Safety control method and device of vehicle, storage medium and electronic device

Abstract

The application provides a safety control method and device of a vehicle, a storage medium and an electronic device, and the method comprises the following steps: collecting body attribute information of a passenger in the vehicle, determining the tolerance of the passenger according to the body attribute information, obtaining historical scene parameters of the vehicle before a crash and obtaining a historical acceleration sequence of the vehicle before the first ignition when the vehicle ignites a passive safety assembly at a first ignition time; calculating waveform intensity according to the historical scene parameters and the historical acceleration sequence, wherein the waveform intensity is used to represent the crash intensity of the vehicle after the first ignition time; and sending a second ignition instruction to the passive safety assembly according to the tolerance and the waveform intensity, wherein the tolerance and the waveform intensity are negatively correlated with the time interval between the second ignition time and the first ignition time. Through the embodiment, the technical problem that the passenger is at high risk of injury in the related art caused by igniting the passive safety assembly at a fixed time after the vehicle crashes is solved.

Description

Technical Field

[0001] This invention relates to the field of vehicle safety technology, and more specifically, to a vehicle safety control method and device, storage medium, and electronic device. Background Technology

[0002] In related technologies, with the continuous advancement of active and passive safety technologies in automobiles, road traffic safety has been significantly improved. Although the fatality rate due to traffic accidents has been declining in recent years, a large number of drivers and passengers still suffer minor, moderate, and severe injuries in accidents every year. Currently, most vehicle passive safety systems adopt a uniform calibration scheme, meaning that under the same conditions, airbags and seat belts for different occupants (e.g., differences in height, build, and age) ignite at the same preset time, and the seat belt force limiter and airbag stiffness are the same. Obviously, this cannot provide optimal protection for different occupants and collision scenarios.

[0003] In related technologies, the first ignition timing is achieved by the airbag controller (ACU) based on pre-calibrated acceleration characteristics. The existing second ignition timing is usually based on the first ignition timing delayed by a fixed time. This design cannot take into account different collision scenarios and occupant parameters, and its response is insufficient to adapt to the complex and ever-changing actual collision environment, affecting occupant safety.

[0004] No efficient and accurate solution has yet been found to address the aforementioned issues in the relevant technologies. Summary of the Invention

[0005] This invention provides a vehicle safety control method and apparatus, storage medium, and electronic device to solve technical problems in related technologies.

[0006] According to an embodiment of the present invention, a vehicle safety control method is provided, comprising: collecting physical attribute information of occupants in the vehicle; determining the occupants' tolerance based on the physical attribute information; after the vehicle ignites a passive safety component at the first ignition moment, acquiring historical scene parameters of the vehicle before a collision and acquiring a historical acceleration sequence of the vehicle before the first ignition; calculating waveform intensity based on the historical scene parameters and the historical acceleration sequence, wherein the waveform intensity is used to characterize the collision intensity of the vehicle after the first ignition moment; and sending a secondary ignition command to the passive safety component based on the tolerance and the waveform intensity, wherein the tolerance and the waveform intensity are both negatively correlated with the time interval between the second ignition moment and the first ignition moment.

[0007] Optionally, calculating waveform intensity based on the historical scene parameters and the historical acceleration sequence includes: normalizing and length-aligning the historical scene parameters and the historical acceleration sequence to obtain corresponding first preprocessed data and second preprocessed data; inputting the first preprocessed data and the second preprocessed data into a pre-trained waveform feature predictor to obtain multiple waveform feature values, wherein each waveform feature value is used to characterize the collision intensity of the vehicle in a corresponding dimension, and the multiple waveform feature values ​​include mean acceleration, peak acceleration time, speed drop, and acceleration change trend index; and calculating waveform intensity using the multiple waveform feature values.

[0008] Optionally, calculating the waveform intensity using the various waveform characteristic values ​​includes: calculating using the following formula. : ;in, The average acceleration, At the peak of acceleration, To reduce speed, As an indicator of acceleration change trend, , , and For the corresponding reference value, , , , These are the corresponding weighting coefficients.

[0009] Optionally, before inputting the first preprocessed data and the second preprocessed data into the pre-trained waveform feature predictor, the method further includes: acquiring multiple sets of sample data, wherein each set of sample data includes a sample acceleration sequence and true feature values, wherein the true feature values ​​include the following true values: mean acceleration, peak acceleration time, velocity drop, and acceleration trend index; calculating feature prediction values ​​using the sample acceleration sequence, wherein the feature prediction values ​​include the following predicted values: mean acceleration, peak acceleration time, velocity drop, and acceleration trend index; constructing a loss function; and training the loss function using the feature prediction values ​​and the true feature values ​​to obtain the waveform feature predictor.

[0010] Optionally, calculating the feature prediction value using the sample acceleration sequence includes: calculating the mean acceleration using the following formula. : ;in, This refers to the first ignition moment in the sample scenario. Let be the acceleration at time t. This is the preset effective collision duration.

[0011] Optionally, calculating the feature prediction value using the sample acceleration sequence includes: calculating the peak acceleration time using the following formula. : ;in, This refers to the first ignition moment in the sample scenario. The preset effective collision duration, Let be the acceleration at time t.

[0012] Optionally, calculating the feature prediction value using the sample acceleration sequence includes: calculating the velocity drop using the following formula. : ;in, This refers to the first ignition moment in the sample scenario. This is the preset effective collision duration.

[0013] Optionally, calculating the feature prediction value using the sample acceleration sequence includes: calculating the acceleration change trend index using the following formula. : ;in, This refers to the first ignition moment in the sample scenario. At the peak of acceleration, for acceleration at any moment for Acceleration at any moment.

[0014] Optionally, constructing the loss function includes using the following formula: : ;in, These are the predicted values ​​for the mean acceleration, peak acceleration time, velocity drop, and acceleration trend index, respectively. These represent the true values ​​of the mean acceleration, peak acceleration, velocity drop, and acceleration trend index, respectively. , , , These are the weight coefficients corresponding to the feature loss. The influence coefficient of physical constraints. These are hyperparameters used to control the sensitivity of physical constraints. It is a constant used to constrain the correlation between the acceleration change trend index and the mean acceleration.

[0015] Optionally, obtaining the historical acceleration sequence of the vehicle before the first ignition includes: calculating the acceleration set of the vehicle at multiple consecutive moments in real time; determining whether the acceleration set is less than a preset threshold; if the acceleration set is less than the preset threshold, determining the current moment as the collision start moment; and obtaining the historical acceleration sequence of the vehicle between the collision start moment and the first ignition moment.

