Vehicle collision test simulation degree prediction method and device and electronic equipment

Abstract

The invention relates to the technical field of mathematical simulation, in particular to a vehicle collision test simulation degree prediction method and device and electronic device.The method comprises the steps that a training set of a vehicle collision test is obtained, and the training set comprises test parameter samples and a simulation degree result of the collision test corresponding to each sample; calculating an entropy value of each subset after each test parameter in the training set divides the training set, and calculating a corresponding overall entropy value after the test parameter divides the training set according to the entropy value; and gradually determining a plurality of division nodes in the training set according to the overall entropy value corresponding to each test parameter after the training set is divided, generating a prediction model of the vehicle collision test simulation degree according to the divided training set, and performing collision test simulation degree prediction on the target vehicle collision test by using the prediction model. Therefore, the problems that the simulation degree prediction cost of the vehicle collision test is high and the vehicle collision test is difficult to popularize and apply in the prior art are solved.

Description

Technical Field

[0001] The present application relates to the field of vehicle technology, and in particular to a method, device and electronic equipment for predicting the simulation degree of a vehicle collision test. Background Art

[0002] Due to the current application of technologies such as intelligent driving technology and zero-gravity seats, as well as the diversification of vehicle appearance, occupant diversity, and vehicle interior diversity, traffic accident forms have become increasingly complex, which in turn leads to vehicle collision accidents being affected by the coupling of multiple key accident parameters. For example, the occurrence of vehicle rollover accidents will be affected by many factors such as environmental conditions, operating parameters, and vehicle parameters. It has the characteristics of multiple factors and high coupling, which leads to the inability to predict whether the set test parameters can meet the simulation requirements when performing rollover tests.

[0003] Related technology The prediction of vehicle collision simulation degree usually adopts the method of combining simulation with real vehicle test. Through simulation test conditions, the approximate parameter setting range is selected, and then further verification is carried out through real vehicle test to improve the accuracy of vehicle collision. However, vehicle collision is a multi-factor, high-coupling test form. The test conditions set in the simulation software are quite different from the actual test, and further verification is required through real vehicle test. However, the cost of real vehicle test is high, and re-verification is required after the vehicle model parameters change, which cannot be promoted and applied. Summary of the invention

[0004] The present application provides a vehicle collision test simulation degree prediction method, device and electronic equipment to solve the problems in the related art that the vehicle collision test simulation degree prediction cost is high and difficult to promote and apply.

[0005] A first aspect embodiment of the present application provides a method for predicting the simulation degree of a vehicle collision test, comprising the following steps: obtaining a training set of a vehicle collision test, wherein the training set includes test parameter samples and a collision test simulation degree result corresponding to each sample; calculating the entropy value of each subset after the training set is divided by each test parameter in the training set, and calculating the overall entropy value corresponding to the division of the training set by the test parameter based on the entropy value; gradually determining multiple division nodes in the training set based on the overall entropy value corresponding to the division of the training set by each test parameter, generating a prediction model for the vehicle collision test simulation degree based on the divided training set, and using the prediction model to predict the collision test simulation degree of a target vehicle collision test.

[0006] Optionally, before gradually determining multiple division nodes in the training set according to the overall entropy value corresponding to each test parameter after the training set is divided, it also includes: identifying the data type of the test parameter sample; if the data type of the test parameter sample is a discrete type, determining the division node according to the sample proportion and weight corresponding to the test parameter under different values ​​and the overall entropy value corresponding to the training set after the test parameter is divided; if the data type of the test parameter sample is a continuous type, discretizing the value of the test parameter sample to obtain multiple discrete values, and determining the division node according to the sample proportion and weight corresponding to each discretized value and the overall entropy value corresponding to the test parameter after the training set is divided.

[0007] Optionally, the calculation formula of the overall entropy value after the discrete type test parameters divide the training set is: ; in, is a discrete type test parameter, D is a training set, For test parameters Divide the training set The overall entropy value of the post-training set, is the entropy value of the training set D, The test parameters The number of values ​​of For the training set Medium test parameters The value of A subset of all test parameter samples of is the total number of samples in the training set D, For subset The proportion of the number of samples in the total number of test parameter samples, For subset The entropy value of .

[0008] Optionally, the calculation formula for the overall entropy value after the continuous type test parameter divides the training set is:

[0009] in, is the test parameter of the continuous type, is the training set, For test parameters Divide the training set The overall entropy value of the post-training set, All possible partition values ​​for the test parameter b A collection of To use the test parameters Divide the value Divide the training set The overall entropy value of the post-training set, For the training set The entropy value of To use the test parameters Divide the value Divide the training set satisfy A subset of conditions, For subset The number of samples in is the total number of samples in the training set D, For subset The entropy value of .

[0010] Optionally, multiple division nodes in the training set are gradually determined according to the corresponding overall entropy value after the training set is divided by each test parameter, including: in the current division stage of the training set, calculating the current overall entropy value of the training set; determining the division node of the training set in the current division stage based on the current overall entropy value; using the division node to divide the training set into multiple sub-training sets, and entering the next division stage until the division stop condition is met, wherein, when calculating the overall entropy value of the training set in each division stage, the division node of the previous division stage is ignored; and multiple division nodes in the training set are determined according to multiple division stages.

[0011] Optionally, before obtaining the training set of the vehicle collision test, it also includes: obtaining multiple test parameters of historical vehicle collision tests and corresponding collision test results; based on the multiple test parameters of historical vehicle collision tests and corresponding collision test simulation results, constructing a training set, a test set and a verification set for the vehicle collision test.

[0012] Optionally, after generating a prediction model for the simulation degree of a vehicle collision test based on the divided training set, the method further includes: using each division node in the training set to divide the test parameter samples in the validation set; calculating a first proportion of the test parameter samples before the verification set is divided in the simulation degree results of the collision test; calculating a second proportion of the test parameter samples after the verification set is divided in the simulation degree results of the collision test; calculating the difference between the first proportion and the second proportion, and adjusting the category of the collision test simulation degree results of the corresponding test parameter samples in the division node according to the difference.

