Automotive airbag control method and system

Abstract

The application provides a kind of automobile safety air bag control method and system, wherein the method comprises: according to the driving data of target vehicle, the personnel data of the people in the vehicle and the air bag data of target vehicle, collision twin is carried out, and the personnel risk degree of air bag control of different pop-up logic is determined;The pre-control instruction of air bag control of pop-up logic corresponding to the minimum personnel risk degree is generated;According to the posterior data of collision twin, collision verification is carried out, and if the verification is passed, the pre-control instruction is executed;If the verification is not passed, according to the actual collision data and the preset air bag control instruction library, the target air bag control instruction is determined and corresponding control is carried out.The application provides a kind of automobile safety air bag control method and system, according to the time window suitable for the personnel risk degree evaluation of two ways of database matching and collision twin and the collision verification result, the final air bag control strategy is determined, and the suitability of air bag control is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automobile airbag control, and in particular to an automobile airbag control method and system. BACKGROUND

[0002] Generally, when a vehicle is subjected to a collision, the airbag deployment logic is as follows: first, the sensor collects data; then, the vehicle system detects and classifies the collision according to the data collection results; finally, according to the collision detection and classification results, it is determined whether the collision condition reaches the triggering threshold of the automobile airbag, and if so, the airbag is deployed. Although the time window of the above-mentioned automobile airbag deployment is relatively short, it relies on the preset threshold rule and simple classification output instruction, which is easy to cause inappropriate deployment, such as: according to the analysis of the collision detection results, it is concluded that the airbag deployment degree of the steering wheel airbag should be the largest, but the driving personnel's sitting posture is not appropriate, which leads to more serious injuries of the driving personnel under the airbag control decision of the airbag with the largest deployment degree than under the airbag control decision of the airbag with the secondary deployment degree.

[0003] Therefore, there is an urgent need for an automobile airbag control method and system to at least solve the above problems. SUMMARY

[0004] One of the purposes of the present application is to provide an automobile airbag control method and system, which performs collision twin of the collision condition of the personnel in the vehicle within the fuzzy perception time, pre-calculates and generates pre-control instructions; introduces posterior data to determine whether the subsequent corresponding collision situation actually occurs, if so, the pre-control instructions are directly triggered, otherwise, the target airbag control instructions matched in the pre-set database are triggered according to the actual collision data, and the final airbag control strategy is determined according to the time window and the collision verification result of the two ways of database matching and personnel risk evaluation of collision twin, thereby improving the suitability of airbag control.

[0005] The automobile airbag control method provided by the embodiment of the present application comprises:

[0006] Performing collision twin according to the driving data of the target vehicle, the personnel data of the personnel in the vehicle and the airbag data of the target vehicle, and determining the personnel risk degree of the airbag control of different deployment logics;

[0007] Generating the pre-control instructions of the airbag control of the deployment logic corresponding to the minimum personnel risk degree;

[0008] Performing collision verification according to the posterior data of the collision twin, and executing the pre-control instructions if the verification is passed;

[0009] If the verification is not passed, determining the target airbag control instructions according to the actual collision data and the pre-set airbag control instruction library, and performing corresponding control.

[0010] Preferably, the driving data of the target vehicle comprises surrounding environment information of the target vehicle and self-state of the target vehicle.

[0011] Preferably, the personnel data of the in-vehicle personnel comprises personnel size, personnel seat position, personnel sitting posture and safety belt wearing condition.

[0012] Preferably, the airbag data comprises airbag distribution and airbag achievable deployment state.

[0013] Preferably, the collision twin is performed according to the driving data of the target vehicle, the personnel data of the in-vehicle personnel and the airbag data of the target vehicle to determine the personnel danger degree of the airbag control of different deployment logics, comprising:

[0014] determining whether a collision will occur within the fuzzy perception duration according to the driving data of the target vehicle;

[0015] predicting the personnel trajectory of the in-vehicle personnel after being impacted by the collision according to the collision situation and the personnel data of the in-vehicle personnel at the current time;

[0016] pre-acting the post-collision state of the in-vehicle personnel according to the personnel trajectory and the airbag control of different deployment logics;

[0017] determining the personnel danger degree according to the preset personnel danger degree evaluation template and the post-collision state.

