A method and equipment for determining critical parameters in vehicle slope rollover tests

Abstract

This application relates to the field of vehicle testing technology, specifically a method and device for determining critical parameters in vehicle slope rollover tests. The method includes: collecting a set of test data from actual vehicle rollovers and dividing it into subsets according to vehicle type; inputting vehicle parameters and test parameters into a forward discriminant model to train the model; calculating a basic inverse model from the forward discriminant model, adding constraints to the output of the basic inverse model, and correcting the basic inverse model using the vehicle parameters and test parameters of the vehicle to be solved; inputting the rollover result into the basic inverse model to obtain an angle range, and filtering test data within the angle range; returning to the operation of training the forward discriminant model until a cutoff condition is met to obtain the target angle range; and finally verifying the rollover critical angle value from the target angle range through actual vehicle angle gradient tests. This application achieves accurate inverse solving from rollover results to test parameters.

Description

Technical Field

[0001] This application relates to the field of vehicle testing technology, and more specifically, to a method and equipment for determining critical parameters in a vehicle slope rollover test. Background Technology

[0002] Due to the high injury and fatality rates associated with rollover accidents on slopes, automakers aim to accurately determine the boundary values ​​of vehicle parameters (speed, center of gravity height, mass distribution ratio, length, width, height, etc.) and environmental parameters (such as slope angle) when a vehicle rolls over on a slope. This is to improve the vehicle's ability to identify and respond to dangerous rollover scenarios and provide timely protection for occupants. However, analysis reveals a complex mapping relationship between vehicle parameters, environmental parameters, and rollover outcomes. Furthermore, the critical value of a specific parameter cannot be directly derived from the rollover result, and currently, there is a lack of effective methods in the industry to solve these problems.

[0003] Currently, some automakers use a trial-and-error approach to solve for the critical parameters of slope rollover. By setting parameter gradient values ​​and combining them into experimental schemes, they conduct extensive real-vehicle tests to verify the optimal critical conditions. However, this method requires large-scale verification tests, leading to a sharp increase in costs. Furthermore, if simulation methods are used for verification, the obtained critical conditions often deviate significantly from those obtained in real-world driving.

[0004] In summary, there is an urgent need for an efficient, accurate, and low-cost method to solve for the critical value of vehicle rollover on slopes. Summary of the Invention

[0005] The purpose of this application is to provide a method and equipment for determining the critical parameters of a vehicle slope rollover test. The method involves establishing a forward discrimination model and calculating the inverse model to obtain a basic inverse model with additional constraints. Then, by using a convergent sample set method, the narrowed target angle range is obtained. Finally, the final rollover critical angle value is obtained by using a real vehicle verification method.

[0006] To achieve the above objectives, this application adopts the following technical solution: Firstly, this application provides a method for determining critical parameters in a vehicle slope rollover test, including: Collect a set of test data on actual vehicle rollovers, and divide the test data set into test data subsets according to vehicle type. The test data subsets include: vehicle parameters, test parameters, and rollover results; the test parameters include the slope angle. The vehicle parameters and test parameters in each subset of test data are input into a forward discrimination model corresponding to the vehicle type, and the rollover result is used as a label to train the forward discrimination model. The basic inverse model is obtained by inverse calculation from the forward discriminant model, and constraints are added to the output of the basic inverse model. Under the constraints described above, the basic inverse model is modified using the vehicle parameters and test parameters of the vehicle to be solved. The results of whether or not the rollover occurs are input into the basic inverse model to obtain the first slope angle and the second slope angle output by the basic inverse model. An angle range is constructed based on the first slope angle and the second slope angle, and test data within the angle range are selected. Return to the operation of training the positive discrimination model using the selected test data until the cutoff condition is met, and obtain the target angle range including the rollover critical angle of the vehicle to be solved; By conducting real-vehicle angle gradient tests, the rollover critical angle value was finally verified from the target angle range.

[0007] Secondly, this application provides an electronic device, comprising: At least one processor, and a memory communicatively connected to at least one of the processors; The memory stores instructions that can be executed by at least one of the processors, which enable the at least one processor to perform the above-described method for determining critical parameters of a vehicle slope rollover test.