[0013] Optionally, after generating a prediction model of the vehicle collision test simulation degree according to the divided training set, it also includes: using a test set to test the prediction accuracy of the prediction model; if the prediction accuracy is lower than a preset threshold, optimizing the division nodes of the prediction model.

[0014] A second aspect of the present application provides a vehicle collision test simulation degree prediction device, including: an acquisition module, used to acquire a training set of a vehicle collision test, wherein the training set includes test parameter samples and the simulation degree results of the collision test corresponding to each sample; a calculation module, used to calculate the entropy value of each subset after the training set is divided by each test parameter in the training set, and calculate the corresponding overall entropy value after the training set is divided by the test parameter based on the entropy value; a generation module, used to gradually determine multiple division nodes in the training set based on the overall entropy value corresponding to the training set after each test parameter is divided, generate a vehicle collision test simulation degree prediction model based on the divided training set, and use the prediction model to predict the collision test simulation degree of a target vehicle collision test.

[0015] Optionally, it also includes: an identification module, which is used to identify the data type of the test parameter sample before gradually determining multiple division nodes in the training set according to the overall entropy value corresponding to each test parameter after the training set is divided; if the data type of the test parameter sample is a discrete type, the division node is determined according to the sample proportion and weight corresponding to the test parameter under different values ​​and the overall entropy value corresponding to the training set after the test parameter is divided; if the data type of the test parameter sample is a continuous type, the value of the test parameter sample is discretized to obtain multiple discrete values, and the division node is determined according to the sample proportion and weight corresponding to each discretized value and the overall entropy value corresponding to the test parameter after the training set is divided.

[0016] Optionally, the calculation formula of the overall entropy value after the discrete type test parameters divide the training set is: ; in, is a discrete type test parameter, D is a training set, For test parameters Divide the training set The overall entropy value of the post-training set, is the entropy value of the training set D, The test parameters The number of values ​​of For the training set Medium test parameters The value of A subset of all test parameter samples of is the total number of samples in the training set D, For subset The proportion of the number of samples in the total number of test parameter samples, For subset The entropy value of .

[0017] Optionally, the calculation formula for the overall entropy value after the continuous type test parameter divides the training set is:

[0018] in, is the test parameter of the continuous type, is the training set, For test parameters Divide the training set The overall entropy value of the post-training set, All possible partition values ​​for the test parameter b A collection of To use the test parameters Divide the value Divide the training set The overall entropy value of the post-training set, For the training set The entropy value of To use the test parameters Divide the value Divide the training set satisfy A subset of conditions, For subset The number of samples in is the total number of samples in the training set D, For subset The entropy value of .

[0019] Optionally, the generation module is further used to: calculate the current overall entropy value of the training set in the current division stage of the training set; determine the division node of the training set in the current division stage based on the current overall entropy value; use the division node to divide the training set into multiple sub-training sets, and enter the next division stage until the division stop condition is met, wherein, when calculating the overall entropy value of the training set in each division stage, the division node of the previous division stage is ignored; and determine multiple division nodes in the training set according to multiple division stages.

[0020] Optionally, it also includes: a construction module, which is used to obtain multiple test parameters of historical vehicle collision tests and corresponding simulation results of collision tests before obtaining the training set of vehicle collision tests; based on the multiple test parameters of historical vehicle collision tests and corresponding simulation results of collision tests, construct a training set, a test set and a verification set for the vehicle collision test.

[0021] Optionally, it also includes: a verification module, which is used to divide the test parameter samples in the verification set by using each division node in the training set after generating a prediction model of the vehicle collision test simulation degree according to the divided training set; calculate the first proportion of the test parameter samples before the verification set is divided in the simulation degree results of the collision test; calculate the second proportion of the test parameter samples after the verification set is divided in the simulation degree results of the collision test; calculate the difference between the first proportion and the second proportion, and adjust the category of the simulation degree results of the collision test of the corresponding test parameter samples in the division node according to the difference.

[0022] Optionally, it also includes: a testing module, which is used to test the prediction accuracy of the prediction model using a test set after generating a prediction model of the vehicle collision test simulation degree according to the divided training set; if the prediction accuracy is lower than a preset threshold, the division nodes of the prediction model are optimized.

[0023] A third aspect of the present application provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to execute a vehicle collision test simulation degree prediction method as described in the above embodiment.

[0024] The fourth aspect of the present application provides a computer-readable storage medium having a computer program or instructions stored thereon, and the computer program or instructions are executed by a processor to perform a vehicle collision test simulation degree prediction method as shown.

[0025] The fifth aspect of the present application provides a computer program product, including a computer program or instructions. When the computer program or instructions are executed, the vehicle collision test simulation degree prediction method as described in the above embodiment is implemented.

[0026] Therefore, this application has at least the following beneficial effects: The embodiment of the present application can calculate the entropy value of each subset after each test parameter in each training set divides the training set, and calculate the overall entropy value corresponding to the training set after the training set is divided by the test parameter according to the entropy value, and then gradually determine multiple division nodes in the training set according to the overall entropy value corresponding to each test parameter after the training set is divided, and generate a prediction model of the vehicle collision test simulation degree according to the divided training set, so that the vehicle collision test simulation degree can be predicted in advance based on the prediction model. When facing changes in test parameters, the collision test simulation degree can be predicted directly based on the prediction model without the need for actual vehicle verification, which is more efficient, less costly, and easy to promote. In addition, in the process of constructing the prediction model, a variety of test parameters related to the vehicle collision simulation degree are comprehensively considered, and the prediction accuracy is higher. Therefore, the technical problems such as the high cost of vehicle collision test simulation degree prediction and the difficulty in popularization and application in the related technology are solved.

