Safety belt staged pre-tightening and force limiting method and system based on passenger signs

Abstract

The invention discloses a safety belt grading type pre-tightening and force limiting method and system based on passenger signs, and the method comprises the steps: obtaining a passenger image, and measuring the weight information of a passenger, so as to calculate the gender, age, upper trunk height, sitting posture angle and weight of the passenger; the height of the safety belt is adjusted according to the upper trunk height and the sitting posture angle; pre-tightening force and a force limiting table are preset according to the gender, the age and the weight of the passenger, different safety belt pre-tightening and safety belt force limiting are matched for the gender, the age and the weight of different passengers, and safety belt pre-tightening is divided into safety belt first-level pre-tightening and safety belt second-level pre-tightening according to the emergency degree. According to the main technical scheme and the main effects, passengers are classified through sign recognition, so that the pre-tightening force and the force limiting value of the safety belt can be dynamically adjusted according to the body weight, the age and other characteristics, and the adaptability of the safety belt system to the passengers with different body types is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of vehicle safety, and in particular to a method and system for graded pre-tightening and force-limiting of a seat belt based on occupant's physical signs. Background Art

[0002] The most common system in modern cars is the three-point seat belt system, which is mainly composed of webbing, retractor, seat belt height adjuster and related accessories, and can effectively restrain the occupant's body.

[0003] When the vehicle's sensors detect an impending collision, the active seatbelt system's pretensioner rapidly tightens the belts through an electromechanical mechanism, reducing the gap between the passenger and the belts and preventing the occupant from moving out of their seat. In the moments following the collision, the force limiter further restricts the occupant's forward motion, preventing a secondary collision and potentially further injury.

[0004] The main drawbacks are:

[0005] (1) The pre-tensioning and force-limiting devices of existing automobile seat belt systems cannot well adapt to the protection of occupants with different physical characteristics;

[0006] (2) Since passengers with different physical characteristics can withstand different webbing tension, restraint protection using the same force-limiting method in the event of a collision is likely to cause secondary injuries;

[0007] (3) The existing active seat belt system is not able to achieve the overall adjustment of the height adjuster and the preload limiter. Summary of the Invention

[0008] Purpose of the invention: The purpose of the present invention is to solve the technical problems in the prior art and to provide a method and system for graded pre-tensioning and force limiting of seat belts based on occupant's physical signs.

[0009] Technical solution:

[0010] In a first aspect, the present application proposes a method for graded pre-tensioning and force limiting of seat belts based on occupant physical signs, comprising the following steps:

[0011] Acquire occupant images and measure occupant weight information to calculate the occupant's gender, age, upper body height, sitting angle, and weight;

[0012] Adjust the height of the seat belt according to the height of the upper body and the sitting angle;

[0013] Preset pretension and force limit tables are created based on the occupant's gender, age, and weight, allowing for different seatbelt pretensions and force limits to be matched to the gender, age, and weight of the occupant. Seatbelt pretensioning is divided into primary and secondary pretensioning based on the degree of urgency.

[0014] The vehicle sensor parameters are obtained and weighted to obtain the comprehensive collision coefficient. The execution of the first-level seat belt pre-tensioning and the second-level seat belt pre-tensioning is determined according to the preset threshold of the comprehensive collision coefficient. The seat belt force limiter is executed when the seat belt is in the second-level pre-tensioning state.

[0015] Preferably, acquiring an image of the occupant and measuring the occupant's weight information to calculate the occupant's gender and age includes:

[0016] Input occupant image, including:

[0017] Acquiring facial images of occupants;

[0018] Perform denoising and light equalization preprocessing;

[0019] Use YOLOv5 to locate the occupant's facial area and crop the facial ROI image;

[0020] Perform feature extraction, which includes:

[0021] The input is fed into the YOLOv5 target detection model to analyze the gender of the occupants and estimate their age.

[0022] Preferably, acquiring an image of the occupant and measuring the occupant's weight information to calculate the occupant's upper torso height includes:

[0023] Obtain system parameters when capturing passenger images, including focal length, installation height, and pitch angle;

[0024] Use the OpenPose algorithm to detect the coordinates of the occupant's neck and waist key points in real time;

[0025] Calculate the pixel height of the upper torso using the following formula:

[0026] ;

[0027] Convert the upper torso pixel height to the upper torso physical height using the following formula:

[0028] ;

[0029] in, is the vertical coordinate of the waist key point, is the vertical coordinate of the neck key point, is the installation height, is the pitch angle, is the focal length.

[0030] Preferably, acquiring an image of the occupant and measuring the occupant's weight information to calculate the occupant's sitting angle includes:

[0031] Input occupant image;

[0032] The OpenPose algorithm is used to detect the coordinates of the occupant's hip and shoulder key points, and the inclination angle formula of the line connecting the hip and shoulder key points is:

[0033] ;

[0034] in,, is the vertical coordinate of the key point of the occupant’s shoulder, is the horizontal coordinate of the key point of the occupant's shoulder, is the ordinate of the occupant’s hip key point, is the horizontal coordinate of the occupant's hip key point,

[0035] The tilt angle thresholds for upright, forward and backward tilt are set respectively to judge the occupant's sitting posture based on the occupant's sitting angle.

[0036] Preferably, acquiring an image of the occupant and measuring the occupant's weight information to calculate the occupant's weight includes:

[0037] Input occupant image;

[0038] Generate 3D human body mesh based on 3D human body reconstruction model;

[0039] The 3D human body mesh is voxelized to obtain the number of voxels and the volume of a single voxel. The human body volume is calculated based on the number of voxels and the volume of a single voxel. The formula is as follows:

[0040] ;

[0041] Calculate visual weight using the following formula:

[0042] :

[0043] Where N is the number of voxels, is the volume of a single voxel, is the average density of the human body.