[0008] Compared with the prior art, the beneficial effects of this application are as follows: This application establishes forward discrimination models for each vehicle segment and calculates a basic inverse model with additional output constraints. This basic inverse model enables the reverse solution from rollover results to test parameters. Then, by converging sample sets, a narrowed target angle range is obtained. Finally, a real-vehicle verification method is used to obtain the final rollover critical angle value. This method effectively solves the boundary condition problem in complex scenarios of vehicle slope rollover testing, providing a feasible test scheme for automakers. It significantly reduces testing costs for automakers while more accurately obtaining the rollover critical angle value under real-vehicle conditions, improving the vehicle's recognition and response to slope rollover boundaries, and providing more efficient protection for occupants. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the specific embodiments of this application or the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0010] Figure 1This is a flowchart illustrating a method for determining critical parameters in a vehicle slope rollover test, as provided in an embodiment of this application. Figure 2 This is a schematic diagram of the structure of the forward discrimination model provided in the embodiments of this application; Figure 3 This is a schematic diagram of the basic inverse model provided in the embodiments of this application; Figure 4 This is a schematic diagram of the electronic device provided in the embodiments of this application. Detailed Implementation

[0011] The following description, in conjunction with the accompanying drawings, illustrates exemplary embodiments of this application, including various details to aid understanding. These should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this application. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0012] Example 1 The method provided in this embodiment is applicable to solving the critical rollover angle of a vehicle on a slope. Figure 1 This is a flowchart of the method for determining critical parameters in a vehicle slope rollover test provided in this application embodiment. See also... Figure 1 This application provides a method for determining critical parameters in a vehicle slope rollover test, comprising: S110. Collect a set of test data of actual vehicle rollover, and divide the test data set into test data subsets according to vehicle type. The test data subsets include: vehicle parameters, test parameters and rollover results; the test parameters include the slope angle.

[0013] Vehicle types include: sedans, SUVs, MPVs, etc.

[0014] Vehicle parameters include: left front wheel mass, right front wheel mass, left rear wheel mass, right rear wheel mass, front axle mass, rear axle mass, vehicle length, vehicle width, and vehicle height.

[0015] The test parameters include: slope angle, test speed (i.e., vehicle speed), and vehicle entry angle. Among them, the slope angle includes the angle of the actual slope and the angle of the rollover test platform.

[0016] The rollover result includes rolling over and not rolling over, which is determined based on the actual rollover state of the vehicle.

[0017] This step divides the test dataset according to vehicle type, obtaining a subset of test data belonging to each vehicle type. This allows for the training of a separate positive discrimination model for each vehicle type, improving adaptability to different vehicle types and thus increasing the accuracy of determining the rollover critical angle value.

[0018] S120. Input the vehicle parameters and test parameters in each subset of test data into the positive discrimination model corresponding to the vehicle type, and train the positive discrimination model using the rollover result as the label.

[0019] Before inputting vehicle parameters and test parameters into the forward discrimination model, the vehicle parameters and test parameters need to be preprocessed to meet the input requirements of the forward discrimination model.

[0020] Optionally, assign a coding value to the text-type data in each subset of experimental data; then de-normalize and normalize the coding value.

[0021] For example, string variables representing vehicle types such as sedans, MPVs, SUVs, and HEVs cannot be directly used for training and prediction in explicit and implicit neural networks. They must be converted to integer values: 1, 2, 3, ... Because the units of experimental parameters are inconsistent and the values ​​vary significantly, methods such as de-normalization and normalization can be used to unify the units and make the values ​​more similar, which is beneficial for accurate network prediction.

[0022] To ensure accurate prediction and simple implementation of the forward discriminant model, the rolling results—rolling and not rolling—are converted to values ​​of 1 and 0, respectively, to facilitate the use of existing nonlinear activation functions.

[0023] Optional, forward discriminant models include: an input layer, one or more intermediate layers, and an output layer. Figure 2 This is a schematic diagram of the forward discrimination model provided in an embodiment of this application. See also... Figure 2 For the first implicit neural network The first layer i For each neuron, we have: ;Formula (1) or, ;Formula (2) in, It is the first Layer i The value of each neuron, express The number of neurons in a layer. Indicates the first Layer i The first neuron and the second The weights of the connections between the j-th neuron in the layer. It indicates the first Layer i The bias values ​​of each neuron, the function Indicates the first Layer iActivation function of a neuron n This represents the total number of floors. for The result of the calculation.