[0027] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which: Figure 1 A flowchart of a vehicle collision test simulation degree prediction method provided according to an embodiment of the present application; Figure 2 A flowchart of a vehicle collision test simulation degree prediction method provided according to a specific embodiment of the present application; Figure 3 This is an example diagram of a vehicle collision test simulation degree prediction device provided according to an embodiment of the present application; Figure 4 It is a schematic diagram of the structure of an electronic device provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0029] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0030] The following describes the vehicle collision test simulation degree prediction method, device and electronic device of the embodiments of the present application with reference to the accompanying drawings. In response to the problem mentioned in the above background technology center that in the prediction of the simulation degree of a collision test based on simulation combined with a real vehicle test in the related technology, the test conditions set by the simulation software are quite different from the actual test, and further verification is required through a real vehicle test. However, the cost of the real vehicle test is high, and re-verification is required after the vehicle model parameters change, and it cannot be promoted and applied. The present application provides a method for predicting the simulation degree of a vehicle collision test. In this method, the entropy value of each subset after each test parameter in each training set is divided into the training set can be calculated, and the overall entropy value corresponding to the training set after the training set is divided into the training set by the test parameter is calculated according to the entropy value, and then the multiple division nodes in the training set are gradually determined according to the overall entropy value corresponding to the training set after each test parameter is divided into the training set, and a prediction model for the simulation degree of a vehicle collision test is generated according to the divided training set, so that the simulation degree of the vehicle collision test can be predicted in advance based on the prediction model. When the test parameters change, the collision test simulation degree can be predicted directly based on the prediction model without the need for real vehicle verification, which is more efficient, less costly, and easier to promote. In addition, in the process of constructing the prediction model, a variety of test parameters related to the vehicle collision simulation degree are comprehensively considered, and the prediction accuracy is higher. This solves the problems in related technologies such as high cost of vehicle collision test simulation prediction and difficulty in popularization and application.

[0031] Specifically, Figure 1 A schematic flow chart of a vehicle collision test simulation degree prediction method provided in an embodiment of the present application.

[0032] like Figure 1 As shown, the vehicle collision test simulation degree prediction method includes the following steps: In step S101, a training set of a vehicle collision test is obtained, wherein the training set includes test parameter samples and a simulation result of the collision test corresponding to each sample.

[0033] Among them, the collision test types may include frontal collision test, side collision test, rollover collision test, etc. The collision test parameters are test setting parameters and vehicle parameters related to the collision test process. The simulation results of the collision test are the simulation results corresponding to the collision test. For example, the simulation results of the frontal collision test may be front end deformation or no deformation, and the simulation results of the rollover test may be rollover or no rollover.

[0034] It is understandable that the embodiment of the present application can obtain a training set of vehicle collision tests in order to subsequently construct a prediction model for the simulation degree of the vehicle collision test.

[0035] The implementation scheme of the present application is described by taking the rollover test in the collision test as an example. For example, the test parameters of the rollover test include test setting parameters and vehicle parameters related to the vehicle rollover test. The test setting parameters include speed, temperature, humidity, ground friction coefficient, slope friction coefficient, slope angle, cut-in angle, single-sided bridge angle, curb height, etc.; vehicle parameters include center of mass height, front and rear axle load distribution, moment of inertia, steering angle, vehicle width, vehicle length, etc.; rollover test simulation results include types such as rollover and no rollover.

[0036] In addition, it should be noted that the sample data in the training set of the embodiment of the present application is collected by relying on fixed data acquisition equipment and test sites to ensure the consistency and repeatability of subsequent real vehicle collision tests.

[0037] In an embodiment of the present application, before obtaining a training set for a vehicle collision test, it also includes: obtaining multiple test parameters of historical vehicle collision tests and corresponding simulation results of the collision tests; based on the multiple test parameters of historical vehicle collision tests and corresponding simulation results of the collision tests, constructing a training set, a test set and a verification set for the vehicle collision test.

[0038] It can be understood that the embodiments of the present application can obtain multiple test parameters under historical vehicle collision tests and the corresponding simulation results of the collision tests, and construct a training set, a test set and a validation set of the vehicle collision test according to the collected historical data, wherein the training set is used to construct a prediction model, the test set is used to test the accuracy of the constructed prediction model, and the validation set is used to verify the generalization ability of the constructed prediction model.

[0039] In step S102, the entropy value of each subset after the training set is divided by each test parameter in the training set is calculated, and the overall entropy value corresponding to the training set divided by the test parameter is calculated based on the entropy value.

[0040] It can be understood that the embodiments of the present application can calculate the entropy value of each subset after the training set is divided by each test parameter in each training set, and calculate the overall entropy value corresponding to the training set after the training set is divided by the test parameter based on the entropy value, wherein the entropy value is used to evaluate the indicators belonging to the same category in the division results when the test parameter is used as the basis for dividing the simulation results of the collision test. The smaller the indicator value, the higher the purity of a category in the division result. The overall entropy value is the entropy value of the overall division strategy of the training set after the division by the test parameter.

[0041] For example, the test parameter in the rollover test is divided into vehicle width, and the categories of the rollover test simulation results include rollover and non-rollover. Then, the entropy value of each subset corresponding to the training set after the training set is divided by vehicle width is calculated, and then the overall entropy value corresponding to the training set after the training set is divided by the vehicle width is calculated.

[0042] In step S103, multiple division nodes in the training set are gradually determined according to the corresponding overall entropy value after the training set is divided according to each test parameter, a prediction model of the vehicle collision test simulation degree is generated according to the divided training set, and the prediction model is used to predict the collision test simulation degree of the target vehicle collision test.

[0043] It can be understood that the embodiment of the present application can gradually determine multiple division nodes in the training set according to the corresponding overall entropy value after the training set is divided by each test parameter, and generate a prediction model for the vehicle collision test simulation degree according to the divided training set, and use the prediction model to predict the collision test simulation degree of the target vehicle collision test, so that the vehicle collision test simulation degree can be predicted in advance based on the prediction model. When facing changes in test parameters, the collision test simulation degree can be predicted directly based on the prediction model without the need for actual vehicle verification, which is more efficient, less costly, and easier to promote. In addition, in the process of constructing the prediction model, a variety of test parameters related to the vehicle collision simulation degree are comprehensively considered, and the prediction accuracy is higher.

[0044] In addition, it should be noted that the prediction model in the embodiment of the present application is similar to being constructed layer by layer through decision numbers, and can also be understood as a classification or judgment model.

