A vehicle dynamic adjustment method for side collision accidents

Abstract

The present application relates to a kind of vehicle dynamic adjustment methods for side collision accident, comprising the following steps: S1: collection vehicle body multidimensional information data information, vehicle environment information data information and vehicle passenger information data information, and through data fusion, real-time updated data model is built;According to the virtual collision data generated by vehicle collision simulation software, neural network is trained, and collision risk prediction model is obtained;S2: the data in model is input into collision risk prediction model to obtain the injury degree of passenger and the position point information of predicted collision, the state between vehicle and potential collision object is divided into different risk levels and risk map is generated;S3: vehicle adjustment system is adjusted according to different risk levels respectively to vehicle suspension system height, speed and seat posture and vehicle safety belt, and risk prompt is carried out to vehicle passenger, to reduce risk level, and risk map is updated in real time.The present application can improve the safety of passenger in side collision accident.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of vehicle safety collision, in particular to a vehicle dynamic adjustment method for side collision accidents. BACKGROUND

[0002] Vehicle side collision accident refers to a traffic accident in which the side of one vehicle collides with another vehicle during driving. This type of accident is common in urban roads, intersections, highways and other scenes, and often leads to serious casualties and vehicle damage.

[0003] In a side collision accident, due to the collision angle and speed of the two vehicles, the vehicle that is hit will often lose its center of gravity and roll over, endangering the safety of the passengers; and because the impact force is directly applied to the side of the vehicle, the kinetic energy will quickly be transferred to the passengers inside the vehicle, and the seating position of the passengers will also affect the extent of their injuries, and even cause serious injuries, especially in the head area.

[0004] With the continuous progress of automobile technology, active safety systems play an increasingly important role in preventing accidents and reducing the loss of collision accidents. Traditional vehicle safety protection measures mainly rely on passive protection, such as reinforcing the vehicle body structure and installing airbags. These measures can provide some protection when an accident occurs, but they have limitations when it comes to side collision accidents. SUMMARY

[0005] The purpose of the present application is to provide a vehicle dynamic adjustment method for side collision accidents to solve the problem of insufficient vehicle protection in vehicle side collision accidents in the prior art.

[0006] In order to achieve the above-mentioned purpose, the technical solution adopted by the present application is as follows: a vehicle dynamic adjustment method for side collision accidents, comprising the following steps:

[0007] S1: collecting vehicle body multi-dimensional information data, vehicle environment information data and vehicle passenger information data, and building a real-time updated data model through data fusion;

[0008] Training the neural network according to the virtual collision data generated by the vehicle collision simulation software to obtain a collision risk prediction model;

[0009] S2: inputting the data in the data model into the collision risk prediction model to obtain passenger injury information and collision prediction position point information, dividing the state between the vehicle and the potential collision object into different risk levels and generating a risk map;

[0010] S3: The vehicle adjustment system adjusts the height of the vehicle suspension system, the speed of the vehicle, the posture of the vehicle seat and the safety belt of the vehicle according to the different risk levels, and prompts the vehicle occupants of the risk, so as to reduce the risk level and update the risk map in real time.

[0011] According to the above technical means, according to different risk levels, the vehicle adjustment system can dynamically adjust the height of the vehicle suspension system, the speed of the vehicle and the posture of the vehicle seat, can adjust the center of gravity of the vehicle by adjusting the height of the vehicle suspension system, prevent the vehicle from rolling in the side collision accident, can adjust the position of the vehicle occupants by adjusting the posture of the vehicle seat, reduce the injury of the vehicle members in the side collision accident, and further can the impact force of the vehicle in the side collision, improve the safety of the occupants;

[0012] On the basis of traditional passive safety measures (such as airbag, vehicle body structure reinforcement), the safety of the vehicle is further improved by actively adjusting the state of the vehicle, and the severity of the vehicle damage caused by the side collision of the vehicle is further reduced, thereby reducing the probability and severity of the accident, and further improving the safety of the passengers;

[0013] By collecting vehicle body multi-dimensional information, vehicle environment information and vehicle occupant information, and then data fusion of each information, and real-time construction of a comprehensive vehicle state model, the system can timely and accurately understand the dynamic changes of the vehicle and its surrounding environment, and provide a solid data foundation for subsequent risk assessment.

[0014] Further, the height adjustment of the vehicle suspension system includes the vehicle adjustment system lowering the suspension of the side of the vehicle impacted by the ground as a reference, and / or the vehicle adjustment system raising the suspension of the side of the vehicle not impacted by the ground as a reference.

[0015] According to the above technical means, the vehicle adjustment system can accurately lower the height of the suspension of the side impacted and raise the height of the suspension of the side not impacted, can maintain the balance of the vehicle, help to reduce the side tilt angle of the vehicle in the collision, and further make the center of gravity of the vehicle stable, ensure that the vehicle can still maintain a stable driving state during the adjustment process, prevent the probability and degree of the vehicle from rolling after the side collision accident, and can adjust the height of the suspension according to the position of the vehicle collision in the side collision accident.

[0016] Further, the adjustment of the speed of the vehicle includes the vehicle adjustment system adjusting the speed and direction of the vehicle.

