Magnetorheological Seat Fault Diagnosis Methods, Systems and Vehicles

Abstract

This application relates to the field of vehicle control technology and discloses a method, system, and vehicle for diagnosing magnetorheological (MR) seat faults. This application involves acquiring a first dataset; extracting features from each data point in the first dataset to obtain frequency domain and time domain data; performing principal component analysis on the frequency and time domain data to determine the variable features associated with different fault types of the MR seat; clustering the variable features associated with each fault type to determine feature clusters formed by the variable features associated with different fault types under each road condition; inputting the feature clusters associated with different fault types under each road condition into a preset support vector machine for training to construct a fault diagnosis model; and using the fault diagnosis model to detect the current vehicle data and seat data to determine whether the MR seat fault diagnosis is valid. In this way, faults can be identified even in complex road conditions, improving the efficiency and accuracy of diagnosis.

Description

Technical Field

[0001] This invention relates to the field of vehicle control technology, and in particular to a magnetorheological seat fault diagnosis method, system, and vehicle. Background Technology

[0002] Magnetorheological seats, as a crucial component of shock and vibration protection systems in transportation vehicles, can minimize the possibility of equipment damage and / or personal injury during shock and vibration loading. They are essential for transportation vehicles and have wide applications, including aircraft, ground vehicles, and marine vehicles.

[0003] In related technologies, the fault diagnosis methods for conventional seats are divided into quantitative analysis and qualitative analysis. However, there is a lack of solutions for diagnosing magnetorheological seat faults. This makes it impossible to determine whether a magnetorheological seat fault has occurred when the vehicle is driving on complex roads, bumpy sections, or sharp turns, which can easily lead to misjudgment. Therefore, there is an urgent need for a magnetorheological seat fault diagnosis method to improve the accuracy of judgment. Summary of the Invention

[0004] To provide a basic understanding of some aspects of the disclosed embodiments, a brief summary is given below. This summary is not intended as a general commentary, nor is it intended to identify key / important components or describe the scope of protection of these embodiments, but rather as a prelude to the detailed description that follows.

[0005] In view of the shortcomings of the prior art described above, this application provides a magnetorheological seat fault diagnosis method, system and vehicle to solve the problem of low accuracy in magnetorheological seat fault diagnosis.

[0006] The first aspect of this application provides a method for diagnosing magnetorheological seat faults, comprising: acquiring a first dataset, the first dataset including vehicle data and seat data corresponding to a vehicle under various road conditions; extracting features from each data in the first dataset to obtain frequency domain data and time domain data; performing principal component analysis on the frequency domain data and the time domain data to determine the variable features associated with the magnetorheological seat under different fault types, wherein the variable features include at least one of the following: vehicle speed, steering wheel angle, steering wheel speed, throttle opening degree, brake pedal opening degree, vehicle height, and seat pressure, displacement, and magnetic field strength; performing clustering processing on the variable features associated with each fault type to determine the feature clusters formed by the variable features associated with different fault types under each road condition; inputting the feature clusters associated with different fault types under each road condition into a preset support vector machine for training to construct a fault diagnosis model; and using the fault diagnosis model to detect the current vehicle data and seat data of the vehicle to determine whether the magnetorheological seat fault diagnosis is faulty.

[0007] In some possible embodiments of this application, before feature extraction is performed on each data in the first dataset, the method further includes: preprocessing the first dataset to obtain a second dataset, and using the second dataset as the first dataset to be extracted and processed. The preprocessing includes at least one of noise processing, missing value processing, outlier processing, data integration processing, and data discretization processing.

[0008] In some possible embodiments of this application, before preprocessing the first dataset, the method further includes: dividing the first dataset into a first data subset representing seat faults and a second data subset representing no seat faults according to the magnetorheological seat faults, and using the first data subset and the second data subset as the first dataset to be preprocessed.

[0009] In some possible embodiments of this application, feature extraction is performed on each data in the first dataset to obtain frequency domain data and time domain data, including: performing feature extraction on vehicle data and seat data in the first dataset based on two dimensions, time domain and frequency domain, respectively, to obtain time domain data containing time domain features and frequency domain data containing frequency domain features. The time domain features include at least one of mean, variance, peak value, and skewness, and the time domain features include at least one of dominant frequency, frequency band energy, and frequency centroid. The variable features are a portion of the data determined in the first dataset.

[0010] In some possible embodiments of this application, principal component analysis is performed on the frequency domain data and the time domain data to determine the variable features associated with the magnetorheological seat under different fault types. This includes: standardizing the frequency domain data and the time domain data respectively to convert them into different feature matrices; calculating the covariance matrix of each feature matrix; calculating the correlation coefficient between each type of feature vector and different faults based on the covariance matrix; calculating the eigenvalues ​​in the covariance matrix based on the correlation coefficients; calculating the principal components of each type of feature vector according to the number of eigenvalues ​​and a preset contribution rate; and determining the variable features associated with the magnetorheological seat under different fault types based on the principal components corresponding to each type of feature vector.

[0011] In some possible embodiments of this application, clustering is performed on the variable features associated with each fault type to determine the feature clusters formed by the variable features associated with different fault types under each road condition. This includes: selecting multiple cluster centers, assigning each variable feature associated with each fault type under different road conditions to the nearest cluster center among the multiple cluster centers; re-clustering the assigned cluster centers until a preset number of clustering is reached to obtain the feature clusters formed by the variable features associated with different fault types under each road condition.

[0012] In some possible embodiments of this application, the feature clusters associated with different fault types under each road condition are input into a preset support vector machine for training to construct a fault diagnosis model, including: pre-setting the value range of the support vector machine; initializing the mathematical optimizer acceleration function, determining the current iteration number, the maximum iteration number, and the acceleration function value; randomly generating an initial population and the position of the initial population in each feature cluster; training the support vector machine using each feature cluster, calculating the current fitness value, selecting the current best fitness value to determine the current optimal population position; determining whether the search phase starts from the exploration phase or the development phase based on the comparison result between the initial population and the acceleration function value, updating the population position in different search phases, calculating the fitness value of all individuals to find the global optimal population position; if the global optimal population position reaches the maximum iteration number, outputting the current Lagrange multiplier and constant as optimal parameters to obtain the fault diagnosis model.

