature fusion is expressed as: ; in, represents the fused interference feature data, is the frequency domain feature data, is the hierarchical feature data, is feature-related data, is the time-varying characteristic data, is the fusion weight coefficient. The weight coefficient is used to balance the influence between different features in order to obtain a comprehensive interference feature description.

[0031] In a specific embodiment, the process of executing step S300 may specifically include the following steps: Perform statistical analysis on the interference characteristic data, calculate the mean and standard deviation of the characteristic components, and obtain characteristic statistical parameters; Construct the Laplace probability density function based on the characteristic statistical parameters and obtain the characteristic distribution function; Perform matrix expansion on the feature distribution function, construct a parameter matrix including position parameters and scale parameters, obtain a distribution parameter matrix, and construct a tapering function based on the distribution parameter matrix, calculate the correlation between any two points in the feature space, and obtain a feature transformation matrix; The interference feature data is rearranged according to the frequency component, the amplitude component, and the phase component to construct an interference signal matrix, and a tensor product operation is performed on the feature transformation matrix and the interference signal matrix to obtain a fusion matrix; The minimum variance distortion-free response is calculated based on the fusion matrix, the weight vector is solved, and the weight vector is multiplied by the fusion matrix to obtain the interference suppression matrix.

[0032] Specifically, statistical analysis is performed on the interference feature data, and the mean and standard deviation of the feature components are calculated to obtain the statistical parameters of the feature. For each feature component in the interference feature data, its mean is calculated. and standard deviation . Mean represents the mean value of the characteristic component, and the standard deviation It reflects the discreteness or volatility of the characteristic component. Based on the characteristic statistical parameters, the Laplace probability density function is constructed to establish the characteristic distribution function. The Laplace distribution is a double exponential distribution, which is suitable for describing the distribution of signals with peaks and long tails. Its probability density function is expressed as: ; in, represents the probability density function of the Laplace distribution, is the location parameter, which represents the mean of the feature. is the scale parameter, which represents the standard deviation of the feature, is a random variable. By using the calculated feature mean and standard deviation, the corresponding Laplace probability density function is constructed to describe the probability distribution of the feature at different values. The feature distribution function is expanded into a matrix, and a parameter matrix containing location parameters and scale parameters is constructed to obtain the distribution parameter matrix. Distribution parameter matrix is a matrix containing the position and scale information of each feature component, which is used to describe the distribution characteristics of the feature component in the entire feature space. Assume that the feature vector has characteristic components, the distribution parameter matrix is ​​expressed as: ; in, Indicates The mean of the feature components, Indicates The scale parameter of the feature component. Based on the distribution parameter matrix, a tapering function is constructed to calculate the correlation between any two points in the feature space. The form of the tapering function is: ; in, Represents the feature vector and The correlation between is the Euclidean distance between feature vectors, is the scale parameter. By calculating the tapering function, we get the similarity relationship between each point in the feature space and construct the feature transformation matrix The feature transformation matrix is ​​used to describe the correlation between different feature components. The interference feature data is rearranged according to the frequency component, amplitude component and phase component to construct the interference signal matrix S. The interference signal matrix S is a structured representation of the interference feature data, where each row corresponds to a feature component and the column represents the feature value at different time or spatial positions. and interference signal matrix Perform tensor product operation to get the fused matrix The tensor product operation is expressed as: ; in, represents the elements in the fusion matrix, is the element of the characteristic transformation matrix, is the element of the interference signal matrix. Through the tensor product operation, the feature relationship in the feature transformation matrix is ​​mapped to the interference signal matrix to achieve the fusion of signal features and obtain a fusion matrix containing multi-dimensional information. Based on the fusion matrix, the minimum variance distortion-free response is calculated to solve the weight vector. The minimum variance distortion-free response is a filter design method whose goal is to minimize the variance of the output signal without distorting the target signal. The weight vector of the minimum variance distortion-free response is solved by the following optimization problem: ; in, is the weight vector, is the covariance matrix of the fusion matrix, is the target signal direction vector, express The conjugate transpose of . By solving the optimization problem, the optimal weight vector is obtained for interference signal suppression. The weight vector is multiplied by the fusion matrix to obtain the interference suppression matrix The interference suppression matrix is ​​the result of optimizing the fusion matrix, in which the target signal is retained and the interference signal is effectively suppressed. The calculation of the interference suppression matrix is ​​expressed as: ; in, is the interference suppression matrix, is the conjugate transpose of the weight vector, is the fusion matrix.