[0017] According to the above technical means, the vehicle adjustment system can adjust the size and direction of the vehicle speed, accurately slow down the vehicle speed according to the risk level, reduce the kinetic energy at the time of collision, thereby reducing the impact force of the collision on the vehicle and the passengers; at the same time, the vehicle adjustment system can adjust the driving direction of the vehicle according to the collected data, thereby preventing the vehicle from colliding.

[0018] According to the above technical means, the vehicle adjustment system can accurately reduce the suspension height of the side that is hit and raise the suspension height of the side that is not hit, maintain the balance of the vehicle, help reduce the roll angle of the vehicle at the time of collision, and ensure that the vehicle can still maintain a stable driving state during the adjustment process; according to the risk level, the vehicle adjustment system can accurately slow down the vehicle speed, reduce the kinetic energy at the time of collision, thereby reducing the impact force of the collision on the vehicle and the passengers; by adjusting the position of the seat along the length direction of the vehicle, the height of the seat, and the angle of the seat back, the lateral displacement of the passengers during side collision can be reduced, and the risk of injury can be reduced.

[0019] Further, the adjusting the vehicle seat posture includes adjusting the distance of the vehicle seat moving along the length direction of the vehicle, the height of the vehicle seat, and the backrest angle of the vehicle seat.

[0020] According to the above technical means, the vehicle adjustment system can adjust the position of the seat along the length direction of the vehicle, the height of the seat, and the angle of the seat back, thereby reducing the lateral displacement of the passengers during side collision and reducing the risk of injury.

[0021] Further, the different risk levels in S3 are divided into low risk level, medium risk level, high risk level, and extremely dangerous risk level according to the degree of injury of the passengers from low to high.

[0022] According to the above technical means, by dividing the risk into low risk level, medium risk level, high risk level, and extremely dangerous risk level from low to high, the system can take different measures according to different risk levels, and ensure that the most appropriate protection strategy can be provided in different situations.

[0023] Further, S3 specifically includes:

[0024] When the risk level is low, the vehicle state is not adjusted and the vehicle passengers are not prompted about the risk;

[0025] When the risk level is medium, the vehicle passengers are prompted about the risk, and the height of the vehicle suspension system and / or the speed of the vehicle and / or the posture of the vehicle seat are adjusted until the risk level is reduced to low risk level;

[0026] At the high risk level, the vehicle occupant is prompted of the risk, and the vehicle suspension system height and / or vehicle speed and / or vehicle seat posture are adjusted until the risk level is reduced to a low risk level.

[0027] At the extremely dangerous risk level, the vehicle occupant is prompted of the risk, and the vehicle seat belt is actively pre-tightened; at the same time, the vehicle suspension system height and / or vehicle speed and / or vehicle seat posture are adjusted until the risk level is reduced to a low risk level.

[0028] According to the above technical means, at the low risk level, the vehicle state is not adjusted, and the occupant is not prompted of the risk. This avoids unnecessary interference and waste of resources, while ensuring that the system can respond quickly when necessary; at the medium risk level, the vehicle speed is adjusted, and the occupant is prompted of the risk. This measure can effectively reduce the severity of the collision and remind the occupant to prepare in advance; at the high risk level, the vehicle speed, seat position and suspension system height are adjusted. These comprehensive measures can significantly reduce the impact of the collision on the occupant and improve safety; at the extremely dangerous risk level, in addition to the above adjustment measures, the safety belt is actively pre-tightened in advance, which can provide additional protection before the collision occurs, further reducing the risk of injury to the occupant.

[0029] Further, the risk prompt is a different degree of vibration of the seat or steering wheel controlled by the vehicle adjustment system according to the risk, to remind the driver and make the driver enter a prepared state in advance.

[0030] According to the above technical means, at the medium risk, high risk and extremely dangerous risk levels, the system prompts the occupant through vibration, so that the occupant can understand the potential collision risk in advance. This not only improves the occupant's alertness, but also provides the occupant with additional reaction time, so that the occupant can take appropriate measures, such as adjusting the posture or performing emergency braking, to reduce the severity of the collision.

[0031] Further, the method of obtaining a collision risk prediction model in S1 specifically includes the following steps:

[0032] S11: Simulate different collision scenarios through a collision simulation software to obtain collision data;

[0033] S12: Classify the un-preprocessed collision data, and select multiple network models according to the classified types, combine the network models to obtain an initial collision risk prediction model;

[0034] The collision data is pre-processed, including data cleaning, data normalization and standardization, and feature extraction to obtain pre-processed collision data;

[0035] S13: training the initial collision risk prediction model according to the pre-processed collision data, and optimizing the trained initial collision risk prediction model to obtain an optimal collision risk prediction model.

[0036] According to the above technical means, different collision scenarios are simulated through collision simulation software to obtain rich collision data. These data cover various possible side collision situations, including different speeds, angles, vehicle types and environmental conditions, ensuring that the model can handle various complex situations and provide more comprehensive information, which helps to more accurately predict collision risks; according to the type of collision data, multiple network models are selected for splicing and combination to form an initial collision risk prediction model; key features are extracted from a large amount of raw data to reduce data dimensionality and improve model training efficiency. At the same time, the extracted features can better reflect the key factors of collision risk, improving the prediction accuracy of the model; this multi-model fusion method can fully utilize the advantages of different models to improve the generalization ability and prediction accuracy of the model.