[0016] Optionally, sending a secondary ignition command to the passive safety component based on the tolerance and the waveform intensity includes: predicting the time interval between the second ignition time and the first ignition time based on the tolerance and the waveform intensity; calculating the second ignition time based on the time interval and the first ignition time; determining whether the second ignition time is within the effective duration of the current collision; if the second ignition time is not within the effective duration of the current collision, calculating a corrected ignition time based on a preset ignition interval threshold, the first ignition time, and the second ignition time, and sending a secondary ignition command to the passive safety component according to the corrected ignition time.

[0017] According to another embodiment of the present invention, a vehicle safety control device is provided, comprising: a first acquisition module, configured to acquire physical attribute information of an occupant in the vehicle, determine the occupant's tolerance based on the physical attribute information, and acquire historical scene parameters of the vehicle before a collision and historical acceleration sequence of the vehicle before the first ignition after the vehicle ignites a passive safety component at the first ignition time; a first calculation module, configured to calculate waveform intensity based on the historical scene parameters and the historical acceleration sequence, wherein the waveform intensity is used to characterize the collision intensity of the vehicle after the first ignition time; and a sending module, configured to send a secondary ignition command to the passive safety component based on the tolerance and the waveform intensity, wherein the tolerance and the waveform intensity are both negatively correlated with the time interval between the second ignition time and the first ignition time.

[0018] Optionally, the first calculation module includes: a processing unit, configured to normalize and length-align the historical scene parameters and the historical acceleration sequence to obtain corresponding first preprocessed data and second preprocessed data; a prediction unit, configured to input the first preprocessed data and the second preprocessed data into a pre-trained waveform feature predictor to obtain multiple waveform feature values, wherein each waveform feature value is used to characterize the collision intensity of the vehicle in a corresponding dimension, and the multiple waveform feature values ​​include mean acceleration, peak acceleration time, speed drop, and acceleration change trend index; and a calculation unit, configured to calculate the waveform intensity using the multiple waveform feature values.

[0019] Optionally, the calculation unit includes: a calculation subunit, used to calculate using the following formula :

[0020] ;in, The average acceleration, At the peak of acceleration, To reduce speed, As an indicator of acceleration change trend, , , and For the corresponding reference value, , , , These are the corresponding weighting coefficients.

[0021] Optionally, the apparatus further includes: a second acquisition module, configured to acquire multiple sets of sample data before the first calculation module inputs the first preprocessed data and the second preprocessed data into the pre-trained waveform feature predictor, wherein each set of sample data includes a sample acceleration sequence and true feature values, wherein the true feature values ​​include the following true values: mean acceleration, peak acceleration time, velocity drop, and acceleration trend index; a second calculation module, configured to calculate feature prediction values ​​using the sample acceleration sequence, wherein the feature prediction values ​​include the following prediction values: mean acceleration, peak acceleration time, velocity drop, and acceleration trend index; a construction module, configured to construct a loss function; and a training module, configured to train the loss function using the feature prediction values ​​and the true feature values ​​to obtain the waveform feature predictor.

[0022] Optionally, the second calculation module includes: a first calculation unit, used to calculate the average acceleration using the following formula. : ;in, This refers to the first ignition moment in the sample scenario. Let be the acceleration at time t. This is the preset effective collision duration.

[0023] Optionally, the second calculation module includes: a second calculation unit, used to calculate the peak acceleration moment using the following formula. : ;in, This refers to the first ignition moment in the sample scenario. The preset effective collision duration, Let be the acceleration at time t.

[0024] Optionally, the second calculation module includes: a third calculation unit, used to calculate the speed reduction using the following formula. : ;in, This refers to the first ignition moment in the sample scenario. This is the preset effective collision duration.

[0025] Optionally, the second calculation module includes: a fourth calculation unit, used to calculate the acceleration change trend index using the following formula. : ;in, This refers to the first ignition moment in the sample scenario. At the peak of acceleration, for acceleration at any moment for Acceleration at any moment.

[0026] Optionally, the construction module includes: a construction unit for constructing a loss function using the following formula. : ;in, These are the predicted values ​​for the mean acceleration, peak acceleration time, velocity drop, and acceleration trend index, respectively. These represent the true values ​​of the mean acceleration, peak acceleration, velocity drop, and acceleration trend index, respectively. , , , These are the weight coefficients corresponding to the feature loss. The influence coefficient of physical constraints. These are hyperparameters used to control the sensitivity of physical constraints. It is a constant used to constrain the correlation between the acceleration change trend index and the mean acceleration.

[0027] Optionally, the first acquisition module includes: a calculation unit, used to calculate the acceleration set of the vehicle at multiple consecutive moments in real time; a judgment unit, used to judge whether the acceleration set is less than a preset threshold; a determination unit, used to determine the current moment as the collision start moment if the acceleration set is less than the preset threshold; and an acquisition unit, used to acquire the historical acceleration sequence of the vehicle between the collision start moment and the first ignition moment.

[0028] Optionally, the sending module includes: a calculation unit, configured to predict the time interval between the second ignition time and the first ignition time based on the tolerance and the waveform intensity; a judgment unit, configured to determine whether the second ignition time is within the effective duration of the current collision; an update unit, configured to calculate a corrected ignition time based on a preset ignition interval threshold, the first ignition time, and the second ignition time if the second ignition time is not within the effective duration of the current collision; and a sending unit, configured to send a second ignition command to the passive safety component according to the corrected ignition time.

[0029] According to another aspect of the embodiments of this application, a storage medium is also provided, the storage medium including a stored program that executes the above steps when the program is run.

[0030] According to another aspect of the embodiments of this application, an electronic device is also provided, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; wherein: the memory is used to store computer programs; and the processor is used to execute the steps in the above method by running the programs stored in the memory.

[0031] This application also provides a computer program product containing instructions that, when run on a computer, cause the computer to perform the steps in the above-described method.

[0032] The beneficial effects of this invention are:

[0033] 1. Not only does it utilize the scene parameters before the collision, but it also combines the acceleration sequence from the start of the collision to the moment of the first ignition. By making full use of the collision acceleration information, it greatly improves the ability to accurately predict the intensity of the current collision. The acceleration sequence can accurately reflect the strength of the current collision. The scene parameters contain rich stiffness information about the collision target in the time dimension, which is closely related to the comprehensive stiffness of the required constraint system. This determines the optimal moment for the second ignition, realizes the adaptive control of the multi-level constraint system, and minimizes the risk of occupant injury.