[0045] In an embodiment of the present application, before gradually determining multiple division nodes in the training set according to the overall entropy value corresponding to each test parameter after the training set is divided, it also includes: identifying the data type of the test parameter sample; if the data type of the test parameter sample is a discrete type, then determining the division node according to the sample proportion and weight corresponding to the test parameter under different values ​​and the overall entropy value corresponding to the training set after the test parameter is divided; if the data type of the test parameter sample is a continuous type, then discretizing the value of the test parameter sample to obtain multiple discrete values, and determining the division node according to the sample proportion and weight corresponding to each discretized value and the overall entropy value corresponding to the test parameter after the training set is divided.

[0046] Among them, data types include continuous type and discrete type. Continuous type data means that the possible values ​​are infinite, such as vehicle speed and temperature, which are continuous type data and have infinite possible values; discrete type data means that the possible values ​​are limited, such as driving mode, which is discrete type data and has limited values, which can be front-wheel drive, rear-wheel drive and four-wheel drive.

[0047] It can be understood that the embodiment of the present application can determine the specific overall entropy value calculation method according to the data type of the test parameter sample, and then determine the division node based on the overall entropy value, and use different methods to determine the division node for test parameters of different data types, thereby improving the accuracy and pertinence of determining the division node. Discrete type parameters are determined according to the value, and continuous type parameters are first discretized and then determined, so that the construction of the prediction model can better adapt to data with different characteristics, solve the problem of how to effectively measure the division effect of continuous parameters in model construction, optimize the model construction process, and help improve the performance of the final prediction model, specifically: If the data type of the test parameter is discrete, the partitioning node is determined according to the sample proportion and weight corresponding to different values ​​of the test parameter and the overall entropy value corresponding to the training set after the test parameter is partitioned; If the data type of the test parameter sample is a continuous type, the value of the test parameter sample is discretized to obtain multiple discretized values, and the division node is determined according to the sample proportion and weight corresponding to each discretized value and the overall entropy value corresponding to the test parameter after the training set is divided.

[0048] In the embodiment of the present application, the calculation formula of the overall entropy value after the discrete type test parameters are divided into training sets is: ; in, is a discrete type test parameter, D is a training set, For test parameters Divide the training set The overall entropy value of the post-training set, is the entropy value of the training set D, The test parameters The number of values ​​of For the training set Medium test parameters The value of A subset of all test parameter samples of is the total number of samples in the training set D, For subset The proportion of the number of samples in the total number of test parameter samples, For subset The entropy value of .

[0049] In the embodiment of the present application, the calculation formula of the overall entropy value after the continuous type test parameter is divided into the training set is:

[0050] in, is the test parameter of the continuous type, is the training set, For test parameters Divide the training set The overall entropy value of the post-training set, All possible partition values ​​for the test parameter b A collection of To use the test parameters Divide the value Divide the training set The overall entropy value of the post-training set, For the training set The entropy value of To use the test parameters Divide the value Divide the training set satisfy A subset of conditions, For subset The number of samples in is the total number of samples in the training set D, For subset The entropy value of .

[0051] In the above calculation formula of overall entropy value ,in, represents the total number of test parameter set samples in the training set, Indicates k The proportion of class samples (i.e. the simulation result category of the rollover test corresponding to the sample) in the collection samples, The above formula is also used for calculation. For example, if it is divided by parameter A, then The total number of samples corresponding to different values ​​of parameter A, and the other parameters are modified similarly.

[0052] In an embodiment of the present application, multiple division nodes in the training set are gradually determined according to the corresponding overall entropy value after the training set is divided by each test parameter, including: in the current division stage of the training set, calculating the current overall entropy value of the training set; determining the division node of the training set in the current division stage based on the current overall entropy value; using the division node to divide the training set into multiple sub-training sets, and entering the next division stage until the division stop condition is met, wherein, when calculating the overall entropy value of the training set in each division stage, the division node of the previous division stage is ignored; and determining multiple division nodes in the training set according to multiple division stages.

[0053] The division node may be the test parameter itself or the value of the test parameter.

[0054] It can be understood that, in the embodiment of the present application, at the current division stage of the training set, the current overall entropy value of the training set is calculated, and the division node of the training set at the current division stage is determined based on the current overall entropy value, and the division node is used to divide the training set into multiple sub-training sets, and enter the next division stage until the division stop condition is met. In which, when calculating the overall entropy value of the training set at each division stage, the division node of the previous division stage is ignored, and multiple division nodes in the training set are determined according to multiple division stages. By calculating the overall entropy value at each division stage and determining the division node according to the overall entropy value, the model structure can be gradually optimized, the blindness of the division can be reduced, and finally a prediction model that is more in line with the data characteristics can be constructed, thereby improving the prediction accuracy of the prediction model.

[0055] The following is a specific example to describe the selection of discrete parameters and continuous parameters for dividing nodes in the embodiment of the present application, taking the vehicle rollover test as an example, wherein the discrete parameters and the corresponding rollover simulation results are shown in Table 1.

[0056] Table 1 Discrete parameter A Discrete parameter B Rollover test simulation results a1 b1 roll a2 b2 No rolling a3 b2 roll a2 b2 No rolling a3 b1 roll a3 b3 No rolling a1 b3 No rolling First, calculate the entropy value of the entire training set (that is, the entire sample): ; When the discrete parameter A is used to divide the entire sample, the values ​​of A are a1, a2, and a3, and the entropy values ​​of the three values ​​are calculated respectively.

[0057] ; ; .

[0058] Then calculate the impact of the entropy value after partitioning using the parameter A: .

[0059] Similarly, calculate the impact of other parameters on the entropy value after division, and select the parameter with the largest entropy value as the division parameter.

[0060] The tumbling results corresponding to the continuous parameters and discrete parameters are shown in Table 2.

