Vehicle risk assessment methods and vehicles

Abstract

This invention discloses a vehicle risk assessment method and a vehicle. The method includes: acquiring vehicle risk assessment data, wherein the risk assessment data includes assessment indicators across multiple dimensions; performing principal component analysis on the assessment indicators across multiple dimensions to obtain a target assessment indicator; and using a risk assessment model to assess the target assessment indicator to obtain a vehicle risk assessment result, wherein the risk assessment result characterizes the level of safety risk posed by the vehicle. This invention solves the technical problem of low accuracy in existing vehicle risk assessment methods.

Description

Technical Field

[0001] This invention relates to the field of vehicle testing, and more specifically, to a vehicle risk assessment method and a vehicle. Background Technology

[0002] With the development of intelligent and connected vehicles, in-vehicle telematics boxes (TBOX), in-vehicle infotainment systems, and instrument clusters are widely used, greatly improving vehicle connectivity. At the same time, vehicle risk points and security access points have also increased, especially with the proliferation of wireless Fidelity (Wi-Fi) hotspots, Bluetooth, and mobile network capabilities. This makes vehicle terminals more vulnerable to attack and poses greater risks. Therefore, risk assessment of vehicles is a crucial means of improving vehicle security.

[0003] However, existing technologies for risk assessment of vehicles involve cumbersome procedures and require a significant amount of manpower and time, resulting in low accuracy in risk assessment.

[0004] There is currently no effective solution to the above problems. Summary of the Invention

[0005] This invention provides a vehicle risk assessment method and a vehicle to at least solve the technical problem of low accuracy in vehicle risk assessment in the prior art.

[0006] According to one aspect of the present invention, a vehicle risk assessment method is provided, comprising: acquiring vehicle risk assessment data, wherein the risk assessment data includes: assessment indicators of multiple dimensions; performing principal component analysis on the assessment indicators of multiple dimensions to obtain a target assessment indicator among the assessment indicators of multiple dimensions; and using a risk assessment model to assess the risk of the target assessment indicator to obtain a risk assessment result of the vehicle, wherein the risk assessment result is used to characterize the level of safety risk of the vehicle.

[0007] Optionally, principal component analysis is performed on the evaluation indicators of multiple dimensions to obtain the target evaluation indicator among the evaluation indicators of multiple dimensions, including: generating an indicator matrix based on the evaluation indicators of multiple dimensions, wherein the elements of each column in the indicator matrix are used to characterize the evaluation indicators of different dimensions; determining the eigenvalues ​​of the evaluation indicators of multiple dimensions based on the indicator matrix; and determining the target evaluation indicator based on the eigenvalues ​​of the evaluation indicators of multiple dimensions.

[0008] Optionally, based on the indicator matrix, the eigenvalues ​​of the evaluation indicators in multiple dimensions are determined, including: determining the covariance matrix of the indicator matrix; and determining the eigenvalues ​​of the evaluation indicators in multiple dimensions based on the covariance matrix.

[0009] Optionally, the target evaluation index is determined based on the feature values ​​of multiple evaluation indicators, including: determining the variance contribution rate of the multiple evaluation indicators based on the feature values ​​of the multiple evaluation indicators; sorting the multiple evaluation indicators in descending order according to the variance contribution rate to obtain the sorting result; and obtaining the target evaluation index from the sorting result, wherein the sum of the variance contribution rates of the target evaluation indicators is greater than the preset contribution rate.

[0010] Optionally, based on the eigenvalues ​​of the evaluation indicators in multiple dimensions, the variance contribution rate of the evaluation indicators in multiple dimensions is determined, including: determining any one of the evaluation indicators in multiple dimensions as the first evaluation indicator; obtaining the sum of the eigenvalues ​​of the evaluation indicators in multiple dimensions to obtain the total eigenvalue; obtaining the ratio of the eigenvalue of the first evaluation indicator to the total eigenvalue to obtain the variance contribution rate of the first evaluation indicator.

[0011] Optionally, the risk assessment model includes an input layer, at least one hidden layer, and an output layer. The risk assessment model is used to assess the risk of the target assessment indicators to obtain the risk assessment result of the vehicle. This includes: obtaining the target assessment indicators using the input layer; extracting features from the target assessment indicators using at least one hidden layer to obtain indicator features; and assessing the risk of the indicator features using the output layer to obtain the risk assessment result.