[0033] In a specific embodiment, the process of executing step S400 may specifically include the following steps: Perform matrix multiplication on the PCBA board structure matrix and the interference suppression matrix to obtain a feature fusion matrix; Construct geometric constraints based on the feature fusion matrix; perform trace operations on the interference suppression matrix to construct energy constraints; perform singular value decomposition on the interference suppression matrix to construct rank constraints; The geometric constraints, energy constraints and rank constraints are combined in series to construct the optimization objective function of multivariable constraints; Find the gradient of the optimization objective function, calculate the first-order derivative and second-order derivative of the optimization objective function with respect to each parameter, and obtain the gradient matrix; The iteration step size and termination condition are set based on the gradient matrix, and the gradient descent method is used for optimization iteration to obtain the optimization solution sequence. The convergence of the optimization solution sequence is judged, and the solution that meets the termination condition is selected as the anti-interference optimization parameter.

[0034] Specifically, matrix multiplication is performed on the structure matrix S of the PCBA board and the interference suppression matrix C to obtain the feature fusion matrix . Matrix multiplication is expressed as: ; in, is the feature fusion matrix, It is the structural matrix of the PCBA board, which contains the physical layout and interconnection relationship of each device on the board. is the interference suppression matrix, which reflects the suppression effect on electromagnetic interference signals. Through matrix multiplication, the feature fusion matrix is ​​obtained Reflects the interaction between the board structure and interference suppression. According to the feature fusion matrix Construct geometric constraints. Geometric constraints are used to describe the geometric characteristics and limitations of the PCBA board in physical layout, ensuring that the adjustment of parameters during the optimization process conforms to the actual physical characteristics of the PCBA board, such as the device positions cannot overlap, and the traces cannot cross. Geometric constraints are described by the specific element relationship of the feature fusion matrix, such as requiring the element value at a specific position to meet a certain distance condition to reflect the relative layout relationship between physical devices. In order to constrain the optimization process, the interference suppression matrix Perform trace operation to construct energy constraints. Trace operation refers to the sum of the diagonal elements of the matrix, which is used to describe the overall energy distribution of the matrix. The trace of the interference suppression matrix is ​​expressed as: ; in, Representation Matrix traces, is the first diagonal elements, is the dimension of the matrix. By calculating the trace, the total energy of interference suppression is obtained. The energy constraint requires that the energy of the suppression matrix is ​​within a specific range, neither too large (to avoid affecting the useful signal) nor too small (to ensure that the interference is effectively suppressed). Perform singular value decomposition to construct the rank constraint. Singular value decomposition is to decompose a matrix into the product of three parts, which is used to describe the rank and degree of freedom of the matrix. The singular value decomposition of the interference suppression matrix is ​​expressed as: ; in, and is the unit orthogonal matrix, It is a diagonal matrix, and the elements on the diagonal are the singular values ​​of the interference suppression matrix. The singular values ​​describe the rank characteristics of the interference suppression matrix. The rank constraint requires that the rank of the matrix cannot exceed a certain range to maintain the sparsity and stability of the matrix, thereby avoiding unnecessary computational redundancy and excessive processing of interference. The geometric constraints, energy constraints, and rank constraints are combined in series to construct an optimization objective function of multivariable constraints. The optimization objective function comprehensively considers the geometric characteristics of the PCBA board, the energy distribution of interference suppression, and the rank characteristics of the matrix. It is a multi-dimensional constrained optimization problem. The objective function is expressed as: ; in, To optimize the objective function, , and denote the cost functions of geometric constraints, energy constraints, and rank constraints respectively, is the weight coefficient of each constraint, which is used to balance the influence between different constraints. In order to optimize the objective function , calculate the gradient of it to calculate the first-order derivative and second-order derivative of each parameter, and obtain the gradient matrix. The gradient matrix describes the change trend of the objective function in the parameter space and is the key information for optimization. The first-order derivative of the objective function is used to describe the change direction of the objective function, while the second-order derivative is used to describe the change speed of the objective function. The combination of the two can more effectively find the minimum value of the objective function. Based on the gradient matrix, set the iteration step size and termination condition, and use the gradient descent method for optimization iteration. The gradient descent method is an optimization method that gradually approaches the minimum value of the function by iterating along the negative gradient direction of the objective function. The iteration step size determines the magnitude of each iteration adjustment, and the termination condition is used to determine whether the optimal solution that meets expectations has been found. When the change amplitude of the objective function is less than a certain set threshold or the maximum number of iterations is reached, the optimization process will terminate. During the iterative process of gradient descent, an optimization solution sequence is generated to describe the evolution process from the initial solution to the final optimal solution. In order to ensure that the solution found is the global optimal, the convergence of the optimization solution sequence is judged. By observing whether the solution sequence gradually approaches the stable value, it is judged whether the iterative process converges to the local or global optimum. When the termination condition is met, the converged solution is selected as the final anti-interference optimization parameter.