[0018] Preferably, the fuzzy perception duration is obtained by the following steps:

[0019] extracting collision derivation factor values according to the driving data;

[0020] determining derivation degrees corresponding to the collision factor value set according to the collision derivation factor values and input values of the collision model;

[0021] obtaining a preselected fuzzy perception duration of a derivation target corresponding to the collision factor value set whose derivation degree is greater than or equal to a preset derivation degree threshold value;

[0022] adjusting the preselected fuzzy perception duration according to the fatigue monitoring result of the driver to obtain the fuzzy perception duration.

[0023] Preferably, the fatigue monitoring result is obtained by the following steps:

[0024] intervening in collecting in-vehicle conversations at the route deviation time of the recommended route and the actual route;

[0025] normalizing and summing the fatigue values corresponding to the fatigue monitoring video stream, the driving duration and the in-vehicle conversation semantics to obtain the fatigue monitoring result.

[0026] Preferably, the fatigue monitoring video stream corresponds to the fatigue value represented by the following steps:

[0027] According to the fatigue monitoring video stream, the blinking frequency change, the eye opening degree change and the head following swing interval change during the steering of the driver are extracted;

[0028] According to the blinking frequency change, the eye opening degree change and the head swing interval change during the steering, the fatigue value represented by the fatigue monitoring video stream is quantified.

[0029] Preferably, the fatigue value represented by the in-vehicle conversation semantics corresponds to the following steps:

[0030] The in-vehicle conversation semantics is unfolded on a preset time axis according to the corresponding extraction time;

[0031] The route change questioning semantics on the time axis is matched, and if the matching is successful, the in-vehicle conversation semantics of another conversation party after the route change questioning semantics is taken as a questioning reply;

[0032] The questioning reply is input into a preset reasonable route change judgment model, and a reasonable value is output;

[0033] According to the reasonable value and a preset fatigue value reading table, the fatigue value corresponding to the reasonable value is read.

[0034] The automobile safety airbag control system provided by the embodiment of the application comprises:

[0035] A collision twin module is configured to perform collision twinning according to driving data of a target vehicle, personnel data of a person in the vehicle and airbag data of the target vehicle, and determine a personnel danger degree of airbag control of different ejection logics;

[0036] A pre-control instruction generation module is configured to generate a pre-control instruction of airbag control of the ejection logic corresponding to the minimum personnel danger degree;

[0037] A first control module is configured to perform collision verification according to posterior data of the collision twinning, and execute the pre-control instruction if the verification is passed;

[0038] A second control module is configured to determine a target airbag control instruction and perform corresponding control according to actual collision data and a preset airbag control instruction library if the verification is not passed.

[0039] The automobile safety airbag control system provided by the embodiment of the application has the following beneficial effects:

[0040] The application carries out collision twin of collision situation of people in the vehicle within the fuzzy perception duration, and generates pre-control instructions in advance; the posterior data is introduced to judge whether the corresponding subsequent collision situation actually occurs, if yes, the pre-control instructions are directly triggered, otherwise, the target airbag control instructions matched in the pre-set database are triggered according to the actual collision data, the final airbag control strategy is determined according to the time window and the collision verification result of the two ways of database matching and collision twin personnel risk evaluation, and the suitability of the airbag control is improved.

[0041] Other features and advantages of the present application will be set forth in the following description, and in part will become apparent to those skilled in the art upon examination of the following or can be learned by practice of the present application. The objects and other advantages of the present application can be realized and attained by the structure particularly pointed out in the written description and claims hereof.

[0042] The technical solutions of the present application are described in detail below with the help of the accompanying drawings and examples. BRIEF DESCRIPTION OF DRAWINGS

[0043] The accompanying drawings are included to provide a further understanding of the present application, and constitute a part of the specification, illustrate embodiments of the present application and explain the present application together with the embodiments, and do not constitute a limitation on the present application. In the drawings:

[0044] Figure 1 It is a schematic diagram of an automobile safety airbag control method in the embodiment of the present application;

[0045] Figure 2 It is a schematic diagram of an automobile safety airbag control system in the embodiment of the present application. DETAILED DESCRIPTION

[0046] The preferred embodiments of the present application are described below in combination with the accompanying drawings, and it should be understood that the preferred embodiments described herein are only used to illustrate and explain the present application, and do not limit the present application.