[0034] 2. Considering the significant differences in energy tolerance among occupants of different heights, weights, and ages during a collision, this application employs a personalized restraint system to achieve more precise safety protection;

[0035] 3. The design of the multi-task loss function that incorporates physical constraints not only improves the accuracy of prediction of each waveform feature value, but also ensures the physical consistency between features. By optimizing and constraining feature prediction, the loss function enhances the interpretability of the system, enabling it to better adapt to complex and ever-changing collision environments. Attached Figure Description

[0036] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0037] Figure 1 This is a hardware structure block diagram of a car according to an embodiment of the present invention;

[0038] Figure 2 This is a flowchart of a vehicle safety control method according to an embodiment of the present invention;

[0039] Figure 3 This is a schematic diagram of the waveform feature predictor framework in an embodiment of the present invention;

[0040] Figure 4 This is a schematic diagram of the waveform feature predictor in an embodiment of the present invention;

[0041] Figure 5 This is a schematic diagram illustrating the calculation of waveform features using acceleration sequences in an embodiment of the present invention;

[0042] Figure 6 This is a schematic diagram of the intelligent control method for secondary ignition in vehicle collision according to an embodiment of the present invention;

[0043] Figure 7 This is a structural block diagram of a vehicle safety control device according to an embodiment of the present invention. Detailed Implementation

[0044] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, and not all of them. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present application. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of the present application can be combined with each other.

[0045] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0046] Example 1

[0047] The method embodiment provided in Embodiment 1 of this application can be executed in an automobile, server, processor, security controller, autonomous driving / assisted driving / intelligent driving controller, or similar processing device. Taking its operation in an automobile as an example, Figure 1 This is a hardware structure block diagram of a car according to an embodiment of the present invention. For example... Figure 1 As shown, a car may include one or more ( Figure 1 Only one is shown in the image. A processor 102 (which may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.) and a memory 104 for storing data are also shown. Optionally, the vehicle may further include a transmission device 106 for communication functions and an input / output device 108. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned automobile. For example, the automobile may also include components that are more... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0048] The memory 104 can be used to store vehicle programs, such as application software programs and modules, like the vehicle program corresponding to a vehicle safety control method in this embodiment of the invention. The processor 102 executes various functional applications and data processing by running the vehicle program stored in the memory 104, thereby implementing the aforementioned method. The memory 104 may include high-speed random access memory and non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the vehicle via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0049] The transmission device 106 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by a vehicle's communication provider. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 may be a Radio Frequency (RF) module used for wireless communication with the Internet.

[0050] This embodiment provides a vehicle safety control method. Figure 2 This is a flowchart of a vehicle safety control method according to an embodiment of the present invention, such as... Figure 2 As shown, the process includes the following steps:

[0051] Step S201: Collect the physical attribute information of the occupants in the vehicle, determine the occupants' tolerance based on the physical attribute information, and after the passive safety components are ignited at the moment of the first ignition of the vehicle, obtain the historical scene parameters of the vehicle before the collision and the historical acceleration sequence of the vehicle before the first ignition.

[0052] Optionally, occupant tolerance characterizes an occupant's ability to resist and withstand external risks (such as impacts). It is related to physical attributes such as occupant weight, age, physique, height, gender, health status, and occupation. For example, the greater the height and weight, the higher the tolerance. Physical attribute information can be obtained by capturing images of the occupant's body through a camera, or by extracting or parsing attribute information such as age and gender input by the occupant.

[0053] Optional passive safety components may include in-vehicle devices such as airbags, seat belts, and side curtain airbags. For example, airbags may inflate on the first and second ignition cycles, and seat belts may tighten on the first and second ignition cycles.

[0054] Optionally, the historical scene parameters before the collision include the type of the collision target, the speeds of the vehicle and the target vehicle, the collision angle, and the overlap rate. These scene parameters determine the state of the vehicle and the target vehicle at the moment of collision and can be divided into two types: categorical variables and numerical variables. The type of target can be classified as a car, truck, bus, pole, guardrail, pedestrian, two-wheeled vehicle, or others, depending on the perception system's classification of the target. The vehicle's speed can be obtained from the CAN bus; the target vehicle's speed is obtained from the radar sensor; the collision angle refers to the relative angle between the target vehicle and the vehicle, which can be obtained from the radar sensor; the overlap rate can be calculated based on a kinematic model to estimate the percentage overlap between the vehicle and the target vehicle at the moment of collision.

[0055] Step S202: Calculate the waveform intensity based on the historical scene parameters and the historical acceleration sequence, wherein the waveform intensity is used to characterize the collision intensity of the vehicle after the first ignition moment;

[0056] Step S203: Send a secondary ignition command to the passive safety component based on the tolerance and the waveform intensity, wherein the tolerance and the waveform intensity are negatively correlated with the time interval between the second ignition time and the first ignition time;

[0057] The time interval in this embodiment can be calculated or predicted by a secondary ignition timing predictor. This predictor determines the time interval between the first and second ignitions based on waveform intensity and occupant type (corresponding to tolerance) classification. It can be calibrated empirically to form a two-dimensional matrix table, which can be directly looked up during application. The specific values ​​in the table are determined based on simulation or experimentation. Alternatively, it can be performed by simulating the constraint system to determine the optimal secondary ignition timing interval under different collision intensities and occupant type classifications. The input (features) are waveform intensity level and occupant type classification, and the output (label) is optimal. Data set B is used, and a machine learning model is trained and fitted to obtain a secondary ignition timing predictor. This machine learning model can be a tree-based decision tree, random forest, XGBoost, etc. In practical applications, the occupant tolerance in the current scenario and the calculated waveform intensity are input into the secondary ignition timing predictor, which then outputs the time interval. This allows for more precise control over the timing of the second ignition. Verification showed that the stronger the occupant's tolerance, the shorter the time interval and the stronger the waveform intensity output by the second ignition timing predictor. Both tolerance and waveform intensity were negatively correlated with the time interval.

[0058] The solution in this embodiment can be applied to any product with secondary functions, such as dual-stage seat belts and dual-stage airbags.

[0059] Through the above steps, the body attribute information of the occupants in the vehicle is collected, and the occupants' tolerance is determined based on the body attribute information. After the passive safety components are ignited at the first ignition moment, the historical scene parameters of the vehicle before the collision and the historical acceleration sequence of the vehicle before the first ignition are obtained. Waveform intensity is calculated based on the historical scene parameters and the historical acceleration sequence, wherein the waveform intensity is used to characterize the collision intensity of the vehicle after the first ignition moment. A second ignition command is sent to the passive safety components based on the tolerance and the waveform intensity, wherein the tolerance and the waveform intensity are negatively correlated with the time interval between the second ignition moment and the first ignition moment. By combining the collision intensity with the body shape information of the occupants, the optimal moment for the second ignition is determined, realizing the adaptive control of the multi-level constraint system, minimizing the risk of occupant injury, and solving the technical problem in related technologies where igniting the passive safety components at a fixed time after a vehicle collision leads to a high risk of occupant injury, thereby maximizing passenger safety after a vehicle collision.