[0035] In a specific embodiment, the process of executing step S500 may specifically include the following steps: The anti-interference optimization parameters and interference signal matrix are normalized and divided into training data set, verification data set and test data set in a ratio of 8:1:1; Construct a three-layer particle swarm-radial basis function neural network structure, where the input layer contains N input nodes, N is equal to the feature vector dimension, the hidden layer contains M Gaussian radial basis function nodes, and the output layer contains 1 output node; The parameters of the neural network structure are encoded into particles. Each particle contains three parts: center vector, width vector and weight vector. They are randomly initialized to obtain an initialized particle swarm. Construct the fitness function of particle swarm optimization, determine the weight coefficient through cross-validation, and obtain the optimization evaluation function; For each particle in the initialized particle swarm, the corresponding encoding parameters are decoded into neural network parameters, the training data is input to calculate the network output, the fitness value is calculated according to the fitness function, the individual optimal position and the group optimal position are updated, the speed update formula is used to calculate the new speed, and the position update formula is used to update the particle position to obtain a new generation of particle swarm; Decode the parameters corresponding to the optimal position of the group into the optimal parameters of the neural network, including the optimal center vector, the optimal width vector and the optimal weight vector, construct the radial basis function, and obtain the optimized neural network structure; The optimized neural network structure is trained by BP. The training data is input into the network. The hidden layer output and network output are calculated by forward propagation. The error gradient is calculated by back propagation. The learning rate η=0.01 is used for parameter update. The training is iterated until the error converges to obtain the trained neural network. The trained neural network is tested on the validation data set, and the validation mean square error is calculated. When the mean square error is less than the preset threshold, the trained neural network is used as an anti-interference neural network model.