[0013] In some possible embodiments of this application, the expression of the fault diagnosis model is:

[0014]

[0015] In the formula, f(x) represents the fault diagnosis model, and K(x) represents the fault diagnosis model. i Let y(x) be the kernel function of the support vector machine, and y(x) be the kernel function of the support vector machine. i It is the label of the i-th sample, a i ' is the Lagrange multiplier, b' is the constant, and P is the total number of samples.

[0016] In some possible embodiments of this application, the fault diagnosis model is used to detect the current vehicle data and seat data of the vehicle to determine whether the magnetorheological seat fault diagnosis is faulty. This includes: inputting the current vehicle data and seat data of the vehicle into the fault diagnosis model to obtain a loss function value; comparing the loss function value with a preset fault threshold; if the loss function value is greater than or equal to the preset fault threshold, then the magnetorheological seat is determined to be faulty; if the loss function value is less than the preset fault threshold, then the magnetorheological seat is determined to be normal.

[0017] A second aspect of this application also provides a magnetorheological seat fault diagnosis system, comprising: an acquisition module for acquiring a first dataset, the first dataset including vehicle data and seat data corresponding to various road conditions; a feature extraction module for extracting features from each data in the first dataset to obtain frequency domain data and time domain data; and a component analysis module for performing principal component analysis on the frequency domain data and the time domain data to determine the variable features associated with the magnetorheological seat under different fault types, wherein the variable features include the vehicle speed, steering wheel angle, steering wheel speed, throttle opening degree, and brake... The system includes: a pedal opening / closing degree, vehicle height, and at least one of the pressure, displacement, and magnetic field strength of the seat; a clustering module for clustering the variable features associated with each fault type to determine the feature clusters formed by the variable features associated with different fault types under each road condition; a model building module for inputting the feature clusters associated with different fault types under each road condition into a preset support vector machine for training to build a fault diagnosis model; and a fault diagnosis module for using the fault diagnosis model to detect the current vehicle data and seat data of the vehicle to determine whether the magnetorheological seat fault diagnosis is faulty.

[0018] A third aspect of this application also provides an electronic device, comprising: a processor and a memory; the memory for storing a computer program, and the processor for executing the computer program stored in the memory to cause the electronic device to perform the methods described above.

[0019] The beneficial effects of this application are:

[0020] By acquiring a first dataset, which includes vehicle data and seat data corresponding to various road conditions, and determining data diversity based on multi-source data, feature extraction is performed on each data point in the first dataset to obtain frequency domain data and time domain data. Principal component analysis is then performed on the frequency domain data and the time domain data to determine the variable features associated with the magnetorheological seat under different fault types, covering the potential associated features and facilitating subsequent accurate judgment. Clustering is then performed on the variable features associated with each fault type to determine the feature clusters formed by the variable features associated with different fault types under each road condition, ensuring the integrity and diversity of the features. The feature clusters associated with different fault types under each road condition are input into a preset support vector machine for training to construct a fault diagnosis model. The fault diagnosis model is then used for detection to determine whether the magnetorheological seat is faulty. In this way, faults can be identified even in complex road conditions, improving the efficiency and accuracy of diagnosis. Accurate diagnosis of magnetorheological seat faults helps prevent potential safety hazards and reduces the risk of traffic accidents caused by seat malfunctions. At the same time, it also improves the reliability and service life of the seat and reduces maintenance costs. Attached Figure Description

[0021] Figure 1 This is a flowchart illustrating a magnetorheological seat fault diagnosis method according to an embodiment of this application;

[0022] Figure 2 This is a schematic diagram of the data cleaning process in a magnetorheological seat fault diagnosis method according to an embodiment of this application;

[0023] Figure 3 This is a schematic diagram illustrating the principle of a magnetorheological seat fault diagnosis method in an embodiment of this application;

[0024] Figure 4 This is a schematic diagram of the structure of a magnetorheological seat fault diagnosis system according to an embodiment of this application;

[0025] Figure 5 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0026] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, unless otherwise specified, the following embodiments and sub-samples in the embodiments can be combined with each other.

[0027] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0028] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the invention. However, it will be apparent to those skilled in the art that embodiments of the invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the invention.

[0029] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate for the embodiments of this disclosure described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion.

[0030] Unless otherwise stated, the term "multiple" means two or more.

[0031] In this embodiment of the disclosure, the character " / " indicates that the objects before and after it are in an "or" relationship. For example, A / B means: A or B.

[0032] The term "and / or" describes an association between objects, indicating that three relationships can exist. For example, A and / or B means: A or B, or A and B.

[0033] Combination Figure 1 The diagram shown is a flowchart illustrating a magnetorheological seat fault diagnosis method according to an embodiment of this application, including:

[0034] Step S101: Obtain the first dataset, which includes vehicle data and seat data corresponding to various road conditions;

[0035] The vehicles include both new energy vehicles and gasoline-powered vehicles. New energy vehicles include, but are not limited to, plug-in hybrid electric vehicles, range-extended electric vehicles, fuel cell vehicles, hybrid electric vehicles, and pure electric vehicles. Gasoline-powered vehicles include, but are not limited to, sedans, SUVs (sports utility vehicles), MPVs (multi-purpose vehicles), pickup trucks, sports cars, trucks, buses, and special-purpose vehicles.

[0036] For example, vehicle data and seat data under various road conditions are collected through on-board sensors, controller area network buses, and other data acquisition methods. Vehicle data includes, but is not limited to, GPS (Global Positioning System) time, vehicle location, vehicle speed, steering wheel angle, steering wheel speed, throttle opening, brake pedal opening, and vehicle height. Seat data includes, but is not limited to, seat pressure, displacement, and magnetic field strength. It should be noted that different sensors are used for data collection depending on the data type, which will not be elaborated here.

[0037] For example, various road conditions, including turning, straight driving, bumpy roads, and uphill driving, are considered, and the data from various types are preprocessed to obtain the first processed dataset A = {A}. D |D=1,2,…M}, where A D This represents the data for the D-th vehicle.