[0047] The embodiment of the present application provides an automobile safety airbag control method, as shown in Figure 1 , comprising:

[0048] Step 1: according to the driving data of the target vehicle, the personnel data of the people in the vehicle and the airbag data of the target vehicle, the collision twin is carried out, and the personnel risk degree of the airbag control of different pop-out logic is determined;

[0049] The target vehicle is a vehicle that needs to be controlled by an airbag; the driving data includes: the surrounding environment information of the target vehicle and the state of the target vehicle itself, the surrounding environment information includes: the relative position relationship between the static object and the target vehicle, the three-dimensional data of the static object, the relative position relationship between the dynamic object and the target vehicle, the three-dimensional data of the dynamic object, the motion speed and direction of the dynamic object, etc., the state of the target vehicle itself is: the three-dimensional data of the target vehicle, the motion speed and direction of the target vehicle; the personnel data is: the body size, seat position, sitting posture and safety belt wearing situation of the personnel; the airbag data is: the airbag distribution and the airbag pop-up state that can be realized;

[0050] The collision twin is carried out according to the driving data of the target vehicle, the personnel data of the personnel in the vehicle and the airbag data of the target vehicle, and the personnel danger degree of the airbag control of different pop-out logic is determined, including:

[0051] According to the driving data of the target vehicle, it is judged whether a collision will occur within the fuzzy perception time;

[0052] Wherein, when it is judged whether a collision will occur, the existing collision model is used to judge according to the surrounding environment information of the target vehicle and the state of the target vehicle itself; the fuzzy perception time is: the time length of the driver's perception of danger and reaction, such as: 2 seconds;

[0053] According to the collision situation and the personnel data of the personnel in the vehicle at the current time, the personnel trajectory of the personnel in the vehicle after the collision impact is predicted;

[0054] Wherein, the collision situation is the collision situation derived by the above collision model, including: the collision position and collision force of different collisions, different collisions refer to all collisions from the first collision derived to the process of the target vehicle completely stopping; the personnel data of the personnel in the vehicle at the current time refers to the body size, seat position, sitting posture and safety belt wearing situation of the personnel in the vehicle at the current time; the personnel trajectory is predicted according to the trajectory prediction model, and the trajectory prediction model is obtained by using machine learning model to learn historical vehicle personnel collision data (annotating historical collision situation, historical personnel data before collision and historical personnel trajectory), the safety airbag in the historical vehicle personnel collision data is not triggered;

[0055] According to the personnel trajectory and the airbag control of different pop-out logic, the post-collision state of the personnel in the vehicle is previewed;

[0056] The airbag control of different ejection logics refers to the ejection time and ejection degree of the airbag at different airbag distribution positions; the post-collision state of the pre-acted vehicle occupant refers to the digital twin of the corresponding vehicle interior situation according to the predicted vehicle occupant trajectory after the occupant is impacted by the collision and the airbag control of different ejection logics, and the state of the post-collision occupant trajectory being blocked by the ejected airbag (post-collision state);

[0057] According to the preset occupant danger degree evaluation template and the post-collision state, the occupant danger degree is determined.

[0058] The preset occupant danger degree evaluation template is a standardized evaluation system for quantifying the injury risk of the vehicle occupant, such as calculating the corresponding danger degree evaluation parameters (head acceleration, chest compression, and neck torque) according to the movement trajectory of each part of the human body of the vehicle occupant in the post-collision state, multiplying the danger degree evaluation parameters by the corresponding preset normalized weights and summing them up to obtain the occupant danger value of the vehicle occupant, and taking the average value of the occupant danger value as the occupant danger degree.