[0060] In one example of this embodiment, obtaining the historical acceleration sequence of the vehicle before the first ignition includes: calculating the acceleration set of the vehicle at multiple consecutive moments in real time; determining whether the acceleration set is less than a preset threshold; if the acceleration set is less than the preset threshold, determining the current moment as the collision start moment; and obtaining the historical acceleration sequence of the vehicle between the collision start moment and the first ignition moment.

[0061] First ignition after collision ( The collision acceleration sequence parameters (waveforms) within the time period. "After the collision" refers to the moment the collision began ( This moment cannot be obtained directly and needs to be calculated. (Collision start moment) Calculation formula:

[0062]

[0063] ;

[0064] in, , These represent the X-axis (longitudinal) and Y-axis (lateral) acceleration sequences from the accelerometer, respectively. Each acceleration in the acceleration set is based on... and Calculated resultant acceleration From the moment the airbag controller first ignites ( Starting from this point, iterate forward at each sampling time. If there exists a certain time... ,and The acceleration values ​​at n consecutive time points are all less than the threshold. Then at that moment Defined as The historical acceleration sequence includes: from arrive The acceleration sequence values ​​within the time interval. A smaller value can be taken; in this example, the value is taken as follows. , The value of 3 is used to filter out noise data in the acceleration data and obtain the acceleration data from the start of the collision (close to the start of the collision) to the first ignition.

[0065] In one embodiment of this example, calculating waveform intensity based on the historical scene parameters and the historical acceleration sequence includes: normalizing and length-aligning the historical scene parameters and the historical acceleration sequence to obtain corresponding first preprocessed data and second preprocessed data; inputting the first preprocessed data and the second preprocessed data into a pre-trained waveform feature predictor to obtain multiple waveform feature values, wherein each waveform feature value is used to characterize the collision intensity of the vehicle in a corresponding dimension, and the multiple waveform feature values ​​include mean acceleration, peak acceleration time, speed drop, and acceleration change trend index; and calculating waveform intensity using the multiple waveform feature values.

[0066] The scene parameters and historical acceleration sequences are preprocessed. This preprocessing includes normalizing the numerical variables (velocity) in the scene parameters and the historical acceleration sequences, padding the historical acceleration sequences, and integer encoding of the categorical variables. The normalization is... Normalization: In the formula These are the training data. The minimum and maximum values ​​of (i.e., the data used to train the waveform feature predictor) The data is normalized. It is worth noting that only training data is used to prevent the leakage of data distribution information. calculate During inference, if the range of input parameters exceeds... Then, a clamping operation is performed on the input, that is: The data padding is to address the issue that the acquired acceleration sequence values ​​may not meet the waveform feature predictor's or its requirements for sequence input length. If the sequence length is less than the predictor's required input length, zeros are padded at the beginning; if it is greater than the predictor's required input length, linear interpolation resampling is performed to ensure that the resampled data length is equal to the waveform feature predictor's required input length.

[0067] The preprocessed historical scene parameters and the historical acceleration sequence are input into a pre-trained waveform feature predictor to obtain waveform feature values. Figure 3 This is a schematic diagram of the waveform feature predictor framework in an embodiment of the present invention. Preprocessed scene parameters and acceleration sequence parameters are input together into a pre-trained waveform feature predictor to obtain the feature values ​​of the vehicle collision waveform after the first ignition. , , These correspond to the mean acceleration, peak acceleration time, velocity drop, and acceleration trend index, respectively. These are then input into the waveform intensity determiner to calculate the waveform intensity S, where... The intensity of the acceleration after the first ignition directly reflects the collision intensity. The rate at which the peak energy arrives after the first ignition is determined; if the peak energy arrives quickly, the system tends to execute the second ignition more quickly in order to absorb more energy. The reaction is the magnitude of the vehicle collision energy after the first ignition; the greater the energy, the greater the collision intensity. The value corresponds to the trend of acceleration sequence changes from the first ignition moment to the moment the acceleration sequence reaches its peak. The larger this value is, the faster the subsequent collision intensity increases, and a second ignition should be performed as soon as possible. The waveform feature values ​​can be mapped to the interval [0,1], so that different feature values ​​are on the same order of magnitude, which facilitates subsequent processing.

[0068] In one example, calculating the waveform intensity using the multiple waveform feature values ​​includes: calculating using the following formula. : ;in, The average acceleration, At the peak of acceleration, To reduce speed, As an indicator of acceleration change trend, , , and For the corresponding reference value, , , , These are the corresponding weighting coefficients.

[0069] Optionally, during the development phase, waveforms for both the standard FRB (100% frontal overlap rigid barrier impact) and MDB (deformable moving barrier side impact) tests can be calculated separately. , , , And take the larger value as , , and This ensures that the evaluation of waveform intensity under all other operating conditions is relative to the standard operating condition. Optional, , , , The weighting coefficients for the corresponding features can be determined empirically or by fitting samples. In one example... , , , The values ​​are 0.3, 0.2, 0.40 and 0.1 respectively.

[0070] In this embodiment, before inputting the first preprocessed data and the second preprocessed data into the pre-trained waveform feature predictor, the method further includes: acquiring multiple sets of sample data, wherein each set of sample data includes a sample acceleration sequence and true feature values, wherein the true feature values ​​include the following true values: mean acceleration, peak acceleration time, velocity drop, and acceleration trend index; calculating feature prediction values ​​using the sample acceleration sequence, wherein the feature prediction values ​​include the following predicted values: mean acceleration, peak acceleration time, velocity drop, and acceleration trend index; constructing a loss function; and training the loss function using the feature prediction values ​​and the true feature values ​​to obtain the waveform feature predictor.

[0071] In one aspect of this embodiment, calculating the feature prediction value using the sample acceleration sequence includes: calculating the mean acceleration using the following formula. : ;in, This refers to the first ignition moment in the sample scenario. Let be the acceleration at time t. This is the preset effective collision duration.

[0072] In another aspect of this embodiment, calculating the feature prediction value using the sample acceleration sequence includes: calculating the peak acceleration time using the following formula. : ;in, This refers to the first ignition moment in the sample scenario. The preset effective collision duration, Let be the acceleration at time t.