[0024] The above equation can be written in matrix form as follows: ;Formula (3) in, Representing the A matrix of values ​​for layer neurons. Indicates the first The matrix of weights of layer neurons, Indicates the first The matrix of biases of layer neurons. for The calculation results. To solve the binary classification problem of tumbling results, the first... The activation function for the layer (i.e., the output layer) is the Sigmoid function, which can receive information from the preceding neural layers and convert it into 0 / 1 values ​​for output, thus making it suitable for solving binary classification problems.

[0025] The expression for the Sigmoid function is: ;Formula (4) In this function, the input is , and If they are equal, they are both outputs of the function.

[0026] The training process of the positive discriminative model is explained in detail below: The sample data of each subset is divided into a training set, a validation set, and a test set, and these three datasets are randomly selected. The training set is input into the forward discriminant model for training (specifically, the vehicle type and test parameters are input into the model, and the rollover result is used as the predicted output value / label). The loss function during training is the difference between the predicted output value and the label of the forward discriminant model.

[0027] Simultaneously, the validation set is used as a benchmark observation input into the forward discriminant model to check whether the loss function output curve of the forward discriminant model (i.e., the curve of the predicted output value changing with the number of iterations) shows overfitting. The network parameters are saved when the loss function decreases below the threshold and no overfitting phenomenon is observed.

[0028] The test set is input into the trained positive discriminant model. This step only calculates the difference between the predicted output value and the label, without retraining or validating the model using the test set. Once the calculated difference is less than a threshold, the positive discriminant model is saved.

[0029] The training and validation process of the positive discrimination model is repeated for each type of vehicle to obtain positive discrimination models for three sub-types (sedan, SUV, MPV), which can be used to determine whether each sub-type has rolled over.

[0030] S130. The basic inverse model is obtained by calculating the inverse model from the forward discriminant model, and constraints are added to the output of the basic inverse model.

[0031] S140. Under the constraints of the constraints, the basic inverse model is modified using the vehicle parameters and test parameters of the vehicle to be solved.

[0032] The relationship between the weights and biases of neurons in each layer of the forward discriminant model is shown in the following equation: The first in the layer i The relationship between a neuron and neurons in layer i, and their corresponding weights and biases.

[0033] ;Formula (5) ;Formula (6) in," "Indicates the identifier of the positive discriminant model, Indicates the first The number of neurons in the layer, In the forward discriminant model, the first... The first in the layer i The first neuron and the second The weights of the j-th neuron in the layer. It is the first in the forward discriminant model Layer i The value of each neuron, It is the first in the forward discriminant model Layer i The bias value of each neuron. Indicates the first Layer i The activation function of a nerve.

[0034] Formula (6) can be expressed in matrix and vector form as follows: ;Formula (7) ;Formula (8) in, It is the diagonal matrix of activation functions, as shown in formula (9), where the elements on the diagonal represent the activation functions of each layer. It should be emphasized that in the direct weight inverse method, the activation function should be an invertible function. It is the first in the forward discriminant model A matrix of values ​​for layer neurons. It is the first in the forward discriminant model The matrix of weights of layer neurons, yes The elements in It is the first in the forward discriminant model The matrix of biases of layer neurons.

[0035] ;Formula (9) The weight matrix is ​​shown in the following formula: ;Formula (10) The explicit expression of the corresponding basic inverse model based on formula (8) is shown in the following formula. Since the weight matrix in the neural network is not invertible, it is subjected to generalized inverse processing.

[0036] ;Formula (11) Standardized regularized least squares method is used in formula (11). This represents the Tikhonov regularization parameter. This represents the identity matrix corresponding to the regularization parameters. The regularization parameters should be determined based on the measurement noise level of the measurable output or the added noise level in the calculated response. Because... Since it becomes a square matrix, the expressions for the weights and biases in the basic inverse model can be represented by formulas (12) and (13): ;Formula (12) ;Formula (13) Using formulas (12) and (13), formula (11) can be expressed by formula (14): ;Formula (14) Here, “←” indicates the identifier for the inverse model. This represents the weight matrix in the basic inverse model. It is the bias matrix of the basic inverse model.

