Airbag control method, application, system and vehicle

Abstract

The application provides an airbag control method, application, system and vehicle; the airbag control method specifically comprises the following steps: randomly obtaining face data samples and collision data samples, and preprocessing and dividing the face data samples and the collision data samples into corresponding training sets and test sets; extracting user feature factors and collision feature factors according to the corresponding training sets, and constructing an airbag control model based on the user feature factors and the collision feature factors; then training the airbag control model through the training sets to obtain an airbag optimization control model; and further performing airbag control test on the test sets according to the airbag optimization control model to control the airbag to pop up under a preset condition. The application can improve the accuracy of user feature data identification, improve the stability, safety and reliability of airbag control, and has a relatively simple implementation manner and low cost.

Description

Technical Field

[0001] This application relates to the field of vehicle safety technology, and in particular to an airbag control method, application, system and vehicle. Background Technology

[0002] In vehicle safety systems, airbags are an important passive safety device, mainly used to cushion the impact force on passengers during a collision and protect them from injury.

[0003] Existing airbag control systems typically consider factors such as the user's facial features and facial turning. However, in practical applications, the acquisition of user facial feature data is often affected by factors such as lighting, occlusion, and facial expressions, leading to a decrease in the accuracy of facial feature recognition and posing a safety risk when controlling airbag deployment. Furthermore, existing airbag control systems are based on multi-source data, which requires the integration and analysis of different types of data sources. This necessitates complex data processing and fusion technologies, placing higher demands on technical developers and increasing development and implementation costs and difficulties to some extent. Summary of the Invention

[0004] To address the aforementioned technical problems, this application provides an airbag control method, application, system, and vehicle that can intelligently adjust airbag deployment based on user characteristic factors and collision characteristic factors, thereby maximizing user protection.

[0005] Specifically, this application provides an airbag control method, which includes the following steps:

[0006] S100: Randomly acquire facial data samples of various user states and collision data samples under various collision factors, and preprocess the facial data samples and collision data samples respectively; divide the preprocessed facial data samples into a first training set and a first test set according to a preset ratio, and divide the preprocessed collision data samples into a second training set and a second test set according to the same preset ratio.

[0007] S200: Extract user feature factors and collision feature factors according to the first training set and the second training set respectively, and construct an airbag control model based on the user feature factors and collision feature factors; train the airbag control model through the first training set and the second training set to obtain an optimized airbag control model.

[0008] S300: Perform airbag control tests on the first test set and the second test set according to the airbag optimization control model, so as to control the airbag deployment under preset conditions.

[0009] The airbag control method can construct an airbag control model based on user characteristic factors and collision characteristic factors, and then use the airbag control model to train the training set of collected facial data samples and collision data samples to obtain an optimized airbag control model that can intelligently adjust the airbag deployment.

[0010] The above method can obtain a more accurate airbag optimization control model for different user characteristics and collision characteristics, improve the accuracy of user characteristic data recognition, and improve the stability, safety and reliability of airbag control; moreover, the above airbag control method is relatively simple to implement and requires low cost.

[0011] The user state includes at least facial expression and facial orientation; the collision factors include at least collision intensity and collision direction.

[0012] Acquiring facial data samples under various user states and collision data samples under various collision factors makes the subsequent training results more representative and accurate.

[0013] The preprocessing in step S100 includes at least: data cleaning, missing value imputation, outlier and inconsistent value smoothing, and normalization.

[0014] The above preprocessing facilitates subsequent training and testing, improving the accuracy of the facial data samples and collision data samples. This, in turn, enhances the accuracy of the extracted user feature factors and collision feature factors, preventing poor training of the airbag control model due to data extraction errors, which could lead to accidental airbag activation.

[0015] The user feature factors include at least the user's facial key points, sitting height, body shape and posture, and head posture; the facial key points include at least the pupils of both eyes and the mouth.

[0016] Acquiring user characteristic factors can serve as a reference for subsequent airbag gas release and adjustment of seat headrest prediction. This intelligent setting can play a role in protecting users to a certain extent when the airbag is triggered.

[0017] The collision characteristic factors include at least the speed at the time of the collision and the angle of the steering wheel.

[0018] The speed and steering wheel angle at the time of a collision indicate the intensity and direction of the collision. By analyzing these two parameters, the current driving status of the vehicle can be obtained.