[0012] Optionally, the input layer contains the same number of neurons as the target evaluation metric, and the hidden layer contains 6 neurons.

[0013] Optionally, the method further includes: acquiring sample data collected in a real environment, wherein the sample data contains sample indicators of multiple dimensions; performing principal component analysis on the sample indicators of multiple dimensions to obtain the target sample indicator among the sample indicators of multiple dimensions; using a risk assessment model to conduct risk assessment on the target sample indicator to obtain the sample assessment result of the vehicle; constructing a loss function of the risk assessment model based on the sample assessment result and the real assessment result of the sample data; and adjusting the model parameters of the risk assessment model based on the loss function of the risk assessment model.

[0014] Optionally, the risk assessment data of the vehicle can be obtained, including: performing an attack feasibility analysis on the vehicle to obtain the attack feasibility level of the vehicle, wherein the attack feasibility level includes multiple first-dimensional indicators; performing an impact rating on the vehicle to obtain the vulnerability impact level of the vehicle, wherein the vulnerability impact level includes multiple second-dimensional indicators; and obtaining risk assessment data based on the attack feasibility level and the vulnerability impact level.

[0015] According to another aspect of the present invention, a vehicle risk assessment device is also provided, comprising: an acquisition module for acquiring vehicle risk assessment data, wherein the risk assessment data includes: assessment indicators of multiple dimensions; an analysis module for performing principal component analysis on the assessment indicators of multiple dimensions to obtain a target assessment indicator among the assessment indicators of multiple dimensions; and an assessment module for performing risk assessment on the target assessment indicator using a risk assessment model to obtain a risk assessment result for the vehicle, wherein the risk assessment result is used to characterize the level of safety risk of the vehicle.

[0016] According to another aspect of the present invention, a vehicle is also provided, comprising: a memory storing an executable program; and a processor for running the program, wherein the program executes the vehicle risk assessment method described above when it runs.

[0017] According to another aspect of the present invention, a non-volatile storage medium is also provided, the non-volatile storage medium including a stored program, wherein, when the program is running, it controls the execution of the vehicle risk assessment method described above in the processor of the device.

[0018] In this embodiment of the invention, risk assessment data of a vehicle is acquired, including: assessment indicators across multiple dimensions; principal component analysis is performed on the assessment indicators across multiple dimensions to obtain target assessment indicators; a risk assessment model is used to assess the target assessment indicators to obtain the risk assessment result of the vehicle. The risk assessment result is used to characterize the level of safety risk posed by the vehicle. It is noteworthy that the risk assessment result is obtained by assessing the target assessment indicators using a risk assessment model, and the target assessment indicators are obtained by principal component analysis of the assessment indicators across multiple dimensions. Principal component analysis and the risk assessment model avoid misjudgments during the assessment process and improve assessment efficiency, achieving the goal of accurately obtaining risk assessment results. This improves the accuracy of risk assessment and solves the technical problem of low accuracy in prior art vehicle risk assessment. Attached Figure Description

[0019] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0020] Figure 1 This is a flowchart of a vehicle risk assessment method according to an embodiment of the present invention;

[0021] Figure 2 This is a schematic diagram of the structure of an optional risk assessment model according to an embodiment of the present invention;

[0022] Figure 3 This is a flowchart of an optional vehicle risk assessment method according to an embodiment of the present invention;

[0023] Figure 4 This is a schematic diagram of the structure of a vehicle risk assessment device according to an embodiment of the present invention. Detailed Implementation

[0024] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0025] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention 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 so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0026] Example 1

[0027] According to an embodiment of the present invention, an embodiment of a vehicle risk assessment method is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0028] Figure 1 This is a flowchart of a vehicle risk assessment method according to an embodiment of the present invention, such as... Figure 1 As shown, the method includes the following steps:

[0029] Step S102: Obtain vehicle risk assessment data, which includes assessment indicators across multiple dimensions.

[0030] Optionally, risk assessment data for vehicles can be obtained through the following methods: conducting an attack feasibility analysis on the vehicle to obtain its attack feasibility level, where the attack feasibility level includes multiple first-dimensional indicators; conducting an impact rating on the vehicle to obtain its vulnerability impact level, where the vulnerability impact level includes multiple second-dimensional indicators; and obtaining risk assessment data based on the attack feasibility level and the vulnerability impact level.