[0044] Preferably, the fusion of visual weight and measured occupant weight information includes:

[0045] Preprocessing the measured passenger weight information;

[0046] The formula for fusing the measured occupant weight information and visual weight is as follows:

[0047] ;

[0048] The calculation method of K is as follows:

[0049] ;

[0050] W pressure To measure the passenger weight information, W vision is the visual weight, K is the weight coefficient, σ vision is the variance of multiple estimations of visual weight, σ pressure The variance of multiple measurements of the pressure sensor.

[0051] Preferably, the height of the safety belt is adjusted according to the height of the upper body and the sitting angle, including:

[0052] ;

[0053] Q1 is the trunk height weight coefficient, Q2 is the sitting angle compensation coefficient, and C is the vehicle model adaptation constant;

[0054] If you are sitting forward, move the seat belt up 5-10mm to avoid strangulation.

[0055] If the seat belt is reclining, lower it by 5-10mm to prevent the shoulder belt from slipping.

[0056] If it is in an upright position, it remains unchanged.

[0057] Preferably, a preload and force limit table is preset according to the gender, age, and weight of the occupants, so as to match different seat belt preloads and seat belt force limits to the gender, age, and weight of different occupants, including:

[0058] The fusion weight class is divided into three levels: light, normal, and heavy; the age class is divided into three levels: juvenile, adult, and senior;

[0059] When the seat belt system activates the first level pre-tensioning protection, the output is a fixed pre-tensioning force F f , where F f is the minimum value among the system classification forces;

[0060] When the seatbelt system activates secondary pretensioning protection:

[0061] According to the weight class and age class, a corresponding gear is selected from the predefined 9 gears of preload force, which satisfies:

[0062] In the same age group, the gear preload of light weight is less than the gear preload of normal weight and less than the gear preload of heavy weight;

[0063] The relationship between gear preloads at different ages within the same weight class is calibrated through collision tests;

[0064] Output the selected secondary preload force;

[0065] According to the weight level, the seat belt stepper motor is controlled to drive the force limiter to the preset slot position, so that the force limit value Fw satisfy:

[0066] F w1 Corresponding to light weight category, F w2 Corresponding to normal weight level, F w3 Corresponding to the heavy weight level, and F w1 <F w2 <F w3 ;

[0067] Differentiated adjustments based on passenger gender:

[0068] If the occupant is a female, the formulas for all seat belt pretensioners and force limiters are as follows:

[0069] ;

[0070] Among them F female Set the seat belt preload and force limiter for women, F male Set the seat belt preload and force limiter for males, with the weighting factor s ranging from 0.85 to 0.95;

[0071] If the occupant is male, keep all strength values unchanged.

[0072] Preferably, vehicle sensor parameters are obtained and weighted to obtain a comprehensive collision coefficient. The execution of the first-level seat belt pre-tensioning and the second-level seat belt pre-tensioning is determined according to a preset threshold of the comprehensive collision coefficient. The seat belt force limiter is executed when the seat belt is in the second-level pre-tensioning state, including:

[0073] The comprehensive collision coefficient is obtained by weighted calculation using the measurement data of the vehicle-mounted acceleration sensor, gyroscope, and FCW forward collision system. The formula is as follows:

[0074]

[0075] Where a is the vehicle acceleration obtained by the acceleration sensor, θ roll is the turning angle obtained by the vehicle gyroscope, and TTC is the collision time threshold obtained by the FCW forward collision system;

[0076] is the weight coefficient;

[0077] When the vehicle sensor detects that the comprehensive collision risk coefficient R is less than the first threshold, the seat belt level 1 pre-tightening signal is triggered, and the seat belt pre-tightening force is F f ;

[0078] When R> the first threshold and R≤ the second threshold, the secondary pre-tensioning of the seat belt is triggered, and the graded pre-tensioning force F is matched according to the occupant's physical signs. e 、F a 、F o；

[0079] When R>the second threshold, the seat belt is mechanically limited according to the preset gear.

[0080] In a first aspect, in some embodiments, a system used in the method described in the above embodiments is provided, comprising a computing unit, a vehicle-mounted camera, a weight sensor, a seat belt height adjuster, a seat belt retractor, and a vehicle sensor;

[0081] The vehicle-mounted camera, weight sensor, seat belt height adjuster, seat belt retractor, and vehicle sensor are respectively connected to the computing unit;

[0082] The computing unit is configured to implement the method according to any one of claims 1 to 9;

[0083] The occupant's image is acquired through the vehicle's onboard camera, and the occupant's weight information is measured through the weight sensor to calculate the occupant's gender, age, upper body height, sitting angle, and weight;

[0084] Adjust the seat belt height adjuster according to the upper body torso height and sitting angle to adjust the seat belt height;

[0085] Preset pretension and force limit tables are created based on the occupant's gender, age, and weight, allowing for matching seatbelt pretensioning, force limiter, and seatbelt retractor configurations to suit the occupant's gender, age, and weight. Seatbelt pretensioning is divided into primary and secondary pretensioning based on the degree of urgency.

[0086] Vehicle sensor parameters are obtained through vehicle sensors, and the comprehensive collision coefficient is obtained by weighting. The execution of the first-level seat belt pre-tensioning and the second-level seat belt pre-tensioning is determined according to the preset threshold of the comprehensive collision coefficient. The seat belt force is limited when the seat belt is in the second-level pre-tensioning state.

[0087] Beneficial effects:

[0088] By identifying occupants through physical sign recognition, the seatbelt preload and force limit can be dynamically adjusted based on weight, age, and other characteristics, significantly improving the seatbelt system's adaptability to occupants of different body types (such as children, the elderly, and lighter individuals).