[0019] In step S200, obtaining the airbag optimized control model specifically includes:

[0020] The airbag control model is trained using any one of decision trees, support vector machines, and neural networks, and optimized in real time by combining user feature factors, collision feature factors, and application scenarios to obtain the final optimized airbag control model; the optimization includes at least cross-validation and grid search.

[0021] Optimizing the airbag control model in real time to obtain an optimized airbag control model can further improve the model's stability, safety, and reliability.

[0022] Following step S200, the method further includes:

[0023] The trained airbag optimization control model is integrated into the system, and the airbag gas release volume and seat headrest height adjustment are predicted based on the user's sitting height, body posture, and head posture.

[0024] The airbag gas release prediction is controlled based on user characteristic factors, enabling intelligent gas release according to user characteristic factors, thus making the airbag optimization control model more widely applicable; the prediction of adjusting the seat headrest height based on user characteristic factors allows the airbag deployment position to adapt to different users, thereby protecting users from injury to a certain extent.

[0025] Based on the same concept, this application also provides an application of an airbag control method for controlling an airbag located at the middle position of the top of a seat back, comprising the following steps:

[0026] S101: Real-time collection of current vehicle driving data and user characteristic data; the user characteristic data includes at least the user's pupils, mouth, sitting height, body shape and posture, and head posture.

[0027] S201: The driving data and user characteristic data are analyzed by the airbag optimization control model to control the airbag deployment under preset conditions.

[0028] The aforementioned application uses driving data and user characteristic data to achieve intelligent protection of users by airbags; and when this application is applied to neck airbags, it can provide strong support for the user's head and neck after the airbag deploys, thereby reducing the probability of head and neck injury in a collision; in addition, the aforementioned application also improves the passive safety of the vehicle.

[0029] It also includes: adjusting the seat headrest to the corresponding height based on the user characteristic data; and determining the user's facial orientation and head movement state based on the head posture, so as to determine the user's vital signs based on the pupils, mouth, facial orientation and head movement state.

[0030] Adjusting the headrest height based on user characteristic data allows the airbag deployment position to follow changes in user characteristic data, making airbag control more intelligent and protecting the user's head and neck from injury to a certain extent.

[0031] The system determines a user's vital signs based on partial data from user characteristic data, enabling detection of the user's condition in the event of a collision and ensuring the user's safety.

[0032] Based on the same concept, this application also provides an application system for an airbag control method, the system comprising:

[0033] Data acquisition module: used to collect real-time driving data of the current vehicle and user characteristic data.

[0034] Analysis module: Used to analyze the driving data and user characteristic data through the airbag optimization control model.

[0035] Control module: Used to control the airbag to deploy under preset conditions based on the analysis results of the analysis module.

[0036] The above system enables airbags to provide intelligent protection for users based on user characteristic data and driving data.

[0037] The analysis module includes:

[0038] Preprocessing unit: Used to preprocess the acquired facial data samples and collision data samples.

[0039] Feature extraction unit: used to extract user feature factors and collision feature factors based on the preprocessing results of the preprocessing unit.

[0040] Training unit: used to construct an airbag control model based on the feature extraction unit and to train the airbag control model.

[0041] Optimization unit: used to optimize the airbag control model to obtain an optimized airbag control model.

[0042] Based on the same concept, this application also provides a vehicle, including at least an airbag and an application system for implementing the airbag control method; wherein the airbag is disposed at the middle position of the top of the seat back.

[0043] Compared with the prior art, the beneficial effects of this application are as follows:

[0044] This application collects facial and collision data samples, divides them into training and testing sets, and then extracts user and collision feature factors from the training set. An airbag control model is then constructed and trained to obtain a final optimized airbag control model. After testing the optimized airbag control model on the testing set, the airbag is deployed under preset conditions. This application can improve the accuracy of user feature data recognition, and enhance the stability, safety, and reliability of airbag control. Furthermore, the implementation method is relatively simple and requires low cost. It solves the technical problems of low accuracy in user feature data recognition, complex data processing, high development difficulty, and high development costs in existing technologies. Attached Figure Description

[0045] Figure 1 This is a flowchart of the airbag control method described in this application.

[0046] Figure 2 for Figure 1 The application flowchart of the airbag control method is shown below.