[0031] The aforementioned attack feasibility level can be a rating of the difficulty of attacking vehicle vulnerabilities across multiple first-dimensional dimensions.

[0032] The aforementioned first dimensions may include, but are not limited to: attack time A1, the attacker's experience and knowledge A2, the attacker's understanding of the vehicle A3, the opportunity to attack A4, and the tools required for the attack A5.

[0033] The aforementioned multiple first-dimensional indicators can be obtained by assigning values ​​of 1 to 5 to multiple first-dimensional indicators. The larger the value of the first-dimensional indicator, the easier it is to attack the vulnerability of the vehicle corresponding to that first-dimensional indicator.

[0034] In one optional embodiment, an attack feasibility analysis is first performed on the vehicle to obtain its attack feasibility level. The attack feasibility of the vehicle may include, but is not limited to: A1 = {1,2,3,4,5}, A2 = {1,2,3,4,5}, A3 = {1,2,3,4,5}, A4 = {1,2,3,4,5}, and A5 = {1,2,3,4,5}. If A1 = 5, it indicates that the vulnerability of the vehicle corresponding to A1 is more easily attacked; if A1 = 1, it indicates that the vulnerability of the vehicle corresponding to A1 is less easily attacked. Similarly, if A2 = 5, it indicates that the vulnerability of the vehicle corresponding to A2 is more easily attacked; if A2 = 1, it indicates that the vulnerability of the vehicle corresponding to A2 is less easily attacked.

[0035] The aforementioned vulnerability impact level can be a hierarchy of the severity of vehicle vulnerabilities across multiple second-dimensional dimensions.

[0036] The aforementioned multiple second dimensions may include, but are not limited to: personal safety B1, property loss B2, vehicle operating status B3, privacy regulations B4, and national and social B5.

[0037] The aforementioned multiple second-dimensional indicators can be obtained by assigning values ​​of 1 to 5 to multiple second-dimensional indicators. The larger the value of the second-dimensional indicator, the deeper the impact of the vulnerability in the vehicle corresponding to that second-dimensional indicator.

[0038] In another optional embodiment, the vulnerability impact level of a vehicle can be obtained by performing an impact rating on the vehicle. The vulnerability impact level may include, but is not limited to: B1 = {1,2,3,4,5}, B2 = {1,2,3,4,5}, B3 = {1,2,3,4,5}, B4 = {1,2,3,4,5}, and B5 = {1,2,3,4,5}. If B1 = 5, it indicates that the vulnerability in the vehicle corresponding to B1 has a deeper impact; if B1 = 1, it indicates that the vulnerability in the vehicle corresponding to B1 has a shallower impact. Similarly, if B2 = 5, it indicates that the vulnerability in the vehicle corresponding to B2 has a deeper impact; if B2 = 1, it indicates that the vulnerability in the vehicle corresponding to B2 has a shallower impact.

[0039] In another optional embodiment, after obtaining the attack feasibility level and the vulnerability impact level of the vehicle, all attack feasibility levels and vulnerability impact levels can be summarized to obtain risk assessment data. That is, the risk assessment data can be A1={1,2,3,4,5}, A2={1,2,3,4,5}, A3={1,2,3,4,5}, A4={1,2,3,4,5}, A5={1,2,3,4,5}, B1={1,2,3,4,5}, B2={1,2,3,4,5}, B3={1,2,3,4,5}, B4={1,2,3,4,5}, B5={1,2,3,4,5}.

[0040] Step S104: Perform principal component analysis on the evaluation indicators of multiple dimensions to obtain the target evaluation indicator among the evaluation indicators of multiple dimensions.

[0041] Optionally, the target evaluation index can be obtained by the following methods: generating an index matrix based on evaluation indexes of multiple dimensions, wherein the elements in each column of the index matrix are used to characterize evaluation indexes of different dimensions; determining the feature values ​​of the evaluation indexes of multiple dimensions based on the index matrix; and determining the target evaluation index based on the feature values ​​of the evaluation indexes of multiple dimensions.

[0042] The aforementioned target evaluation indicators can be used as inputs to a risk assessment model.

[0043] The principal component analysis mentioned above can be Principal Component Analysis (PCA), but is not limited to this.

[0044] The aforementioned indicator matrix can be a matrix where rows represent multiple different dimensions, and columns represent different indicators for each dimension. For example, it could be R... m×n Where R represents the indicator matrix, m represents risk assessment data with m rows and different dimensions, and n represents different assessment indicators with n columns.