[0037] Further, the data fusion in S1 includes the following steps:

[0038] S11': extracting key features from the vehicle body multi-dimensional information data, the vehicle environment information data and the vehicle occupant information data;

[0039] S12': using a multi-modal deep learning network to fuse the key features to generate a data model including the vehicle body multi-dimensional information data, the vehicle environment information data and the vehicle occupant information data and real-time updating.

[0040] According to the above technical means, key features are extracted from vehicle body multi-dimensional information data, vehicle environment information data and vehicle occupant information data, including the distance between the test vehicle and the collision vehicle, the speed of the test vehicle and the occupant posture of the test vehicle. These key features cover various aspects of information about vehicle driving state, surrounding environment and occupant state, providing comprehensive data support for subsequent data fusion and risk assessment; a multi-modal deep learning network is used to fuse the key features to generate a data model including vehicle body multi-dimensional information data, vehicle environment information data and vehicle occupant information data and real-time updating. The multi-modal deep learning network can process different modal data (such as time series data, image data and point cloud data) and extract deep features to improve the prediction accuracy of the model; the real-time updating data model and the application of the multi-modal deep learning network enable the system to flexibly adjust the protection strategy according to different risk levels.

[0041] Further, in S1:

[0042] The vehicle body multi-dimensional information data includes vehicle driving speed and direction, vehicle suspension height and vehicle driving acceleration;

[0043] The vehicle environment information data includes the speed and relative position information of the surrounding objects with the vehicle as the reference;

[0044] The vehicle occupant information data includes the vehicle occupant size, vehicle occupant posture and seat position information.

[0045] According to the above technical means, the vehicle body multi-dimensional information data includes vehicle driving speed and direction, vehicle suspension height and vehicle driving acceleration, which can comprehensively reflect the driving state of the vehicle, provide the system with dynamic information of the vehicle, help the system accurately evaluate the current state and potential risks of the vehicle, and the vehicle environment information data includes the speed and relative position information of the surrounding objects with the vehicle as the reference. These data can provide detailed information of the environment around the vehicle, help the system identify potential collision objects and risk areas, and provide early warning and adjustment; the vehicle occupant information data includes the vehicle occupant size, occupant posture and seat position information. These data can provide the current state of the occupant, help the system adjust the seat posture according to the specific situation of the occupant, and ensure that the occupant is in the best protection position in the collision.

[0046] The present application realizes the beneficial effects:

[0047] 1. According to different risk levels, the vehicle adjustment system can dynamically adjust the vehicle suspension system height, vehicle speed and vehicle seat posture, can adjust the center of gravity of the vehicle by adjusting the height of the vehicle suspension system, prevent the vehicle from rolling over in a side collision accident, can adjust the position of the vehicle occupant by adjusting the vehicle seat posture, reduce the injury of the vehicle member in a side collision accident, and further reduce the impact force of the vehicle in a side collision, improve the safety of the occupant;

[0048] 2. On the basis of traditional passive safety measures (such as airbags and vehicle body structure reinforcement), the safety of the vehicle is further improved by actively adjusting the vehicle state, thereby reducing the severity of vehicle damage caused by side collision, thereby reducing the probability and severity of accidents, and thereby improving the safety of the passengers;

[0049] 3. By collecting vehicle body multi-dimensional information, vehicle environment information and vehicle occupant information, and then data fusion of each information, a comprehensive vehicle state model is constructed in real time, so that the system can timely and accurately understand the dynamic changes of the vehicle and its surrounding environment, and provide a solid data foundation for subsequent risk assessment. BRIEF DESCRIPTION OF DRAWINGS

[0050] Fig. 1 It is the overall flowchart of the present application;

[0051] Fig. 2 A training flowchart for a collision model prediction of the present application;

[0052] Fig. 3 A flowchart for fusing collected vehicle data of the present application.

[0053] The accompanying drawings are only intended to illustrate the present application, and should not be construed as limiting the present application; in order to better illustrate the present embodiment, some components in the drawings may be omitted, enlarged or reduced, and do not represent the actual product size; for those skilled in the art, it is understandable that some well-known structures and their descriptions in the drawings may be omitted; the same or similar reference numerals correspond to the same or similar components; the terms describing the positional relationship in the drawings are only intended to illustrate the present application, and should not be construed as limiting the present application. DETAILED DESCRIPTION

[0054] It should be noted that the embodiments and technical features in the present application can be combined with each other without conflict, and the detailed description in the specific embodiments should be understood as the explanation and description of the purpose of the present application, and should not be regarded as improper limitation of the present application.

[0055] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the specific technical scheme of the present application will be further described in detail below with reference to the drawings of the embodiments of the present application. The following embodiments are used to illustrate the present application, but not to limit the scope of the present application.

[0056] In the embodiments of the present application, the terms "first" and "second" are only used for descriptive purposes, and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features limited by "first" and "second" can explicitly or implicitly include one or more features. In the description of the embodiments of the present application, unless otherwise specified, the meaning of "a plurality of" is two or more.

[0057] In the embodiments of the present application, unless otherwise specified and limited, the term "connection" should be understood broadly, for example, "connection" can be fixed connection, or detachable connection, or integral; can be directly connected, or indirectly connected through intermediate medium.