[0061] Table 2 Continuous parameter A Discrete parameter B Rollover test simulation results 1.01 b1 roll 1.02 b2 No rolling 1.03 b2 roll 1.04 b2 No rolling 1.05 b1 roll 1.06 b3 No rolling 1.07 b3 No rolling The internal values ​​of continuous parameters are continuous. Unlike discrete parameters, the entropy value corresponding to a certain value cannot be obtained through statistical calculation, because it is impossible to count the corresponding proportion of the results of the rollover test simulation. Therefore, it is necessary to discretize the continuous values ​​and divide the continuous parameters into two or more intervals. Each interval is equivalent to a discrete value, that is, the division node is determined by setting the internal separation point.

[0062] For example, as shown in Table 2, if it is known through calculation that when 1.05 is used as the separation point, those less than or equal to 1.05 are grouped as one group, and those greater than 1.05 are grouped as one group, then the interior of the continuous parameter is divided into two groups. In this way, the entropy values ​​corresponding to the two groups can be calculated separately according to the entropy value calculation method of the discrete parameter, and the interval with the larger entropy value can be selected as the partition parameter. If the group is less than or equal to 1.05 and has a large entropy value, then the calculation and selection of the partition point can still be continued within it. For example, if the entropy value of a group less than or equal to 1.03 is large, then 1.03 can be used as the partition point, and the partition point with the maximum entropy value can be continued within it.

[0063] In an embodiment of the present application, after generating a prediction model for the simulation degree of a vehicle collision test based on the divided training set, it also includes: using each division node in the training set to divide the test parameter samples in the verification set; calculating a first proportion of the test parameter samples before the verification set is divided in the simulation degree results of the collision test; calculating a second proportion of the test parameter samples after the verification set is divided in the simulation degree results of the collision test; calculating the difference between the first proportion and the second proportion, and adjusting the category of the simulation degree results of the collision test of the corresponding test parameter samples in the division node according to the difference.

[0064] Since in the process of dividing the sample data in the training set, that is, in the process of building the prediction model, the division nodes are determined simply by calculating the parameter entropy value and the overall entropy value, but the generalization ability of the prediction model is not improved after some division nodes are divided, that is, there are inefficient or invalid divisions for the prediction accuracy of the entire model using some division nodes. Therefore, the embodiment of the present application can use the sample data in the validation set to verify whether the selection of the division nodes is reasonable, and calculate the improvement of the generalization ability of each node of the prediction model built according to the training set before and after the division. If there is no improvement, the division node should be directly converted into the category with the largest proportion of categories in its sample number. By calculating the difference in the proportion of the test parameter samples in the simulation results of the collision test before and after the division, the division nodes are optimized, and the model can be fine-tuned according to the actual data distribution, so that the division nodes are more in line with the actual characteristics of the data, and the accuracy and adaptability of the prediction model are further improved. Specifically: Each division node in the training set is used to divide the test parameter samples in the validation set, and a first proportion of the test parameter samples in the validation set before division in the simulation result of the collision test and a second proportion of the test parameter samples after division in the simulation result of the collision test are calculated. According to the difference between the first proportion and the second proportion, the category of the simulation result of the collision test of the corresponding test parameter sample in the division node is adjusted. For example, if the difference is greater than a set value, the category of the simulation result of the collision test of the corresponding test parameter sample in the division node is adjusted, otherwise it is not adjusted.

[0065] In an embodiment of the present application, after generating a prediction model for the vehicle collision test simulation degree according to the divided training set, it also includes: using a test set to test the prediction accuracy of the prediction model; if the prediction accuracy is lower than a preset threshold, optimizing the division nodes of the prediction model.

[0066] Among them, the preset threshold can be set according to the specific situation and is not specifically limited.

[0067] It can be understood that the embodiments of the present application can utilize the prediction accuracy of the prediction model in the test set; if the prediction accuracy is lower than the preset threshold, the partitioning nodes of the prediction model are optimized, and the model is improved by optimizing the partitioning nodes to ensure that the prediction model has a higher prediction accuracy and improve the practicality of the model.

[0068] The following describes the vehicle collision test simulation result prediction method of the present application through a specific embodiment, taking the vehicle rollover test simulation result prediction as an example. Figure 2 As shown, including: S110, tumbling data collection and division.

[0069] Step 1: Determine the test setup parameters and vehicle parameters related to various rollover tests. Test setup parameters include: speed, temperature, humidity, ground friction factor, slope friction factor, slope angle, cut-in angle, single-side bridge angle, curb height, etc. All test parameters related to the test setup. Vehicle parameters include: center of mass height, front and rear axle load distribution, moment of inertia, steering angle, vehicle width, vehicle length, etc. All vehicle parameters related to the vehicle.

[0070] Step 2: The collection of test parameters (including test setting parameters and vehicle parameters, hereinafter referred to as parameters) in step 1 should be related to the type of rollover test determined, and parameters not involved may not be collected. For example, in a slope rollover, the single-side bridge angle in the rollover test is not collected.

[0071] Step 3: All parameters should be collected to ensure that there are no missing values. Each set of test parameter collection has a corresponding tumbling simulation result.

[0072] Step 4: Preliminary classification of all parameters, according to the characteristics of the parameter values ​​into two categories: discrete and continuous. Discrete data means that the possible values ​​are limited, and continuous data means that the possible values ​​are unlimited.

[0073] Step 5: List all parameters and their values ​​in a statistical table to facilitate the calculation of subsequent statistical steps. As shown in Table 3, Table 3 is a table of test parameters and corresponding rollover test simulation results. Discrete and continuous parameters can be coded in sequence, regardless of order.

[0074] Table 3

[0075] Step 6: Group the sample collection into training set, test set and validation set.

[0076] S120: Define optimization strategy.

[0077] Step 1: Determine the entropy value of a parameter. The entropy value is used to evaluate the indicators belonging to the same category in the division results when the parameter is used as the basis for division of the tumbling results. The smaller the indicator value, the higher the purity of a category in the division results.

[0078] ; in, Indicates the number of samples in the tumbling sample set (training set), Indicates k The proportion of class samples in the collection samples, Indicates the total number of types corresponding to the simulation results of the rollover test.

[0079] Step 2: Determine the entropy value of the overall partition. In addition to determining the entropy value of a certain parameter partition in step 1, it is also necessary to determine the entropy value of the overall partition strategy after using the parameter partition. The overall partition strategy is accumulated based on various parameter partitions, and the result of the previous parameter partition will have a direct impact on the remaining sample size and category ratio. Therefore, each selection of a parameter partition will have a direct and dynamic impact on the overall entropy value.