[0045] In one optional embodiment, risk assessment data from different dimensions can be processed using PCA, with rows representing the data and multiple indicators for each dimension representing the columns, resulting in an indicator matrix X, where X∈R. m×n .

[0046] Alternatively, eigenvalues ​​can be obtained by: determining the covariance matrix of the indicator matrix; and determining the eigenvalues ​​of multiple dimensions of the evaluation indicators based on the covariance matrix.

[0047] The covariance matrix mentioned above can be Y, and the specific calculation formula is as follows:

[0048]

[0049] Where X is the index matrix, X T Let R be the transpose of X, m be the number of risk assessment data points in different dimensions, i.e., the number of rows in the indicator matrix, and R be the transpose of X. n*n This represents risk assessment data from n different dimensions of behavior, listed as n assessment indicators.

[0050] The characteristic values ​​of the above-mentioned multiple evaluation indicators can be λ, and the specific calculation formula is as follows:

[0051] λ(IY)α=0,

[0052] Where I represents the identity matrix, and α is the eigenvector corresponding to the eigenvalue λ, which can take the value α. i , i = 0, ..., n.

[0053] In an alternative embodiment, the index matrix X and the transpose of X can be used. T The covariance matrix Y of the indicator matrix X is determined from the risk assessment data of m different dimensions. Then, the eigenvalues ​​λ of the assessment indicators of multiple dimensions can be determined based on the identity matrix I, the covariance matrix Y, and the eigenvectors α corresponding to the eigenvalues ​​λ.

[0054] Optionally, the target evaluation index can be determined by the following method: based on the feature values ​​of the evaluation indexes in multiple dimensions, determine the variance contribution rate of the evaluation indexes in multiple dimensions; sort the evaluation indexes in multiple dimensions in descending order according to the variance contribution rate to obtain the sorting result; obtain the target evaluation index from the sorting result, wherein the sum of the variance contribution rates of the target evaluation indexes is greater than the preset contribution rate.

[0055] Optionally, the variance contribution rate can be determined by the following method: determine any one of the evaluation indicators among multiple dimensions as the first evaluation indicator; obtain the sum of the eigenvalues ​​of the evaluation indicators of multiple dimensions to obtain the total eigenvalue; obtain the ratio of the eigenvalue of the first evaluation indicator to the total eigenvalue to obtain the variance contribution rate of the first evaluation indicator.

[0056] The first evaluation indicator mentioned above can be λ z , where z≤n.

[0057] The total eigenvalues ​​mentioned above can be λ0 + ... + λ n .

[0058] The aforementioned variance contribution rate can be obtained by dividing the eigenvalue λ of a certain eigenvector α by the sum of the eigenvalue λ of all eigenvectors α, and can be expressed as follows: Where z≤n.

[0059] The aforementioned preset contribution rate can be a contribution rate that the user sets in advance to filter out the target evaluation indicators. When the sum of the variance contribution rates of multiple evaluation indicators is greater than the preset contribution rate, multiple evaluation indicators can be identified as target evaluation indicators.

[0060] In another alternative embodiment, one of the evaluation indicators from multiple dimensions can first be determined as the first evaluation indicator λ. z Secondly, the sum of the eigenvalues ​​of multiple evaluation indicators can be obtained to get the total eigenvalue λ0 + ... + λ n Finally, the ratio of the eigenvalue of the first evaluation indicator to the total eigenvalue can be obtained, thus yielding the variance contribution rate of the first evaluation indicator. Where z≤n.

[0061] In another optional embodiment, the variance contribution rates of multiple evaluation indicators can first be sorted in descending order to obtain the sorting results. Finally, the multiple variance contribution rates can be added together in sequence. In response to the sum of variance contribution rates being greater than the preset contribution rate, the evaluation indicator corresponding to each variance contribution rate in the sum of variance contribution rates can be determined as the target evaluation indicator.

[0062] Step S106: Use the risk assessment model to conduct a risk assessment on the target assessment indicators to obtain the risk assessment results of the vehicle. The risk assessment results are used to characterize the level of safety risk of the vehicle.

[0063] Optionally, the risk assessment model includes an input layer, at least one hidden layer, and an output layer. The risk assessment model is used to assess the risk of the target assessment indicators to obtain the risk assessment result of the vehicle. This includes: obtaining the target assessment indicators using the input layer; extracting features from the target assessment indicators using at least one hidden layer to obtain indicator features; and assessing the risk of the indicator features using the output layer to obtain the risk assessment result.

