Automobile fault detection method, device, equipment and storage medium

Abstract

The present invention relates to the field of automobile fault detection technology, and specifically to an automobile fault detection method, device, equipment and storage medium. The present invention first collects data at each moment of the detected automobile in real time, uses the current moment data among the data at each moment to form a current sample matrix, then applies a dynamic principal component analysis algorithm to the current sample matrix to obtain an augmented matrix of the current sample matrix, then applies a kernel principal component analysis algorithm to the augmented matrix to obtain a kernel matrix, and then constructs a principal component matrix. Finally, the eigenvectors of the principal component matrix and the kernel matrix work together to determine the detection result of the automobile. The present invention uses data for judging the detection result to come from the detected automobile itself, inputs the objective data of the automobile data into a mathematical model composed of a dynamic principal component analysis algorithm and a kernel principal component analysis algorithm, and can quantitatively detect automobile faults based on the data output by the data model, thereby improving the detection accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of automobile fault detection, and in particular to an automobile fault detection method, device, equipment and storage medium. Background Art

[0002] Automotive fault detection includes both logic testing (which uses the logical relationships between statistical data to verify errors. The logical relationships within statistical data reflect the objective existence and inherent logic of things) and physical testing. The former involves extensive repetition and extensive content, and the functional logic must be modified for each vehicle. Due to the complexity of the functional logic, errors can easily occur during the modification process, reducing the accuracy of fault detection using the modified logic. Physical testing methods include experience-based testing and diagnosis, instrument-based testing and diagnosis, and self-diagnosis. Experience-based testing and diagnosis rely on maintenance personnel's prior experience and simple tests to locate faults. This lack of objective data leads to less accurate results. Instrument-based testing and diagnosis utilizes instrument-measured data parameters, states, curves, and waveforms to diagnose faults. However, maintenance personnel still need to address many fault types that have yet to be incorporated into mathematical models. Self-diagnosis utilizes the vehicle's self-diagnostic system to cross-check signal levels within the electronic control unit (ECU), with reference values stored in memory. If the signal level exceeds the permitted limit, the ECU identifies the signal as a fault and stores a fault code in memory. However, the vehicle's self-diagnosis system has not yet reached a level that can guarantee an accurate detection rate.

[0003] In summary, the existing automobile fault detection methods have low accuracy.

[0004] Therefore, the existing technology needs to be improved and enhanced. Summary of the Invention

[0005] To solve the above technical problems, the present invention provides a vehicle fault detection method, device, equipment and storage medium, which solves the problem of low accuracy of some vehicle fault detection methods.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] In a first aspect, the present invention provides a vehicle fault detection method, comprising:

[0008] Applying a dynamic principal component analysis algorithm to a current sample matrix to obtain an augmented matrix corresponding to the current sample matrix, wherein the current sample matrix is composed of current moment data of the detected vehicle;

[0009] Applying a kernel principal component analysis algorithm to the augmented matrix to obtain a kernel matrix corresponding to the augmented matrix;

[0010] Obtaining a principal component matrix according to each eigenvector of the kernel matrix, each eigenvalue of the kernel matrix, and the kernel matrix;

[0011] A detection result of the detected vehicle is obtained according to the principal component matrix and each eigenvector of the kernel matrix.

[0012] In one implementation, the dynamic principal component analysis algorithm is applied to the current sample matrix to obtain an augmented matrix corresponding to the current sample matrix, wherein the current sample matrix is composed of the current moment data of the detected vehicle, including:

[0013] Setting the model order of the dynamic principal component analysis algorithm;

[0014] According to the model order, the number of matrix columns is obtained;

[0015] The number of matrix rows is obtained according to the model order and the total number of samples, where the total number of samples is the number of samples contained in the sample library where the current sample is located;

[0016] Sequentially selecting samples preceding the current sample from the sample library until the number of selected samples reaches the number of matrix columns, wherein the generation time of data included in the samples preceding the current sample is before the generation time of data included in the current sample, and the selected samples are recorded as a first sample group;

[0017] Sequentially selecting samples that are located before the current sample from the sample library until the number of selected samples reaches the number of rows in the matrix, and the selected samples are recorded as a second sample group;

[0018] An augmented matrix is constructed according to the first sample group and the second sample group.