[0080] ; Assuming that parameter a has V possible values ​​(discrete parameters are finite; continuous parameters are infinite), if a is used to divide the sample set, V branch points will be generated, where the vth branch point contains all the values ​​of parameter a in the sample set D. The sample is denoted as , calculate After the parameter entropy value is obtained, the weight is assigned according to the number of samples in the sample set. , indicating that the more samples there are, the greater the impact of the parameter on the tumbling result. The overall entropy value of the sample set after the parameter is divided is calculated.

[0081] S130: Selection of partitioning parameters.

[0082] Step 1: According to the difference between finite and infinite values ​​of parameters, they can be divided into discrete parameters and continuous parameters. The value of discrete parameters and its sample ratio Can be calculated directly.

[0083] Step 2: Continuous parameters need to be discretized. That is, the continuous parameters are divided into two or more intervals, each interval is equivalent to a discretized value. The division strategy is based on the calculation of the overall entropy value.

[0084]

[0085] in, t It is expressed as the partition value of the continuous parameter. According to the formula, the partition value that can maximize the overall entropy value is selected as the internal partition value of the continuous parameter.

[0086] Step 3: One or more division points of the continuous parameter may be selected. The number of division points actually selected needs to consider the overall entropy value and the entropy value after the model is established in S140.

[0087] S140, gradually calculating and establishing a determination model (ie, the prediction model of the present application).

[0088] Step 1: In the sample collection, calculate the overall entropy value of each discrete parameter and continuous parameter in turn, and select the largest parameter as the partition parameter.

[0089] Step 2: In the sample set of the partition parameter selected in step 1, calculate the overall entropy value of the remaining discrete parameters and continuous parameters, and select the largest parameter as the next partition parameter.

[0090] Step 3: Calculate the overall partition model in sequence. Note: The continuous parameters are based on the partition points in step 2 of S130. t After the division, the obtained parameters are still continuous, so the new division values ​​still need to be calculated again inside it. t , select the largest overall entropy value as the next partition node.

[0091] Step 4: After all the training set data has been divided, the test set is introduced to test the judgment model. If the test accuracy reaches the ideal value, the judgment model is successfully constructed.

[0092] S150, validation set to test generalization ability.

[0093] Step 1: Establish verification accuracy index. In the process of dividing the training set sample data according to S120 and S130, the selection of the partitioning parameters is determined simply by calculating the parameter entropy value and the overall entropy value, because the generalization ability of the model cannot be improved before and after the selection of certain parameters, that is, there are some parameter partitions that are inefficient or invalid partition parameters for the prediction accuracy of the entire model. Therefore, the validation set data should be used to calculate the improvement of the generalization ability of each node of the classification model established according to the training set before and after the partition. If there is no improvement, the partition point should be directly converted to the category with the largest proportion of its sample number.

[0094] Step 2: The generalization ability can be calculated by comparing the proportion of a certain category in the number of samples before and after the division.

[0095] In summary, this embodiment is based on the actual vehicle rollover test statistical data accumulated in the past, and uses the probability statistics principle to analyze and process the test data. In the process of building the vehicle rollover test simulation prediction model, the overall entropy value corresponding to the training set divided by different test parameters is calculated (for example, for discrete parameter a, using the formula ; For the continuous parameter b, use the formula ), is essentially a measure of the impact of different test parameters on the uncertainty of the sample set. This process is closely related to calculating the probability values ​​of the categories corresponding to different test parameters. For example, and The calculation of isentropic values ​​involves the proportion of each type of sample in the sample set, which reflects the probability of samples belonging to different categories. By comparing the classification accuracy under different paths, the best test simulation prediction method is established. The probability values ​​of the corresponding categories of different test parameters are calculated, the classification accuracy under different paths is compared, and the best test simulation prediction method is established.

[0096] The vehicle collision test simulation degree prediction method of the embodiment of the present application is based on the probability statistics of the test data and is not affected by the size and unit of the data value. Therefore, the preliminary data processing is simple and easy, and there is no need for complex unitization and normalization processing; based on the inherent probabilistic characteristics of the test statistical data, it has the characteristics of strong interpretability, rapid application, and convenient prediction; it can make mixed predictions of discrete and continuous data variables, and is suitable for collision result judgment under complex scene conditions; the data based on it comes from the actual vehicle collision test, and can more accurately predict the test simulation degree results compared with simulation technology, making up for the limitations of judgment based on experience; the data used relies on fixed data acquisition equipment and test sites, which can ensure the consistency and repeatability of subsequent actual vehicle collision tests, and has strong robustness and generalization ability; the process of establishing the prediction model introduces verification accuracy indicators, reduces inefficient or invalid division nodes, can reduce the risk of overfitting, and reduces training and prediction time.

[0097] According to the vehicle collision test simulation degree prediction method proposed in the embodiment of the present application, the entropy value of each subset after each test parameter in each training set is divided into the training set can be calculated, and the overall entropy value corresponding to the training set after the training set is divided into the training set by the test parameter is calculated based on the entropy value, and then the multiple division nodes in the training set are gradually determined according to the overall entropy value corresponding to each test parameter after the training set is divided into the training set, and a prediction model for the vehicle collision test simulation degree is generated according to the divided training set, so that the vehicle collision test simulation degree can be predicted in advance based on the prediction model. When facing changes in test parameters, the collision test simulation degree can be predicted directly based on the prediction model without the need for actual vehicle verification, which is more efficient, less costly, and easier to promote. In addition, in the process of constructing the prediction model, a variety of test parameters related to the vehicle collision simulation degree results are comprehensively considered, and the prediction accuracy is higher.

[0098] Next, a vehicle collision test simulation degree prediction device proposed according to an embodiment of the present application is described with reference to the accompanying drawings.

[0099] Figure 3 It is a block diagram of a vehicle collision test simulation degree prediction device according to an embodiment of the present application.