[0047] Figure 3 for Figure 1 The application system framework diagram of the airbag control method described above. Detailed Implementation

[0048] This application provides an airbag control method, application, system, and vehicle to solve the technical problems of low accuracy in user feature data recognition, complex data processing, high development and implementation difficulty, and high development cost in the prior art.

[0049] This application provides an airbag control method, the general idea of ​​which is as follows:

[0050] Facial data samples from various user states and collision data samples under various collision factors are randomly acquired. These samples undergo data cleaning, missing value imputation, outlier and inconsistency value smoothing, and normalization. The processed facial data samples are then divided into a first training set and a first test set according to a preset ratio, and the processed collision data samples are divided into a second training set and a second test set according to the same preset ratio. User feature factors and collision feature factors are extracted from the first and second training sets, respectively, and an airbag control model is constructed based on these factors. The airbag control model is trained on the first and second training sets using any one of decision trees, support vector machines, and neural networks. The model is then optimized in real-time by incorporating user feature factors, collision feature factors, and application scenarios to obtain a final optimized airbag control model. Finally, airbag control tests are conducted on the first and second test sets based on the optimized airbag control model to control airbag deployment under preset conditions.

[0051] The following detailed description of an airbag control method, application, system, and vehicle according to this application, with reference to specific embodiments and accompanying drawings, provides further insight. Example 1:

[0052] Please see Figure 1 This application provides an airbag control method, which includes the following steps:

[0053] S100: Randomly acquire facial data samples of various user states and collision data samples under various collision factors, and preprocess the facial data samples and collision data samples respectively; divide the preprocessed facial data samples into a first training set and a first test set according to a preset ratio, and divide the preprocessed collision data samples into a second training set and a second test set according to the same preset ratio.

[0054] The user state includes at least facial expression and facial orientation; the collision factors include at least collision intensity and collision direction.

[0055] The facial data samples are acquired by camera sensors installed on the vehicle's central control panel and at the connection point between the top of the seat back and the headrest; typically, they collect facial image data of different users in the vehicle with various facial expressions and orientations; in this embodiment, the camera sensor is a 360-degree surround view camera, which can collect facial image data over a large range.

[0056] It should be noted that the first training set is a collection of facial image data.

[0057] The collision data samples are collected through vehicle body self-sensing devices, which collect at least the current vehicle operating status, surrounding environment information, driver habits, and steering wheel information; the vehicle body self-sensing sensors are usually installed on the steering wheel and / or at any position on the vehicle body.

[0058] It should be noted that the current vehicle operating status includes, for example, vehicle speed and acceleration; the surrounding environment information includes, for example, obstacle information and road conditions; the driver's habits include at least the cockpit configuration habits; and the steering wheel information includes at least the steering wheel angle.

[0059] The preprocessing in step S100 includes at least: data cleaning, missing value imputation, outlier and inconsistent value smoothing, and normalization.

[0060] After completing the above preprocessing, the preprocessed facial data samples and collision data samples are divided into test sets and training sets according to a ratio of, for example, 8:2. Feature factors are extracted and used as input for machine learning algorithms. Step S200 is then executed.

[0061] S200: Extract user feature factors and collision feature factors according to the first training set and the second training set respectively, and construct an airbag control model based on the user feature factors and collision feature factors; train the airbag control model through the first training set and the second training set to obtain an optimized airbag control model.

[0062] A deep learning-based convolutional neural network extracts user feature factors from a first training set of collected facial data samples. This includes: inputting the first training set (a collection of facial image data), extracting user feature factors through convolution operations, then performing a sliding window operation on the corresponding facial image data using a filter, and weighting and summing the pixel values ​​of each window with the weights in the convolution kernel to obtain the output result. The user feature factors include at least the user's facial key points, sitting height, body posture, and head posture. The facial key points include at least the pupils of both eyes and the mouth.

[0063] Furthermore, pooling operations can be used to reduce the resolution of user feature factors, thereby reducing the number of parameters and improving computational efficiency.

[0064] A preset processor extracts collision feature factors from the collected collision data samples; the collision feature factors include at least the speed at the time of the collision and the angle of the steering wheel.

[0065] It should be noted that the type of processor can be selected by those skilled in the art based on the actual application, and no limitation is made here.

[0066] Specifically, obtaining the airbag optimized control model includes:

[0067] The airbag control model is trained using any one of decision trees, support vector machines, and neural networks, and optimized in real time by combining user feature factors, collision feature factors, and application scenarios to obtain the final optimized airbag control model; the optimization includes at least cross-validation and grid search.