[0089] The force limiter outputs force according to the set gear. When a collision occurs, the controlled mechanical deformation of the force limiter's force limiter effectively prevents the seatbelt tension from exceeding the occupant's tolerance threshold, reducing the risk of chest injury and avoiding secondary injuries caused by excessive restraint.

[0090] The seatbelt system combines onboard multi-source sensors (such as accelerometers, gyroscopes, and forward collision warning systems) to dynamically assess collision risk and implements primary pretensioning, secondary pretensioning, and force limiting protection in stages, forming an integrated feedforward-feedback adaptive protection mechanism.

[0091] Both the preload and force-limiting modules are controlled by electric drive actuators, and combined with control algorithms, they achieve automatic adjustment without perception. Occupants can obtain customized protection without manual intervention, taking into account both safety and comfort. BRIEF DESCRIPTION OF THE DRAWINGS

[0092] Figure 1 Provide a schematic diagram of the method framework of the present invention;

[0093] Figure 2 Provides preset preload and force limit indication for the present invention;

[0094] Figure 3 Provides a schematic diagram of the image and gravity processing flow for the present invention;

[0095] Figure 4 A schematic diagram of a hierarchical pre-tightening process is provided for the present invention;

[0096] Figure 5 A schematic diagram of an existing safety belt device is provided for the present invention;

[0097] Figure 6 A schematic diagram of an existing safety belt device is provided for the present invention;

[0098] Figure 7 Provide a schematic diagram of the system of this application for the present invention;

[0099] Figure 8 It is a structural diagram of a device provided in one embodiment of the present application;

[0100] Figure 9 This is a structural block diagram of an electronic device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0101] In order to make the technical solution of the present invention clearer, the present invention is further described in detail below with reference to the specific embodiments of the drawings.

[0102] Example 1

[0103] In order to make the purpose, technical solutions and advantages of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. Unless otherwise defined, the technical terms or scientific terms used herein should be the common meanings understood by people with ordinary skills in the field to which the present invention belongs. The words "including" and similar words used in this article mean that the elements or objects appearing before the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects.

[0104] In view of the problems existing in the existing technology, such as Figure 1-Figure 4 As shown, a method for graded pre-tensioning and force limiting of seat belts based on occupant physical signs includes the following steps:

[0105] S101. Acquire occupant images and measure occupant weight information to calculate the occupant's gender, age, upper torso height, sitting angle, and weight. Perform multi-dimensional physical identification of the occupant, integrating facial image recognition (gender, age, body shape) with seat pressure data (weight) to provide basic parameter support for subsequent personalized seatbelt adjustment. The camera collects occupant image data in real time, including front, side, and overhead angles. For multi-seat vehicles, each seat is equipped with an independent camera, or a wide-angle camera is used to cover multiple seats.

[0106] Through facial recognition technology and ECU processing, the occupant's gender, age, and body shape (upper torso height, sitting angle, and visual weight) are determined.

[0107] Upper torso height and sitting angle serve as the basis for adjusting the guide ring height of the seatbelt height adjuster. The occupant's gender and age are directly derived by the facial recognition module of the computing unit. Visual weight is weighted and fused with the initial weight data from the weight sensor. The resulting weight data, along with gender and age, serves as the basis for seatbelt preload and force limit grading.

[0108] S102. Adjust the seat belt height according to upper torso height and sitting angle, control the guide ring position, and adjust the seat belt fit point to avoid blind spots such as strangulation and shoulder slippage caused by uncomfortable wearing height. Ensure the seat belt is always in the optimal position for use, optimize force distribution, improve user comfort, and enhance adaptability to complex vehicle operating conditions.

[0109] S103: Preset a preload and force limit table based on the occupant's gender, age, and weight to match different seatbelt preloads and force limits to the gender, age, and weight of the occupant. The seatbelt preload is divided into primary and secondary preloads based on the degree of urgency. The physical sign parameters are mapped to the graded preload / force limit rules to differentiate between primary and secondary preloads, and the force limit level is set.

[0110] S104. Obtain vehicle sensor parameters, perform weighted calculations to obtain a comprehensive collision coefficient, and decide whether to implement level 1 or level 2 seatbelt pretensioning based on a preset threshold of the comprehensive collision coefficient. Implement seatbelt force limiting in level 2 seatbelt pretensioning. Real-time perception of the vehicle status and forward collision risk is achieved, and different protection strategies are activated in stages based on the risk level.

[0111] In some specific embodiments, acquiring an image of an occupant and measuring the occupant's weight information to calculate the occupant's gender and age includes:

[0112] Input occupant image, including:

[0113] Acquiring facial images of occupants;

[0114] Perform denoising and light equalization preprocessing;

[0115] Use YOLOv5 to locate the occupant's facial area and crop the facial ROI image;

[0116] Perform feature extraction, which includes:

[0117] The input is fed into the YOLOv5 target detection model to analyze the gender of the occupants and estimate their age.

[0118] Specifically, in this embodiment of the present invention, an onboard camera is used to capture real-time facial image data of vehicle occupants at a resolution of 1280×720 and a frame rate of 30 FPS. After denoising and illumination equalization preprocessing, the original image is pre-processed using the YOLOv5 object detection model to locate and extract the occupant's facial region. A 224×224 pixel facial region of interest (ROI) image is cropped to serve as input for subsequent feature extraction and classification.

[0119] The extracted facial ROI image is analyzed using the YOLOv5 model to identify key features such as facial contour, eyelid opening and closing, and jaw shape. Combined with the classification capabilities of the deep learning model, the occupant's gender is automatically determined, outputting a male or female category.