[0037] Formula (14) can also be expressed in the form of formula (15): ;Formula (15) The inverse expression of the sigmoid activation function is: ;Formula (16) in, , which is the output value of the activation function.

[0038] Figure 3This is a schematic diagram of the basic inverse model provided in this application embodiment. The training process of the basic inverse model is as follows: Since the forward discrimination model has already been trained, and the batch size and learning rate during training are set to small values, the basic inverse model calculated by the forward discrimination model only needs parameter fine-tuning. Label values ​​are assigned to the output parameters of the basic inverse model, excluding the slope angle. The rollover result is input into the basic inverse model, which outputs the vehicle parameters and test parameters of the vehicle to be solved. Since label values ​​have already been assigned to the output parameters other than the slope angle (i.e., constraints have been added), a loss function is constructed based on the output parameters other than the slope angle and their corresponding label values ​​in the basic inverse model. By minimizing the loss function, the model parameters in the basic inverse model are updated.

[0039] This application employs a method of adding partial output constraints. At the output of the basic inverse model, label values ​​are set for output neurons other than the parameter to be determined (i.e., the slope angle), thereby limiting and constraining these output values. Simultaneously, a new loss function is designed for the basic inverse model. This loss function only calculates the loss for neurons with labeled values, excluding the slope angle. That is, the weight and bias updates of the basic inverse model are caused by the loss calculation of neurons with labeled values, while the slope angle passively follows the updates of the basic inverse model and outputs a predicted value, without feeding back the loss calculation or affecting the gradient and network training. For example, the loss function is as follows: ;Formula (17) in, This represents the output value of a neuron with a label. The label values ​​represent the labels of neurons with labels, including all parameters in the vehicle parameters and test parameters except for the slope angle.

[0040] S150. Input the results of whether or not the rollover occurs into the basic inverse model to obtain the first slope angle and the second slope angle output by the basic inverse model.

[0041] First, select the corresponding basic inverse model based on the type of vehicle to be solved. Input the non-rollover result into the basic inverse model to obtain the first slope angle output by the basic inverse model; input the rollover result into the basic inverse model to obtain the second slope angle output by the basic inverse model; wherein, the second slope angle is greater than the first slope angle.

[0042] The input to the basic inverse model is the result of whether the vehicle rolled over (0 or 1), and the output is vehicle parameters and test parameters. It is evident that the basic inverse model has the exact opposite structure to the forward discriminant model.

[0043] S160. Construct an angle range based on the first slope angle and the second slope angle, and filter the test data within the angle range.

[0044] An angle range is constructed by using the first slope angle as the lower limit and the second slope angle as the upper limit.

[0045] In the subset of test data corresponding to the vehicle to be solved, determine whether the slope angle falls within the angle range, and filter the target slope angle that falls within the angle range; then filter out the test data corresponding to the target slope angle.

[0046] S170. Determine whether the cutoff condition has been met.

[0047] The cutoff criteria include: the selected test data contains only one type of roll result, or the angle range is less than the set angle range threshold.

[0048] According to S150, the results of whether or not the rollover occurs are input into the updated base inverse model to obtain the new first slope angle and the new second slope angle output by the base inverse model. If the angle range constructed by the new first slope angle and the new second slope angle is less than the set angle range threshold, such as the minimum error value of the test conditions, or the minimum travel that the angle can reach, such as 7 degrees, then the loop stops.

[0049] If the angle range formed by the new first slope angle and the new second slope angle is greater than or equal to the set angle range threshold, then continue to filter the test data within the new angle range.

[0050] If the selected test data only show rolling results or only show non-rolling results, then stop the loop.

[0051] S180. If not, then use the selected experimental data to train the positive discrimination model and return to S130.

[0052] The selected experimental data is used to train the forward discriminant model, and the base inverse model is then updated. The training process of the forward discriminant model is described in the above embodiment and will not be repeated here.

[0053] Since the experimental data (training samples) has been reduced, the parameters of the forward discriminant model have also been updated after training. Therefore, the trained forward discriminant model focuses more on simulating the nonlinear mapping relationship within the "slope angle interval".

[0054] Then, based on the above description, a new basic inverse model is obtained by calculating the inverse model according to the new forward discrimination model. The maximum range of the output angle of this basic inverse model is the "slope angle interval".

[0055] S190. If so, obtain the target angle range including the rollover critical angle of the vehicle to be solved.