[0064] Optionally, the input layer contains the same number of neurons as the target evaluation metric, and the hidden layer contains 6 neurons.

[0065] The risk assessment model described above can be a neural network model, but is not limited to it. The risk assessment model may include, but is not limited to, an input layer, at least one hidden layer, and an output layer. In this embodiment, an example with one input layer, two hidden layers, and one output layer is used, but it is not limited to this.

[0066] It should be noted that the number of neurons in the input layer is the same as the number of target evaluation metrics, and the hidden layer can contain multiple neurons. In this embodiment, 6 neurons are used as an example, but it is not limited to this.

[0067] Figure 2 This is a schematic diagram of the structure of an optional risk assessment model according to an embodiment of the present invention, such as... Figure 2 As shown, the risk assessment model includes an input layer containing z neurons, where the neurons are C1, ..., C1... z It consists of two hidden layers, each containing six neurons, and an output layer, where the output layer contains one neuron, SL. 11 W represents the connection between neuron C1 and the first neuron in the hidden layer. 12 The line connecting neuron C1 to the second neuron in the hidden layer, and so on, can be used to obtain W. 16 Connect the lines.

[0068] The aforementioned indicator characteristic can be H, and the specific calculation formula is as follows:

[0069]

[0070] Where H represents the indicator characteristic, w 1i c represents the weight. i b represents the input to the hidden layer. i This represents the bias value of the current layer.

[0071] It should be noted that the hidden layer uses the ReLU activation function, with the expression f(x) = max(0,x), which means that the output result is the maximum value among multiple values.

[0072] The above risk assessment results d SL (m) can represent the level of risk of a vehicle. A higher risk assessment result means that the vehicle is at greater risk, and a lower risk assessment result means that the vehicle is at lower risk.

[0073] In one optional embodiment, the target assessment index can first be obtained through the input layer of the risk assessment model. Then, features of the target assessment index can be extracted through at least one hidden layer to obtain index features H. Finally, the risk assessment result d can be obtained by performing risk assessment on the index features through the output layer. SL (m).

[0074] Optionally, the method further includes: acquiring sample data collected in a real environment, wherein the sample data contains sample indicators of multiple dimensions; performing principal component analysis on the sample indicators of multiple dimensions to obtain the target sample indicator among the sample indicators of multiple dimensions; using a risk assessment model to conduct risk assessment on the target sample indicator to obtain the sample assessment result of the vehicle; constructing a loss function of the risk assessment model based on the sample assessment result and the real assessment result of the sample data; and adjusting the model parameters of the risk assessment model based on the loss function of the risk assessment model.

[0075] The aforementioned sample data can be multiple data points collected by the user in advance from multiple real vehicles, and the sample data includes sample indicators of multiple dimensions.

[0076] The target sample index M mentioned above can be a sample index obtained through the PCA process.

[0077] The above sample evaluation results y SL (m) can be the assessment result obtained after risk assessment of the target sample indicators through a risk assessment model.

[0078] The loss function mentioned above can be L(Θ), and the specific calculation formula is as follows:

[0079]

[0080] In this embodiment of the invention, risk assessment data of a vehicle is acquired, including: assessment indicators across multiple dimensions; principal component analysis is performed on the assessment indicators across multiple dimensions to obtain target assessment indicators; a risk assessment model is used to assess the target assessment indicators to obtain the risk assessment result of the vehicle. The risk assessment result is used to characterize the level of safety risk posed by the vehicle. It is noteworthy that the risk assessment result is obtained by assessing the target assessment indicators using a risk assessment model, and the target assessment indicators are obtained by principal component analysis of the assessment indicators across multiple dimensions. Principal component analysis and the risk assessment model avoid misjudgments during the assessment process and improve assessment efficiency, achieving the goal of accurately obtaining risk assessment results. This improves the accuracy of risk assessment and solves the technical problem of low accuracy in prior art vehicle risk assessment.

[0081] This invention proposes a cybersecurity risk assessment method for vehicle controllers based on PCA and neural networks. By combining PCA principal component analysis with neural networks, it overcomes the one-sidedness and subjectivity of traditional assessment methods, and also solves the problem that traditional automotive cybersecurity risk assessment requires a lot of manual labor and time.