[0120] By analyzing facial skin texture features (such as fine line density, nasolabial fold depth, and skin color distribution) and pigmentation patterns, combined with the image regression capabilities of the YOLOv5 model, we estimate the age range of the occupants. The prediction error is controlled within a range of ±3 years, and the occupants are divided into age groups based on the estimated age:

[0121] Juvenile: <18 years old;

[0122] Adult: 18 to 50 years old;

[0123] Elderly: >50 years old.

[0124] The age classification result serves as one of the bases for subsequent classification of seat belt pretensioner and force limiter parameters.

[0125] In some specific embodiments, acquiring an image of the occupant and measuring the occupant's weight information to calculate the occupant's upper torso height includes:

[0126] Obtain system parameters when capturing passenger images, including focal length, installation height, and pitch angle;

[0127] Use the OpenPose algorithm to detect the coordinates of the occupant's neck and waist key points in real time;

[0128] Calculate the pixel height of the upper torso using the following formula:

[0129] ;

[0130] Convert the upper torso pixel height to the upper torso physical height using the following formula:

[0131] ;

[0132] in, is the vertical coordinate of the waist key point, is the ordinate of the neck key point, is the installation height, is the pitch angle, is the focal length.

[0133] Specifically, the present invention uses the OpenPose pose estimation algorithm to detect key points in the occupant's image. By extracting the coordinates of the skeletal key points of the occupant's neck and waist, the vertical distance between them in the image is calculated to infer the actual height of the occupant's upper torso. This height serves as the core reference parameter for adjusting the vertical position of the seatbelt guide ring.

[0134] Camera calibration parameters: including known focal length f, installation height h cam , installation pitch angle θ, used for subsequent image pixel and physical quantity mapping conversion;

[0135] OpenPose skeleton keypoint output: can stably identify the neck keypoints and waist keypoints in the occupant image and obtain their pixel coordinate values.

[0136] In some specific embodiments, acquiring an image of an occupant and measuring the occupant's weight information to calculate the occupant's sitting angle includes:

[0137] Input occupant image;

[0138] The OpenPose algorithm is used to detect the coordinates of the occupant's hip and shoulder key points, and the inclination angle formula of the line connecting the hip and shoulder key points is:

[0139] ;

[0140] in, is the ordinate of the key point of the occupant's shoulder, is the horizontal coordinate of the key point of the occupant's shoulder, is the ordinate of the occupant’s hip key point, is the horizontal coordinate of the occupant's hip key point,

[0141] The tilt angle thresholds for upright, forward and backward tilt are set respectively to judge the occupant's sitting posture based on the occupant's sitting angle.

[0142] Specifically, this module uses the OpenPose pose estimation algorithm to extract key points from the occupant image, obtain the coordinates of the hip and shoulder key points, and calculate the inclination angle of the occupant's upper body posture based on the inclination angle calculation formula of the line connecting the two points.

[0143] The current sitting posture of the occupant can be determined by the inclination angle θ:

[0144] upright sitting posture (θ is approximately 90°);

[0145] forward-leaning sitting posture (θ is greater than a certain positive threshold);

[0146] Reclining sitting posture (θ is less than a certain negative threshold);

[0147] This posture information will serve as an important input variable for adjusting the vertical height of the seat belt guide ring.

[0148] In some specific embodiments, acquiring an image of an occupant and measuring the occupant's weight information to calculate the occupant's weight includes:

[0149] Input occupant image;

[0150] Generate 3D human body mesh based on 3D human body reconstruction model;

[0151] The 3D human body mesh is voxelized to obtain the number of voxels and the volume of a single voxel. The human body volume is calculated based on the number of voxels and the volume of a single voxel. The formula is as follows:

[0152] ;

[0153] Calculate visual weight using the following formula:

[0154] :

[0155] Where N is the number of voxels, is the volume of a single voxel, is the average density of the human body, usually around 1.0-1.1 g / cm 3 .

[0156] Specifically, the present invention uses an image-based 3D human body reconstruction model (such as SMPL) to reconstruct the occupant's body shape. Weight is calculated by estimating the 3D volume and combining it with a human body density formula. To improve estimation accuracy, this process is repeated multiple times.

[0157] After acquiring occupant posture information using an image recognition system, an SMPL model is constructed to fit the occupant's three-dimensional form, generating a 3D mesh model. The 3D mesh is discretized into a number of voxel units using a voxelization method. The occupant's total volume V is calculated. Assuming the average human body density to be ρ, and ignoring individual differences in muscle-to-fat ratios based on statistical and medical data, a visual weight estimate is generated. The actual error in the estimate is approximately ±8%. This estimated visual weight is then weighted and fused with the pressure weight measured by the pressure sensor to obtain a more accurate true weight.

[0158] In some specific embodiments, the fusion of visual weight and measured occupant weight information includes:

[0159] The measured occupant weight information is preprocessed. In order to improve the stability and reliability of the pressure sensor measurement results, the following preprocessing steps are performed on the raw output data:

[0160] Noise filtering: A low-pass filter is used to filter the original pressure signal to eliminate high-frequency noise interference caused by instantaneous fluctuations, seat vibrations, etc.

[0161] Pressure distribution calculation: Based on the distributed pressure sensor array under the seat, the occupant's force area and total pressure are calculated by multiple iterations according to the distribution matrix, thereby obtaining a relatively accurate overall pressure value;

[0162] The formula for fusing the measured occupant weight information and visual weight is as follows:

[0163] ;

[0164] The calculation method of K is as follows:

[0165] ;

[0166] W pressure To measure the passenger's weight information, the total pressure directly obtained by the pressure sensor is converted into weight, W vision is the visual weight, K is the weight coefficient, which is dynamically adjusted according to the sensor confidence, σ vision is the variance of multiple estimations of visual weight, σ pressure The variance of multiple measurements of the pressure sensor;

[0167] This results in a more accurate passenger weight W real , and is divided into three levels: light, normal, and severe (based on pre-set thresholds).