[0056] Through the above iterative scheme, the values ​​of rolling and not rolling are gradually converged to a certain small range.

[0057] S191. By conducting real vehicle angle gradient tests, the rollover critical angle value is finally verified from the target angle range.

[0058] For example, the target angle range is divided according to a set travel value (e.g., 1 degree as a step) to obtain multiple angle values; at each of the multiple angle values, a real vehicle verification test is carried out, and the rollover critical angle value is obtained based on whether the real vehicle rolls over.

[0059] For example, after dividing the vehicle into three angle values ​​of 45, 46, and 47 degrees, a slope test platform with these values ​​is designed. The actual vehicle is then driven onto the slope test platform for testing, and it is observed whether the vehicle rolls over. If the vehicle does not roll over at angles of 45 and 46, but rolls over at angle 47, then the critical rollover angle for the actual vehicle is 47 degrees.

[0060] This application establishes forward discrimination models for each vehicle segment and calculates a basic inverse model with additional output constraints. This basic inverse model enables the reverse solution from rollover results to test parameters. Then, by converging sample sets, a narrowed target angle range is obtained. Finally, a real-vehicle verification method is used to obtain the final rollover critical angle value. This method effectively solves the boundary condition problem in complex scenarios of vehicle slope rollover testing, providing a feasible test scheme for automakers. It significantly reduces testing costs for automakers while more accurately obtaining the rollover critical angle value under real-vehicle conditions, improving the vehicle's recognition and response to slope rollover boundaries, and providing more efficient protection for occupants.

[0061] like Figure 4 As shown, this embodiment provides an electronic device, including: At least one processor; and A memory communicatively connected to at least one of the processors; wherein, The memory stores instructions executable by at least one of the processors to enable the processor to perform the described method. Since at least one processor in the electronic device is capable of performing the described method, it thus possesses at least the same advantages as the described method.

[0062] Optionally, the electronic device also includes interfaces for connecting the various components, including high-speed interfaces and low-speed interfaces. The components are interconnected using different buses and can be mounted on a common motherboard or otherwise installed as needed. The processor can process instructions executed within the electronic device, including instructions stored in or on memory to display graphical information of a GUI (Graphical User Interface) on an external input / output device (such as a display device coupled to the interface). In other embodiments, multiple processors can be used with multiple memories, and / or multiple buses can be used with multiple memories, if desired. Similarly, multiple electronic devices (e.g., as a server array, a group of blade servers, or a multiprocessor system) can be connected, each providing some of the necessary operations. Figure 4 Take processor 301 as an example.

[0063] The memory 302, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as the program instructions / modules corresponding to the vehicle slope rollover test critical parameter determination method in this embodiment of the application. The processor 301 executes various functional applications and data processing of the electronic device by running the software programs, instructions, and modules stored in the memory 302, thereby realizing the above-mentioned vehicle slope rollover test critical parameter determination method.

[0064] The memory 302 may primarily include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a given function; the data storage area may store data created based on the use of the electronic device. Furthermore, the memory 302 may include high-speed random access memory and non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 302 may further include memory remotely located relative to the processor 301, which can be connected to the electronic device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0065] The electronic device may also include an input device 303 and an output device 304. The processor 301, memory 302, input device 303, and output device 304 can be connected via a bus or other means; the figure shows an example of a connection via a bus.

[0066] Input device 303 can receive input digital or character information, and output device 304 may include a display device, an auxiliary lighting device (e.g., an LED), and a haptic feedback device (e.g., a vibration motor). The display device may include, but is not limited to, a liquid crystal display (LCD), a light-emitting diode (LED) display, and a plasma display. In some embodiments, the display device may be a touchscreen.

[0067] This embodiment provides a computer-readable storage medium storing computer instructions for instructing a computer to perform the methods described above. The computer instructions on this computer-readable storage medium, used to instruct a computer to perform the methods described above, thus possess at least the same advantages as the methods described above.

[0068] The medium in this application may be any combination of one or more computer-readable media. The medium may be a computer-readable signal medium or a computer-readable storage medium. The medium may be, for example, but not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples of the medium (a non-exhaustive list) include: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this document, the medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0069] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of sending, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0070] Program code contained on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wire, optical fiber, RF (Radio Frequency), or any suitable combination thereof.