[0019] In one implementation, applying a kernel principal component analysis algorithm to the augmented matrix to obtain a kernel matrix corresponding to the augmented matrix includes:

[0020] Subtracting the column matrix from the row matrix of the augmented matrix to obtain an intermediate matrix;

[0021] Calculating the second norm of the intermediate matrix;

[0022] A kernel matrix is constructed according to the second norm of the intermediate matrix.

[0023] In one implementation, a subtraction operation is performed on the row matrices of the augmented matrix to obtain an intermediate matrix.

[0024] In one implementation, obtaining a principal component matrix based on each eigenvector of the kernel matrix, each eigenvalue of the kernel matrix, and the kernel matrix includes:

[0025] Accumulating each of the characteristic values to obtain a total cumulative sum;

[0026] Arrange the eigenvalues in descending order to obtain a sequence consisting of the eigenvalues;

[0027] The eigenvalues are sequentially screened from the head of the sequence until the ratio of the local cumulative sum corresponding to the screened eigenvalue to the total cumulative sum is greater than a threshold;

[0028] A principal component matrix is obtained based on the eigenvectors and the kernel matrix corresponding to the filtered eigenvalues.

[0029] In one implementation, obtaining the detection result of the detected vehicle based on the principal component matrix and each eigenvector of the kernel matrix includes:

[0030] Obtaining a data fluctuation index of the detected vehicle based on each eigenvector of the kernel matrix and the kernel matrix, wherein the data fluctuation index represents the data stability of the detected vehicle;

[0031] A detection result of the detected vehicle is obtained according to the principal component matrix and the data fluctuation index.

[0032] In one implementation, obtaining the detection result of the detected vehicle based on the principal component matrix and the data fluctuation index includes:

[0033] Multiplying the principal component matrix by the transposed matrix of the principal component matrix to obtain a first matrix;

[0034] Multiplying the kernel matrix by a matrix formed by each eigenvector of the kernel matrix to obtain an intermediate matrix;

[0035] Multiplying the intermediate matrix by the transposed matrix of the intermediate matrix to obtain a second matrix;

[0036] subtracting the first matrix from the second matrix to obtain a correlation matrix of the detected cars, wherein elements in the correlation matrix are used to represent the degree of correlation between the data of the detected cars;

[0037] Comparing the data fluctuation index with a set fluctuation control limit to obtain a first comparison result;

[0038] Comparing each element of the correlation matrix of the detected vehicle with a set correlation control limit to obtain a second comparison result;

[0039] A detection result of the detected vehicle is obtained according to the first comparison result and the second comparison result.

[0040] In one implementation, the calculation method of the set associated control limit includes:

[0041] Calculating the mean and standard deviation of each element value in the association matrix of the fault-free cars;

[0042] The associated control limit is obtained according to the mean value and the standard deviation.

[0043] In a second aspect, an embodiment of the present invention further provides a vehicle fault detection device, wherein the device includes the following components:

[0044] an augmented matrix calculation module, configured to apply a dynamic principal component analysis algorithm to a current sample matrix to obtain an augmented matrix corresponding to the current sample matrix, wherein the current sample matrix is composed of the current moment data of the detected vehicle;

[0045] A kernel matrix calculation module, configured to apply a kernel principal component analysis algorithm to the augmented matrix to obtain a kernel matrix corresponding to the augmented matrix;

[0046] A principal component matrix calculation module, configured to obtain a principal component matrix based on each eigenvector of the kernel matrix, each eigenvalue of the kernel matrix, and the kernel matrix;

[0047] The detection module is used to obtain the detection result of the detected vehicle according to the principal component matrix and each eigenvector of the kernel matrix.

[0048] In a third aspect, an embodiment of the present invention further provides a terminal device, wherein the terminal device includes a memory, a processor, and a vehicle fault detection program stored in the memory and executable on the processor, and when the processor executes the vehicle fault detection program, the steps of the above-mentioned vehicle fault detection method are implemented.

[0049] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, on which a vehicle fault detection program is stored. When the vehicle fault detection program is executed by a processor, the steps of the above-mentioned vehicle fault detection method are implemented.