[0082] The specific implementation steps of this invention include the following:

[0083] 1. Attack feasibility and vulnerability impact level jointly determine the vehicle risk level. Five factors determine attack feasibility: attack time, attacker's experience and knowledge, attacker's understanding of the target, attack opportunity, and tools required for the attack. These five factors are assigned values ​​from 1 to 5. For example, attack time A1 = {1, 2, 3, 4, 5}, with higher values ​​indicating shorter attack time and easier vulnerability exploitation. Similarly, the other four factors determining attack feasibility are assigned values ​​according to the same rules: attacker's experience and knowledge A2 = {1, 2, 3, 4, 5}, attacker's understanding of the target A3 = {1, 2, 3, 4, 5}, attack opportunity A4 = {1, 2, 3, 4, 5}, and tools required for the attack A5 = {1, 2, 3, 4, 5}. The vulnerability impact level is determined by the following factors: personal safety B1, property damage B2, vehicle operating status B3, privacy regulations B4, and national / social factors B5. In addition to considering the impact of vulnerabilities on individuals, the impact on the nation and society should also be considered. Therefore, the impact on the nation and society is also included as one of the factors determining the vulnerability impact level, denoted as B5. Similarly, the five factors are assigned values ​​from 1 to 5: Personal safety B1 = {1, 2, 3, 4, 5}, Property loss B2 = {1, 2, 3, 4, 5}, Vehicle operating status B3 = {1, 2, 3, 4, 5}, Privacy regulations B4 = {1, 2, 3, 4, 5}, and National and social impact B5 = {1, 2, 3, 4, 5}. The higher the value, the deeper the impact of the vulnerability.

[0084] 2. The training set uses a large number of real-world risk assessment datasets, derived by experienced experts. Principal Component Analysis (PCA) is used to reduce the dimensionality of the training set, and the selected principal components are used as input to the neural network. The steps of PCA are as follows:

[0085] (1) The training set has m samples, each with dimension n. These samples are arranged in matrix form, where each row represents a sample and each column represents a dimension, resulting in an m×n sample matrix X, X∈R. m×n Calculate the covariance matrix of X. Y∈R n×n ;

[0086] (2) Calculate the eigenvalues ​​λ and corresponding eigenvectors α of the covariance matrix Y using the formula λ(IY)α=0. i , i = 1,...,n, I represents the identity matrix, and each eigenvector α represents a principal component;

[0087] (3) Calculate the variance contribution rate of each principal component, which is the sum of the eigenvalues ​​of all eigenvectors divided by the eigenvalues ​​of a certain eigenvector; calculate the cumulative variance contribution rate, which is the sum of the variance contribution rates of all current eigenvectors. Extract the first z principal components in descending order of variance contribution rate.

[0088] 3. The network constructed in this invention is as follows: Figure 2 As shown, the network consists of an input layer, an output layer, and hidden layers. The input layer has z neurons, the hidden layers have two layers (each with six neurons), and the output layer has one neuron, which outputs the vehicle risk level. The output of each layer serves as the input to the next layer, and the output of each layer is determined by the weights and input of the current layer. The output expression of the first hidden layer neuron is as follows: w 1i For the weights, C i b1 is the input, and b1 is the bias value of the current layer. The hidden layer uses the ReLU activation function, f(x) = max(0,x).

[0089] 4. Network Loss Function M is the number of samples in the dataset, y SL (m) represents the expected output of the vulnerability risk level, and d SL (m) represents the actual output of the vulnerability risk level. The loss function L(Θ) is the mean square error between the actual output and the expected output. The smaller the loss function, the better the parameters are trained, and the closer the actual output of the network is to the expected output.

[0090] Figure 3 This is a flowchart of an optional vehicle risk assessment method according to an embodiment of the present invention, such as... Figure 3 As shown, the method includes the following steps:

[0091] Step S301: Construct a training dataset using risk assessment data from real-world cases;

[0092] Step S302: Principal component analysis is used to reduce the dimensionality of the training set and extract principal components;

[0093] Step S303: The selected principal component is used as the input to the neural network, and machine learning is performed on the neural network model to train a risk assessment model.

[0094] Step S304: Use a risk assessment model to perform a risk assessment on the risk assessment data of the vehicle to be assessed.