[0058] In the embodiments of the present application, the terms "comprising", "containing" or any other variant thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements not only includes those elements, but can also include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without more limitations, the element defined by the phrase "comprising a" does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0059] The technical solutions of the embodiments are described in detail below with reference to specific drawings.

[0060] As shown in the drawings, Fig. 1 A vehicle dynamic adjustment method for side collision accidents in the embodiments includes the following steps:

[0061] S1: Collect vehicle body multi-dimensional information data, vehicle environment information data, and vehicle occupant information data, and build a real-time updated data model through data fusion; train a neural network according to virtual collision data generated by vehicle collision simulation software to obtain a collision risk prediction model; S2: input the data in the data model into the collision risk prediction model to obtain occupant injury degree information and collision prediction position point information, divide the state between the vehicle and the potential collision object into different risk levels and generate a risk map; S3: the vehicle adjustment system adjusts the vehicle suspension system height, vehicle speed, vehicle seat posture, and vehicle safety belt according to the different risk levels, and gives a risk prompt to the vehicle occupant to reduce the risk level and update the risk map in real time.

[0062] According to different risk levels, the vehicle adjustment system can dynamically adjust the vehicle suspension system height, vehicle speed, and vehicle seat posture, can adjust the center of gravity of the vehicle by adjusting the height of the vehicle suspension system to prevent the vehicle from rolling over in a side collision accident, can adjust the position of the vehicle occupant by adjusting the vehicle seat posture to reduce the injury of the vehicle occupant in a side collision accident, and can further reduce the impact force of the vehicle in a side collision to improve the safety of the occupant.

[0063] On the basis of traditional passive safety measures (such as airbags and reinforced vehicle body structure), the vehicle state is actively adjusted to further improve the safety of the vehicle, thereby reducing the severity of vehicle damage caused by vehicle side collision, reducing the probability and severity of accidents, and thereby improving the safety of the vehicle occupants.

[0064] By collecting multi-dimensional information of the vehicle body, vehicle environment information and vehicle passenger information, and then performing data fusion on each information, and constructing a comprehensive vehicle state model in real time, the system can timely and accurately understand the dynamic changes of the vehicle and its surrounding environment, and provide a solid data foundation for subsequent risk assessment.

[0065] Preferably, the vehicle adjustment system can also adjust the release time of the vehicle airbag, and can release the airbag in advance according to the passenger information and the collision information, thereby reducing the harm to the passengers during the collision.

[0066] Preferably, the multi-dimensional information of the vehicle is collected in real time by the vehicle-mounted sensor system, including speed, acceleration, direction angle, suspension height, and relative position and speed of surrounding vehicles and obstacles, and the body size, seat position and other information of the passengers in the vehicle are comprehensively analyzed. The vehicle-mounted sensor system includes radar, camera, laser radar, suspension height sensor, inertial measurement unit and GPS. These sensors work cooperatively to collect and analyze vehicle and environment information in real time; the passenger state monitoring system continuously monitors the body size, posture and seat position of the passengers through the in-vehicle camera and seat pressure sensor, to provide accurate data for collision prediction and vehicle dynamic adjustment.

[0067] Preferably, the vehicle collision simulation software includes one or more of multi-rigid-body dynamics analysis software (Mathematical DYnamic MOdel MADYMO), dynamic analysis finite element program (LS-DYNA) and computer three-dimensional collision impact simulation system (PAM-CRASH).

[0068] In this embodiment, the vehicle suspension system height adjustment includes lowering the suspension of the side of the vehicle that is hit by the vehicle adjustment system with the ground as a reference, and / or raising the suspension of the side of the vehicle that is not hit by the vehicle adjustment system with the ground as a reference. The vehicle adjustment system can accurately lower the suspension height of the side of the vehicle that is hit and raise the suspension height of the side of the vehicle that is not hit, thereby maintaining the balance of the vehicle, helping to reduce the side tilt angle of the vehicle during the collision, and further stabilizing the center of gravity of the vehicle, ensuring that the vehicle can still maintain a stable driving state during the adjustment process, preventing the probability and degree of rollover of the vehicle after a side collision accident, and being able to adjust the suspension height according to the position of the vehicle collision in the side collision accident.

[0069] Preferably, the vehicle suspension system height adjustment includes lowering the suspension height of the side of the vehicle that is hit, or raising the suspension height of the side of the vehicle that is not hit, or lowering the suspension height of the side of the vehicle that is hit and raising the suspension height of the side of the vehicle that is not hit at the same time.

[0070] In this embodiment, adjusting the vehicle speed includes adjusting the magnitude and direction of the vehicle speed by the vehicle adjustment system. The vehicle adjustment system can slow down the vehicle according to the risk level by adjusting the magnitude and direction of the vehicle speed, thereby reducing the kinetic energy at the time of collision and reducing the impact on the vehicle and passengers. At the same time, the vehicle adjustment system can adjust the driving direction of the vehicle according to the collected data, thereby preventing the vehicle from colliding.

[0071] The vehicle adjustment system can accurately lower the suspension height on the side that is hit and raise the suspension height on the side that is not hit, thereby maintaining the balance of the vehicle, reducing the roll angle of the vehicle at the time of collision, and ensuring that the vehicle can still maintain a stable driving state during adjustment. According to the risk level, the vehicle adjustment system can accurately slow down the vehicle, thereby reducing the kinetic energy at the time of collision and reducing the impact on the vehicle and passengers. By adjusting the position of the seat along the length of the vehicle, the height of the seat, and the angle of the seat back, the lateral displacement of the passengers during side impact can be reduced, and the risk of injury can be reduced.