[0036] Specifically, the anti-interference optimization parameters and interference signal matrix are processed for data standardization, and the data is mapped to a range of mean 0 and standard deviation 1 to reduce the impact of scale differences between different features on model training. The standardized data set is divided into training data set, verification data set and test data set in a ratio of 8:1:1 to facilitate model training, verification and testing. During the training process, 80% of the data is used to train the model, 10% of the data is used to verify the model performance, and the remaining 10% of the data is used to test the generalization ability of the model. A three-layer particle swarm-radial basis neural network structure is constructed. The neural network consists of three parts: input layer, hidden layer and output layer, where the input layer contains N input nodes, where N is equal to the dimension of the feature vector. The hidden layer contains M Gaussian radial basis function nodes, each of which transforms the input features through the Gaussian radial basis function to enhance the model's representation ability for nonlinear features. The output layer contains 1 output node, which is used to generate the final prediction value. In the hidden layer, the Gaussian radial basis function is expressed as: ; in, Indicates The output of hidden layer nodes, is the input vector, For the The center vector of the node, For the The width parameter of each node. By using the Gaussian radial basis function, the nonlinear mapping of the input data to the hidden layer is realized. After the network structure is constructed, the parameters of the neural network are encoded by particles, and each particle contains three parts: the center vector, the width vector and the weight vector. The center vector is used to represent the center position of the Gaussian radial basis function, the width vector represents the scale of the radial basis function, and the weight vector represents the connection weight between the hidden layer and the output layer. These particles are randomly initialized to obtain the initial particle swarm. Particle swarm optimization is a swarm intelligence algorithm that simulates the information exchange between individuals in the swarm to find the optimal solution. In order to guide the process of particle swarm optimization, a fitness function is constructed. The fitness function is used to evaluate the quality of each particle (that is, each set of neural network parameter configuration). The weight coefficient is determined by cross-validation to avoid overfitting of the model and obtain an optimized evaluation function. The fitness function is defined as the mean square error of the model on the training data set, and its formula is: ; in, is the sample size, is the actual output, is the predicted output of the model. By minimizing the fitness function, the optimal model parameters that can minimize the prediction error are found. For each particle in the initialized particle swarm, the corresponding encoding parameters are decoded into the parameters of the neural network, and the training data is input to calculate the network output. Then the fitness value is calculated according to the fitness function, and the individual optimal position of each particle and the global optimal position of the entire swarm are updated. In order to update the position and speed of the particles, the speed update formula and the position update formula are used. The speed update formula is: ; in, For the Particles in The speed of time, is the inertia weight, and is the learning factor, and is a random number, For the The individual optimal position of each particle is is the global optimal position, For particles in The position at the moment. The position update formula is: ; By updating the speed and position, the particle swarm gradually approaches the optimal solution, and after multiple iterations, a new generation of particle swarm is obtained. After the particle swarm optimization is completed, the parameters corresponding to the optimal position of the swarm are decoded into the optimal parameters of the neural network, including the optimal center vector, the optimal width vector and the optimal weight vector. These parameters are used to construct the radial basis function to obtain the optimized neural network structure. The optimized neural network structure is trained by back propagation. The training data is input into the network, and forward propagation is performed to calculate the output of the hidden layer and the final output of the network, and then back propagation is performed to calculate the error gradient. Use the learning rate The parameters of the network are updated gradually. The goal of back propagation is to minimize the loss function (such as mean square error). Through continuous iterative training until the error converges, a trained neural network model is finally obtained. After the training is completed, the neural network is tested on the validation data set to verify the performance and generalization ability of the model. The validation data set is input into the neural network, and the mean square error on the validation set is calculated. If the mean square error on the validation set is less than the preset threshold, it means that the model has met the expected performance requirements. At this time, the trained neural network is used as the final anti-interference neural network model.

[0037] In a specific embodiment, the process of executing step S600 may specifically include the following steps: Normalize the original input signal of the PCBA board to obtain a standardized input signal, and segment the standardized input signal into time windows to obtain a segmented signal sequence; The segmented signal sequence is input into the input layer of the anti-interference neural network model to extract signal features, and the activation value of each input node is calculated using the Gaussian radial basis function to obtain the input layer feature map; According to the input layer feature map, nonlinear transformation is performed in the hidden layer of the anti-interference neural network model, and the output value of each hidden layer node is calculated to obtain the interference feature representation; Based on the interference feature representation, the output layer weight vector of the anti-interference neural network model is linearly combined to obtain the filtered signal data; The compensation coefficient is calculated for the error between the filtered signal data and the expected signal, and the signal amplitude and phase are corrected by a dynamic compensation function to obtain a compensated signal; The compensation signal is denormalized, the signal value is mapped back to the original amplitude range to obtain the initial output signal, and the adjacent time windows of the initial output signal are smoothly connected to obtain the target output signal of the PCBA board.