[0095] The proposed method for assessing network security risks of vehicle controllers based on PCA and neural networks quantifies the network security risks of vehicle controllers, avoids misjudgments during the assessment process, and greatly improves assessment efficiency, making the risk assessment more objective and comprehensive, and has strong practicality.

[0096] Example 2

[0097] According to another aspect of the present invention, a vehicle risk assessment device is also provided. This device can perform the vehicle risk assessment method provided in Embodiment 1 above. The specific implementation method and preferred application scenario are the same as those in Embodiment 1 above, and will not be repeated here.

[0098] Figure 4 This is a schematic diagram of the structure of a vehicle risk assessment device according to an embodiment of the present invention, such as... Figure 4 As shown, the device includes: an acquisition module 40 for acquiring vehicle risk assessment data, wherein the risk assessment data includes assessment indicators of multiple dimensions; an analysis module 42 for performing principal component analysis on the assessment indicators of multiple dimensions to obtain the target assessment indicator among the assessment indicators of multiple dimensions; and an assessment module 44 for using a risk assessment model to perform risk assessment on the target assessment indicator to obtain the risk assessment result of the vehicle, wherein the risk assessment result is used to characterize the level of safety risk of the vehicle.

[0099] Optionally, the analysis module includes: a generation unit for generating an indicator matrix based on multiple dimensions of evaluation indicators, wherein the elements in each column of the indicator matrix are used to characterize evaluation indicators of different dimensions; a first determination unit for determining the feature values ​​of the multiple dimensions of evaluation indicators based on the indicator matrix; and a second determination unit for determining the target evaluation indicator based on the feature values ​​of the multiple dimensions of evaluation indicators.

[0100] Optionally, the first determining unit includes: a first determining subunit for determining the covariance matrix of the index matrix; and a second determining subunit for determining the eigenvalues ​​of the evaluation indexes of multiple dimensions based on the covariance matrix.

[0101] Optionally, the second determining unit includes: a third determining subunit, used to determine the variance contribution rate of the evaluation indicators of multiple dimensions based on the feature values ​​of the evaluation indicators of multiple dimensions; a sorting subunit, used to sort the evaluation indicators of multiple dimensions in descending order according to the variance contribution rate to obtain the sorting result; and an obtaining subunit, used to obtain the target evaluation indicator from the sorting result, wherein the sum of the variance contribution rates of the target evaluation indicators is greater than the preset contribution rate.

[0102] Optionally, the third determining subunit is further configured to: determine any one of the evaluation indicators among the multiple dimensions as the first evaluation indicator; obtain the sum of the eigenvalues ​​of the second evaluation indicator to obtain the total eigenvalue, wherein the second evaluation indicator is the evaluation indicator other than the second evaluation indicator among the multiple dimensions of evaluation indicators; obtain the ratio of the eigenvalue of the first evaluation indicator to the total eigenvalue to obtain the variance contribution rate of the first evaluation indicator.

[0103] Optionally, the risk assessment model includes an input layer, at least one hidden layer, and an output layer. The assessment module includes a first acquisition unit for acquiring target assessment indicators using the input layer; an extraction unit for extracting features from the target assessment indicators using at least one hidden layer to obtain indicator features; and a first assessment unit for performing risk assessment on the indicator features using the output layer to obtain risk assessment results.

[0104] Optionally, the input layer contains the same number of neurons as the target evaluation metric, and the hidden layer contains 6 neurons.

[0105] Optionally, the evaluation module further includes: a second acquisition unit for acquiring sample data collected in a real environment, wherein the sample data contains sample indicators of multiple dimensions; a first analysis unit for performing principal component analysis on the sample indicators of multiple dimensions to obtain the target sample indicator among the sample indicators of multiple dimensions; a second evaluation unit for using a risk assessment model to perform risk assessment on the target sample indicator to obtain the sample evaluation result of the vehicle; a construction unit for constructing the loss function of the risk assessment model based on the sample evaluation result and the real evaluation result of the sample data; and an adjustment unit for adjusting the model parameters of the risk assessment model based on the loss function of the risk assessment model.

[0106] Optionally, the acquisition module includes: a second analysis unit for performing an attack feasibility analysis on the vehicle to obtain the attack feasibility level of the vehicle, wherein the attack feasibility level includes multiple first-dimensional indicators; a rating unit for performing an impact rating on the vehicle to obtain the vulnerability impact level of the vehicle, wherein the vulnerability impact level includes multiple second-dimensional indicators; and a processing unit for obtaining risk assessment data based on the attack feasibility level and the vulnerability impact level.