[0038] Step S102: Extract features from each data in the first dataset to obtain frequency domain data and time domain data;

[0039] For example, signal processing techniques, such as Fourier transform and wavelet transform, are used to convert various types of data in the first dataset (i.e., vehicle data and seat data) into frequency domain data. Simultaneously, statistical analysis methods, such as mean, variance, and standard deviation, are used to extract time domain data. Through feature extraction, useful information for fault diagnosis, such as frequency components and amplitude variations, can be extracted from the first dataset, providing a foundation for subsequent principal component analysis.

[0040] Step S103: Perform principal component analysis on the frequency domain data and the time domain data to determine the variable characteristics associated with the magnetorheological seat under different fault types. The variable characteristics include at least one of the following: vehicle speed, steering wheel angle, steering wheel speed, throttle opening degree, brake pedal opening degree, vehicle height, and seat pressure, displacement, and magnetic field strength.

[0041] For example, principal component analysis transforms the data into a new coordinate system through linear transformation, so that the first principal component in the new dataset has the largest variance (i.e., the most information), and the subsequent principal components decrease in size accordingly; in this way, the dimensionality of the data is reduced by retaining fewer principal components, while retaining most of the information.

[0042] For example, principal component analysis (PCA) can be used to reduce the dimensionality of the extracted frequency and time domain data, identifying the variable characteristics associated with different fault types of the magnetorheological seat. Of course, other dimensionality reduction algorithms, such as linear discriminant analysis and independent component analysis, can also be used to achieve different dimensionality reduction effects.

[0043] Step S104: Perform clustering processing on the variable features associated with each fault type to determine the feature clusters formed by the variable features associated with different fault types under each road condition;

[0044] For example, clustering algorithms divide data points into different clusters (feature clusters) by calculating the distance or similarity between them. This allows fault types with similar characteristics to be grouped together. For instance, clustering algorithms such as K-means and hierarchical clustering can be used to cluster the variable features associated with each fault type, forming feature clusters. Furthermore, other clustering algorithms can be used, such as noisy spatial clustering methods and spectral clustering, to adapt to different data distributions and characteristics.

[0045] Step S105: Input the feature clusters associated with different fault types under each road condition into a preset support vector machine for training to construct a fault diagnosis model;

[0046] Support Vector Machine (SVM) is a binary classification algorithm based on the maximum margin principle. It finds a hyperplane to separate data points of different classes. During training, the SVM searches for an optimal hyperplane that maximizes the margin between the two classes of data points.

[0047] For example, feature clusters associated with different fault types under each road condition can be used as training samples and input into a pre-defined support vector machine (SVM) for training to construct a fault diagnosis model. Of course, other machine learning algorithms, such as decision trees and neural networks, can also be sampled to construct different fault diagnosis models.

[0048] Step S106: Use the fault diagnosis model to detect the current vehicle data and seat data of the vehicle to determine whether the magnetorheological seat fault diagnosis is faulty.

[0049] Among them, the fault diagnosis model can determine whether the data belongs to a known fault type by analyzing and processing the input data.

[0050] For example, real-time vehicle data and seat data are collected, and a fault diagnosis model is used to detect the data to determine whether the magnetorheological seat has a fault.

[0051] Through the above method, useful information for fault diagnosis is extracted from the original data (the first dataset) by means of feature extraction, principal component analysis, clustering processing, etc., and an accurate fault diagnosis model is constructed. By constructing a fault diagnosis model based on machine learning to replace the traditional diagnosis method based on experience and rules, the labor cost and time cost of fault diagnosis are reduced. This application can not only detect and handle the faults of the magnetorheological seat in a timely manner, but also ensure the riding comfort and safety of passengers, and improve the overall performance and quality of the vehicle.

[0052] Optionally, in some embodiments, before performing feature extraction on each data in the first dataset, it further includes:

[0053] Preprocessing the first dataset to obtain a second dataset, where the second dataset is the dataset to be subjected to feature extraction processing, that is, the original dataset, and the preprocessing includes at least one of noise processing, missing value processing, outlier processing, data integration processing, and data discretization processing.

[0054] Exemplarily, please refer to Figure 2 , which is a schematic diagram of the data cleaning process in a magnetorheological seat fault diagnosis method in an embodiment of the present application. From top to bottom, it includes noise processing, missing value processing, outlier processing, and data discretization processing respectively.

[0055] For example, for noise processing, the moving average method is used for denoising.

[0056] Assume that the existing signal sequence is x1,…x T , the window size of the moving average calculation is N, N<T, and T is the data length. Then the value at time t is calculated as shown in formula (1):

[0057]

[0058] For example, for missing value processing, different methods can be adopted for filling different data types, such as mean filling, mode filling, or fixed value filling.

[0059] For example, for outlier processing, outliers are detected and processed using statistical methods, such as the 3σ principle and the box plot method. According to the characteristics of outliers, their processing includes 3 cases: deleting outliers; treating them as missing values and then using missing value filling; average value correction, that is, using methods such as the mean method and the interpolation method to replace the outlier with the previous and next two observed values and treating it as a missing value.

[0060] Apply an advanced data cleaning algorithm based on machine learning to automatically identify and remove complex noise and outliers. The data is standardized using the representation score normalization method, and at the same time, principal component analysis is used for feature extraction and selection to extract M high-dimensional features, effectively reducing the data dimension and retaining key information.

[0061] For example, data discretization divides the range of data values ​​into intervals of equal width.

[0062] For example, data integration processing involves cleaning, transforming, and merging data to form a unified dataset.

[0063] The above preprocessing methods significantly improve the quality and usability of sample data, and reduce the errors in fault diagnosis models caused by sample data quality issues.

[0064] Optionally, in some embodiments, before preprocessing the first dataset, the following steps are further included:

[0065] The first dataset is divided into a first data subset representing seat faults and a second data subset representing no seat faults according to the magnetorheological seat faults. The first data subset and the second data subset are used as the first dataset to be preprocessed.

[0066] For example, the collected data from each vehicle is divided into a faulty seat dataset A′ and a fault-free seat dataset A″ according to seat diagnostic criteria. In this embodiment, the division of the first dataset is based on the data labels or features. In this example, the data label is the fault state of the magnetorheological seat. By examining the label of each data point, it can be assigned to the corresponding data subset.

[0067] By using the above method, data points with similar characteristics are grouped together through data segmentation, which facilitates subsequent data analysis and modeling. At the same time, it also helps to reduce data redundancy and noise, and improves the accuracy and efficiency of data analysis.