[0059] Step 2: Generate pre-control instructions of the airbag control of the ejection logic corresponding to the minimum occupant danger degree;

[0060] The pre-control instructions are the airbag control instructions that may be executed according to the collision pre-acting;

[0061] Step 3: Perform collision verification according to the posterior data of the collision twin, and if the verification passes, execute the pre-control instructions;

[0062] The posterior data are related data for verifying whether the collision derived based on the collision model actually occurs, such as the collision sensor data of the target vehicle within twice the fuzzy perception time; the verification passing means that the collision situation obtained based on the posterior data analysis is consistent with the collision situation derived based on the collision model;

[0063] Step 4: If the verification does not pass, determine the target airbag control instructions according to the actual collision data and the preset airbag control instruction library and perform corresponding control.

[0064] The verification not passing means that the collision situation obtained based on the posterior data analysis is inconsistent with the collision situation derived based on the collision model, or it is determined based on the posterior data analysis that no collision has occurred; the actual collision data are the collision sensor data when the driver still performs vehicle operation after perceiving danger and a collision occurs; the preset airbag control instruction library stores one-to-one corresponding collision sensor data and airbag control instructions, and when the actual collision data and a certain set of collision sensor data in the airbag control instruction library are consistent, the corresponding associated airbag control instructions are taken as the target airbag control instructions.

[0065] The working principle and beneficial effects of the above technical solutions are:

[0066] The present application carries out collision twin of the collision situation of the in-vehicle personnel within the fuzzy perception duration according to the driving data of the target vehicle, the personnel data of the in-vehicle personnel and the airbag data of the target vehicle, and pre-calculates to generate a pre-control instruction.

[0067] In actual car collision accidents, the driver may not be able to perceive the danger and make a decision in time within the fuzzy perception duration, so when the collision verification passes, the pre-control instruction is directly triggered, which overcomes the problem that the time window of determining the airbag control instruction is too long after the actual collision, resulting in that the instruction cannot be determined and triggered in time. In addition, compared with directly matching the target airbag control instruction according to the actual collision data, the pre-control instruction does not excessively rely on the preset threshold rule and simple classification output instruction, and the airbag control instruction is more suitable. If the driver can perceive the danger and make a decision in time within the fuzzy perception duration, the decision result may be successful risk avoidance, or the collision may still occur. If the risk avoidance is successful, the airbag is not triggered, otherwise, the target airbag control instruction is determined based on the airbag control instruction library.

[0068] The present application carries out collision twin of the collision situation of the in-vehicle personnel within the fuzzy perception duration, and pre-calculates to generate a pre-control instruction. The posterior data is introduced to judge whether the subsequent corresponding collision situation actually occurs. If yes, the pre-control instruction is directly triggered, otherwise, the target airbag control instruction matched in the pre-set database is triggered according to the actual collision data. The final airbag control strategy is determined according to the time window and the collision verification result of the two ways of database matching and personnel risk evaluation of collision twin, which improves the suitability of airbag control.

[0069] In one embodiment, the fuzzy perception duration is obtained as follows:

[0070] According to the driving data, the collision derivation factor value is extracted;

[0071] The collision derivation factor value is a quantitative index related to the potential collision risk extracted from the driving data, such as the following distance when following, the lane deviation standard deviation, etc.

[0072] According to the collision derivation factor value and the input value of the collision model, the derivation degree corresponding to the collision factor value set is determined.

[0073] The input value of the collision model is the standardized parameter of the input parameter required for the collision model to perform collision derivation, such as when the following distance reaches a certain value and the lane deviation standard deviation reaches a certain value, the collision model can be input to perform collision derivation. The derivation degree is the inverse of the average value of the deviation degree of the collision derivation factor value and the input value of the corresponding factor type.

[0074] obtaining a fuzzy perception time length according to the preselected fuzzy perception time length of the derivation target of the corresponding collision factor value set with the derivation degree greater than or equal to the preset derivation degree threshold and the fatigue monitoring result of the driver;

[0075] The preset derivation degree threshold is set by a person in advance. The derivation target of the corresponding collision factor value set refers to a collision risk scenario described by the collision factor value set together, such as a rear-end collision risk scenario with a preceding vehicle or a risk scenario of collision with a roadside flower bed. The preselected fuzzy perception time length of the derivation target is a time preset for the driver to perceive and react to a collision accident type corresponding to the derivation target, such as a preselected fuzzy perception time length of 1 second for a collision accident type of collision with a flower bed and a preselected fuzzy perception time length of 2 seconds for a collision accident type of rear-end collision.