[0050] Beneficial effects: The present invention first collects data of the detected car at each moment in real time, uses the current moment data in the data at each moment to form a current sample matrix, then applies a dynamic principal component analysis algorithm to the current sample matrix to obtain an augmented matrix of the current sample matrix, and then applies a kernel principal component analysis algorithm to the augmented matrix to obtain a kernel matrix, and then constructs a principal component matrix based on the eigenvectors, eigenvalues and the kernel matrix itself. Finally, the principal component matrix and the eigenvectors of the kernel matrix work together to determine the detection result of the car.

[0051] Because the correlation between the individual data in the augmented matrix of the present invention is low, using data with low correlation as the basis for detecting vehicle faults can improve the accuracy of detection results. Furthermore, the present invention combines the dynamic principal component analysis algorithm and the kernel principal component analysis algorithm to improve detection accuracy for the following reasons: In actual industrial processes, most system processes are nonlinear. Ignoring their nonlinearity only rarely has a significant impact, but for most systems, their inherent nonlinearity cannot be ignored. Directly using linearization methods will introduce significant errors, while kernel principal component analysis can address data nonlinearity. The same applies to dynamics: the impact of the temporal nature of process variables on the system should be considered. Traditional principal component analysis considers static models, where the observed value of a variable at a given moment is independent of the values before and after, and is mutually independent. This does not conform to the characteristics of actual process industries, where processes are expected to be slowly changing and dynamic relationships are prevalent. Dynamic principal component analysis can address data dynamics. In summary, the present invention combines the two algorithms to improve detection accuracy. Moreover, the data used by the present invention to judge the detection results comes from the detected vehicle itself. The objective data of the vehicle is input into a mathematical model composed of a dynamic principal component analysis algorithm and a kernel principal component analysis algorithm. According to the data output by the data model, the vehicle fault can be quantitatively detected, thereby improving the detection accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is the overall flow chart of the present invention;

[0053] Figure 2 T is the simulated temperature of the automobile engine in the embodiment of the present invention. 2 Schematic diagram;

[0054] Figure 3 This is a schematic diagram of an SPE simulated for automobile engine temperature in an embodiment of the present invention;

[0055] Figure 4 T is the temperature simulation result of the automobile motor in the embodiment of the present invention. 2 Schematic diagram;

[0056] Figure 5This is a schematic diagram of an SPE simulated for automobile motor temperature in an embodiment of the present invention;

[0057] Figure 6 This is a block diagram of the internal structure of a terminal device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0058] The following is a clear and complete description of the technical solutions of the present invention in conjunction with the embodiments and the accompanying drawings. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0059] Research has found that automotive fault detection methods include test logic and physical testing. The former involves extensive repetition and extensive content, and requires changes to the functional logic for each vehicle. Due to the complexity of the functional logic, errors can easily occur during the modification process, reducing the accuracy of fault detection using the modified functional logic. Physical testing methods include experience-based testing and diagnosis, instrument-based testing and diagnosis, and self-diagnosis. Experience-based testing and diagnosis requires maintenance personnel to rely on previous work experience and simple tests to locate the fault. This lack of objective data leads to poor test accuracy. Instrument-based testing and diagnosis utilizes instrument-measured data, including parameters, states, curves, and waveforms, to diagnose the fault. However, maintenance personnel still need to address many fault types that have not yet been incorporated into mathematical models. Self-diagnosis methods utilize the vehicle's self-diagnostic system to cross-check signal levels in the electronic control unit (ECU), with reference values stored in memory. If the signal level exceeds the allowable limit, the ECU identifies the signal as a fault and sends a fault code to memory. However, vehicle self-diagnostic systems have not yet reached a level that guarantees accurate detection rates.

[0060] To address the aforementioned technical issues, the present invention provides a vehicle fault detection method, apparatus, device, and storage medium, addressing the low accuracy of some vehicle fault detection methods. In specific implementation, a dynamic principal component analysis algorithm is first applied to the current sample matrix to obtain the augmented matrix corresponding to the current sample matrix. A kernel principal component analysis algorithm is then applied to the augmented matrix to obtain the kernel matrix corresponding to the augmented matrix. A principal component matrix is then obtained based on the eigenvectors and eigenvalues of the kernel matrix and the kernel matrix. Finally, the detection results of the detected vehicle are obtained based on the principal component matrix and the eigenvectors of the kernel matrix.