[0100] like Figure 3 As shown, the vehicle collision test simulation degree prediction device 10 includes: an acquisition module 100 , a calculation module 200 and a generation module 300 .

[0101] Among them, the acquisition module 100 is used to obtain a training set of a vehicle collision test, wherein the training set includes test parameter samples and the simulation results of the collision test corresponding to each sample; the calculation module 200 is used to calculate the entropy value of each subset after the training set is divided by each test parameter in the training set, and calculate the corresponding overall entropy value after the training set is divided by the test parameter based on the entropy value; the generation module 300 is used to gradually determine multiple division nodes in the training set according to the overall entropy value corresponding to the training set after each test parameter is divided, generate a prediction model for the simulation degree of the vehicle collision test according to the divided training set, and use the prediction model to predict the simulation degree of the collision test for the target vehicle collision test.

[0102] In the embodiment of the present application, the vehicle collision test simulation degree prediction device 10 of the embodiment of the present application further includes: an identification module.

[0103] Among them, the identification module is used to identify the data type of the test parameter sample before gradually determining multiple division nodes in the training set according to the overall entropy value corresponding to each test parameter after the training set is divided; if the data type of the test parameter sample is a discrete type, the division node is determined according to the sample proportion and weight corresponding to the test parameter under different values ​​and the overall entropy value corresponding to the training set after the test parameter is divided; if the data type of the test parameter sample is a continuous type, the value of the test parameter sample is discretized to obtain multiple discrete values, and the division node is determined according to the sample proportion and weight corresponding to each discretized value and the overall entropy value corresponding to the test parameter after the training set is divided.

[0104] In the embodiment of the present application, the calculation formula of the overall entropy value after the discrete type test parameters are divided into training sets is: ; in, is a discrete type test parameter, D is a training set, For test parameters Divide the training set The overall entropy value of the post-training set, is the entropy value of the training set D, The test parameters The number of values ​​of For the training set Medium test parameters The value of A subset of all test parameter samples of is the total number of samples in the training set D, For subset The proportion of the number of samples in the total number of test parameter samples, For subset The entropy value of .

[0105] In the embodiment of the present application, the calculation formula of the overall entropy value after the continuous type test parameter is divided into the training set is:

[0106] in, is the test parameter of the continuous type, is the training set, For test parameters Divide the training set The overall entropy value of the post-training set, All possible partition values ​​for the test parameter b A collection of To use the test parameters Divide the value Divide the training set The overall entropy value of the post-training set, For the training set The entropy value of To use the test parameters Divide the value Divide the training set satisfy A subset of conditions, For subset The number of samples in is the total number of samples in the training set D, For subset The entropy value of .

[0107] In an embodiment of the present application, the generation module 300 is further used to: calculate the current overall entropy value of the training set in the current division stage of the training set; determine the division node of the training set in the current division stage based on the current overall entropy value; use the division node to divide the training set into multiple sub-training sets, and enter the next division stage until the division stop condition is met, wherein, when calculating the overall entropy value of the training set in each division stage, the division node of the previous division stage is ignored; and determine multiple division nodes in the training set according to multiple division stages.

[0108] In the embodiment of the present application, the vehicle collision test simulation degree prediction device 10 of the embodiment of the present application further includes: a construction module.

[0109] Among them, the construction module is used to obtain multiple test parameters of historical vehicle collision tests and corresponding simulation results of collision tests before obtaining the training set of vehicle collision tests; based on the multiple test parameters of historical vehicle collision tests and corresponding simulation results of collision tests, the training set, test set and verification set of the vehicle collision test are constructed.

[0110] In the embodiment of the present application, the vehicle collision test simulation degree prediction device 10 of the embodiment of the present application further includes: a verification module.

[0111] Among them, the verification module is used to divide the test parameter samples in the verification set by using each division node in the training set after generating a prediction model of the vehicle collision test simulation degree according to the divided training set; calculate the first proportion of the test parameter samples before the verification set is divided in the simulation degree results of the collision test; calculate the second proportion of the test parameter samples after the verification set is divided in the simulation degree results of the collision test; calculate the difference between the first proportion and the second proportion, and adjust the category of the simulation degree results of the collision test of the corresponding test parameter samples in the division node according to the difference.

[0112] In the embodiment of the present application, the vehicle collision test simulation degree prediction device 10 of the embodiment of the present application further includes: a testing module.

[0113] Among them, the testing module is used to generate a prediction model of the vehicle collision test simulation degree according to the divided training set, and then use the test set to test the prediction accuracy of the prediction model; if the prediction accuracy is lower than the preset threshold, the division nodes of the prediction model are optimized.

[0114] It should be noted that the above explanations of the embodiment of the vehicle collision test simulation degree prediction method are also applicable to the vehicle collision test simulation degree prediction device of this embodiment, and will not be repeated here.

[0115] According to the vehicle collision test simulation degree prediction device proposed in the embodiment of the present application, the entropy value of each subset after the training set is divided by each test parameter in each training set can be calculated, and the overall entropy value corresponding to the training set after the training set is divided by the test parameter is calculated based on the entropy value, and then the multiple division nodes in the training set are gradually determined according to the overall entropy value corresponding to the training set after each test parameter is divided. A prediction model of the vehicle collision test simulation degree is generated according to the divided training set, so that the vehicle collision test simulation degree can be predicted in advance based on the prediction model. When facing changes in test parameters, the collision test simulation degree can be predicted directly based on the prediction model without the need for actual vehicle verification, which is more efficient, less costly, and easier to promote. In addition, in the process of constructing the prediction model, a variety of test parameters related to the vehicle collision simulation degree are comprehensively considered, and the prediction accuracy is higher.

[0116] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device may include: Memory 401 , processor 402 , and a computer program stored in the memory 401 and executable on the processor 402 .

[0117] When the processor 402 executes the program, the vehicle collision test simulation degree prediction method provided in the above embodiment is implemented.

[0118] Furthermore, the electronic device further comprises: The communication interface 403 is used for communication between the memory 401 and the processor 402 .

[0119] The memory 401 is used to store computer programs that can be executed on the processor 402 .