[0107] Example 3

[0108] According to another aspect of the present invention, a vehicle is also provided, comprising: a memory storing an executable program; and a processor for running the program, wherein the program executes the vehicle risk assessment method described above when it runs.

[0109] Example 4

[0110] According to another aspect of the present invention, a non-volatile storage medium is also provided, the non-volatile storage medium including a stored program, wherein, when the program is running, it controls the execution of the vehicle risk assessment method described above in the processor of the device.

[0111] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0112] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0113] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0114] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0115] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0116] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0117] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A vehicle risk assessment method, characterized by, The method comprises: obtaining risk assessment data of the vehicle, wherein the risk assessment data comprises a plurality of dimensions of evaluation indicators; generating an indicator matrix based on the plurality of dimensions of evaluation indicators, wherein the elements of each column in the indicator matrix are used to represent different dimensions of evaluation indicators; determining eigenvalues of the plurality of dimensions of evaluation indicators based on the indicator matrix; determining a target evaluation indicator based on the eigenvalues of the plurality of dimensions of evaluation indicators; performing risk assessment on the target evaluation indicator by using a risk assessment model to obtain a risk assessment result of the vehicle, wherein the risk assessment result is used to represent the level of safety risk of the vehicle; wherein, based on the eigenvalues of the plurality of dimensions of evaluation indicators, determining the target evaluation indicator comprises: determining any one of the plurality of dimensions of evaluation indicators as a first evaluation indicator; obtaining the sum of the eigenvalues of the plurality of dimensions of evaluation indicators to obtain a total eigenvalue; obtaining the ratio of the eigenvalue of the first evaluation indicator to the total eigenvalue to obtain the variance contribution rate of the first evaluation indicator; sorting the plurality of dimensions of evaluation indicators in descending order according to the variance contribution rate to obtain a sorting result; obtaining the target evaluation indicator from the sorting result, wherein the sum of the variance contribution rates of the target evaluation indicators is greater than a preset contribution rate.

2. The vehicle risk assessment method of claim 1, wherein, Based on the indicator matrix, determining the eigenvalues of the plurality of dimensions of evaluation indicators comprises: determining the covariance matrix of the indicator matrix; determining the eigenvalues of the plurality of dimensions of evaluation indicators based on the covariance matrix.

3. The vehicle risk assessment method of claim 1, wherein, The risk assessment model comprises: an input layer, at least one hidden layer and an output layer, and performing risk assessment on the target evaluation indicator by using the risk assessment model to obtain the risk assessment result of the vehicle, comprising: obtaining the target evaluation indicator by using the input layer; extracting features of the target evaluation indicator by using the at least one hidden layer to obtain indicator features; performing risk assessment on the indicator features by using the output layer to obtain the risk assessment result.

4. The vehicle risk assessment method of claim 3, wherein, The number of neurons included in the input layer is the same as the number of target evaluation indicators, and the hidden layer includes 6 neurons.

5. The vehicle risk assessment method of claim 3, wherein, The method further comprises: obtaining sample data collected in a real environment, wherein the sample data comprises a plurality of dimensions of sample indicators; performing principal component analysis on the plurality of dimensions of sample indicators to obtain a target sample indicator among the plurality of dimensions of sample indicators; performing risk assessment on the target sample indicator by using the risk assessment model to obtain a sample evaluation result of the vehicle; constructing a loss function of the risk assessment model based on the sample evaluation result and a real evaluation result of the sample data; adjusting the model parameters of the risk assessment model based on the loss function of the risk assessment model.

6. The vehicle risk assessment method of claim 1, wherein, Obtaining risk assessment data of the vehicle comprises: performing attack feasibility analysis on the vehicle to obtain an attack feasibility level of the vehicle, wherein the attack feasibility level comprises a plurality of first dimension indicators; performing impact rating on the vehicle to obtain a vulnerability impact rating of the vehicle, wherein the vulnerability impact rating comprises indicators of a plurality of second dimensions; obtaining the risk assessment data based on the attack feasibility rating and the vulnerability impact rating.

7. A vehicle characterized by comprising: comprise: a memory storing an executable program; a processor configured to execute the program, wherein the program, when executed, performs the vehicle risk assessment method according to any one of claims 1 to 6.

Images