[0068] Optionally, in some embodiments, feature extraction is performed on each data point in the first dataset to obtain frequency domain data and time domain data, including:

[0069] Feature extraction is performed on the vehicle data and seat data in the first dataset based on both the time domain and frequency domain dimensions to obtain time domain data containing time domain features and frequency domain data containing frequency domain features. The time domain features include at least one of mean, variance, peak value, and skewness, and the time domain features include at least one of dominant frequency, frequency band energy, and frequency centroid. The variable features are a portion of the data in the first dataset. In other words, by determining a portion of the data in the first dataset, it is possible to determine which of the following are variable features: vehicle speed, steering wheel angle, steering wheel speed, throttle opening, brake pedal opening, vehicle height, and the pressure, displacement, and magnetic field strength of the seat.

[0070] For example, the mean is calculated to reflect the overall level of a signal by averaging the values ​​of a class of signals. The variance is calculated to reflect the volatility of a signal by measuring the dispersion of its values. The peak value is determined by calculating the maximum value of a class of signals to represent the strongest part of the signal. The skewness is calculated to measure the degree of asymmetry in the distribution of a class of signals around the mean.

[0071] It should be noted that the mean, variance, and peak value are features calculated from the signal data, while the calculation of skewness is relatively complex and requires the use of the signal's distribution function or probability density function. By extracting time-domain features, the overall characteristics and trends of the signal over time are revealed. The extracted time-domain features are derived from vehicle data and seat data.

[0072] For example, the Fourier transform converts a signal from the time domain to the frequency domain to obtain its spectral information. It calculates the dominant frequency by identifying the frequency component with the highest energy. It then calculates the band energy within a specific frequency band. Finally, it determines the frequency centroid by calculating the signal's frequency centroid based on the spectral information, reflecting the overall distribution characteristics of the signal in the frequency domain.

[0073] It should also be noted that the calculation of the main frequency, band energy and frequency centroid needs to be based on the signal's spectrum information, which is obtained through Fourier transform. By extracting frequency domain features, the overall characteristics and trends of the signal in the frequency domain are revealed. The extracted frequency domain features also come from vehicle data and seat data.

[0074] Optionally, in some embodiments, principal component analysis is performed on the frequency domain data and the time domain data to determine the variable characteristics associated with the magnetorheological seat under different fault types, including:

[0075] The frequency domain data and the time domain data are standardized respectively and converted into different feature matrices;

[0076] Calculate the covariance matrix of each of the feature matrices, and calculate the correlation coefficient between each type of feature vector and different faults based on the covariance matrix;

[0077] The eigenvalues ​​in the covariance matrix are calculated based on the correlation coefficient, and the principal components of each type of eigenvector are calculated according to the number of eigenvalues ​​and the preset contribution rate.

[0078] Based on the principal components corresponding to the feature vectors of each class, the variable features associated with the magnetorheological seat in different fault types are determined.

[0079] For example, frequency domain data and time domain data are acquired, and the frequency domain data and time domain data are standardized to eliminate differences in dimensions and numerical ranges between different features. Standardization typically includes converting the data into normally distributed data with a mean of 0 and a standard deviation of 1, or scaling the data to a specific range. The standardized data is then converted into a feature matrix, where each row of the feature matrix represents a data sample, and each column represents a feature (i.e., a frequency domain feature or a time domain feature).

[0080] For each feature matrix, its covariance matrix is ​​calculated, reflecting the correlation between features. Based on the covariance matrix, the correlation coefficient between each type of feature vector and different faults is calculated, or mutual information, Pearson correlation coefficient, etc., where the correlation coefficient is used to measure the linear relationship between feature vectors and faults. Based on the covariance matrix, its eigenvalues ​​are calculated, where the eigenvalues ​​reflect the contribution of the feature vectors to the covariance matrix. Principal components are calculated for each type of feature vector according to the number of eigenvalues ​​and a preset contribution rate (e.g., 80%). Based on the principal components corresponding to each type of feature vector, the performance under different fault types is analyzed, and the variable features associated with different fault types are determined based on the principal component analysis results.

[0081] By employing the above methods, standardization and feature matrix transformation were used to eliminate dimensional and numerical range differences in the original data. By calculating the covariance matrix and correlation coefficient, the linear relationship between the eigenvectors and the faults was determined. Principal components were obtained through PCA analysis, reducing the dimensionality of the data and simplifying subsequent analysis and processing. At the same time, by determining the characteristics of the associated variables, the accuracy and reliability of the model were improved.

[0082] Optionally, in some embodiments, clustering is performed on the variable features associated with each fault type to determine feature clusters formed by the variable features associated with different fault types under each road condition, including:

[0083] Select multiple cluster centers, and assign each variable feature associated with each fault type under different road conditions to the nearest cluster center among the multiple cluster centers;

[0084] The assigned cluster centers are re-clustered until a preset number of clusterings is reached, thereby obtaining feature clusters formed by the variable features associated with different fault types under each road condition.

[0085] For example, based on the characteristics of the data and the expected number of clusters, multiple initial cluster centers are randomly selected or selected based on a certain strategy, such as using K-Means++. The distance from each variable feature to each cluster center is calculated, for example, using Euclidean distance. Each variable feature is assigned to the nearest cluster center, forming a preliminary clustering result. Based on the preliminary clustering result, the cluster centers of each cluster are recalculated, for example, by the mean, median, etc. The above steps are repeated until the preset number of clustering is reached or the cluster centers no longer change significantly. Through iterative optimization, the accuracy and stability of the clustering results are improved. After reaching the preset number of clustering, the final clustering result is output, which is the feature clusters formed by the variable features associated with different fault types under each road condition.

[0086] The above methods ensure the accuracy and stability of the clustering results, while also improving the computational efficiency of the entire clustering process.

[0087] Optionally, in some embodiments, the feature clusters associated with different fault types under each road condition are input into a preset support vector machine for training to construct a fault diagnosis model, including:

[0088] The value range of the support vector machine is preset;

[0089] Initialize the mathematical optimizer acceleration function, and determine the current iteration number, the maximum iteration number, and the acceleration function value;

[0090] An initial population and its location are randomly generated in each of the feature clusters.