[0168] In some specific embodiments, adjusting the height of the safety belt according to the upper body torso height and sitting angle includes:

[0169] ;

[0170] Q1 is the trunk height weight coefficient, Q2 is the sitting angle compensation coefficient, and C is the vehicle model adaptation constant;

[0171] If you are sitting forward, move the seat belt up 5-10mm to avoid strangulation.

[0172] If the seat belt is reclining, lower it by 5-10mm to prevent the shoulder belt from slipping.

[0173] If it is in an upright position, it remains unchanged.

[0174] Specifically, Q1: Torso height weight coefficient (default value is 0.8);

[0175] Q2: Sitting angle compensation coefficient, set according to sitting posture status:

[0176] If sitting forward: Q2=+0.1;

[0177] If the sitting posture is reclining: Q2=−0.1;

[0178] If sitting upright: Q2=0;

[0179] C: Vehicle adaptation constant, calibrated based on collision test data and used for universal adjustment between different vehicle structures.

[0180] In the prior art, a seat belt height adjuster includes an adjustment mechanism (not shown in the figure), which includes a guide ring assembly, a stepping motor, and a limit sensor;

[0181] Guide ring assembly: A slidable guide ring integrated into the B-pillar (main driver's seat), which achieves vertical movement (travel range: 150-250 mm) by driving the rack and pinion mechanism via a stepper motor.

[0182] Stepper motor: Use closed-loop control stepper motor, equipped with absolute encoder, to provide real-time feedback of guide ring position.

[0183] Limit sensor: Hall sensors are set at both ends of the guide ring track to prevent overtravel.

[0184] In some specific embodiments, such as Figure 2 As shown, a table of preload and force limits is preset based on the gender, age, and weight of the occupants, so that different seat belt preloads and force limits are matched to the gender, age, and weight of different occupants, including:

[0185] The fusion weight class is divided into three levels: light, normal, and heavy; the age class is divided into three levels: juvenile, adult, and senior;

[0186] Weight dimension classification: The passenger weight is divided into three levels:

[0187] Lightweight: <50kg;

[0188] Normal: 50kg–80kg;

[0189] Heavyweight:>80kg.

[0190] Age dimension classification: The age of passengers is divided into three levels:

[0191] Junior level: <18 years old;

[0192] Adult age: 18–50 years old;

[0193] Old age: >50 years old;

[0194] When the seat belt system activates the first level pre-tensioning protection, the output is a fixed pre-tensioning force F f , where F f is the minimum value among the system classification forces;

[0195] When the system detects a low level of collision risk (such as emergency braking, slight collision tendency), it starts the first level pre-tensioning program. At this time, the seat belt system outputs a fixed pre-tensioning force F f This force value is the minimum protective tension common to all occupants, which plays a role in initially eliminating gaps in the webbing and improving the fit, and does not change due to differences in the occupants' physical signs.

[0196] When the seatbelt system activates secondary pretensioning protection:

[0197] According to the weight class and age class, a corresponding gear is selected from the predefined 9 gears of preload force, which satisfies:

[0198] In the same age group, the gear preload of light weight is less than the gear preload of normal weight and less than the gear preload of heavy weight;

[0199] The relationship between gear preloads at different ages within the same weight class is calibrated through collision tests;

[0200] Output the selected secondary preload force;

[0201] This mechanism ensures that the occupant is fixed within an appropriate tension range to avoid physical injuries caused by excessive restraint;

[0202] According to the weight level, the seat belt stepper motor is controlled to drive the force limiter to the preset slot position, so that the force limit value F w satisfy:

[0203] F w1 Corresponding to light weight category, F w2 Corresponding to normal weight level, F w3 corresponds to the heavy weight category, and ;

[0204] Differentiated adjustments based on passenger gender:

[0205] If the occupant is a female, the formulas for all seat belt pretensioners and force limiters are as follows:

[0206] ;

[0207] Among them F female Set the seat belt preload and force limiter for women, F male Seatbelt preload and force limiter limits are set for men, with a weighting factor s ranging from 0.85 to 0.95. To further enhance the system's individual adaptability, the system automatically introduces a gender correction factor when it determines the occupant is female. This correction factor is set based on ergonomic and clinical data to compensate for physiological differences in bone density and muscle tension in female occupants, preventing injuries caused by excessive tension.

[0208] If the occupant is male, keep all strength values unchanged.

[0209] In some specific embodiments, vehicle sensor parameters are obtained and weighted to obtain a comprehensive collision coefficient. The execution of the first-level seat belt pretensioning and the second-level seat belt pretensioning is determined based on a preset threshold of the comprehensive collision coefficient. The seat belt force limiter is executed in the second-level seat belt pretensioning state, including:

[0210] The comprehensive collision coefficient is obtained by weighted calculation using the measurement data of the vehicle-mounted acceleration sensor, gyroscope, and FCW forward collision system. The formula is as follows:

[0211]

[0212] Where a is the vehicle acceleration obtained by the acceleration sensor, θ roll is the turning angle obtained by the vehicle gyroscope, and TTC is the collision time threshold obtained by the FCW forward collision system;

[0213] is the weight coefficient, obtained based on actual vehicle calibration, satisfying μ1+μ2+μ3=1;

[0214] When the vehicle sensor detects that the comprehensive collision risk coefficient R is less than the first threshold, the seat belt level 1 pre-tightening signal is triggered, and the seat belt pre-tightening force is F f ;

[0215] When R> the first threshold and R≤ the second threshold, the secondary pre-tensioning of the seat belt is triggered, and the graded pre-tensioning force F is matched according to the occupant's physical signs. e 、F a 、F o；

[0216] When R>the second threshold, the seat belt is mechanically limited according to the preset gear.