[0038] Specifically, the original input signal of the PCBA board is normalized to obtain a standardized input signal. The standardized input signal is segmented into time windows, and the entire signal is divided according to a time window of a fixed length to obtain multiple segmented signal sequences. The segmented signal sequence is input into the input layer of the anti-interference neural network model for signal feature extraction. Each node in the input layer calculates the activation value through the Gaussian radial basis function to achieve feature mapping of the input signal. The formula of the Gaussian radial basis function is: ; in, For the The activation value of the input node, is the input vector, For the The center vector of the Gaussian function, For the The width parameter of the Gaussian function. Through the calculation of the Gaussian radial basis function, the input layer realizes the mapping of the standardized signal features, so that the input signal can be converted into a feature representation in a high-dimensional space, which is convenient for nonlinear processing in the subsequent hidden layer. According to the feature mapping of the input layer, the signal is passed to the hidden layer of the anti-interference neural network model, and a nonlinear transformation is performed in the hidden layer. The hidden layer performs a weighted summation of the input features through a series of nodes, and calculates the output value of each node through the activation function to obtain the representation of the interference feature. The calculation of the hidden layer is expressed as: ; in, For the The output value of hidden layer nodes, is an activation function, usually a nonlinear function (such as ReLU or sigmoid function), is the weight connecting the input node and the hidden layer node, is the bias term. Next, based on the interference feature representation, the output layer of the anti-interference neural network model is linearly combined to obtain the filtered signal data. The calculation of the output layer is expressed as: ; in, is the filtered signal data of the output layer, is the weight between the hidden layer nodes and the output layer nodes, is the bias term of the output layer. The filtered signal data is compensated to ensure that the amplitude and phase of the output signal are consistent with the original signal. The error between the filtered signal data and the expected signal is calculated, and the compensation coefficient is calculated based on the error. The calculation of the compensation coefficient is realized by the dynamic compensation function, which can accurately correct the amplitude and phase of the signal to compensate for the phase distortion and amplitude change introduced by the filtering process. The compensation signal is corrected by the following formula: ; in, is the compensated signal, is the amplitude compensation coefficient, is the filtered signal, is the phase compensation coefficient. Through the compensation operation, it is ensured that the final output signal not only suppresses interference, but also keeps the amplitude and phase consistent with the original signal, improving the authenticity of the signal and system performance. After the compensation process is completed, the compensated signal is denormalized to map the amplitude of the signal back to the original amplitude range. The signal is restored from the standardized range to the actual physical range for subsequent use and output. The denormalization formula is: ; in, is the signal value after denormalization, is the normalized signal after compensation, and are the mean and standard deviation of the original signal respectively. The adjacent time windows of the initial output signal are smoothly connected. Since the time window segmentation causes boundary discontinuity between signals, the boundary effect is eliminated by smooth connection. The weighted average method or overlapping windowing method is used to weight the signals of adjacent time windows to achieve seamless connection of the signals and obtain the target output signal of the PCBA board.

[0039] See also Figure 2 , Figure 2 A schematic block diagram of the structure of the vehicle display PCBA board anti-interference device 200 provided in an embodiment of the present application, as shown in FIG. Figure 2 As shown, the vehicle-mounted display PCBA board anti-interference device 200 includes: The acquisition module 210 is used to collect electromagnetic interference data of the PCBA board of the vehicle display, obtain an electromagnetic interference data set, and construct a PCBA board structure matrix containing device layout data; An extraction module 220 is used to perform fast Fourier transform and interference feature extraction on the electromagnetic interference data set to obtain interference feature data; The operation module 230 is used to construct a probability distribution function based on the interference feature data, generate a feature transformation matrix, and perform a joint operation on the feature transformation matrix and the interference signal matrix to obtain an interference suppression matrix; A construction module 240 is used to construct multivariable constraints according to the PCBA board structure matrix and the interference suppression matrix, and solve the anti-interference optimization parameters based on the multivariable constraints; A training module 250 is used to input the anti-interference optimization parameters and the interference signal matrix into a particle swarm-radial basis function neural network for training to obtain an anti-interference neural network model; The correction module 260 is used to filter and compensate the original input signal of the PCBA board through the anti-interference neural network model to obtain the target output signal of the PCBA board.

[0040] Through the collaborative cooperation of the above components, electromagnetic probes are arranged at the four vertices of the PCBA board, and the Maxwell equations and space coordinate mapping technology are combined to accurately obtain the electromagnetic field distribution data, fully grasp the interference propagation characteristics, and adopt a method combining fast Fourier transform and wavelet decomposition to realize multi-scale analysis of interference signals, accurately extract frequency domain features and time-varying features, and effectively identify the main interference components; the Laplace probability distribution function and tapered matrix transformation are introduced to construct feature space mapping relationships, which improves the modeling ability and suppression effect of nonlinear interference; based on the multivariable constrained optimization method, the structural characteristics of the PCBA board are organically combined with the interference suppression strategy, and the particle swarm algorithm is combined with the radial basis neural network, and the learning ability and generalization performance of the network are improved through adaptive parameter optimization; a dynamic compensation correction mechanism is designed to adjust the compensation coefficient in real time according to the signal error to ensure the stability and accuracy of the output signal; through a multi-level and multi-dimensional signal processing strategy, the compensation coefficient is adjusted in real time to ensure the stability and accuracy of the output signal, which significantly improves the display quality.