[0076] The preselected fuzzy perception time length is increased according to the fatigue monitoring result of the driver, and the fuzzy perception time length is obtained.

[0077] The fatigue monitoring result is a perception result of the driving state of the driver in the vehicle based on a multi-modal perception device such as a camera preset on a vehicle A-pillar. When the driver is tired, the reaction time will be longer, and therefore, the preselected fuzzy perception time length is extended to obtain the fuzzy perception time length.

[0078] The working principle and beneficial effects of the above technical solution are as follows:

[0079] The reflection of the driver on the collision risk is disturbed by the driving state of the driver. Long-distance driving will cause the driver to be slow in reaction and unable to avoid danger in time. Therefore, a quantitative index related to a potential collision risk is extracted from driving data, compared with an input value of a collision model, and a derivation degree corresponding to a collision factor value set is calculated. The fuzzy perception time length is determined according to the preselected fuzzy perception time length of the derivation target of the corresponding collision factor value set with the derivation degree greater than or equal to the preset derivation degree threshold and the fatigue monitoring result of the driver, and the suitability of the derivation time length setting of subsequent collision derivation is improved.

[0080] In one embodiment, the fatigue monitoring result is obtained as follows:

[0081] The in-vehicle conversation is collected at a moment of route deviation of the recommended route and the actual route;

[0082] The recommended route is a navigation route of the vehicle machine. The actual route is a real-time route of the vehicle. When the recommended route deviates from the actual route, the driver may be distracted and miss an intersection, and the deviation moment is used as a collection time of the in-vehicle conversation;

[0083] The fatigue monitoring result is obtained by normalizing and summing the fatigue monitoring video stream, the driving time length, and the fatigue value corresponding to the in-vehicle conversation semantics.

[0084] Wherein, the more obvious the fatigue behavior of the driver analyzed from the image in the fatigue monitoring video stream, the greater the fatigue value corresponding to the representation; the longer the driving time, the greater the fatigue value corresponding to the representation; the greater the fatigue confirmation degree of the in-vehicle conversation semantics corresponding to the discussion content on the driver, the greater the fatigue value corresponding to the representation; the normalization rule is set by an artificial beforehand.

[0085] The working principle and beneficial effects of the above technical solution are:

[0086] The present application introduces fatigue monitoring video stream, driving time and in-vehicle conversation semantics at the moment of route deviation to determine the fatigue monitoring result, which is more comprehensive.

[0087] In one embodiment, the fatigue value corresponding to the representation of the fatigue monitoring video stream is obtained as follows:

[0088] According to the fatigue monitoring video stream, the blinking frequency change, the eye opening degree change and the following time interval change of the head swing when turning of the driver are extracted;

[0089] Wherein, the blinking frequency change is the change of the number of blinks per minute; the eye opening degree change is the change of the eye exposure area per minute; the following time interval of the head swing when turning is the time interval of the head closely turning to check the rearview mirror when the vehicle body turns;

[0090] According to the blinking frequency change, the eye opening degree change and the head swing time interval change when turning, the fatigue value represented by the fatigue monitoring video stream is quantified.

[0091] Wherein, the greater the blinking frequency increment, the smaller the eye opening degree increment and the greater the following time interval increment, the greater the fatigue value represented by the fatigue monitoring video stream, and the specific quantitative proportional relationship is set by an artificial beforehand.

[0092] The working principle and beneficial effects of the above technical solution are:

[0093] When determining the fatigue value represented by the fatigue monitoring video stream, the present application extracts the blinking frequency change, the eye opening degree change and the following time interval change of the head swing when turning of the driver from the fatigue monitoring video stream to quantify the fatigue value represented by the fatigue monitoring video stream, which improves the comprehensiveness of the representation basis and the representation accuracy.