[0072] Preferably, in a side impact scenario, by adjusting the suspension height, the contact angle and force between the side structure of the vehicle and the colliding object can be changed, thereby minimizing the damage to the passengers in the vehicle. For a height-adjustable suspension system, common examples include air suspension or electromagnetic suspension. Taking air suspension as an example, when receiving a signal from the electronic control unit (ECU), the air compressor or exhaust valve will work accordingly. If it is necessary to raise the suspension height, the air compressor will fill more air into the air spring, causing the spring to elongate and thereby increasing the height of the vehicle body. If it is necessary to lower the height (in some special collision scenarios, lowering the height may help to stabilize the vehicle posture), the exhaust valve will open to release some air in the air spring. Electromagnetic suspension adjusts the height and stiffness of the suspension by changing the electromagnetic force, and its response speed is faster.

[0073] Preferably, automatic emergency steering is an advanced active safety technology for vehicles. It mainly works by coordinating the sensors, ECU, and steering system of the vehicle to automatically control the steering of the vehicle to avoid collision or reduce the severity of the collision when the vehicle faces potential collision danger. According to the collision prediction results, the direction of the vehicle is automatically adjusted to avoid the high damage risk area as much as possible. For example, when the risk map predicts that the vehicle is in a high risk and extreme risk, the system will quickly adjust the angle of the steering wheel to change the driving direction of the vehicle, so that the collision point deviates from the high risk area where the passengers are located, such as moving the collision force away from the passenger compartment, or guiding the vehicle to move in a relatively open and safe direction.

[0074] Preferably, the vehicle automatically adjusts the seat by adjusting a plurality of motors installed inside and outside the vehicle seat, each motor being capable of adjusting the height of the vehicle seat, the backrest angle of the vehicle seat and the position along the length direction of the vehicle, so that the vehicle adjustment system can adjust each motor, thereby adjusting the height of the vehicle seat, the backrest angle of the vehicle seat and the position along the length direction of the vehicle.

[0075] In this embodiment, adjusting the vehicle seat posture includes adjusting the distance of the vehicle seat moving along the length direction of the vehicle, the height of the vehicle seat and the backrest angle of the vehicle seat by the vehicle adjustment system. The vehicle adjustment system can reduce the lateral displacement of the occupant in a side impact by adjusting the position of the seat along the length direction of the vehicle, the height of the seat and the angle of the seat backrest, thereby reducing the risk of injury.

[0076] In this embodiment, the different risk levels in S3 are classified from low to high as low risk level, medium risk level, high risk level and extremely dangerous risk level according to the degree of injury of the occupant. By classifying the risk from low to high as low risk level, medium risk level, high risk level and extremely dangerous risk level, the system can take different measures according to different risk levels to ensure that the most appropriate protection strategy is provided in different situations.

[0077] In this embodiment, S3 specifically includes:

[0078] In the low risk level, the vehicle state is not adjusted and the vehicle occupant is not prompted of the risk;

[0079] In the medium risk level, the vehicle occupant is prompted of the risk, and the height of the vehicle suspension system and / or the speed of the vehicle and / or the posture of the vehicle seat are adjusted until the risk level is reduced to the low risk level;

[0080] In the high risk level, the vehicle occupant is prompted of the risk, and the height of the vehicle suspension system and / or the speed of the vehicle and / or the posture of the vehicle seat are adjusted until the risk level is reduced to the low risk level;

[0081] In the extremely dangerous risk level, the vehicle occupant is prompted of the risk, and the vehicle seat belt is actively pre-tightened; at the same time, the height of the vehicle suspension system and / or the speed of the vehicle and / or the posture of the vehicle seat are adjusted until the risk level is reduced to the low risk level.

[0082] In the low risk level, no adjustment is made to the vehicle state and no risk warning is given to the passengers. This avoids unnecessary interference and waste of resources, while ensuring that the system can respond quickly when necessary; in the medium risk level, the vehicle speed is adjusted and the passengers are warned of the risk. This measure can effectively reduce the severity of the collision and remind the passengers to prepare in advance; in the high risk level, the vehicle speed, seat position and suspension height are adjusted. These comprehensive measures can significantly reduce the impact of the collision on the passengers and improve safety; in the extreme risk level, in addition to the above adjustment measures, the seat belt is also activated in advance to provide additional protection and further reduce the risk of injury to the passengers.

[0083] Preferably, in the low risk state, the vehicle has sufficient safety distance from the potential collision object (such as other vehicles, pedestrians, obstacles, etc.) and the vehicle's driving state is stable. For example, in normal traffic flow, the vehicle maintains a following distance of several seconds from the front vehicle, and the vehicle speed is within a reasonable range. At this time, the sensor (such as radar or ultrasonic sensor) shows that the distance between the vehicle and the surrounding objects is greater than the safe braking distance, for example, for a vehicle traveling at a general speed, if the following distance is greater than 30 meters and the vehicle speed is less than 60 km / h, the acceleration sensor shows that the vehicle acceleration is stable, there is no sudden acceleration or deceleration, and the steering angle sensor shows that the vehicle driving direction is stable, there is no sudden steering action.