[0091] The support vector machine is trained using each of the feature clusters, the current fitness value is calculated, and the current best fitness value is selected to determine the current optimal population position.

[0092] Based on the comparison results between the initial population and the acceleration function value, it is determined whether the search phase starts from the exploration phase or the development phase. The population position is updated in different search phases, and the fitness value of all individuals is calculated in order to find the globally optimal population position.

[0093] If the global optimal population position reaches the maximum number of iterations, then the current Lagrange multipliers and constants are output as the optimal parameters to obtain the fault diagnosis model.

[0094] For example, before training the Support Vector Machine (SVM), the value ranges of parameters in the SVM (such as Lagrange multipliers α, penalty parameter C, kernel function parameter γ, etc.) are pre-defined based on experience or data range. These parameters will affect the classification effect and performance of the SVM. A mathematical optimizer acceleration function is selected, such as momentum, root mean square optimizer, adaptive moment estimate, etc., and its parameters are initialized. For example, the current iteration number, the maximum iteration number is set to a preset value, and relevant parameters of the acceleration function are set, such as learning rate, momentum coefficient, etc.

[0095] Based on the obtained feature clusters, a certain number of initial populations are randomly generated in each feature cluster, and their initial positions are determined. Within each feature cluster, an SVM is trained using the parameter combinations from the current population, and its fitness value (e.g., classification accuracy, recall) is calculated. The individual with the best fitness value is selected as the current optimal population position. Based on the fitness values ​​and acceleration function values ​​of the individuals in the initial population, it is determined whether the current search phase should be the exploration phase or the development phase. The exploration phase focuses on searching for new solutions within a larger range, while the development phase focuses on a refined search near already found, relatively optimal solutions.

[0096] During the exploration phase, the population position is updated using random or heuristic methods. During the development phase, methods such as gradient descent or local search are used to update the population position. The fitness values ​​of individuals are calculated at different search phases, and the globally optimal population position is selected. When the globally optimal population position reaches the preset maximum number of iterations, the search process stops, and the currently optimal Lagrange multipliers and constants are output as the optimal parameters of the SVM, thus obtaining the fault diagnosis model.

[0097] By employing the above methods, combining feature clusters, acceleration functions, and dynamic search strategies, the optimal combination of SVM parameters can be efficiently found, improving the model's classification performance and generalization ability. Simultaneously, by optimizing SVM parameters and training an accurate fault diagnosis model, accurate classification and diagnosis of fault types can be achieved, improving the stability and reliability of fault diagnosis. Adopting intelligent fault diagnosis methods reduces reliance on human experience, lowers the cost and time of fault diagnosis, and, combined with optimization strategies such as feature clusters and acceleration functions, enables rapid identification and diagnosis of fault types, improving the real-time performance and accuracy of fault diagnosis.

[0098] Optionally, in some embodiments, the expression of the fault diagnosis model is:

[0099]

[0100] In equation (1), f(x) is the fault diagnosis model, and K(x) is the fault diagnosis model. i Let y(x) be the kernel function of the support vector machine, and y(x) be the kernel function of the support vector machine. iIt is the label of the i-th sample, a i ' is the Lagrange multiplier, b' is the constant, and P is the total number of samples.

[0101] The fault diagnosis model is obtained by solving the following formula:

[0102]

[0103] In equation (2), L is the Lagrangian function, ω is the normal vector of the hyperplane; b is a constant term, and a penalty function C and a slack variable ξ are used to improve the performance of SVM and increase its fault tolerance; α i ≥0, μ i ≥0 is a Lagrange multiplier; ξ i ≥0; C is a constant greater than 0. The value of the penalty function C is positively correlated with the degree of penalty of the function. The larger the penalty parameter, the smaller the value of the slack variable is required, mainly playing a trade-off role. The set is represented as A={(x i y i )|i=1,,2,…,n},xi∈Rn,y i ∈{+1,-1}.

[0104] Optionally, in some embodiments, the fault diagnosis model is used to detect the current vehicle data and seat data of the vehicle to determine whether the magnetorheological seat fault diagnosis is faulty, including:

[0105] Input the current vehicle data and seat data of the vehicle into the fault diagnosis model to obtain the loss function value;

[0106] The loss function value is compared with a preset fault threshold.

[0107] If the loss function value is greater than or equal to a preset fault threshold, then the magnetorheological seat is determined to be faulty.

[0108] If the loss function value is less than the preset fault threshold, then the magnetorheological seat is determined to be normal.

[0109] For example, the loss function value reflects the difference between the current state and the normal state learned by the model. When the loss function value exceeds the set fault threshold, it indicates that the difference between the current state and the normal state learned by the model is too large, and the seat is determined to be faulty; when the loss function value does not exceed the set fault threshold, it indicates that the difference between the current state and the normal state learned by the model is either too large or too small, and the seat is determined to be normal.

[0110] The above methods enable real-time detection of fault conditions in magnetorheological (MRT) seats, improving the timeliness and accuracy of fault detection. Fault diagnosis models based on machine learning and statistical methods can intelligently assess seat status, reducing the possibility of human intervention and misjudgment. This solution can be easily extended to fault detection in other vehicles and components, improving the overall reliability and safety of transportation. Furthermore, by promptly detecting faults and outputting relevant information, it can guide maintenance personnel to quickly locate and repair problems, thereby reducing maintenance costs and time.

[0111] Please see Figure 3 This is a schematic diagram illustrating the principle of a magnetorheological seat fault diagnosis method according to an embodiment of this application, detailed below:

[0112] Step 1: Collect vehicle data and seat fault information under different operating conditions, such as turning, straight driving, bumpy driving, and uphill driving, and preprocess the data to obtain M processed vehicle datasets A = {A...} D |D=1,2,…M}, where M is the number of vehicles, A D This represents the data for the D-th vehicle, which includes, but is not limited to, GPS time, vehicle location, speed, steering angle, or throttle opening. The magnetorheological seat dataset collects seat data through sensors on the seat, including but not limited to pressure sensors, displacement sensors, or magnetic field strength sensors.

[0113] Step 2: The collected data from each vehicle is divided into faulty seat dataset A′ and non-faulty seat dataset A″ according to the seat diagnostic criteria.

[0114] Step 3 involves preprocessing the collected dataset, including data cleaning and feature extraction, as detailed above.