[0217] Specifically, to achieve early detection and response to external vehicle risks, the present invention constructs a comprehensive collision risk assessment mechanism based on multi-source sensor signal fusion technology. Its core is to calculate the real-time collision risk coefficient R, and use this information to trigger seatbelt pretensioning and force limiting control.

[0218] The vehicle's longitudinal acceleration a is monitored in real time by the onboard acceleration sensor to identify the following dynamic behaviors: emergency braking, acceleration shock, and initial impulse changes in a collision.

[0219] Use the gyroscope component in the inertial measurement unit (IMU) to obtain the vehicle's roll attitude deviation angle θ roll , used to assist in judging: high-risk conditions such as vehicle rollover and rolling, abnormal dynamic turning posture, and the possibility of asymmetric collision;

[0220] The distance d and relative speed v between the vehicle and the obstacle ahead are detected in real time by external sensing devices such as the on-board forward-looking camera and millimeter-wave radar, and the collision time threshold is calculated according to the following formula:

[0221] ;

[0222] The on-board ECU dynamically fuses the multi-source sensor data based on the Kalman filter algorithm to calculate the real-time collision risk coefficient R;

[0223] Level 1 preload protection (low-risk level), triggered when: When the real-time calculated comprehensive collision risk coefficient R is at a low level (R ≤ the first threshold), the system determines that the current vehicle is in a low-risk operating condition, such as minor emergency braking or small-angle steering;

[0224] Execution: The calculation unit ECU receives and processes the collision risk data, converts it into a "pre-tensioning signal", and sends it to the retractor motor control module. The system outputs a PWM signal with a fixed duty cycle, controls the motor to drive the retractor at low power, and tightens the seat belt to the first level pre-tensioning force F. f , thereby slightly tightening the webbing, reducing the gap between the occupant and the seatbelt, and improving the fit before emergency response.

[0225] Secondary preload protection (medium and high risk levels)

[0226] Trigger condition: When the risk factor R exceeds the first threshold, and when R> the first threshold and R≤ the second threshold, the system identifies the current driving state as medium-to-high risk (such as sudden obstacles or obvious collision trends);

[0227] Execution operation: The system selects a matching graded preload gear (such as Fe, Fa, Fo, etc.) based on the identified occupant's physical information (gender, age, weight).

[0228] The ECU outputs a PWM duty cycle signal corresponding to the gear position, controlling the retractor motor to output different torques, achieving precise secondary preload tension output and enhancing the occupant fixation effect.

[0229] Force limiter protection (high risk / crash)

[0230] Trigger conditions: When the vehicle is in the second-level preload state and the comprehensive collision risk coefficient R continues to increase or the vehicle actually collides.

[0231] Execute operation: Start the mechanical protection mechanism of the force limiter. According to the result of the physical sign matching, the force limiter is in the corresponding preset gear (such as F w1 、F w2 、F w3 ), when the tension exceeds the threshold, controlled yielding / torsion deformation occurs, limiting the maximum tension of the seat belt to no more than the human body tolerance value, avoiding excessive acceleration of the occupant's chest, and thus reducing secondary injuries.

[0232] In the present invention, the preload force and the force limiting rod are realized by the following methods:

[0233] Force limiting is achieved through a force limiting rod (not shown in the figure, installed on the seat belt retractor). After the occupant gets on the vehicle and fastens the seat belt, the occupant's vital signs are identified and the mechanical force limiting rod is designed as a three-speed adjustable structure (low / medium / high) according to the seat belt graded preload and force limiting rules. Force limiting rods in different gears correspond to different force limiting values, and the force limiting rod can be driven by a stepper motor to the preset slot position.

[0234] Position mapping: Convert the gear position to the target number of stepper motor steps (such as F w1 =500 steps, F w2 =800 steps, F w3 =1100 steps, actually calibrated by experiment), after driving the force limit rod to the relative slot position, the stepper motor stops working, the force limit rod is fixed and works at the moment of car collision.

[0235] The preload is achieved by:

[0236] According to the known rules for graded preload and force limit of seat belts, force classification has been made for the first and second preloads. The preload of the belt can be controlled by controlling the output torque of the DC motor.

[0237] Actuator - DC motor (not shown in the figure, installed in the seat belt retractor)

[0238] A high-response brushless DC motor (rated torque 2.5 N·m, response time ≤ 15 ms) is used to drive the retractor shaft through a gearbox (deceleration and torque increase), and a Hall sensor is equipped to monitor the amount of webbing tightening in real time.

[0239] Motor control: A dedicated control module is integrated into the ECU, with a built-in preload grading algorithm and PWM signal generator, supporting 0-100% duty cycle adjustment. The ECU outputs a PWM signal (duty cycle 0-100%) according to the preload gear size, adjusting the motor current (0-5A) in real time, thereby changing the torque and thus the webbing stress.

[0240] (1) First level preload:

[0241] The ECU sends a fixed duty cycle PWM signal, and the motor tightens the webbing to F with a fixed power. f .

[0242] (2) Secondary preload:

[0243] The target preload force is calculated based on the gear table, and the ECU outputs a PWM duty cycle signal corresponding to the preload force gear. The PID algorithm dynamically adjusts the motor power until the force sensor feedback value matches the target.