[0073] In another aspect of this embodiment, calculating the feature prediction value using the sample acceleration sequence includes: calculating the velocity drop using the following formula. : ;in, This refers to the first ignition moment in the sample scenario. This is the preset effective collision duration.

[0074] In another aspect of this embodiment, calculating the feature prediction value using the sample acceleration sequence includes: calculating the acceleration change trend index using the following formula. : ;in, This refers to the first ignition moment in the sample scenario. At the peak of acceleration, for acceleration at any moment for Acceleration at any moment.

[0075] During the training phase of the waveform feature predictor, it is also necessary to preprocess the sample data based on the first ignition moment of the sample scene. ,Sure , obtain from to The corresponding acceleration sequences within the time interval are padded to the same length, where the padded length is the maximum length of all unpadded acceleration sequences. In this example, the sequence length is defined as a data point with a duration of 35ms. The numerical variables in the scene parameters and the acceleration sequences are normalized, and the categorical variables are encoded with consecutive positive integers starting from 0. The final sample input (features) consists of the sample scene parameters and the sample acceleration sequences, and the output is a dataset A of predicted waveform feature values. Furthermore, the database obtained from the simulation can be expanded, considering the first ignition time of each sample. The acceleration is determined based on the airbag controller algorithm, but this results in only one sample entering dataset A for each collision. To increase data diversity, in this application case, the synthesized acceleration of the simulation results for each collision is calculated. The range of values ​​is [ -5ms, [+5ms], each collision simulation result generates 11 training samples into dataset A. This increases data diversity to some extent, making the coverage of training data wider than the data distribution in actual use, while also solving the problem of different operating conditions. The problem of model instability caused by fluctuations.

[0076] Regarding the model building and structure of the waveform feature predictor, this example adopts a deep learning model based on an encoder-decoder architecture. Figure 4This is a schematic diagram of the waveform feature predictor in this embodiment of the invention. The inputs include categorical variables, numerical variables, and sequence variables (categorical and numerical variables come from scene parameters, and sequence variables come from acceleration sequences). The encoder inputs the categorical variables into the embedding layer, concatenates them along the feature dimension, and then inputs them into a multilayer perceptron (MLP1) to obtain the hidden representation of the categorical variables. The numerical variables are directly fed into a multilayer perceptron (MLP2) to obtain their hidden representations. Each sequence variable is input into a linear projection layer, summed along the feature dimension, and then fed into a sequence feature extractor (RNN, CNN, Transformer, etc.) to obtain the hidden representation of the sequence variables, which contains the time series features along the time dimension. Finally, the hidden representations of the categorical variables, numerical variables, and sequence variables are concatenated along the feature dimension to obtain the encoder's hidden representation. The decoder consists of four heads, each responsible for predicting one feature value. Each head is composed of a multilayer perceptron and projects the encoder's hidden representation onto the semantic space of the feature values ​​to obtain the feature values. The multilayer perceptron includes an input layer, one or more hidden layers, and an output layer. The input and output layers are composed of a linear projection layer, and the hidden layers are composed of a linear projection layer, a batch normalization layer, and an activation function layer. Optionally, MLP1 and MLP2 each have only one hidden layer, and the decoder consists of two hidden layers.

[0077] In the training process of the waveform feature predictor, in order to improve the prediction accuracy of the vehicle waveform intensity, a multi-task loss function combining physical constraints was constructed. ), used for training the waveform feature predictor.

[0078] In this embodiment, constructing the loss function includes: constructing the loss function using the following formula. :

[0079] ;

[0080] in, These are the predicted values ​​for the mean acceleration, peak acceleration time, velocity drop, and acceleration trend index, respectively. These represent the true values ​​of the mean acceleration, peak acceleration, velocity drop, and acceleration trend index, respectively. , , , These are the weight coefficients corresponding to the feature loss. The influence coefficient of physical constraints. These are hyperparameters used to control the sensitivity of physical constraints. It is a constant used to constrain the correlation between the acceleration change trend index and the mean acceleration.

[0081] The loss function in this embodiment aims to simultaneously predict the mean acceleration ( Peak time () ), speed difference ( )and Specifically, the loss function By weighting the prediction errors of each feature, an additional physical constraint is added ( The loss term corresponding to the physical constraints in the loss function is:

[0082]

[0083] ;

[0084] in, This is used to adjust the relationship between the trend indicator and the mean acceleration, ensuring that the mean acceleration also tends to increase as the trend indicator increases rapidly. The design of the loss function and the implementation of physical constraints significantly improve the model's prediction accuracy and interpretability, with an approximately 4.5% improvement in prediction accuracy compared to a loss function without physical constraints. Without a loss function incorporating physical constraints, although the model exhibits high prediction accuracy during training, some predictions may show significant errors and contradict physical laws during testing. The model might predict mean or peak acceleration values ​​that exceed the limits of actual vehicle physics, leading to distorted results, or show large peak accelerations with minimal speed drops. With the introduction of physical constraints, all errors of the trained model on the test set are controlled within 10%, avoiding illogical situations. This improvement not only enhances the model's reliability but also increases the accuracy of collision safety assessments.

[0085] During training, the specific physical constraints are implemented in the following aspects: First, there is an inverse relationship between the mean acceleration and the peak acceleration time; when the mean acceleration is high, the peak acceleration time should be earlier accordingly. Virtual constraint relationships are then set. First, ensure that the predicted peak moment satisfies physical logic; second, there is a positive correlation between the mean acceleration and the velocity drop, the greater the acceleration, the more significant the velocity drop after the collision, and the constraint relationship is as follows: This ensures that the velocity decrease increases with the increase of the mean acceleration; finally, the relationship between the trend indicator and the mean acceleration requires that the trend indicator increases rapidly under high acceleration conditions, and the constraint is expressed as follows: This is to ensure that the trend continues to increase even under high acceleration.

[0086] The overall training process of the waveform feature predictor includes database construction, model building, and model training. Sample data can be obtained from actual or simulated scenarios. In this embodiment, the database is constructed based on more efficient simulated scenarios. The database construction includes: (1) establishing a high-precision finite element model of the whole vehicle. (2) randomly sampling in a space composed of different scenario parameters to generate a simulation matrix table. (3) updating the finite element model according to the simulation matrix table, completing the simulation calculation, and processing the calculation results to obtain the X and Y accelerations of the vehicle during the collision process, and further calculating the composite acceleration. The sample acceleration sequence was obtained, and the synthetic acceleration was analyzed. The speed can be obtained by integrating it. This will be needed later when calculating the feature prediction value of the speed drop. Figure 5 This is a schematic diagram illustrating the calculation of waveform features using acceleration sequences in an embodiment of the present invention. The horizontal axis represents time, the vertical axis represents acceleration, the solid line represents known quantities, and the dashed line represents predicted quantities. , At the moment of first ignition, its acceleration is: , The moment the collision begins. The moment the collision ends, At the peak of acceleration, its acceleration is: .