[0115] Step 4, Feature Extraction and Selection: For time-domain data, such as time series of vibration signals, extract statistical features such as mean, variance, peak value, and skewness. For example, the peak value of a vibration signal may be related to abnormal impacts on the equipment. For frequency-domain data, such as time-domain data converted to the frequency domain using Fourier transform, extract features such as dominant frequency, frequency band energy, and frequency centroid. In fault diagnosis, different fault types may lead to changes in specific frequency components. For multivariate data, features such as correlation and covariance between variables can be calculated to reflect the relationships between different parameters.

[0116] Specifically, the feature vector F = {f} is extracted from the vehicle dataset A using principal component analysis. D |D=1,2,…,M}, where f D Let represent the feature vector of vehicle D, and This represents the nth dimension of the data for vehicle D, where n represents the number of data types collected for each vehicle. The specific steps for feature selection are as follows:

[0117] Step 4.1, convert the original data f D Standardization is performed to eliminate the influence of dimensions and numerical values ​​on the analysis results, denoted as the standardized data matrix. Where N is the number of data types and T is the total number of time points.

[0118] Step 4.2, calculate the correlation coefficient matrix of the standardized data. Where, r ij The correlation coefficient between the data in column i and column j is calculated according to formula (3):

[0119]

[0120] Where k is the current time, T is the total time, and x ki Let x be the data matrix of the i-th column at time k. kj Let j be the data matrix at time k. Let be the mean of the data matrix in the i-th column. Let r be the mean of the data matrix in column j, and let r be the mean of the data matrix in column j. ij =r ji and r ii =1.

[0121] Step 4.3, find the characteristic roots of the characteristic equation of R, det(R-λE)=0, where λ1≥λ2≥…≥λ N >0, where R is the correlation coefficient matrix of formula (3), E is the identity matrix, and λ N It is the Nth characteristic root.

[0122] Step 4.4, determine the number m of principal components: α is the cumulative contribution rate, typically taken as 80%, and λ is the cumulative contribution rate. b Let b be the b-th eigenvalue, and N be the number of data types.

[0123] Step 4.5: Calculate the m corresponding unit eigenvectors and principal components.

[0124] Step 4.6, Data Transformation: Transform the original data into a new coordinate system composed of the first m principal components to obtain the new test sample of the D-th vehicle after dimensionality reduction. And the new eigenvector F′={f′ D |D=1,2,…,M}.

[0125] Step 5: Use an improved clustering algorithm (such as K-means clustering) to cluster F′={f′ D |D=1,2,…,M} are clustered into different categories.

[0126] (1) Initialization: Randomly select Q initial cluster center coordinates

[0127]

[0128] in, Let q be the cluster center. Represents the q-th cluster center The coordinates of the Nth dimension.

[0129] (2) Calculation: Calculate the distance for each data sample.

[0130] For each data point in the dataset, calculate its distance to each cluster center and assign it to the nearest cluster center. For example, using Euclidean distance as the distance metric, the distance is calculated according to formula (5):

[0131]

[0132] in, The test sample f′ is the Dth vehicle series. D The nth eigenvalue in the middle. It is the nth cluster center The nth dimension coordinate value.

[0133] (3) Allocation: Based on the new test sample f′ D The distance to each cluster center determines the distance to the new test sample f′. D The vehicles are assigned to the nearest cluster, resulting in Q clusters. The data feature set of each vehicle in each cluster is then obtained. Let represent the feature vector of the J-th vehicle in the q-th cluster, and S represents the nth dimension feature of the j-th vehicle in the q-th cluster. q This represents the number of vehicles in the q-th cluster.

[0134] (4) Update: Recalculate the center of each cluster using equation (6). And assign to

[0135]

[0136] (5) Iteration: Repeat the assignment and update steps until the cluster centers remain unchanged, and output the final cluster centers, thereby grouping data with similar characteristics into one class, and obtaining...

[0137] Step 6, Construct a fault diagnosis model: This involves processing the data in Step 5. As input, a fault diagnosis model is constructed using a support vector machine (SVM) based on an intelligent optimization algorithm, such as... Figure 3 As shown, the details are as follows:

[0138] Step 6.1: Calculate the decision function using equation (7) to determine whether a fault has occurred. If it does, a malfunction has occurred; otherwise, it is considered normal.

[0139]

[0140] in, It is the kernel function of SVM. It represents the data feature vector of the J-th vehicle in the q-th cluster. Let y be the nth dimension feature of the j-th vehicle in the q-th cluster, p be the data category, and y be the nth dimension feature. n It is the label for the nth dimension, with a value of 1 or -1, α n is the Lagrange multiplier for the support vector machine, b is a constant, and sgn is the sign function.

[0141] Step 6.2: Initialize the parameters of the mathematical optimizer's acceleration function. Min and Max are the minimum and maximum values ​​of the acceleration function, respectively, with values ​​of 0.2 and 1. iter is the number of iterations, with a maximum iteration count of iter. max And the initialization is determined by equation (8) to determine the current search stage of the i-th time:

[0142] MOA(i)=Min+iter*(Max-Min) / iter max (8)

[0143] Step 6.3: Initialize the population size npop and the population positions X = [α1α2,…,α] n r1∈[0,1] is generated randomly.

[0144] If r1 > MOA(iter) value, proceed to step 6.4 exploration phase; otherwise, proceed to step 6.5 development phase.

[0145] Step 6.4, Exploration Phase: Randomly generate r2∈[0,1], and calculate the position update according to equation (9):

[0146]

[0147] Where μ is the control parameter for adjusting the search process, and X i,j (iter+1) represents the solution in iter+1 iterations, best(X) j ) is the current optimal position, ε is the minimum value to avoid a denominator of 0, UB j and LB jThese are the upper and lower limits of the search, and MOP is the coefficient, calculated according to equation (10):

[0148]

[0149] Here, η is a sensitive parameter that defines the development accuracy of the iterative process, and MOP(i) represents the coefficient of the i-th iteration.

[0150] Step 6.5, Development Phase: Randomly generate r3∈[0,1], and calculate the position update according to equation (11):

[0151]

[0152] Where μ is the control parameter for adjusting the search process, and X i,j (iter+1) represents the solution in iter+1 iterations, best(X) j ) is the current optimal position, ε is the minimum value to avoid a denominator of 0, UB j and LB j These are the upper and lower limits of the search, and MOP is the coefficient.