[0068] The above process enables automatic learning of the relationships between various features to determine when to trigger the airbag. To improve the generalization ability and testing accuracy of the algorithm, techniques such as cross-validation and grid search can be used for optimization. In addition, those skilled in the art can adaptively select other optimization algorithms for optimization.

[0069] Ensemble learning methods can also be used to improve the robustness and generalization ability of airbag control models, such as random forests and AdaBoost. In addition to ensemble learning, pruning and feature selection methods can also be used.

[0070] Through the above training and optimization, the final airbag optimization control model is obtained. For example, when the speed is greater than or equal to the first speed threshold and less than the second speed threshold, and / or the angle is greater than or equal to the first angle threshold and less than the second angle threshold, it is considered that the current vehicle has only experienced a minor collision and there is no need to trigger the airbag inflation and deployment. When the speed is greater than or equal to the second speed threshold, and / or the angle is greater than or equal to the second angle threshold, it is considered that the vehicle collision is more serious and there is a risk of injury to the occupants. In this case, it is necessary to trigger the airbag inflation and deployment to reduce the degree of injury to the occupants.

[0071] Following step S200, the method further includes:

[0072] The trained airbag optimization control model is integrated into the system, and the airbag gas release volume and seat headrest height adjustment are predicted based on the user's sitting height, body posture, and head posture.

[0073] The integration method is not specified here.

[0074] Three different algorithms are used to predict the airbag gas release volume and seat headrest height based on the user's sitting height, body posture, and head posture. The prediction results of the three algorithms are then integrated to obtain a comprehensive prediction result, enabling subsequent control of airbag gas release and seat headrest height adjustment based on the user's sitting height, body posture, and head posture.

[0075] S300: Perform airbag control tests on the first test set and the second test set according to the airbag optimization control model, so as to control the airbag deployment under preset conditions.

[0076] The reserved test data, namely the first test set and the second test set, are used to evaluate the accuracy, recall, and F1 score of the airbag optimized control model.

[0077] The preset conditions include, for example, a speed greater than or equal to a second speed threshold, and / or an angle greater than or equal to a second angle threshold. When the current vehicle meets these preset conditions, the relevant controller controls the airbag to deploy. Example 2:

[0078] Please see Figure 2 This application also provides an application of an airbag control method for controlling an airbag located at the center of the top of a seat back, comprising the following steps:

[0079] S101: Real-time collection of current vehicle driving data and user characteristic data.

[0080] The user feature data includes at least the user's pupils, mouth, sitting height, body shape and posture, and head posture.

[0081] S201: The driving data and user characteristic data are analyzed by the airbag optimization control model to control the airbag deployment under preset conditions.

[0082] After the user is seated, the airbag optimization control model can adjust the headrest of the seat to the corresponding height based on the user's characteristic data.

[0083] The headrest height is adjusted based on the user's sitting height and head posture to align the center of the headrest with the center of the back of the user's head. Adjusting the headrest height is achieved by adjusting the distance between the seat back and the headrest. This process is completed automatically after the user sits down.

[0084] Throughout the user's time inside the vehicle, the user's facial orientation and head movement can be determined based on the head posture, and the user's vital signs can be determined based on the pupils, mouth, facial orientation, and head movement.

[0085] For example, if the vital signs state is unconscious, those skilled in the art can also set up an alarm system. When the user's vital signs state is detected to be unconscious or in a dangerous state, the alarm system can be activated to make an emergency call and / or turn on the warning lights.

[0086] The airbag optimization control model determines the current driving status of the vehicle based on the driving data, which includes at least the current vehicle speed and steering wheel angle.

[0087] In one feasible implementation, when the speed is greater than or equal to a first speed threshold and less than a second speed threshold, and / or the angle is greater than or equal to a first angle threshold and less than a second angle threshold, it is considered that the vehicle has only experienced a minor collision and there is no need to trigger the airbag deployment; however, when the speed is greater than or equal to the second speed threshold, and / or the angle is greater than or equal to the second angle threshold, it is considered that the vehicle collision is more serious and there is a risk of injury to the occupants, in which case it is necessary to trigger the airbag deployment to reduce the degree of injury to the occupants.