[0087] This embodiment uses adaptive moment estimation (ADAM) as the optimizer. L2 regularization, dropout, and early stopping are employed to prevent overfitting. After training, cases are randomly selected from the test dataset for testing. The average loss is 0.065, and the prediction results for each feature value generally fall within ±10% of the error threshold, achieving an accuracy close to 90%.

[0088] Optionally, sending a secondary ignition command to the passive safety component based on the tolerance and the waveform intensity includes: predicting the time interval between the second ignition time and the first ignition time based on the tolerance and the waveform intensity; calculating the second ignition time based on the time interval and the first ignition time; determining whether the second ignition time is within the effective duration of the current collision; if the second ignition time is not within the effective duration of the current collision, calculating a corrected ignition time based on a preset ignition interval threshold, the first ignition time, and the second ignition time, and sending a secondary ignition command to the passive safety component according to the corrected ignition time.

[0089] In this embodiment, the time interval between the second ignition time and the first ignition time The second ignition timing predictor adaptively predicts the collision conditions and occupant body types (corresponding to tolerance) based on the collision conditions. Training samples are pre-built, multiple waveform intensity parameters are pre-set, and corresponding simulation condition combinations are established for different occupant body types (e.g., categorized by weight, height, or dummy type). Under each waveform intensity parameter and occupant body type combination, the time interval Δt between the second and first ignition timings is within a preset range. Random sampling is performed to obtain multiple candidate values. For each candidate value, a corresponding finite element simulation is performed to calculate the occupant damage evaluation index under that candidate value condition. Under the same collision waveform (intensity) and occupant body shape conditions, the occupant damage indices corresponding to different candidate values ​​are compared, and the candidate value that minimizes the occupant damage index is selected as the optimal time interval under the combination of collision condition and occupant body shape. When configuring labels, the waveform intensity parameter and occupant body shape classification are used as input features, and the optimal time interval determined above is used as the output label, thereby forming a sample dataset for model training. Based on the training dataset constructed above, a pre-selected machine learning algorithm is used to train the secondary ignition timing predictor. To improve the interpretability of the model, the prediction model is preferably trained using a tree-based machine learning model.

[0090] In practical applications, the secondary ignition timing predictor receives the waveform intensity and the tolerance corresponding to the current occupant's physical attributes as input, and outputs the time interval between the second ignition timing and the first ignition timing, thereby determining the second ignition timing of the restraint system.

[0091] Based on waveform intensity and occupant body type classification, the interval between the second ignition timing and the first ignition timing is output. Then, the second ignition moment It is necessary to ensure The following constraints must be met: If the conditions are not met, a corrected ignition timing needs to be calculated.

[0092] ,in To ensure that the interval between the second ignition time and the first ignition time is not less than a pre-set threshold. and no later than Based on the simulation results of the damage, calculations show that... A 5ms delay is acceptable and will have virtually no impact on the damage increase. A value of 100ms can be adopted to ensure that the function of the second ignition is released when the first collision has basically ended, thus meeting regulatory requirements.

[0093] This embodiment proposes a secondary ignition intelligent control method based on a comprehensive decision-making process using pre-collision data and occupant characteristics. Figure 6 This is a schematic flowchart of the intelligent control method for secondary ignition in a vehicle collision according to an embodiment of the present invention, including:

[0094] Step 601: Obtain the scene parameters before the collision. The scene parameters before the collision include the type of the collision target, the speed of the vehicle and the target vehicle, the collision angle, and the overlap rate. These scene parameters can uniquely determine the state of the vehicle and the target vehicle at the zero moment of the collision.

[0095] Step 602: Obtain the acceleration sequence from the time of collision to the first ignition;

[0096] Step 603: Preprocess the scene parameters and acceleration sequence;

[0097] Step 604: Input the preprocessed parameters into the pre-trained waveform feature predictor to obtain waveform feature values;

[0098] Step 605: Input the waveform feature values ​​into the waveform intensity determiner to obtain the waveform intensity;

[0099] Step 606: The secondary ignition timing predictor classifies the timing based on waveform intensity and occupant type, and outputs the interval between the second ignition timing and the first ignition timing. ;

[0100] Step 607, based on the calculated secondary ignition timing The input is given to the product that needs secondary functions (such as dual-stage seat belts and dual-stage airbags) to perform the secondary ignition function.

[0101] This embodiment analyzes collision scene parameters acquired before the collision (including collision target type, relative collision speed, collision angle, and overlap rate) and the acceleration sequence from the start of the collision to the first ignition. Using deep learning, it predicts waveform feature values ​​(acceleration trend indicators, speed drop, peak time, etc.) in real time to determine the collision intensity. Furthermore, by combining this with body shape information such as the size and weight of the occupants, it determines the optimal time for the second ignition, achieving adaptive control of the multi-level constraint system and minimizing the risk of occupant injury. The predicted waveform feature values ​​involve multiple values ​​with physical constraints. A physically constrained loss function design method for training the waveform feature predictor model is also proposed, significantly improving the model's prediction accuracy.

[0102] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0103] Example 2

[0104] This embodiment also provides a vehicle safety control device for implementing the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0105] Figure 7 This is a structural block diagram of a vehicle safety control device according to an embodiment of the present invention, such as... Figure 5 As shown, the device includes:

[0106] The first acquisition module 71 is used to collect the physical attribute information of the occupants in the vehicle, determine the occupants' tolerance based on the physical attribute information, and after the passive safety components are ignited at the moment of the first ignition of the vehicle, acquire the historical scene parameters of the vehicle before the collision and acquire the historical acceleration sequence of the vehicle before the first ignition.

[0107] The first calculation module 72 is used to calculate the waveform intensity based on the historical scene parameters and the historical acceleration sequence, wherein the waveform intensity is used to characterize the collision intensity of the vehicle after the first ignition moment;

[0108] The sending module 73 is used to send a secondary ignition command to the passive safety component based on the tolerance and the waveform intensity, wherein the tolerance and the waveform intensity are negatively correlated with the time interval between the second ignition time and the first ignition time.