[0061] For example, the current data of the car is collected, such as the engine torque, engine speed, and engine coolant temperature of the car at the current moment. These three are used to form the current sample matrix. The dynamic principal component analysis algorithm is applied to the current sample matrix to generate an augmented matrix. The current sample matrix is one of the elements of the augmented matrix, and the other elements of the augmented matrix are matrices composed of the torque, speed, and coolant temperature generated by the engine at the moment before the current moment. Then, the kernel matrix of the augmented matrix is calculated and decentralized. The element value of the kernel matrix is the difference between the values of each element in the augmented matrix. Finally, the principal component matrix is obtained based on the eigenvectors, eigenvalues, and kernel matrix of the kernel matrix. The engine is judged whether it has a fault based on the elements in the principal component matrix.

[0062] Exemplary Methods

[0063] The automobile fault detection method of this embodiment can be applied to a terminal device, which can be a terminal product with computing functions, such as a computer. Figure 1 As shown in , the vehicle fault detection method specifically includes the following steps S100, S200, S300, and S400:

[0064] S100 , applying a dynamic principal component analysis algorithm to a current sample matrix to obtain an augmented matrix corresponding to the current sample matrix, wherein the current sample matrix is composed of current moment data of the detected vehicle.

[0065] For example, calculate the torque a of the car's engine at the current time t t , speed v t , coolant temperature w t , the torque a at time t-1 before the current time t t-1 , speed v t-1 , coolant temperature w t-1 , torque a at time t-2 t-2 , speed v t-2 , coolant temperature w t-2 Torque a t , speed v t , coolant temperature w t Arrange in order to form a row matrix x t ; Torque a t-1 , speed v t-1 , coolant temperature w t-1 Arrange in order to form a row matrix x t-1 ; Torque a t-2 , speed v t-2 , coolant temperature w t-2 Arrange in order to form a row matrix x t-2 .x t-2 、xt-1 、x t Arrange in order to form the total sample matrix x=[x t-2 ,x t-1 ,x t ].

[0066] a t 、v t 、w t 、a t-1 、v t-1 、w t-1 、a t-2 、v t-2 、w t-2 These are all values after standardization, such as calculation in a′ t , a′ t-1 , a′ t-2 The average value of σ is a′ t , a′ t-1 , a′ t-2 The standard deviation of the three, a′ t , a′ t-1 , a′ t-2 A t 、a t-1 、a t-2 The value before normalization. t 、w t 、a t-1 、v t-1 、w t-1 、a t-2 、v t-2 、w t-2 The same calculation method as above is used.

[0067] In one embodiment, step S100 includes the following steps S101, S102, S103, S104, S105, and S106:

[0068] S101, setting the model order of the dynamic principal component analysis algorithm DPCA.

[0069] The DPCA model order S is determined to consider the dynamic relationship between the current data and the past data. In this embodiment, S=2 is taken into consideration from the perspective of computational time cost.

[0070] S102: Obtain the number of matrix columns according to the model order S.

[0071] In this embodiment, the number of matrix columns included in the augmented matrix to be constructed is equal to S+1.

[0072] S103 , obtaining the number of matrix rows according to the model order and the total number of samples n, where the total number of samples is the number of samples included in the sample library where the current sample is located.

[0073] In this embodiment, the number of matrix rows included in the augmented matrix to be constructed is equal to n-S+1.

[0074] S104, sequentially selecting samples that precede the current sample from the sample library until the number of selected samples reaches the number of matrix columns, the generation time of the data included in the samples that precede the current sample is before the generation time of the data included in the current sample, and the selected samples are recorded as the first sample group.

[0075] If the current sample is x t , then the samples included in the first sample group are x t ,x t-1 ,…,x t-S .

[0076] S105 , sequentially selecting samples located before the current sample from the sample library until the number of selected samples reaches the number of matrix rows, and the selected samples are recorded as a second sample group.

[0077] If the current sample is x t , then the samples included in the first sample group are x t ,x t-1 ,…,x t+S-n

[0078] S106 : Construct an augmented matrix X(S) according to the first sample group and the second sample group.

[0079]

[0080] S200 , applying a kernel principal component analysis algorithm to the augmented matrix to obtain a kernel matrix corresponding to the augmented matrix.

[0081] Step S200 includes the following steps S201, S202, and S203:

[0082] S201 , performing a subtraction operation on the row matrix of the augmented matrix to obtain an intermediate matrix.