[0244] In some embodiments, a system as used in the above embodiments is proposed, combined with Figure 8, including a computing unit 9, a vehicle-mounted camera 6, a weight sensor 8, a seat belt height adjuster 4, a seat belt retractor 5, and a vehicle sensor;

[0245] The vehicle-mounted camera 6, weight sensor 8, seat belt height adjuster 4, seat belt retractor 5, and vehicle sensor 10 are respectively connected to the calculation unit;

[0246] The computing unit is used to implement the method described in the above embodiment;

[0247] The image of the occupant 7 is obtained by the vehicle-mounted camera 6 and the weight information of the occupant 7 is measured by the weight sensor 8 to calculate the gender, age, upper body height, sitting angle and weight of the occupant 7;

[0248] Adjust the seat belt height adjuster 4 according to the upper body trunk height and sitting angle to adjust the seat belt height;

[0249] A pre-tensioning force and force limit table is preset based on the gender, age, and weight of the occupant 7, so that different seat belt pre-tensioning and seat belt force limiter configurations are matched to the gender, age, and weight of the occupant 7. Seat belt pre-tensioning is divided into primary and secondary pre-tensioning according to the degree of urgency.

[0250] Vehicle sensor parameters are obtained through vehicle sensors, and the comprehensive collision coefficient is obtained by weighting. The execution of the first-level seat belt pre-tensioning and the second-level seat belt pre-tensioning is determined according to the preset threshold of the comprehensive collision coefficient. The seat belt force is limited when the seat belt is in the second-level pre-tensioning state.

[0251] Specifically, the computing unit includes a face recognition module, a posture analysis module, and a data fusion module;

[0252] (1) Face recognition module: Based on the YOLOv5 object detection model (input resolution greater than or equal to 1280×720), identify the facial features of occupant 7 and extract gender and age.

[0253] (2) Posture analysis module: Combined with the OpenPose pose estimation algorithm (COCO key point model), it outputs the coordinates of key skeletal points and calculates the upper torso height (error ≤±2cm), sitting angle, and visual weight of occupant 7.

[0254] (3) Data fusion module: The vehicle ECU performs weighted fusion of the visual weight recognized by the camera and the pressure weight data of the weight sensor 8 (the weight coefficient is dynamically adjusted according to the sitting posture of the occupant 7) to comprehensively obtain the actual weight information of the occupant 7.

[0255] like Figure 5-7, a seat belt buckle 31 assembly is provided on one side of the seat 1, and a seat belt height adjuster 4 and a seat belt retractor 5 are provided on the other side of the seat belt;

[0256] The safety belt includes a webbing 2, a safety belt buckle 31, a waist belt fixing buckle 32, a buckle belt fixing buckle 33, a locking tongue 34, a safety belt height adjuster 4, and a safety belt retractor 5;

[0257] The vehicle-mounted camera 6 is arranged in front of the seat 1 on which the passenger 7 sits, and the weight sensor 8 is arranged on the lower side of the seat 1 .

[0258] In other embodiments of the present invention, an electronic device 400 is disclosed. Figure 9 As shown, the electronic device may include: one or more computing units; a memory 402; a display 403; one or more applications (not shown); and one or more computer programs 404. The above components may be connected via one or more communication buses 405. The one or more computer programs 404 are stored in the memory 402 and configured to be executed by the one or more computing units. The one or more computer programs 404 include instructions, which may be used to execute the following: Figures 1 to 5 and each step in the corresponding embodiment.

[0259] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions within the technical scope disclosed in the present invention should be covered by the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A graded pre-tensioning and force-limiting method for seat belts based on occupant vital signs, characterized in that: Including steps: Acquire occupant images and measure occupant weight information to calculate the occupant's gender, age, upper body height, sitting angle, and weight; Adjust the height of the seat belt according to the height of the upper body and the sitting angle; Preset pretension and force limit tables are created based on the occupant's gender, age, and weight, allowing for different seatbelt pretensions and force limits to be matched to the gender, age, and weight of the occupant. Seatbelt pretensioning is divided into primary and secondary pretensioning based on the degree of urgency. The vehicle sensor parameters are obtained and weighted to obtain the comprehensive collision coefficient. The execution of the first-level seat belt pre-tensioning and the second-level seat belt pre-tensioning is determined according to the preset threshold of the comprehensive collision coefficient. The seat belt force limiter is executed when the seat belt is in the second-level pre-tensioning state.

2. The method according to claim 1, characterized in that Acquiring an image of an occupant and measuring the occupant's weight information to calculate the occupant's gender and age includes: Input occupant image, including: Acquiring facial images of occupants; Perform denoising and light equalization preprocessing; Use YOLOv5 to locate the occupant's facial area and crop the facial ROI image; Perform feature extraction, including: The input is fed into the YOLOv5 target detection model to analyze the gender of the occupants and estimate their age.

3. The method according to claim 1, characterized in that Acquiring an image of an occupant and measuring the occupant's weight information to calculate the occupant's upper torso height includes: Obtain system parameters when capturing passenger images, including focal length, installation height, and pitch angle; Use the OpenPose algorithm to detect the coordinates of the occupant's neck and waist key points in real time; Calculate the pixel height of the upper torso using the following formula: ； Convert the upper torso pixel height to the upper torso physical height using the following formula: ； in, is the vertical coordinate of the waist key point, is the vertical coordinate of the neck key point, is the installation height, is the pitch angle, is the focal length.

4. The method according to claim 3, characterized in that Acquiring an image of an occupant and measuring the occupant's weight information to calculate the occupant's sitting angle includes: Input occupant image; The OpenPose algorithm is used to detect the coordinates of the occupant's hip and shoulder key points, and the inclination angle formula of the line connecting the hip and shoulder key points is: ； in, is the ordinate of the key point of the occupant's shoulder, is the horizontal coordinate of the key point of the occupant's shoulder, is the ordinate of the occupant’s hip key point, is the horizontal coordinate of the occupant's hip key point, The tilt angle thresholds for upright, forward and backward tilt are set respectively to judge the occupant's sitting posture based on the occupant's sitting angle.