[0041] See also Figure 3 , Figure 3 A schematic block diagram of the structure of a vehicle display PCBA board anti-interference device 300 provided in an embodiment of the present application, wherein the vehicle display PCBA board anti-interference device 300 includes a processor 301 and a memory 302, wherein 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.

[0042] 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 vehicle-mounted display PCBA board anti-interference methods.

[0043] The processor 301 is used to provide computing and control capabilities to support the operation of the entire vehicle display PCBA board anti-interference device 300.

[0044] 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 vehicle display PCBA board anti-interference methods.

[0045] 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 present application scheme, and does not constitute a limitation on the vehicle display PCBA board anti-interference device 300 involved in the present application scheme. The specific vehicle display PCBA board anti-interference device 300 may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0046] 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.

[0047] It should be noted that technicians in the relevant field can clearly understand that for the convenience and simplicity of description, the specific working process of the vehicle display PCBA board anti-interference device 300 described above can refer to the corresponding process of the aforementioned vehicle display PCBA board anti-interference method, and will not be repeated here.

[0048] 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 method for the vehicle display PCBA board provided in the embodiment of the present application.

[0049] The computer-readable storage medium may be an internal storage unit of the vehicle display PCBA board anti-interference device 300 of the aforementioned embodiment, such as a hard disk or memory of the vehicle display PCBA board anti-interference device 300. The computer-readable storage medium may also be an external storage device of the vehicle display PCBA board anti-interference 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 vehicle display PCBA board anti-interference device 300.

[0050] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0051] 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 a computer device (which can be a personal computer, server, or network device, etc.) to execute 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.

[0052] 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. A vehicle display PCBA board anti-interference method, characterized in that: include: Collect electromagnetic interference data of the PCBA board of the vehicle display to obtain an electromagnetic interference data set, and construct a PCBA board structure matrix containing device layout data; Performing fast Fourier transform and interference feature extraction on the electromagnetic interference data set to obtain interference feature data; Constructing a probability distribution function based on the interference feature data, generating a feature transformation matrix, and performing a joint operation on the feature transformation matrix and an interference signal matrix to obtain an interference suppression matrix; Constructing multivariable constraints according to the PCBA board structure matrix and the interference suppression matrix, and solving anti-interference optimization parameters based on the multivariable constraints; Inputting the anti-interference optimization parameters and the interference signal matrix into a particle swarm-radial basis function neural network for training to obtain an anti-interference neural network model; The original input signal of the PCBA board is filtered and compensated by the anti-interference neural network model to obtain the target output signal of the PCBA board.

2. The anti-interference method for the PCBA board of the vehicle display according to claim 1 is characterized in that: The electromagnetic interference data of the PCBA board of the vehicle display is collected to obtain an electromagnetic interference data set, and a PCBA board structure matrix containing device layout data is constructed, including: Installing electromagnetic probes at four vertex positions of the PCBA board, setting the sampling frequency of the electromagnetic probes to 1 MHz, sampling the PCBA board, and obtaining an electromagnetic interference data set; The electromagnetic interference data set is digitized, the sampling time is set to 10s, the electromagnetic interference data set is analog-to-digital converted to obtain an electromagnetic interference digital signal, and the electromagnetic interference digital signal is spatially mapped to establish a two-dimensional plane grid including x-coordinates and y-coordinates to obtain the electromagnetic field distribution data of the PCBA board; Performing intensity analysis on the electromagnetic field distribution data of the PCBA board, calculating the electric field intensity and magnetic field intensity of each grid point, and obtaining electromagnetic field intensity distribution parameters; Constructing electromagnetic wave conduction parameters according to the electromagnetic field intensity distribution parameters, calculating the propagation direction and attenuation coefficient of the electromagnetic wave through the Maxwell equations, and obtaining an electromagnetic wave conduction model; Scan the PCBA board for device layout, record the center coordinates, occupied area and device type of each device, obtain device layout data, establish connection relationships between devices according to the device layout data, calculate the length, width and layer distribution of the connection, and obtain device interconnection data; The electromagnetic field intensity distribution parameters, the electromagnetic wave conduction model, the device layout data and the device interconnection data are matrix synthesized to obtain the PCBA board structure matrix.