[0083] S202: Calculate the binary norm of the intermediate matrix.

[0084] S203: Construct a kernel matrix K according to the second norm of the intermediate matrix.

[0085] The kernel matrix K is the matrix after the matrix k is decentralized, and the element k(i,j) in the i-th row and j-th column of the matrix k is:

[0086]

[0087] k(i,j) is the element of matrix k in the form of matrix, X(S) i is the i-th row of the augmented matrix X(S), X(S) j is the jth row of the augmented matrix X(S), c is a constant, ||||2 is the second norm, x(S) ι -x(S) j is the intermediate matrix.

[0088] Decentralize the matrix k to obtain the kernel matrix K:

[0089] K=kI n k-kI n +I n kI n

[0090] I n The representative element is An n×n matrix.

[0091] S300 , obtaining a principal component matrix according to each eigenvector of the kernel matrix, each eigenvalue of the kernel matrix, and the kernel matrix.

[0092] In one embodiment, step S300 includes the following steps S301, S302, S303, and S304:

[0093] S301: Accumulate the characteristic values to obtain a total cumulative sum.

[0094] S302: Arrange the eigenvalues in descending order to obtain a sequence consisting of the eigenvalues.

[0095] S303 , screening the eigenvalues in sequence from the head of the sequence until the ratio of the local cumulative sum corresponding to the screened eigenvalue to the total cumulative sum is greater than a threshold.

[0096] S304: Obtain a principal component matrix according to the eigenvectors corresponding to the filtered eigenvalues and the kernel matrix.

[0097] S301 to S304 form the principal component matrix T according to the following formula A :

[0098]

[0099] T A =α A K

[0100] λ iis the i-th eigenvalue of the kernel matrix K, s is the total number of eigenvalues, is the total cumulative sum, is the local cumulative sum, λ j The jth eigenvalue in the sequence obtained in descending order, 85% is the threshold, α A is a matrix consisting of A eigenvectors.

[0101] For example, the kernel matrix K has a total of five eigenvalues λ1, λ2, λ3, λ4, λ5, which are arranged in descending order as λ5, λ2, λ4, λ1, λ3. When the sum of λ5 and λ2 (local cumulative sum) and λ1, λ2, λ3, λ4, λ5 (total cumulative sum) is less than 85%, and when the sum of λ5, λ2, λ4 (local cumulative sum) and λ1, λ2, λ3, λ4, λ5 (total cumulative sum) is greater than 85%, then the selected eigenvalues (denoted as principal elements) are λ5, λ2, λ4, and the three eigenvectors of the kernel matrix corresponding to these three eigenvalues (the value of A is 3 at this time) constitute α A .

[0102] S400 , obtaining a detection result of the detected vehicle according to the principal component matrix and each eigenvector of the kernel matrix.

[0103] In one embodiment, step S400 includes the following steps S401 to S408:

[0104] S401, obtaining the data fluctuation index T of the detected vehicle based on each eigenvector α of the kernel matrix and the kernel matrix K 2 The data fluctuation index user characterizes the data stability of the detected vehicle.

[0105] T=αK

[0106] T 2 =T A Λ -1 T A T

[0107] The diagonal matrix composed of Λ pivot elements.

[0108] For example, the current sample includes engine torque, engine speed, and engine coolant temperature, and the matrix T 2 The size of the internal element value corresponding to the engine torque reflects the fluctuation of the engine torque at each moment. The larger the element value is, the more unstable the engine torque is, which is reflected in the engine as an engine failure.

[0109] S402, the principal component matrix T A Multiply by the transposed matrix T of the principal component matrix AT , get the first matrix T A T A T .

[0110] S403 : Multiply the matrix α formed by each eigenvector of the kernel matrix by the kernel matrix K to obtain an intermediate matrix T=αK.

[0111] S404, multiplying the intermediate matrix by the transposed matrix T of the intermediate matrix T T , get the second matrix TT T .

[0112] S405 , subtracting the first matrix from the second matrix to obtain a correlation matrix SPE of the detected car, where elements in the correlation matrix are used to represent the degree of correlation between the data of the detected car.

[0113]

[0114] S406, the data fluctuation index T 2 and the set fluctuation control limit Compare and obtain a first comparison result.

[0115] In one embodiment, the value of β is 99%.