[0094] In one embodiment, the fatigue value corresponding to the representation of the in-vehicle conversation semantics is obtained as follows:

[0095] The in-vehicle conversation semantics is unfolded on the preset time axis according to the corresponding extraction time;

[0096] When the in-vehicle dialogue semantics are unfolded on the timeline, they are unfolded in the order of the extraction time corresponding to the in-vehicle dialogue semantics; each point on the preset timeline corresponds to a specific moment.

[0097] Match the route change challenge semantics on the timeline. If the match is successful, use the in-car conversation semantics of the other party after the route change challenge semantics as the challenge response.

[0098] Among them, the semantics of questioning route changes are: semantics that question why the route has been changed, such as: "Did we take the wrong turn?" or "Shouldn't we have just gotten off at that intersection?";

[0099] Input the response to the question into the preset reasonable route change judgment model, and output a reasonable value;

[0100] The preset reasonable route change judgment model is an AI model that judges whether the route change is reasonable. It will combine real-time traffic conditions and responses to questions to judge the reasonableness (reasonable value) of the route change.

[0101] Based on the reasonable value and a preset fatigue value reading table, the fatigue value corresponding to the reasonable value is read. The preset fatigue value reading table stores the conversion relationship between the reasonable value and the fatigue value represented by the semantics of the in-vehicle dialogue. The conversion relationship satisfies the relationship that the smaller the reasonable value, the larger the fatigue value. The specific conversion ratio is preset manually.

[0102] The working principle and beneficial effects of the above technical solution are as follows:

[0103] When the recommended route deviates from the actual route, it is not always because the driver is fatigued and takes the wrong turn. It is also possible that the driver has chosen a better route on their own. Therefore, specific analysis is required.

[0104] During analysis, the semantics of the in-vehicle dialogue are expanded along a timeline and matched with the semantics of route change objections. Upon successful matching, the in-vehicle dialogue from the other party following the route change objection on the timeline is used as the objection response. This response is input into a reasonable route change determination model that combines real-time traffic conditions and the objection response to assess the reasonableness of the route change. The model outputs a reasonable value. Based on this reasonable value and a pre-set fatigue value table, the corresponding fatigue value is retrieved, improving the accuracy of driver fatigue assessment when the recommended route deviates from the actual route.

[0105] This invention provides an automotive airbag control system, such as... Figure 2 As shown, it includes:

[0106] The collision twin module 1 is used to perform collision twinning based on the target vehicle's driving data, the occupants' data, and the target vehicle's airbag data, and to determine the occupant risk level of airbag control with different deployment logics.

[0107] a pre-control instruction generating module 2 for generating a pre-control instruction of airbag control corresponding to the pop-up logic when the personnel danger degree is the minimum;

[0108] a first control module 3 for performing a collision verification according to the collision twin posterior data, and executing the pre-control instruction if the verification is passed;

[0109] a second control module 4 for determining a target airbag control instruction and performing corresponding control according to the actual collision data and the preset airbag control instruction library if the verification is not passed.

[0110] Obviously, various modifications and changes can be made to the present application by those skilled in the art without departing from the spirit and scope of the present application. Thus, if these modifications and changes of the present application fall within the scope of the present application claims and their equivalents, it is intended to include these modifications and changes in the present application.

Claims

1. A method for controlling automotive airbags, characterized in that, include: Collision twins are generated based on the target vehicle's driving data, the occupants' data, and the target vehicle's airbag data to determine the occupant risk level for airbags with different deployment logics. Generate the pre-control command for the airbag control of the ejection logic when the personnel risk is minimized; Collision verification is performed based on the posterior data of the collision twin. If the verification passes, the pre-control command is executed. If the verification fails, the target airbag control command is determined and corresponding control is performed based on the actual collision data and the preset airbag control command library. This includes creating a collision twin based on the target vehicle's driving data, occupant data, and airbag data to determine the occupant risk associated with different airbag deployment logics, including: Based on the target vehicle's driving data, determine whether a collision will occur within the time frame of the fuzzy perception. Based on the collision scenario and the current occupant data, predict the trajectory of the occupants after the impact of the collision. Based on the trajectory of the people and the airbag control with different deployment logic, the post-collision state of the people in the vehicle is simulated; The level of personnel risk is determined based on the preset personnel risk assessment template and the post-collision state. The steps for obtaining the duration of fuzzy perception are as follows: Based on driving data, extract collision inference factor values; Based on the collision derivation factor values ​​and the input values ​​of the collision model, determine the derivation degree corresponding to the collision factor value set; Obtain the pre-selected fuzzy perception duration of the derivation target for the set of collision factor values ​​whose derivation degree is greater than or equal to the preset derivation degree threshold; The pre-selected fuzzy perception duration is adjusted upwards based on the driver's fatigue monitoring results to obtain the fuzzy perception duration.