[0153] Step 6.6: Calculate the fitness value and record the optimal solution.

[0154] Step 6.7, determine if i satisfies the termination condition; if iter <iter max Repeat steps 6.2-6.7; otherwise, output the optimal parameters of the SVM model, i.e., by obtaining a. i Let b' be the Lagrange multiplier and b' be the constant, and establish the optimal detection model.

[0155] Compared with the prior art, this application has the following technical effects:

[0156] First, by building machine learning models, we can achieve effective fusion of multiple vehicle data, fully explore the potential information of multi-source data by utilizing the correlation between data, and improve the comprehensiveness and accuracy of seat fault diagnosis.

[0157] Secondly, by using machine learning methods, seat malfunctions under different operating conditions can be diagnosed. As vehicle operation changes over time, the machine learning model can be retrained by updating the training dataset to adapt to dynamic changes and maintain the effectiveness of fault diagnosis.

[0158] Third, by training with a large amount of labeled data, the complex mapping relationship between the normal state and fault state of the seat is learned, and the information is mapped to a high-dimensional feature space. The best model parameters are selected through intelligent optimization algorithms to improve the accuracy of fault diagnosis. At the same time, compared with traditional fault diagnosis, the diagnosis time is significantly shortened, the fault is prevented from deteriorating, and the efficiency and quality of fault resolution are improved.

[0159] Fourth, by providing early warnings of potential malfunctions, users can better plan equipment maintenance time and resources. Users are no longer forced to respond passively to malfunctions and carry out emergency repairs. They can rationally allocate maintenance personnel, reduce the inconvenience caused by equipment failures, minimize the impact on users' daily work and life, and improve user satisfaction and trust.

[0160] Please see Figure 4 The diagram below shows a structural schematic of a magnetorheological seat fault diagnosis system according to an embodiment of this application, including:

[0161] The acquisition module 401 is used to acquire a first dataset, which includes vehicle data and seat data corresponding to various road conditions.

[0162] Feature extraction module 402 is used to extract features from each data in the first dataset to obtain frequency domain data and time domain data;

[0163] The component analysis module 403 is used to perform principal component analysis on the frequency domain data and the time domain data to determine the variable characteristics associated with the magnetorheological seat under different fault types. The variable characteristics include at least one of the following: vehicle speed, steering wheel angle, steering wheel speed, throttle opening degree, brake pedal opening degree, vehicle height, and seat pressure, displacement, and magnetic field strength.

[0164] Clustering module 404 is used to perform clustering processing on the variable features associated with each fault type, and determine the feature clusters formed by the variable features associated with different fault types under each road condition;

[0165] The model building module 405 is used to input the feature clusters associated with different fault types under each road condition into a preset support vector machine for training, and build a fault diagnosis model.

[0166] The fault diagnosis module 406 uses the fault diagnosis model to detect the current vehicle data and seat data of the vehicle to determine whether the magnetorheological seat fault diagnosis is faulty.

[0167] Among them, the acquisition module 401 is used to perform the aforementioned step 101, the feature extraction module 402 is used to perform the aforementioned step 102, the component analysis module 403 is used to perform the aforementioned step 103, the clustering module 404 is used to perform the aforementioned step 104, the model construction module 405 is used to perform the aforementioned step 105, and the fault diagnosis module 406 is used to perform the aforementioned step 106. Each of the above modules corresponds to a step in the magnetorheological seat fault diagnosis method, and its execution process corresponds to the corresponding step, which will not be described in detail here.

[0168] The magnetorheological seat fault diagnosis system provided in this embodiment acquires a first dataset, which includes vehicle data and seat data corresponding to various road conditions. Based on multi-source data, data diversity is determined. Feature extraction is performed on each data point in the first dataset to obtain frequency domain data and time domain data. Principal component analysis is performed on the frequency domain data and the time domain data to determine the variable features associated with the magnetorheological seat under different fault types, covering the potential associated features, which is beneficial for subsequent accurate judgment. Clustering is performed on the variable features associated with each fault type to determine different fault types under each road condition. The feature clusters formed by the associated variable features ensure the integrity and diversity of the features. The feature clusters associated with different fault types under each road condition are input into a preset support vector machine for training to construct a fault diagnosis model. The fault diagnosis model is used for detection to determine whether the magnetorheological seat fault diagnosis is faulty. In this way, faults can be identified even in complex road conditions, improving the efficiency and accuracy of diagnosis. Accurate diagnosis of magnetorheological seat faults helps prevent potential safety hazards and reduces the risk of traffic accidents caused by seat malfunctions. At the same time, it also improves the reliability and service life of the seat and reduces maintenance costs.

[0169] In another embodiment, this application also provides a vehicle that includes the magnetorheological seat fault diagnosis system provided in any of the above embodiments.

[0170] For specific limitations regarding the vehicle, please refer to the limitations of the magnetorheological seat fault diagnosis method mentioned above, which will not be repeated here. The various modules in the aforementioned vehicle can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in hardware within or independently of the processor in the electronic device, or stored in software within the memory of the electronic device, so that the processor can call and execute the corresponding operations of each module.

[0171] This disclosure also provides an electronic device, including: a processor and a memory; the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the electronic device performs the above-described method.

[0172] Figure 5 A schematic diagram of a computer system suitable for implementing the embodiments of this application is shown. It should be noted that... Figure 5 The computer system 500 of the electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0173] like Figure 5As shown, the computer system 500 includes a Central Processing Unit (CPU) 501, which can perform various appropriate actions and processes according to a program stored in Read-Only Memory (ROM) 502 or a program loaded from storage portion 508 into Random Access Memory (RAM) 503, such as performing the methods described in the above embodiments. The RAM 503 also stores various programs and data required for system operation. The CPU 501, ROM 502, and RAM 503 are interconnected via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0174] The following components are connected to I / O interface 505: an input section 506 including a keyboard, mouse, etc.; an output section 507 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a LAN (Local Area Network) card, modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to I / O interface 505 as needed. Removable media 511, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 510 as needed so that computer programs read from them can be installed into storage section 508 as needed.