[0120] The memory 401 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0121] If the memory 401, the processor 402 and the communication interface 403 are implemented independently, the communication interface 403, the memory 401 and the processor 402 can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 4 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0122] Optionally, in a specific implementation, if the memory 401, the processor 402 and the communication interface 403 are integrated on a chip, the memory 401, the processor 402 and the communication interface 403 can communicate with each other through an internal interface.

[0123] The processor 402 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0124] An embodiment of the present application also provides a computer-readable storage medium having a computer program or instruction stored thereon. When the computer program or instruction is executed by a processor, the above-mentioned vehicle collision test simulation degree prediction method is implemented.

[0125] The embodiment of the present application also provides a computer program product, including a computer program or instructions, which, when executed, implements the above vehicle collision test simulation degree prediction method.

[0126] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0127] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0128] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.

[0129] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above embodiment, N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one or a combination of multiple of the following technologies known in the art: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0130] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

Claims

1. A method for predicting the degree of simulation of a vehicle collision test, characterized in that: The following steps are involved: Acquire a training set of vehicle collision tests, wherein the training set includes test parameter samples and a simulation result of the collision test corresponding to each sample; Calculate the entropy value of each subset after the training set is divided by each test parameter in the training set, and calculate the corresponding overall entropy value after the training set is divided by the test parameter according to the entropy value; A plurality of division nodes in the training set are gradually determined according to the overall entropy value corresponding to each test parameter after the training set is divided, a prediction model of the vehicle collision test simulation degree is generated according to the divided training set, and the prediction model is used to predict the collision test simulation degree of the target vehicle collision test.

2. The vehicle collision test simulation degree prediction method according to claim 1, characterized in that: Before gradually determining a plurality of division nodes in the training set according to the overall entropy value corresponding to the division of the training set by each test parameter, the method further includes: Identifying a data type of the test parameter sample; If the data type of the test parameter sample is a discrete type, the division node is determined according to the sample proportion and weight corresponding to different values ​​of the test parameter and the overall entropy value corresponding to the training set after the test parameter is divided; If the data type of the test parameter sample is a continuous type, the value of the test parameter sample is discretized to obtain multiple discrete values, and the division node is determined according to the sample proportion and weight corresponding to each discretized value and the overall entropy value corresponding to the test parameter after the training set is divided.

3. The vehicle collision test simulation degree prediction method according to claim 2, characterized in that: The calculation formula of the overall entropy value after the discrete type of test parameters are divided into training sets is: ； in, is a discrete type test parameter, D is a training set, For test parameters Divide the training set The overall entropy value of the post-training set, is the entropy value of the training set D, The test parameters The number of values ​​of For the training set Medium test parameters The value of A subset of all test parameter samples of is the total number of samples in the training set D, For subset The proportion of the number of samples in the total number of test parameter samples, For subset The entropy value of .

4. The vehicle collision test simulation degree prediction method according to claim 2, characterized in that: The calculation formula of the overall entropy value after the continuous type of test parameters are divided into training sets is: in, is the test parameter of the continuous type, is the training set, For test parameters Divide the training set The overall entropy value of the post-training set, All possible partition values ​​for the test parameter b A collection of To use the test parameters Divide the value Divide the training set The overall entropy value of the post-training set, For the training set The entropy value of To use the test parameters Divide the value Divide the training set satisfy A subset of conditions, For subset The number of samples in is the total number of samples in the training set D, For subset The entropy value of .

5. The vehicle collision test simulation degree prediction method according to claim 2, characterized in that: The step of gradually determining a plurality of division nodes in the training set according to the overall entropy value corresponding to the division of the training set according to each test parameter comprises: At the current partitioning stage of the training set, calculating the current overall entropy value of the training set; Determine the partition node of the training set at the current partition stage based on the current overall entropy value; The training set is divided into a plurality of sub-training sets by using the division node, and the next division stage is entered until the division stop condition is met, wherein when calculating the overall entropy value of the training set in each division stage, the division node of the previous division stage is ignored; A plurality of partition nodes in the training set are determined according to a plurality of partition stages.

6. The vehicle collision test simulation degree prediction method according to claim 1, characterized in that: Before obtaining the training set of vehicle collision tests, it also includes: Obtain multiple test parameters of historical vehicle collision tests and corresponding collision test simulation results; Based on multiple test parameters of the historical vehicle collision test and corresponding collision test simulation results, a training set, a test set and a validation set of the vehicle collision test are constructed.

7. The vehicle collision test simulation degree prediction method according to claim 6, characterized in that: After generating the prediction model of the vehicle collision test simulation degree according to the divided training set, the method further includes: Using each partition node in the training set to partition the test parameter samples in the validation set; Calculating a first proportion of the test parameter samples before the verification set is divided in the simulation result of the collision test; Calculating a second proportion of the test parameter samples after the verification set is divided in the simulation result of the collision test; The difference between the first proportion and the second proportion is calculated, and the category of the simulation result of the collision test corresponding to the test parameter sample in the division node is adjusted according to the difference.

8. The vehicle collision test simulation degree prediction method according to claim 7, characterized in that: After generating the prediction model of the vehicle collision test simulation degree according to the divided training set, the method further includes: Using the test set to test the prediction accuracy of the prediction model; If the prediction accuracy is lower than a preset threshold, the partitioning nodes of the prediction model are optimized.

9. A vehicle collision test simulation prediction device, characterized in that: include: An acquisition module, used to acquire a training set of a vehicle collision test, wherein the training set includes test parameter samples and a simulation result of a collision test corresponding to each sample; A calculation module, used to calculate the entropy value of each subset after the training set is divided by each test parameter in the training set, and calculate the corresponding overall entropy value after the training set is divided by the test parameter according to the entropy value; A generation module is used to gradually determine multiple division nodes in the training set according to the overall entropy value corresponding to each test parameter after the training set is divided, generate a prediction model of the vehicle collision test simulation degree according to the divided training set, and use the prediction model to predict the collision test simulation degree of the target vehicle collision test.

10. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the vehicle collision test simulation degree prediction method according to any one of claims 1 to 8.

Images