[0088] When the airbag needs to be deployed, the airbag deployment position can effectively protect the user's head and neck because the seat height has been adjusted after the user sits down. Example 3:

[0089] Please see Figure 3 This application also provides an application system for an airbag control method, the system comprising:

[0090] Data acquisition module: used to collect real-time driving data of the current vehicle and user characteristic data.

[0091] In the data acquisition module, the vehicle's current driving data is collected through the vehicle's self-sensing device, and user feature data is collected through the camera sensor; the user feature data includes at least the user's pupils, mouth, sitting height, body shape and posture, and head posture.

[0092] The vehicle body self-sensing sensor is typically installed on the steering wheel and / or at any location on the vehicle body; the camera sensor is installed on the vehicle's central control panel and at the connection point between the top of the seat back and the seat headrest.

[0093] Analysis module: Used to analyze the driving data and user characteristic data through the airbag optimization control model.

[0094] The analysis module includes:

[0095] Preprocessing unit: Used to preprocess the acquired facial data samples and collision data samples.

[0096] Feature extraction unit: used to extract user feature factors and collision feature factors based on the preprocessing results of the preprocessing unit.

[0097] Training unit: used to construct an airbag control model based on the feature extraction unit and to train the airbag control model.

[0098] Optimization unit: used to optimize the airbag control model to obtain an optimized airbag control model.

[0099] The functions of the preprocessing unit, feature extraction unit, training unit, and optimization unit are all implemented through computer programs.

[0100] Control module: Used to control the airbag to deploy under preset conditions based on the analysis results of the analysis module.

[0101] The control module is equivalent to a controller or airbag control unit. It determines whether the current vehicle's driving data meets preset conditions based on the analysis results, and controls the airbag to deploy based on the user characteristic data.

[0102] When the controller or airbag controller determines that the airbag needs to be inflated, it sends a signal to the gas generator to trigger the gas generator to release gas; then it controls the airbag to inflate for a first preset time and pop out, or controls the airbag to be fully inflated and pop out.

[0103] It should be noted that when a vehicle is involved in a collision, a person may be thrown from their seat and endangered in just 0.09 seconds. This 0.09 seconds is equivalent to the blink of an eye. Therefore, for an airbag to protect the human body, it must successfully deploy within this 0.09-second time.

[0104] In this embodiment, the gas generator uses a single-stage gas generator or a bipolar solid gas generator to release gas. The gas enters the airbag through a transmission pipe and completes inflation and ejection within 30-35ms.

[0105] It should also be noted that the airbag is a multi-layered inflatable neck airbag, which is installed in the middle of the top of the seat back. When the neck airbag deploys, it deploys diagonally upwards to maximize the protection of the user's head and neck.

[0106] After the airbag is deployed, the airbag releases gas after a second preset time and a third preset time, or releases a preset volume of gas, based on the user's characteristic data.

[0107] The second preset duration, the third preset duration, and the preset volume all follow the changes in body shape and posture in the user's feature data. The body shape and posture reflect the user's body width. Users with different body widths experience different impact forces. In order to prevent the airbag from causing further damage to the user, a certain volume of gas needs to be released after the airbag is fully inflated. Taking user A's body width as L1 and user B's body width as L2 as an example, where L1 is greater than L2, user A's airbag releases gas faster than user B's, and the volume of gas released is also greater than user B's.

[0108] It should be noted that after the airbag inflates and deploys, the seat belt will also play its role in protecting the user from being trapped or injured by the airbag; only when the airbag and seat belt work together can the user's personal safety be protected to the greatest extent. Example 4:

[0109] This application also provides a vehicle that includes at least an airbag and an application system for implementing the airbag control method; wherein the airbag is disposed at the middle position of the top of the seat back.

[0110] In summary, this application provides an airbag control method, application, system, and vehicle. It involves collecting facial data samples and collision data samples, preprocessing them to obtain corresponding training and testing sets, extracting user feature factors and collision feature factors from the training set, constructing an airbag control model, and then training the airbag control model using the training set to obtain an optimized airbag control model. Finally, the optimized airbag control model is tested using the testing set to control airbag deployment under preset conditions. This application can improve the accuracy of user feature data recognition, enhance the stability, safety, and reliability of airbag control, and is relatively simple to implement with low cost.