2. The method for controlling an automotive airbag as described in claim 1, characterized in that, The target vehicle's driving data includes information about the target vehicle's surrounding environment and the target vehicle's own status.

3. The method for controlling an automotive airbag as described in claim 1, characterized in that, The data on the occupants of the vehicle includes: occupant's body size, seat position, posture, and whether the seatbelt is being worn.

4. The method for controlling an automotive airbag as described in claim 1, characterized in that, Airbag data includes airbag distribution and the achievable deployment states of the airbags.

5. The method for controlling an automotive airbag as described in claim 1, characterized in that, The steps for obtaining fatigue monitoring results are as follows: Intervene to collect in-vehicle conversations when the recommended route deviates from the actual route; The fatigue monitoring results are obtained by normalizing and summing the fatigue values ​​represented by the fatigue monitoring video stream, driving time, and in-vehicle dialogue semantics.

6. The method for controlling an automotive airbag as described in claim 5, characterized in that, The steps for obtaining the fatigue values ​​represented by the fatigue monitoring video stream are as follows: Based on the fatigue monitoring video stream, extract the driver's blink frequency changes, eye opening changes, and head following time interval changes during steering. The fatigue values ​​characterized by the fatigue monitoring video stream are quantified based on changes in blink frequency, eye opening degree, and head sway time interval during turning.

7. The method for controlling an automotive airbag as described in claim 5, characterized in that, The steps for obtaining the fatigue value corresponding to the semantic representation of in-vehicle dialogue are as follows: The semantics of in-vehicle conversations are expanded on a preset timeline according to the corresponding extraction times. Match the route change challenge semantics on the timeline. If the match is successful, use the in-car conversation semantics of the other party after the route change challenge semantics as the challenge response. Input the response to the question into the preset reasonable route change judgment model, and output a reasonable value; Based on the reasonable value and the preset fatigue value reading table, read the fatigue value corresponding to the reasonable value.

8. A car airbag control system, characterized in that, include: The collision twin module is used to perform collision twinning based on the target vehicle's driving data, the occupants' data, and the target vehicle's airbag data, and to determine the occupant risk level of airbag control with different deployment logics. The pre-control instruction generation module is used to generate pre-control instructions for the airbag control of the ejection logic when the personnel risk is minimized. The first control module is used to perform collision verification based on the posterior data of the collision twin. If the verification passes, the pre-control command is executed. The second control module is used to determine the target airbag control command and perform corresponding control based on the actual collision data and the preset airbag control command library if the verification fails. The collision twin module performs the following operations: Based on the target vehicle's driving data, determine whether a collision will occur within the time frame of the fuzzy perception. Based on the collision scenario and the current occupant data, predict the trajectory of the occupants after the impact of the collision. Based on the trajectory of the people and the airbag control with different deployment logic, the post-collision state of the people in the vehicle is simulated; The level of personnel risk is determined based on the preset personnel risk assessment template and the post-collision state. The steps for obtaining the duration of fuzzy perception are as follows: Based on driving data, extract collision inference factor values; Based on the collision derivation factor values ​​and the input values ​​of the collision model, determine the derivation degree corresponding to the collision factor value set; Obtain the pre-selected fuzzy perception duration of the derivation target for the set of collision factor values ​​whose derivation degree is greater than or equal to the preset derivation degree threshold; The pre-selected fuzzy perception duration is adjusted upwards based on the driver's fatigue monitoring results to obtain the fuzzy perception duration.

Images