[0116] S407, each element of the correlation matrix SPE of the detected vehicle is respectively associated with the set correlation control limit Compare and obtain a second comparison result.

[0117] For example, one element of the correlation matrix SPE represents the correlation between the engine speeds at different times. When the element is greater than the speed corresponding to the When the value is greater than 0.05, it indicates that the speed is abnormal, that is, the engine is faulty.

[0118] In one embodiment, the associated control limits is calculated as follows:

[0119] The mean value mSPE and the standard deviation vSPE corresponding to each element value in the incidence matrix of the fault-free car are calculated.

[0120] According to the mean value mSPE and the standard deviation vSPE, the associated control limit is obtained

[0121] Where g = vSPE / (2*msSPE), h =(2*mSPE 2 ) / vSPE.

[0122] For example, collect data of cars that have not broken down at all times (such as engine temperature and torque), calculate the augmented matrix, kernel matrix, principal component matrix, and correlation matrix corresponding to the cars that have not broken down in turn, and find the mSPE and vSPE corresponding to the same data variable in multiple correlation matrices, such as the mSPE and vSPE of temperature, and bring the mSPE and vSPE of temperature into In the calculation formula, the associated control limit of temperature is obtained

[0123] S408: Obtain a detection result of the detected vehicle according to the first comparison result and the second comparison result.

[0124] (First comparison result), SPE is greater than (Second comparison result), if any one of the conditions is met, it means that the detected car has broken down.

[0125] The accuracy of the vehicle fault detection method of the present invention is illustrated below using the XMQ6127AGCHEVN61 hybrid bus as an example:

[0126] The sampling period is 1 second. 10,000 normal samples are used for offline training and 9,262 samples are used for online testing. Faults are introduced starting from the 1,001th online sample. Figures 2 to 5 The horizontal axis represents the sampling time, and the vertical axis represents T 2 Or SPE value. The straight line represents T 2 The control limit of the SPE or the control limit of the SPE, the curve represents the T of the tested car 2 Or SPE value.

[0127] Figure 2 and Figure 3 This is the fault detection diagram when the engine is overheated. It can be seen that Figure 2 and Figure 3 Starting from the 1001th sample, the curve begins to be higher than the straight line, indicating that the method of the present invention successfully detects the fault.

[0128] Figure 4 and Figure 5 This is the fault detection diagram when the motor is overheated. It can be seen that Figure 4 and Figure 5 Starting from the 1001th sample, at least the curve line where the SPE statistic is located begins to be higher than the straight line, indicating that the method of the present invention successfully detects the fault.

[0129] In summary, because the correlation between the individual data in the augmented matrix of the present invention is low, using data with low correlation as the basis for detecting vehicle faults can improve the accuracy of the detection results. Furthermore, the present invention combines the dynamic principal component analysis algorithm and the kernel principal component analysis algorithm. Furthermore, the data used to determine the detection results in the present invention comes from the vehicle being tested. By inputting this objective vehicle data into the mathematical model constructed by the dynamic principal component analysis algorithm and the kernel principal component analysis algorithm, the vehicle fault can be quantitatively detected based on the data output by the data model, thereby improving detection accuracy.

[0130] The present invention is also applicable to fault detection of hybrid vehicles. It adopts the principal component analysis algorithm in data-driven and establishes an accurate data model. Compared with the automobile self-diagnosis system based on manual experience and non-real-time, the present invention optimizes the timeliness and accuracy of fault detection.

[0131] When the detected vehicle is a hybrid vehicle, the present invention takes into account the dynamic and nonlinear problems of hybrid vehicle data, adopts a dynamic kernel principal component analysis algorithm, and improves the accuracy of fault detection.

[0132] Exemplary devices

[0133] This embodiment also provides a vehicle fault detection device, which includes the following components:

[0134] an augmented matrix calculation module, configured to apply a dynamic principal component analysis algorithm to a current sample matrix to obtain an augmented matrix corresponding to the current sample matrix, wherein the current sample matrix is composed of the current moment data of the detected vehicle;

[0135] A kernel matrix calculation module, configured to apply a kernel principal component analysis algorithm to the augmented matrix to obtain a kernel matrix corresponding to the augmented matrix;

[0136] A principal component matrix calculation module, configured to obtain a principal component matrix based on each eigenvector of the kernel matrix, each eigenvalue of the kernel matrix, and the kernel matrix;

[0137] The detection module is used to obtain the detection result of the detected vehicle according to the principal component matrix and each eigenvector of the kernel matrix.