[0084] Preferably, in the medium risk state, the distance between the vehicle and the potential collision object gradually approaches the safe braking distance, or the vehicle has some unstable factors that may lead to a collision. For example, the front vehicle suddenly decelerates, the distance between the vehicle and the front vehicle rapidly shortens, but it has not reached the level of emergency braking; or the vehicle is in the process of changing lanes, and the distance between the vehicle and the vehicle in the adjacent lane is close. At this time, the sensor shows that the distance between the vehicle and the front object is between 50%-90% of the safe braking distance. For example, when the vehicle speed is 80 km / h, the following distance is shortened to about 30-50 meters. The acceleration sensor may detect that the vehicle has a certain degree of sudden deceleration or the vehicle is performing a steering action, and the steering angle sensor shows that the steering wheel has a certain angle of rotation, which may cause the vehicle to approach other objects.

[0085] For example, the vehicle suddenly encounters an obstacle in front, or the distance to the vehicle in front is sharply shortened to almost the extent of a rear-end collision. The distance sensor shows that the distance between the vehicle and the obstacle in front is less than 50% of the safe braking distance. For example, when the vehicle is traveling at 100 km / h, the distance to the vehicle in front is less than 30 meters. The acceleration sensor detects that the vehicle does not have enough deceleration to avoid a collision, and the steering angle sensor shows that even if the vehicle turns, it is difficult to avoid the collision object.

[0086] For example, the vehicle suddenly encounters an obstacle in front, or the distance to the vehicle in front is sharply shortened to almost the extent of a rear-end collision. The distance sensor shows that the distance between the vehicle and the obstacle in front is less than 50% of the safe braking distance. For example, when the vehicle is traveling at 100 km / h, the distance to the vehicle in front is less than 30 meters. The acceleration sensor detects that the vehicle does not have enough deceleration to avoid a collision, and the steering angle sensor shows that even if the vehicle turns, it is difficult to avoid the collision object.

[0087] In this embodiment, the risk prompt is to adjust the system to control the seat or steering wheel to vibrate from high to low in vibration degree according to the risk level from high to low, so as to remind the driver and make the driver enter the preparation state in advance. In the medium risk, high risk and extreme danger levels, the system prompts the occupant through vibration, so that the occupant can understand the potential collision risk in advance. This not only improves the alertness of the occupant, but also provides additional reaction time for the occupant to take appropriate measures, such as adjusting the posture or making emergency braking, so as to reduce the severity of the collision.

[0088] For example, the vehicle suddenly encounters an obstacle in front, or the distance to the vehicle in front is sharply shortened to almost the extent of a rear-end collision. The distance sensor shows that the distance between the vehicle and the obstacle in front is less than 50% of the safe braking distance. For example, when the vehicle is traveling at 100 km / h, the distance to the vehicle in front is less than 30 meters. The acceleration sensor detects that the vehicle does not have enough deceleration to avoid a collision, and the steering angle sensor shows that even if the vehicle turns, it is difficult to avoid the collision object.

[0089] For example, the vehicle suddenly encounters an obstacle in front, or the distance to the vehicle in front is sharply shortened to almost the extent of a rear-end collision. The distance sensor shows that the distance between the vehicle and the obstacle in front is less than 50% of the safe braking distance. For example, when the vehicle is traveling at 100 km / h, the distance to the vehicle in front is less than 30 meters. The acceleration sensor detects that the vehicle does not have enough deceleration to avoid a collision, and the steering angle sensor shows that even if the vehicle turns, it is difficult to avoid the collision object.

[0090] Preferably, for high risk, when the risk map on the display screen or head-up display (HUD) of the vehicle shows orange, the automatic emergency steering is triggered when there are no obstacles and sufficient space in front and obliquely in front of the vehicle, the collision point is deviated from the high-risk area where the occupants are located, the degree of side collision is maximized, and the dynamic adjustment system adjusts the suspension of the vehicle on the collision side and the seat where the occupants are located to be lowered, improving the smoothness of driving, and if the backrest angle of the front passenger seat is large, the normal sitting posture can be quickly restored until the risk map shows no risk, and the suspension and seat are automatically restored.

[0091] Preferably, for extremely dangerous risk, when the risk map on the display screen or head-up display (HUD) of the vehicle shows red, the automatic emergency steering is triggered when there are no obstacles and sufficient space in front and obliquely in front of the vehicle, the collision point is deviated from the high-risk area where the occupants are located, the degree of side collision is maximized, and the dynamic adjustment system adjusts the suspension of the vehicle on the collision side and the seat where the occupants are located to be lowered, improving the smoothness of driving, and adjusts the seat to move away from the collision point until the risk is reduced to no risk, and the suspension and seat are automatically restored, and the safety belt active pre-tightening and side airbag can be triggered in advance if necessary.

[0092] In this embodiment, the risk prompt is to adjust the system to control the seat or steering wheel to vibrate at different levels to remind the driver, so that the driver can enter a prepared state in advance. In the medium risk, high risk and extremely dangerous risk levels, the system vibrates to prompt the occupants to understand the potential collision risk in advance, which not only improves the alertness of the occupants, but also provides additional reaction time for the occupants to take appropriate measures, such as adjusting the posture or performing emergency braking, thereby reducing the severity of the collision.