[0111] Although exemplary embodiments have been described herein with reference to the accompanying drawings, it should be understood that the above exemplary embodiments are merely illustrative and are not intended to limit the scope of this application. Various changes and modifications can be made therein by those skilled in the art without departing from the scope and spirit of this application. All such changes and modifications are intended to be included within the scope of this application as claimed in the appended claims.

[0112] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0113] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed.

[0114] The various component embodiments of this application can be implemented in hardware, or as software modules running on one or more processors, or a combination thereof. Those skilled in the art will understand that microprocessors or digital signal processors (DSPs) can be used in practice to implement some or all of the functions of some modules according to the embodiments of this application. This application can also be implemented as an apparatus program (e.g., a computer program and computer program product) for performing part or all of the methods described herein. Such an implementation of this application can be stored on a computer-readable medium, or can be in the form of one or more signals. Such signals can be downloaded from an Internet website, provided on a carrier signal, or provided in any other form.

[0115] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0116] Although the description of this application has been made in conjunction with the specific embodiments described above, it will be apparent to those skilled in the art that many substitutions, modifications, and variations can be made based on the foregoing. Therefore, all such substitutions, modifications, and variations are included within the spirit and scope of the appended claims.

Claims

1. An airbag control method characterized by, The airbag control method comprises the following steps: S100: randomly obtaining face data samples of various user states and collision data samples under various collision factors, and respectively preprocessing the face data samples and the collision data samples; dividing the preprocessed face data samples into a first training set and a first test set according to a preset ratio, and dividing the preprocessed collision data samples into a second training set and a second test set according to the same preset ratio; S200: extracting user feature factors and collision feature factors according to the first training set and the second training set respectively, and constructing an airbag control model based on the user feature factors and the collision feature factors; training the airbag control model through the first training set and the second training set to obtain an airbag optimized control model; S300: performing airbag control test on the first test set and the second test set according to the airbag optimized control model to control the airbag to pop up under a preset condition.

2. The airbag control method according to claim 1 characterized by, The user state at least includes facial expression and facial orientation; The collision factor at least includes collision intensity and collision direction; The preprocessing in the step S100 at least includes data cleaning, missing value filling, abnormal value and inconsistent value data smoothing, and normalization processing.

3. The airbag control method according to claim 1, characterized by, The user feature factor at least includes user's facial key points, sitting height, body shape and head posture; The facial key points at least include both eyes and mouth; The collision feature factor at least includes speed and steering wheel angle when the collision occurs.

4. The airbag control method according to claim 3, characterized by, In the step S200, obtaining the airbag optimized control model specifically comprises: training the airbag control model through any one of decision tree, support vector machine and neural network, and combining user feature factors and collision feature factors in real time, and optimizing the airbag control model according to application scenarios to obtain the final airbag optimized control model; the optimization at least includes cross-validation and grid search.

5. The airbag control method according to claim 4, characterized by, After the step S200, it further comprises: integrating the trained airbag optimized control model into the system, and predicting the airbag gas release volume and the seat headrest height adjustment according to the user's sitting height, body shape and head posture.

6. An application system for implementing the airbag control method according to any one of claims 1 to 5, characterized by The system is applied to control the airbag arranged at the middle position of the top of the seat back, and the system comprises: A data acquisition module is configured to acquire driving data of a current vehicle and user feature data in real time; the user feature data at least includes both eyes, mouth, sitting height, body shape and head posture of the user; An analysis module is configured to analyze the driving data and the user feature data through the airbag optimized control model; and is further configured to determine the facial orientation and head movement state of the user according to the head posture, and determine the physical state of the user according to the both eyes, mouth, facial orientation and head movement state; A control module is configured to control the airbag to pop up under a preset condition according to the analysis result of the analysis module; and is further configured to adjust the seat headrest to a corresponding height according to the user feature data.

7. The system of claim 6, wherein, The analysis module comprises a preprocessing unit configured to preprocess the obtained face data samples and collision data samples; The feature extraction unit is configured to extract user feature factors and collision feature factors according to the preprocessing result of the preprocessing unit. The training unit is configured to construct an airbag control model based on the feature extraction unit and train the airbag control model. The optimization unit is configured to optimize the airbag control model to obtain an optimized airbag control model.

8. A vehicle characterized by comprising: An application system comprising at least an airbag and the airbag control method according to claim 6. The airbag is arranged at a middle position on the top of the seat back.

Images