3. The anti-interference method for the PCBA board of the vehicle display according to claim 2 is characterized in that: The performing fast Fourier transform and interference feature extraction on the electromagnetic interference data set to obtain interference feature data includes: The electromagnetic interference data set is segmented, the data is segmented according to a window length of 1024 points, and each segment of data is windowed using a Hanning window function to obtain a windowed data sequence; Performing a fast Fourier transform operation on the windowed data sequence to calculate a frequency spectrum distribution diagram, and performing energy density analysis on the frequency spectrum distribution diagram to obtain frequency domain feature data; A frequency threshold is set for the frequency domain feature data, and the main interference frequency band data is obtained by detecting the frequency point where the spectrum energy density exceeds -60dBm. The main interference frequency band data is input into the db4 wavelet function, and the signal is decomposed into 5 layers to obtain wavelet coefficients of different scales. Reconstructing the wavelet coefficients of different scales, extracting the frequency component, amplitude component and phase component of each layer of signals to obtain hierarchical feature data, constructing a feature vector matrix based on the hierarchical feature data, calculating the correlation coefficient between each feature component, and obtaining feature correlation data; Performing time-varying analysis on the characteristic correlation data, calculating the time domain change rate and change period of the characteristic component, and obtaining time-varying characteristic data; The frequency domain feature data, the hierarchical feature data, the feature association data and the time-varying feature data are feature-fused to obtain the interference feature data.

4. The anti-interference method for the PCBA board of the vehicle display according to claim 3 is characterized in that: The constructing a probability distribution function based on the interference feature data, generating a feature transformation matrix, and performing a joint operation on the feature transformation matrix and an interference signal matrix to obtain an interference suppression matrix includes: Performing statistical analysis on the interference characteristic data, calculating the mean and standard deviation of characteristic components, and obtaining characteristic statistical parameters; Constructing a Laplace probability density function based on the characteristic statistical parameters to obtain a characteristic distribution function; Performing matrix expansion on the feature distribution function, constructing a parameter matrix including position parameters and scale parameters, obtaining a distribution parameter matrix, constructing a tapering function according to the distribution parameter matrix, calculating the correlation between any two points in the feature space, and obtaining a feature transformation matrix; The interference feature data is rearranged according to frequency components, amplitude components, and phase components to construct an interference signal matrix, and a tensor product operation is performed on the feature transformation matrix and the interference signal matrix to obtain a fusion matrix; The minimum variance undistorted response is calculated based on the fusion matrix, a weight vector is solved, and a matrix multiplication operation is performed on the weight vector and the fusion matrix to obtain the interference suppression matrix.

5. The anti-interference method for the PCBA board of the vehicle display according to claim 4 is characterized in that: The constructing of multivariable constraint conditions according to the PCBA board structure matrix and the interference suppression matrix, and solving anti-interference optimization parameters based on the multivariable constraint conditions, includes: Performing a matrix multiplication operation on the PCBA board structure matrix and the interference suppression matrix to obtain a feature fusion matrix; Constructing geometric constraints according to the feature fusion matrix; performing trace operation on the interference suppression matrix to construct energy constraints; performing singular value decomposition on the interference suppression matrix to construct rank constraints; The geometric constraint condition, the energy constraint condition and the rank constraint condition are combined in series to construct an optimization objective function of multivariable constraints; Obtaining a gradient for the optimization objective function, calculating the first-order derivative and the second-order derivative of the optimization objective function with respect to each parameter, and obtaining a gradient matrix; The iteration step size and termination condition are set based on the gradient matrix, and the gradient descent method is used for optimization iteration to obtain an optimization solution sequence. The convergence of the optimization solution sequence is judged, and the solution that meets the termination condition is selected as the anti-interference optimization parameter.