[0093] As shown in Fig. 2 In this embodiment, the method for obtaining the collision risk prediction model in S1 specifically includes the following steps:

[0094] S11: Simulate different collision scenarios by a collision simulation software to obtain collision data;

[0095] S12: Classify the collision data, select a plurality of network models according to the classified types, combine each network model to obtain an initial collision risk prediction model, and pre-process the collision data, including data cleaning, data normalization and standardization, and feature extraction to obtain pre-processed collision data;

[0096] S13: Train the initial collision risk prediction model according to the pre-processed collision data, and optimize the trained initial collision risk prediction model to obtain an optimal collision risk prediction model.

[0097] The collision simulation software is used to simulate different collision scenarios and obtain rich collision data. These data cover various possible side collision situations, including different speeds, angles, vehicle types and environmental conditions, ensuring that the model can handle various complex situations and provide more comprehensive information to help more accurately predict collision risks; according to the types of collision data, multiple network models are selected for splicing and combination to form an initial collision risk prediction model; key features are extracted from a large amount of raw data to reduce data dimensionality and improve model training efficiency. At the same time, the extracted features can better reflect the key factors of collision risk, improving the prediction accuracy of the model; this multi-model fusion method can fully utilize the advantages of different models to improve the generalization ability and prediction accuracy of the model.

[0098] Preferably, the types of collision data include time series data, image data and point cloud data.

[0099] Preferably, the time series data (such as vehicle speed, acceleration) data selects a recurrent neural network (RNN), a long short-term memory network (LSTM) or a gated recurrent unit (GRU); the image data (such as camera image) selects a convolutional neural network (CNN); the point cloud data (such as LiDAR) selects a point cloud processing network (PointNet), a voxel network (VoxelNet) or a graph neural network (GNN); the multi-modal data fusion selects a multi-modal deep learning network, such as a joint CNN, RNN or Transformer; the damage prediction selects a regression neural network or a multi-task learning network.

[0100] In this embodiment, the types of collision data in S22 include time series data, image data and point cloud data. The obtained collision data includes time series data, image data and point cloud data, and these multi-modal data provide more comprehensive information to help more accurately predict collision risks.

[0101] As shown in Fig. 3 In this embodiment, the data fusion in S1 includes the following steps:

[0102] S11': extracting key features from vehicle body multi-dimensional information data, vehicle environment information data and vehicle occupant information data;

[0103] S12': using a multi-modal deep learning network to fuse the key features to generate a data model including vehicle body multi-dimensional information data, vehicle environment information data and vehicle occupant information data and updated in real time.

[0104] The key features are extracted from the vehicle body multi-dimensional information data, vehicle environment information data and vehicle occupant information data, including the distance between the test vehicle and the collision vehicle, the speed of the test vehicle and the occupant posture of the test vehicle. These key features cover the multi-aspect information of vehicle driving state, surrounding environment and occupant state, providing comprehensive data support for subsequent data fusion and risk assessment; the key features are fused using a multi-modal deep learning network to generate a data model that includes vehicle body multi-dimensional information data, vehicle environment information data and vehicle occupant information data and is updated in real time. The multi-modal deep learning network can process data of different modalities (such as time series data, image data and point cloud data) and extract deep features to improve the prediction accuracy of the model; the real-time updated data model and the application of the multi-modal deep learning network enable the system to flexibly adjust the protection strategy according to different risk levels.

[0105] In this embodiment, in S1: the vehicle body multi-dimensional information data includes vehicle driving speed and direction, vehicle suspension height and vehicle driving acceleration; the vehicle environment information data includes the speed and relative position information of the surrounding objects with the vehicle as the reference; the vehicle occupant information data includes the vehicle occupant body size, the vehicle occupant posture and the seat position information.

[0106] The vehicle body multi-dimensional information data includes vehicle driving speed and direction, vehicle suspension height and vehicle driving acceleration, which can comprehensively reflect the driving state of the vehicle and provide the system with dynamic information of the vehicle, helping the system to accurately assess the current state and potential risks of the vehicle. The vehicle environment information data includes the speed and relative position information of the surrounding objects with the vehicle as the reference. These data can provide detailed information of the vehicle's surrounding environment, helping the system to identify potential collision objects and risk areas and to provide early warning and adjustment; the vehicle occupant information data includes the vehicle occupant body size, the occupant posture and the seat position information. These data can provide the current state of the occupant, helping the system to adjust the seat posture according to the specific situation of the occupant to ensure that the occupant is in the best protection position in the event of a collision.

[0107] In summary, as shown in Figs. 1-3 The vehicle dynamic adjustment method for side collision accidents in this embodiment includes the following steps:

[0108] S1: Collect vehicle body multi-dimensional information data, vehicle environment information data and vehicle occupant information data, and build a real-time updated data model through data fusion;

[0109] The different collision scenes are simulated by a collision simulation software to obtain collision data; the collision data is preprocessed, including data cleaning, data normalization and standardization and feature extraction to obtain preprocessed collision data; the collision data is classified, and a plurality of network models are selected according to the types of the collision data, the network models are combined to obtain an initial collision risk prediction model; the initial collision risk prediction model is trained according to the preprocessed collision data, and the trained initial collision risk prediction model is optimized to obtain an optimal collision risk prediction model;

[0110] S2: inputting data in the data model into the collision risk prediction model to obtain occupant injury degree information and collision prediction position point information, dividing the state between the vehicle and the potential collision object into different risk levels and generating a risk map;

[0111] S3: the vehicle adjustment system adjusts the vehicle suspension system height, the vehicle speed and the vehicle seat posture according to different risk levels respectively, and gives a risk prompt to the vehicle occupant to reduce the risk level, and updates the risk map in real time.