[0138] Based on the above embodiment, the present invention further provides a terminal device, whose principle block diagram can be shown as follows: Figure 6As shown. The terminal device includes a processor, a memory, a network interface, a display screen, and a temperature sensor connected via a system bus. The processor of the terminal device is used to provide computing and control capabilities. The memory of the terminal device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the terminal device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for detecting automobile faults is implemented. The display screen of the terminal device can be a liquid crystal display screen or an electronic ink display screen, and the temperature sensor of the terminal device is pre-set inside the terminal device to detect the operating temperature of the internal device.

[0139] Those skilled in the art will understand that Figure 6 The principle block diagram shown in the figure is only a block diagram of a partial structure related to the solution of the present invention, and does not constitute a limitation on the terminal device to which the solution of the present invention is applied. The specific terminal device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0140] In one embodiment, a terminal device is provided. The terminal device includes a memory, a processor, and an automobile fault detection program stored in the memory and executable on the processor. When the processor executes the automobile fault detection program, the following operating instructions are implemented:

[0141] Applying a dynamic principal component analysis algorithm to a current sample matrix to obtain an augmented matrix corresponding to the current sample matrix, wherein the current sample matrix is composed of current moment data of the detected vehicle;

[0142] Applying a kernel principal component analysis algorithm to the augmented matrix to obtain a kernel matrix corresponding to the augmented matrix;

[0143] Obtaining a principal component matrix according to each eigenvector of the kernel matrix, each eigenvalue of the kernel matrix, and the kernel matrix;

[0144] A detection result of the detected vehicle is obtained according to the principal component matrix and each eigenvector of the kernel matrix.

[0145] Those skilled in the art will appreciate that all or part of the processes in the above-described embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the above-described embodiments. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct RAM bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM).

[0146] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention 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 various embodiments of the present invention.

Claims

1. A method for detecting automobile faults, characterized in that: include: Applying a dynamic principal component analysis algorithm to a current sample matrix to obtain an augmented matrix corresponding to the current sample matrix, wherein the current sample matrix is composed of current moment data of the detected vehicle; Applying a kernel principal component analysis algorithm to the augmented matrix to obtain a kernel matrix corresponding to the augmented matrix; Obtaining a principal component matrix according to each eigenvector of the kernel matrix, each eigenvalue of the kernel matrix, and the kernel matrix; Obtaining a detection result of the detected vehicle according to the principal component matrix and each eigenvector of the kernel matrix; The dynamic principal component analysis algorithm is applied to the current sample matrix to obtain an augmented matrix corresponding to the current sample matrix, wherein the current sample matrix is composed of the current moment data of the detected vehicle, including: Setting the model order of the dynamic principal component analysis algorithm; According to the model order, the number of matrix columns is obtained; The number of matrix rows is obtained according to the model order and the total number of samples, where the total number of samples is the number of samples contained in the sample library where the current sample is located; Sequentially selecting samples preceding the current sample from the sample library until the number of selected samples reaches the number of matrix columns, wherein the generation time of data included in the samples preceding the current sample is before the generation time of data included in the current sample, and the selected samples are recorded as a first sample group; Sequentially selecting samples that are located before the current sample from the sample library until the number of selected samples reaches the number of rows in the matrix, and the selected samples are recorded as a second sample group; constructing an augmented matrix based on the first sample group and the second sample group; The step of applying a kernel principal component analysis algorithm to the augmented matrix to obtain a kernel matrix corresponding to the augmented matrix includes: Performing a subtraction operation on the row matrix of the augmented matrix to obtain an intermediate matrix; Calculating the second norm of the intermediate matrix; A kernel matrix is constructed according to the second norm of the intermediate matrix.

2. The vehicle fault detection method according to claim 1, wherein: The principal component matrix is obtained based on each eigenvector of the kernel matrix, each eigenvalue of the kernel matrix, and the kernel matrix, including: Accumulating each of the characteristic values to obtain a total cumulative sum; Arrange the eigenvalues in descending order to obtain a sequence consisting of the eigenvalues; The eigenvalues are sequentially screened from the head of the sequence until the ratio of the local cumulative sum corresponding to the screened eigenvalue to the total cumulative sum is greater than a threshold; A principal component matrix is obtained based on the eigenvectors and the kernel matrix corresponding to the filtered eigenvalues.