[0175] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the embodiments of this disclosure. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0176] The above embodiments are merely illustrative of the principles and effects of this application and are not intended to limit this application. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of this application. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this application should still be covered by the claims of this application.

Claims

1. A method for diagnosing faults in a magnetorheological seat, characterized in that, include: Obtain the first dataset, which includes vehicle data and seat data corresponding to various road conditions; Feature extraction is performed on each data point in the first dataset to obtain frequency domain data and time domain data; Principal component analysis is performed on the frequency domain data and the time domain data to determine the variable features associated with the magnetorheological seat under different fault types. These variable features include at least one of the following: vehicle speed, steering wheel angle, steering wheel speed, throttle opening, brake pedal opening, vehicle height, and seat pressure, displacement, and magnetic field strength. The frequency domain data and the time domain data are standardized to convert them into different feature matrices. The covariance matrix of each feature matrix is ​​calculated, and the correlation coefficient between each type of feature vector and different faults is calculated based on the covariance matrix. Eigenvalues ​​in the covariance matrix are calculated based on the correlation coefficients, and principal components of each type of feature vector are calculated according to the number of eigenvalues ​​and a preset contribution rate. Based on the principal components corresponding to each type of feature vector, the variable features associated with the magnetorheological seat under different fault types are determined. Clustering is performed on the variable features associated with each fault type to determine the feature clusters formed by the variable features associated with different fault types under each road condition; The feature clusters associated with different fault types under each road condition are input into a preset support vector machine for training to construct a fault diagnosis model; The fault diagnosis model is used to detect the current vehicle data and seat data of the vehicle to determine whether the magnetorheological seat fault diagnosis is faulty.

2. The method according to claim 1, characterized in that, Before performing feature extraction on each data point in the first dataset, the following steps are also included: The first dataset is preprocessed to obtain a second dataset, and the second dataset is used as the first dataset to be extracted and processed. The preprocessing includes at least one of noise processing, missing value processing, outlier processing, data integration processing, and data discretization processing.

3. The method according to claim 2, characterized in that, Before preprocessing the first dataset, the process also includes: The first dataset is divided into a first data subset representing seat faults and a second data subset representing no seat faults according to the magnetorheological seat faults. The first data subset and the second data subset are used as the first dataset to be preprocessed.

4. The method according to claim 1, characterized in that, Feature extraction is performed on each data point in the first dataset to obtain frequency domain data and time domain data, including: The vehicle data and seat data in the first dataset are subjected to feature extraction based on two dimensions: time domain and frequency domain, respectively, to obtain time domain data containing time domain features and frequency domain data containing frequency domain features. The time domain features include at least one of mean, variance, peak value, and skewness, and the frequency domain features include at least one of dominant frequency, frequency band energy, and frequency centroid. The variable features are a portion of the data determined in the first dataset.

5. The method according to claim 1, characterized in that, Clustering is performed on the variable features associated with each fault type to determine the feature clusters formed by the variable features associated with different fault types under each road condition, including: Select multiple cluster centers, and assign each variable feature associated with each fault type under different road conditions to the nearest cluster center among the multiple cluster centers; The assigned cluster centers are re-clustered until a preset number of clusterings is reached, thereby obtaining feature clusters formed by the variable features associated with different fault types under each road condition.

6. The method according to claim 1, characterized in that, The feature clusters associated with different fault types under each road condition are input into a preset support vector machine for training to construct a fault diagnosis model, including: The value range of the support vector machine is preset; Initialize the mathematical optimizer acceleration function, and determine the current iteration number, the maximum iteration number, and the acceleration function value; An initial population and its location are randomly generated in each of the feature clusters. The support vector machine is trained using each of the feature clusters, the current fitness value is calculated, and the current best fitness value is selected to determine the current optimal population position. Based on the comparison results between the initial population and the acceleration function value, it is determined whether the search phase starts from the exploration phase or the development phase. The population position is updated in different search phases, and the fitness value of all individuals is calculated in order to find the globally optimal population position. If the global optimal population position reaches the maximum number of iterations, then the current Lagrange multipliers and constants are output as the optimal parameters to obtain the fault diagnosis model.

7. The method according to any one of claims 1 to 6, characterized in that, The fault diagnosis model is used to detect the current vehicle data and seat data of the vehicle to determine whether the magnetorheological seat fault diagnosis is faulty, including: Input the current vehicle data and seat data of the vehicle into the fault diagnosis model to obtain the loss function value; The loss function value is compared with a preset fault threshold. If the loss function value is greater than or equal to a preset fault threshold, then the magnetorheological seat is determined to be faulty. If the loss function value is less than the preset fault threshold, then the magnetorheological seat is determined to be normal.

8. A magnetorheological seat fault diagnosis system, characterized in that, include: The acquisition module is used to acquire a first dataset, which includes vehicle data and seat data corresponding to various road conditions. The feature extraction module is used to extract features from each data in the first dataset to obtain frequency domain data and time domain data; The component analysis module is used to perform principal component analysis on the frequency domain data and the time domain data to determine the variable features associated with the magnetorheological seat under different fault types. These variable features include at least one of the following: vehicle speed, steering wheel angle, steering wheel speed, throttle opening, brake pedal opening, vehicle height, and seat pressure, displacement, and magnetic field strength. The module standardizes the frequency domain data and the time domain data, converting them into different feature matrices. It calculates the covariance matrix of each feature matrix and, based on the covariance matrix, calculates the correlation coefficient between each type of feature vector and different faults. It calculates the eigenvalues ​​in the covariance matrix based on the correlation coefficients and calculates the principal components of each type of feature vector according to the number of eigenvalues ​​and a preset contribution rate. Based on the principal components corresponding to each type of feature vector, it determines the variable features associated with the magnetorheological seat under different fault types. The clustering module is used to perform clustering processing on the variable features associated with each fault type, and to determine the feature clusters formed by the variable features associated with different fault types under each road condition; The model building module is used to input the feature clusters associated with different fault types under each road condition into a preset support vector machine for training, and to build a fault diagnosis model. The fault diagnosis module uses the fault diagnosis model to detect the current vehicle data and seat data of the vehicle to determine whether the magnetorheological seat fault diagnosis is faulty.

9. A vehicle, characterized in that, Including the magnetorheological seat fault diagnosis system as described in claim 8.

Images