[0112] The above-mentioned sequence numbers of the embodiments of the present application are only for description, and do not represent the advantages and disadvantages of the embodiments. The above is only the preferred embodiment of the present application, and does not limit the patent scope of the present application, and any equivalent structure or equivalent process transformation using the content of the present application specification and drawings, or direct or indirect application in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A method of dynamically adjusting a vehicle for a side collision accident, characterized by, The method comprises the following steps: S1: collecting multi-dimensional information data of a vehicle body, vehicle environment information data, and vehicle occupant information data, and building a real-time updated data model through data fusion; training a neural network according to virtual collision data generated by vehicle collision simulation software to obtain a collision risk prediction model; S2: inputting data in the data model into the collision risk prediction model to obtain occupant injury degree information and collision prediction position point information, dividing the state between the vehicle and a potential collision object into different risk levels, and generating a risk map; S3: adjusting a vehicle suspension system height, a vehicle speed, a vehicle seat posture, and a vehicle safety belt according to the different risk levels, and giving a risk prompt to a vehicle occupant to reduce the risk level and update the risk map in real time.

2. The method of dynamically adjusting a vehicle for a side collision accident according to claim 1, wherein, The vehicle suspension system height adjustment comprises lowering a suspension on a side of the vehicle that is hit by the vehicle adjustment system taking the ground as a reference, and / or raising a suspension on a side of the vehicle that is not hit by the vehicle adjustment system taking the ground as a reference.

3. The method of claim 1, wherein, The adjustment of the vehicle speed comprises adjusting a vehicle speed size and a vehicle speed direction by the vehicle adjustment system.

4. The method of claim 1, wherein, The adjustment of the vehicle seat posture comprises adjusting a distance that a vehicle seat moves along a length direction of the vehicle, a height of the vehicle seat, and a backrest angle of the vehicle seat by the vehicle adjustment system.

5. The method of dynamically adjusting a vehicle for a side impact collision of claim 1, wherein, The different risk levels in S3 are divided into a low risk level, a medium risk level, a high risk level, and an extremely dangerous risk level from low to high according to the injury degree of the occupant.

6. The method of dynamically adjusting a vehicle for a side collision according to claim 5, wherein, S3 specifically comprises: in the low risk level, no adjustment is made to the vehicle state and no risk prompt is given to the vehicle occupant; in the medium risk level, a risk prompt is given to the vehicle occupant, and the vehicle suspension system height and / or the vehicle speed and / or the vehicle seat posture are adjusted until the risk level is reduced to the low risk level; in the high risk level, a risk prompt is given to the vehicle occupant, and the vehicle suspension system height and / or the vehicle speed and / or the vehicle seat posture are adjusted until the risk level is reduced to the low risk level; in the extremely dangerous risk level, a risk prompt is given to the vehicle occupant, and the vehicle safety belt is actively pre-tightened; at the same time, the vehicle suspension system height and / or the vehicle speed and / or the vehicle seat posture are adjusted until the risk level is reduced to the low risk level.

7. The method of dynamically adjusting a vehicle for a side collision according to claim 6, wherein, The risk prompt is that the vehicle adjustment system controls the seat or the steering wheel to vibrate to different degrees according to the risk to remind the driver, so that the driver enters a prepared state in advance.

8. The method of dynamically adjusting a vehicle for a side impact collision of claim 1, wherein, The method for obtaining the collision risk prediction model in S1 specifically comprises the following steps: S11: simulating different collision scenarios by a collision simulation software to obtain collision data; S12: classifying the collision data, selecting a plurality of network models according to the classified types, combining the network models to obtain an initial collision risk prediction model; The collision data is preprocessed, including data cleaning, data normalization and standardization, and feature extraction to obtain preprocessed collision data; S13: training the initial collision risk prediction model according to the preprocessed collision data, and optimizing the trained initial collision risk prediction model to obtain an optimal collision risk prediction model.

9. The method of dynamically adjusting a vehicle for a side impact collision of claim 1, wherein, The data fusion in S1 includes the following steps: S11': extracting key features from the vehicle body multi-dimensional information data, the vehicle environment information data, and the vehicle occupant information data; S12': using a multi-modal deep learning network to fuse the key features to generate a data model including the vehicle body multi-dimensional information data, the vehicle environment information data, and the vehicle occupant information data and being updated in real time.

10. The method of dynamically adjusting a vehicle for a side collision according to claim 1, wherein, In S1: The vehicle body multi-dimensional information data includes vehicle driving speed and direction, vehicle suspension height, and vehicle driving acceleration; The vehicle environment information data includes the speed and relative position information of the surrounding objects with the vehicle as the reference; The vehicle occupant information data includes vehicle occupant body size, vehicle occupant posture, and seat position information.

Images