[0109] Optionally, the first calculation module includes: a processing unit, configured to normalize and length-align the historical scene parameters and the historical acceleration sequence to obtain corresponding first preprocessed data and second preprocessed data; a prediction unit, configured to input the first preprocessed data and the second preprocessed data into a pre-trained waveform feature predictor to obtain multiple waveform feature values, wherein each waveform feature value is used to characterize the collision intensity of the vehicle in a corresponding dimension, and the multiple waveform feature values ​​include mean acceleration, peak acceleration time, speed drop, and acceleration change trend index; and a calculation unit, configured to calculate the waveform intensity using the multiple waveform feature values.

[0110] Optionally, the calculation unit includes: a calculation subunit, used to calculate using the following formula :

[0111] ;in, The average acceleration, At the peak of acceleration, To reduce speed, As an indicator of acceleration change trend, , , and For the corresponding reference value, , , , These are the corresponding weighting coefficients.

[0112] Optionally, the apparatus further includes: a second acquisition module, configured to acquire multiple sets of sample data before the first calculation module inputs the first preprocessed data and the second preprocessed data into the pre-trained waveform feature predictor, wherein each set of sample data includes a sample acceleration sequence and true feature values, wherein the true feature values ​​include the following true values: mean acceleration, peak acceleration time, velocity drop, and acceleration trend index; a second calculation module, configured to calculate feature prediction values ​​using the sample acceleration sequence, wherein the feature prediction values ​​include the following prediction values: mean acceleration, peak acceleration time, velocity drop, and acceleration trend index; a construction module, configured to construct a loss function; and a training module, configured to train the loss function using the feature prediction values ​​and the true feature values ​​to obtain the waveform feature predictor.

[0113] Optionally, the second calculation module includes: a first calculation unit, used to calculate the average acceleration using the following formula. : ;in, This refers to the first ignition moment in the sample scenario. Let be the acceleration at time t. This is the preset effective collision duration.

[0114] Optionally, the second calculation module includes: a second calculation unit, used to calculate the peak acceleration moment using the following formula. : ;in, This refers to the first ignition moment in the sample scenario. The preset effective collision duration, Let be the acceleration at time t.

[0115] Optionally, the second calculation module includes: a third calculation unit, used to calculate the speed reduction using the following formula. : ;in, This refers to the first ignition moment in the sample scenario. This is the preset effective collision duration.

[0116] Optionally, the second calculation module includes: a fourth calculation unit, used to calculate the acceleration change trend index using the following formula. : ;in, This refers to the first ignition moment in the sample scenario. At the peak of acceleration, for acceleration at any moment for Acceleration at any moment.

[0117] Optionally, the construction module includes: a construction unit for constructing a loss function using the following formula. : ;in, These are the predicted values ​​for the mean acceleration, peak acceleration time, velocity drop, and acceleration trend index, respectively. These represent the true values ​​of the mean acceleration, peak acceleration, velocity drop, and acceleration trend index, respectively. , , , These are the weight coefficients corresponding to the feature loss. The influence coefficient of physical constraints. These are hyperparameters used to control the sensitivity of physical constraints. It is a constant used to constrain the correlation between the acceleration change trend index and the mean acceleration.

[0118] Optionally, the first acquisition module includes: a calculation unit, used to calculate the acceleration set of the vehicle at multiple consecutive moments in real time; a judgment unit, used to judge whether the acceleration set is less than a preset threshold; a determination unit, used to determine the current moment as the collision start moment if the acceleration set is less than the preset threshold; and an acquisition unit, used to acquire the historical acceleration sequence of the vehicle between the collision start moment and the first ignition moment.

[0119] Optionally, the sending module includes: a calculation unit, configured to predict the time interval between the second ignition time and the first ignition time based on the tolerance and the waveform intensity; a judgment unit, configured to determine whether the second ignition time is within the effective duration of the current collision; an update unit, configured to calculate a corrected ignition time based on a preset ignition interval threshold, the first ignition time, and the second ignition time if the second ignition time is not within the effective duration of the current collision; and a sending unit, configured to send a second ignition command to the passive safety component according to the corrected ignition time.

[0120] It should be noted that the above modules can be implemented by software or hardware. For the latter, they can be implemented in the following ways, but are not limited to: all the above modules are located in the same processor; or, the above modules are located in different processors in any combination.

[0121] Example 3

[0122] Embodiments of the present invention also provide a storage medium storing a computer program, wherein the computer program is configured to execute the steps in any of the above method embodiments when running.

[0123] Optionally, in this embodiment, the storage medium may be configured to store a computer program for performing the following steps:

[0124] S1, collect the physical attribute information of the occupants in the vehicle, determine the occupants' tolerance based on the physical attribute information, and after the passive safety components are ignited at the moment of the first ignition of the vehicle, obtain the historical scene parameters of the vehicle before the collision and the historical acceleration sequence of the vehicle before the first ignition.

[0125] S2, calculate the waveform intensity based on the historical scene parameters and the historical acceleration sequence, wherein the waveform intensity is used to characterize the collision intensity of the vehicle after the first ignition moment;

[0126] S3, a secondary ignition command is sent to the passive safety component based on the tolerance and the waveform intensity, wherein the tolerance and the waveform intensity are negatively correlated with the time interval between the second ignition time and the first ignition time.

[0127] Optionally, in this embodiment, the storage medium may include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0128] Embodiments of the present invention also provide an electronic device including a memory and a processor, the memory storing a computer program and the processor being configured to run the computer program to perform the steps in any of the above method embodiments.

[0129] Optionally, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor and the input / output device is connected to the processor.

[0130] Optionally, in this embodiment, the processor can be configured to perform the following steps via a computer program:

[0131] S1, collect the physical attribute information of the occupants in the vehicle, determine the occupants' tolerance based on the physical attribute information, and after the passive safety components are ignited at the moment of the first ignition of the vehicle, obtain the historical scene parameters of the vehicle before the collision and the historical acceleration sequence of the vehicle before the first ignition.

[0132] S2, calculate the waveform intensity based on the historical scene parameters and the historical acceleration sequence, wherein the waveform intensity is used to characterize the collision intensity of the vehicle after the first ignition moment;

[0133] S3, a secondary ignition command is sent to the passive safety component based on the tolerance and the waveform intensity, wherein the tolerance and the waveform intensity are negatively correlated with the time interval between the second ignition time and the first ignition time.

[0134] Optionally, specific examples in this embodiment can refer to the examples described in the above embodiments and optional implementations, and will not be repeated here.