6. The vehicle-mounted display PCBA board anti-interference method according to claim 5, characterized in that: The step of inputting the anti-interference optimization parameters and the interference signal matrix into a particle swarm-radial basis function neural network for training to obtain an anti-interference neural network model comprises: Performing data standardization processing on the anti-interference optimization parameters and the interference signal matrix, and dividing them into a training data set, a verification data set, and a test data set in a ratio of 8:1:1; Construct a three-layer particle swarm-radial basis function neural network structure, where the input layer contains N input nodes, N is equal to the feature vector dimension, the hidden layer contains M Gaussian radial basis function nodes, and the output layer contains 1 output node; The parameters of the neural network structure are encoded into particles, each particle includes three parts: a center vector, a width vector and a weight vector, and are randomly initialized to obtain an initialized particle group; Construct the fitness function of particle swarm optimization, determine the weight coefficient through cross-validation, and obtain the optimization evaluation function; For each particle in the initialized particle swarm, the corresponding encoding parameters are decoded into neural network parameters, the training data is input to calculate the network output, the fitness value is calculated according to the fitness function, the individual optimal position and the group optimal position are updated, the speed update formula is used to calculate the new speed, and the position update formula is used to update the particle position to obtain a new generation of particle swarm; Decoding the parameters corresponding to the optimal position of the group into optimal parameters of the neural network, including an optimal center vector, an optimal width vector and an optimal weight vector, constructing a radial basis function, and obtaining an optimized neural network structure; The optimized neural network structure is subjected to BP training, the training data is input into the network, the hidden layer output and the network output are calculated by forward propagation, the error gradient is calculated by back propagation, the parameters are updated by using a learning rate η=0.01, and the training is iterated until the error converges to obtain a trained neural network; The trained neural network is tested on a validation data set, and a validation mean square error is calculated. When the mean square error is less than a preset threshold, the trained neural network is used as the anti-interference neural network model.

7. The vehicle-mounted display PCBA board anti-interference method according to claim 6, characterized in that: The filtering and compensating of the original input signal of the PCBA board by the anti-interference neural network model to obtain the target output signal of the PCBA board includes: Normalizing the original input signal of the PCBA board to obtain a standardized input signal, and segmenting the standardized input signal into time windows to obtain a segmented signal sequence; Inputting the segmented signal sequence into the input layer of the anti-interference neural network model to extract signal features, using Gaussian radial basis function to calculate the activation value of each input node to obtain an input layer feature map; According to the input layer feature map, a nonlinear transformation is performed in the hidden layer of the anti-interference neural network model, and an output value of each hidden layer node is calculated to obtain an interference feature representation; Based on the interference feature representation, a linear combination is performed through the output layer weight vector of the anti-interference neural network model to obtain filtered signal data; Calculating a compensation coefficient for the error between the filtered signal data and the expected signal, and correcting the signal amplitude and phase through a dynamic compensation function to obtain a compensated signal; The compensation signal is denormalized, the signal value is mapped back to the original amplitude range to obtain an initial output signal, and adjacent time windows of the initial output signal are smoothly connected to obtain the target output signal of the PCBA board.

8. An anti-interference device for a PCBA board of a vehicle display, characterized in that: The method for preventing interference of a PCBA board of a vehicle-mounted display according to any one of claims 1 to 7 comprises: The acquisition module is used to collect electromagnetic interference data of the PCBA board of the vehicle display, obtain an electromagnetic interference data set, and construct a PCBA board structure matrix containing device layout data; An extraction module, used for performing fast Fourier transform and interference feature extraction on the electromagnetic interference data set to obtain interference feature data; A calculation module, used for constructing a probability distribution function based on the interference feature data, generating a feature transformation matrix, and performing a joint operation on the feature transformation matrix and an interference signal matrix to obtain an interference suppression matrix; A construction module, used to construct multivariable constraints according to the PCBA board structure matrix and the interference suppression matrix, and solve the anti-interference optimization parameters based on the multivariable constraints; A training module, used for inputting the anti-interference optimization parameters and the interference signal matrix into a particle swarm-radial basis function neural network for training to obtain an anti-interference neural network model; The correction module is used to filter and compensate the original input signal of the PCBA board through the anti-interference neural network model to obtain the target output signal of the PCBA board.

9. An anti-interference device for a vehicle-mounted display PCBA board, characterized in that: The vehicle-mounted display PCBA board anti-interference device comprises: a memory and at least one processor, wherein the memory stores instructions; The at least one processor calls the instructions in the memory to enable the vehicle-mounted display PCBA board anti-interference device to execute the vehicle-mounted display PCBA board anti-interference method according to any one of claims 1-7.

10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the vehicle display PCBA board anti-interference method according to any one of claims 1 to 7 is implemented.

Images