3. The vehicle fault detection method according to claim 1, wherein: The detecting result of the detected vehicle is obtained based on the principal component matrix and each eigenvector of the kernel matrix, including: Obtaining a data fluctuation index of the detected vehicle based on each eigenvector of the kernel matrix and the kernel matrix, wherein the data fluctuation index represents the data stability of the detected vehicle; A detection result of the detected vehicle is obtained according to the principal component matrix and the data fluctuation index.

4. The vehicle fault detection method according to claim 3, wherein: Obtaining the detection result of the detected vehicle based on the principal component matrix and the data fluctuation index includes: Multiplying the principal component matrix by the transposed matrix of the principal component matrix to obtain a first matrix; Multiplying the kernel matrix by a matrix formed by each eigenvector of the kernel matrix to obtain an intermediate matrix; Multiplying the intermediate matrix by the transposed matrix of the intermediate matrix to obtain a second matrix; subtracting the first matrix from the second matrix to obtain a correlation matrix of the detected cars, wherein elements in the correlation matrix are used to represent the degree of correlation between the data of the detected cars; Comparing the data fluctuation index with a set fluctuation control limit to obtain a first comparison result; Comparing each element of the correlation matrix of the detected vehicle with a set correlation control limit to obtain a second comparison result; A detection result of the detected vehicle is obtained according to the first comparison result and the second comparison result.

5. The vehicle fault detection method according to claim 4, wherein: The calculation method of the associated control limit is set, including: Calculating the mean and standard deviation of each element value in the association matrix of the fault-free cars; The associated control limit is obtained according to the mean value and the standard deviation.

6. An automobile fault detection device, characterized in that: The device comprises the following components: an augmented matrix calculation module, configured to apply a dynamic principal component analysis algorithm to a current sample matrix to obtain an augmented matrix corresponding to the current sample matrix, wherein the current sample matrix is composed of the current moment data of the detected vehicle; A kernel matrix calculation module, configured to apply a kernel principal component analysis algorithm to the augmented matrix to obtain a kernel matrix corresponding to the augmented matrix; A principal component matrix calculation module, configured to obtain a principal component matrix based on each eigenvector of the kernel matrix, each eigenvalue of the kernel matrix, and the kernel matrix; A detection module, configured to obtain a detection result of the detected vehicle based on the principal component matrix and each eigenvector of the kernel matrix; The dynamic principal component analysis algorithm is applied to the current sample matrix to obtain an augmented matrix corresponding to the current sample matrix, wherein the current sample matrix is composed of the current moment data of the detected vehicle, including: Setting the model order of the dynamic principal component analysis algorithm; According to the model order, the number of matrix columns is obtained; The number of matrix rows is obtained according to the model order and the total number of samples, where the total number of samples is the number of samples contained in the sample library where the current sample is located; Sequentially selecting samples preceding the current sample from the sample library until the number of selected samples reaches the number of matrix columns, wherein the generation time of data included in the samples preceding the current sample is before the generation time of data included in the current sample, and the selected samples are recorded as a first sample group; Sequentially selecting samples that are located before the current sample from the sample library until the number of selected samples reaches the number of rows in the matrix, and the selected samples are recorded as a second sample group; constructing an augmented matrix based on the first sample group and the second sample group; The step of applying a kernel principal component analysis algorithm to the augmented matrix to obtain a kernel matrix corresponding to the augmented matrix includes: Performing a subtraction operation on the row matrix of the augmented matrix to obtain an intermediate matrix; Calculating the second norm of the intermediate matrix; A kernel matrix is constructed according to the second norm of the intermediate matrix.

7. A terminal device, characterized in that: The terminal device includes a memory, a processor, and a vehicle fault detection program stored in the memory and running on the processor. When the processor executes the vehicle fault detection program, the steps of the vehicle fault detection method according to any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium, characterized in that The computer-readable storage medium stores an automobile fault detection program, and when the automobile fault detection program is executed by the processor, the steps of the automobile fault detection method according to any one of claims 1 to 5 are implemented.

Images