[0071] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages—such as Java, Smalltalk, and C++—as well as conventional procedural programming languages—such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0072] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this application can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this application can be achieved, and this is not limited herein.

[0073] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for determining critical parameters in a vehicle slope rollover test, characterized in that, include: Collect a set of test data on actual vehicle rollovers, and divide the test data set into test data subsets according to vehicle type. The test data subsets include: vehicle parameters, test parameters, and rollover results; the test parameters include the slope angle. The vehicle parameters and test parameters in each subset of test data are input into a forward discrimination model corresponding to the vehicle type, and the rollover result is used as a label to train the forward discrimination model. The basic inverse model is obtained by inverse calculation from the forward discriminant model, and constraints are added to the output of the basic inverse model. Under the constraints described above, the basic inverse model is modified using the vehicle parameters and test parameters of the vehicle to be solved. The results of whether or not the rollover occurs are input into the basic inverse model to obtain the first slope angle and the second slope angle output by the basic inverse model. An angle range is constructed based on the first slope angle and the second slope angle, and test data within the angle range are selected. Return to the operation of training the positive discrimination model using the selected test data until the cutoff condition is met, and obtain the target angle range including the rollover critical angle of the vehicle to be solved; By conducting real-vehicle angle gradient tests, the rollover critical angle value was finally verified from the target angle range.

2. The method according to claim 1, characterized in that, Vehicle parameters include: left front wheel mass, right front wheel mass, left rear wheel mass, right rear wheel mass, front axle mass, rear axle mass, vehicle length, vehicle width, and vehicle height; The test parameters include: slope angle, test speed, and vehicle entry angle.

3. The method according to claim 2, characterized in that, Before inputting the vehicle parameters and test parameters from each subset of test data into the forward discrimination model corresponding to the vehicle type, the following steps are also included: Assign a coding value to the text-type data in each subset of experimental data; The encoded values ​​are de-normalized and normalized.

4. The method according to claim 1, characterized in that, The forward discrimination model includes: an input layer, one or more intermediate layers, and an output layer; The activation function of the output layer is the Sigmoid function.

5. The method according to claim 1, characterized in that, Add constraints to the output of the basic inverse model, including: Set label values ​​for the output parameters of the basic inverse model, except for the slope angle; Under the constraints described above, the basic inverse model is modified using the vehicle parameters and test parameters of the vehicle to be solved, including: The rollover results are input into the basic inverse model. Based on the output parameters of the basic inverse model other than the slope angle and the corresponding label values, a loss function is constructed. The model parameters in the underlying inverse model are updated by minimizing the loss function.

6. The method according to claim 1, characterized in that, The results of whether or not the slope has rolled are input into the basic inverse model to obtain the first slope angle and the second slope angle output by the basic inverse model, including: The result of no rollover is input into the basic inverse model to obtain the first slope angle output by the basic inverse model. The result of the rollover is input into the basic inverse model to obtain the second slope angle output by the basic inverse model; The second slope angle is greater than the first slope angle.

7. The method according to claim 6, characterized in that, An angle range is constructed based on the first slope angle and the second slope angle, and test data within the angle range are selected, including: The angle range is constructed by taking the first slope angle as the lower limit and the second slope angle as the upper limit. In the subset of test data corresponding to the vehicle to be solved, determine whether the slope angle falls within the angle range, and filter the target slope angle that falls within the angle range. The test data corresponding to the target slope angle were selected.

8. The method according to claim 1, characterized in that, The cutoff conditions include: the selected test data contains only one type of tumbling result, or the angle range is less than a set angle range threshold.

9. The method according to any one of claims 1-8, characterized in that, By conducting real-vehicle angle gradient tests, the rollover critical angle value was finally verified from the target angle range, including: The target angle range is divided according to the set travel value to obtain multiple angle values; For each of the multiple angle values, a real-vehicle verification test was conducted; The rollover critical angle value is obtained based on whether the actual vehicle rolls over.

10. An electronic device, characterized in that, include: At least one processor, and a memory communicatively connected to at least one of the processors; The memory stores instructions executable by at least one of the processors, which, when executed by at least one of the processors, enable the at least one of the processors to perform the method for determining critical parameters of a vehicle slope rollover test as described in any one of claims 1-9.

Images