5. The method according to claim 1, wherein Acquiring an image of an occupant and measuring the occupant's weight information to calculate the occupant's weight includes: Input occupant image; Generate 3D human body mesh based on 3D human body reconstruction model; The 3D human body mesh is voxelized to obtain the number of voxels and the volume of a single voxel. The human body volume is calculated based on the number of voxels and the volume of a single voxel. The formula is as follows: ； Calculate visual weight using the following formula: ： Where N is the number of voxels, is the volume of a single voxel, is the average density of the human body.

6. The method according to claim 5, characterized in that Fusion of visual weight and measured occupant weight information, including: Preprocessing the measured passenger weight information; The formula for fusing the measured occupant weight information and visual weight is as follows: ； The calculation method of K is as follows: ； W pressure To measure the passenger weight information, W vision is the visual weight, K is the weight coefficient, σ vision is the variance of multiple estimations of visual weight, σ pressure The variance of multiple measurements of the pressure sensor.

7. The method according to claim 4, characterized in that Adjust the height of the seat belt according to the height of the upper body and the sitting angle, including: ； Q1 is the trunk height weight coefficient, Q2 is the sitting angle compensation coefficient, and C is the vehicle model adaptation constant; If you are sitting forward, move the seat belt up 5-10mm to avoid strangulation. If the seat belt is reclining, lower it by 5-10mm to prevent the shoulder belt from slipping. If it is in an upright position, it remains unchanged.

8. The method according to claim 1, characterized in that Preset preload and force limit tables based on the occupant's gender, age, and weight to match different seatbelt preloads and force limits to the gender, age, and weight of different occupants, including: The fusion weight class is divided into three levels: light, normal, and heavy; the age class is divided into three levels: juvenile, adult, and senior; When the seat belt system activates the first level pre-tensioning protection, the output is a fixed pre-tensioning force F f , where F f is the minimum value among the system classification forces; When the seatbelt system activates secondary pretensioning protection: According to the weight class and age class, a corresponding gear is selected from the predefined 9 gears of preload force, which satisfies: In the same age group, the gear preload of light weight is less than the gear preload of normal weight and less than the gear preload of heavy weight; The relationship between gear preloads at different ages within the same weight class is calibrated through collision tests; Output the selected secondary preload force; According to the weight level, the seat belt stepper motor is controlled to drive the force limiter to the preset slot position, so that the force limit value F w satisfy: F w1 Corresponding to light weight category, F w2 Corresponding to normal weight level, F w3 Corresponding to the heavy weight level, and F w1 <F w2 <F w3 ; Differentiated adjustments based on passenger gender: If the occupant is a female, the formulas for all seat belt pretensioners and force limiters are as follows: ； Among them F female Set the seat belt preload and force limiter for women, F male Set the seat belt preload and force limiter for males, with the weighting factor s ranging from 0.85 to 0.95; If the occupant is male, keep all strength values unchanged.

9. The method according to claim 1, characterized in that Obtain vehicle sensor parameters and weight them to obtain a comprehensive collision coefficient. Determine whether to implement primary or secondary seat belt pretensioning based on the preset threshold of the comprehensive collision coefficient. When the seat belt is in secondary pretensioning, implement seat belt force limiting, including: The comprehensive collision coefficient is obtained by weighted calculation using the measurement data of the vehicle-mounted acceleration sensor, gyroscope, and FCW forward collision system. The formula is as follows: Where a is the vehicle acceleration obtained by the acceleration sensor, θ roll is the turning angle obtained by the vehicle gyroscope, and TTC is the collision time threshold obtained by the FCW forward collision system; is the weight coefficient; When the vehicle sensor detects that the comprehensive collision risk coefficient R is less than the first threshold, the seat belt level 1 pre-tightening signal is triggered, and the seat belt pre-tightening force is F f ; When R> the first threshold and R≤ the second threshold, the secondary pre-tensioning of the seat belt is triggered, and the graded pre-tensioning force F is matched according to the occupant's physical signs. e 、F a 、F o； When R>the second threshold, the seat belt is mechanically limited according to the preset gear.

10. A system used in the method according to any one of claims 1 to 9, characterized in that: Including computing unit, vehicle-mounted camera, weight sensor, seat belt height adjuster, seat belt retractor, vehicle sensor; The vehicle-mounted camera, weight sensor, seat belt height adjuster, seat belt retractor, and vehicle sensor are respectively connected to the computing unit; The computing unit is configured to implement the method according to any one of claims 1 to 9; The occupant's image is acquired through the vehicle's onboard camera, and the occupant's weight information is measured through the weight sensor to calculate the occupant's gender, age, upper body height, sitting angle, and weight; Adjust the seat belt height adjuster according to the upper body torso height and sitting angle to adjust the seat belt height; Preset pretension and force limit tables are created based on the occupant's gender, age, and weight, allowing for matching seatbelt pretensioning, force limiter, and seatbelt retractor configurations to suit the occupant's gender, age, and weight. Seatbelt pretensioning is divided into primary and secondary pretensioning based on the degree of urgency. Vehicle sensor parameters are obtained through vehicle sensors, and the comprehensive collision coefficient is obtained by weighting. The execution of the first-level seat belt pre-tensioning and the second-level seat belt pre-tensioning is determined according to the preset threshold of the comprehensive collision coefficient. The seat belt force is limited when the seat belt is in the second-level pre-tensioning state.

Images