[0135] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0136] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented using software plus a general-purpose hardware platform, or of course, using hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0137] It should be understood that the terminology used herein is for the purpose of describing particular exemplary embodiments only and is not intended to be limiting. Unless the context clearly indicates otherwise, the singular forms “a,” “an,” and “described” as used herein may also include the plural forms. The terms “comprising,” “including,” “containing,” and “having” are inclusive and therefore indicate the presence of the stated features, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, elements, components, and / or combinations thereof. The method steps, processes, and operations described herein are not construed as requiring them to be performed in a particular order described or illustrated unless the order of performance is explicitly indicated. It should also be understood that additional or alternative steps may be used.

[0138] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A vehicle safety control method, characterized in that, include: The system collects the physical attribute information of the occupants in the vehicle, determines the occupants' tolerance based on the physical attribute information, and obtains the historical scene parameters of the vehicle before the collision and the historical acceleration sequence of the vehicle before the first ignition after the passive safety components are ignited at the moment of the first ignition of the vehicle. The occupants' tolerance is used to characterize the occupants' ability to resist and withstand external risks. The waveform intensity is calculated based on the historical scene parameters and the historical acceleration sequence, wherein the waveform intensity is used to characterize the collision intensity of the vehicle after the first ignition moment; A secondary ignition command is sent to the passive safety component based on the tolerance and the waveform intensity, wherein the tolerance and the waveform intensity are negatively correlated with the time interval between the second ignition time and the first ignition time.

2. The method according to claim 1, characterized in that, Calculating waveform intensity based on the historical scene parameters and the historical acceleration sequence includes: The historical scene parameters and the historical acceleration sequence are normalized and length-aligned to obtain the corresponding first preprocessed data and second preprocessed data. The first preprocessed data and the second preprocessed data are input into a pre-trained waveform feature predictor to obtain multiple waveform feature values. Each waveform feature value is used to characterize the collision intensity of the vehicle in the corresponding dimension. The multiple waveform feature values ​​include the mean acceleration, the peak acceleration time, the speed drop, and the acceleration change trend index. The waveform intensity is calculated using the aforementioned multiple waveform characteristic values.

3. The method according to claim 2, characterized in that, Calculating waveform intensity using the aforementioned multiple waveform feature values ​​includes: Calculate using the following formula : ； in, The average acceleration, At the peak of acceleration, To reduce speed, As an indicator of acceleration change trend, , , and For the corresponding reference value, , , , These are the corresponding weighting coefficients.

4. The method according to claim 2, characterized in that, Before inputting the first preprocessed data and the second preprocessed data into the pre-trained waveform feature predictor, the method further includes: Multiple sets of sample data are acquired, wherein each set of sample data includes a sample acceleration sequence and feature true values, wherein the feature true values ​​include the following true values: mean acceleration, peak acceleration time, velocity drop, and acceleration change trend index; The sample acceleration sequence is used to calculate feature prediction values, which include the following prediction values: mean acceleration, peak acceleration time, velocity drop, and acceleration change trend index. Construct the loss function; The loss function is trained using the predicted feature values ​​and the true feature values ​​to obtain the waveform feature predictor.

5. The method according to claim 4, characterized in that, The calculation of feature prediction values ​​using the sample acceleration sequence includes: The mean acceleration is calculated using the following formula. : ； in, This refers to the first ignition moment in the sample scenario. Let be the acceleration at time t. This is the preset effective collision duration.

6. The method according to claim 4, characterized in that, The calculation of feature prediction values ​​using the sample acceleration sequence includes: The peak acceleration time is calculated using the following formula. : ； in, This refers to the first ignition moment in the sample scenario. The preset effective collision duration, Let be the acceleration at time t.

7. The method according to claim 4, characterized in that, The calculation of feature prediction values ​​using the sample acceleration sequence includes: The speed drop is calculated using the following formula. : ； in, This refers to the first ignition moment in the sample scenario. This is the preset effective collision duration.

8. The method according to claim 4, characterized in that, The calculation of feature prediction values ​​using the sample acceleration sequence includes: The acceleration change trend index is calculated using the following formula. : ； in, This refers to the first ignition moment in the sample scenario. At the peak of acceleration, for acceleration at any moment for Acceleration at any moment.

9. The method according to claim 4, characterized in that, Constructing the loss function includes: The loss function is constructed using the following formula. : in, These are the predicted values ​​for the mean acceleration, peak acceleration time, velocity drop, and acceleration trend index, respectively. These represent the true values ​​of the mean acceleration, peak acceleration, velocity drop, and acceleration trend index, respectively. , , , These are the weight coefficients corresponding to the feature loss. The influence coefficient of physical constraints. These are hyperparameters used to control the sensitivity of physical constraints. It is a constant used to constrain the correlation between the acceleration change trend index and the mean acceleration.

10. The method according to claim 1, characterized in that, Obtaining the vehicle's historical acceleration sequence before the first ignition includes: The acceleration set of the vehicle at multiple consecutive moments is calculated in real time. Determine whether all of the acceleration sets are less than a preset threshold; If all of the acceleration sets are less than a preset threshold, the current moment is determined as the collision start moment; Obtain the historical acceleration sequence of the vehicle from the moment the collision begins to the moment the first ignition is initiated.

11. The method according to claim 1, characterized in that, Sending a secondary ignition command to the passive safety component based on the tolerance and the waveform intensity includes: The time interval between the second ignition time and the first ignition time is predicted based on the tolerance and the waveform intensity. The second ignition time is calculated based on the time interval and the first ignition time. Determine whether the second ignition time is within the effective duration of the current collision; If the second ignition time is not within the effective duration of the current collision, a corrected ignition time is calculated based on the preset ignition interval threshold, the first ignition time, and the second ignition time, and a second ignition command is sent to the passive safety component according to the corrected ignition time.

12. A vehicle safety control device, characterized in that, include: The first acquisition module is used to collect the physical attribute information of the occupants in the vehicle, determine the occupants' tolerance based on the physical attribute information, and after the passive safety components are ignited at the moment of the first ignition of the vehicle, acquire the historical scene parameters of the vehicle before the collision and acquire the historical acceleration sequence of the vehicle before the first ignition. The occupants' tolerance is used to characterize the occupants' ability to resist and withstand external risks. The first calculation module is used to calculate the waveform intensity based on the historical scene parameters and the historical acceleration sequence, wherein the waveform intensity is used to characterize the collision intensity of the vehicle after the first ignition moment; The sending module is used to send a secondary ignition command to the passive safety component based on the tolerance and the waveform intensity, wherein the tolerance and the waveform intensity are negatively correlated with the time interval between the second ignition time and the first ignition time.

13. A storage medium, characterized in that, The storage medium stores a computer program, wherein the computer program is configured to execute the method described in any one of claims 1 to 11 when it is run.

14. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to run the computer program to perform the method as described in any one of claims 1 to 11.

Images