A vehicle safety assurance method and system based on artificial intelligence

Abstract

This application provides a vehicle safety assurance method and system based on artificial intelligence. Specifically, it uses piezoelectric elements in the vehicle to collect real-time correlation characteristics between tire tread deformation amplitude and vibration frequency, simultaneously acquiring driving operation data to form a multi-dimensional operation sequence. The tire tread vibration characteristics are dynamically coupled with the steering wheel angular rate to generate composite safety parameters reflecting the road adhesion coefficient and steering inertia deviation. The sampling frequency and signal gain of the piezoelectric elements are dynamically adjusted according to the parameter change amplitude, ensuring that the matching relationship between tire tread deformation and steering wheel operation adapts to road conditions in real time, forming a parameter adaptation state. When the parameter change amplitude exceeds limits during continuous steering or emergency braking, the power torque growth slope is limited and the braking pulse intensity is increased based on the parameter adaptation state, simultaneously optimizing the control priority of lateral slippage and longitudinal braking. The technical solution provided in this application improves handling stability and braking safety under complex road conditions.

Description

Technical Field

[0001] This application relates to the field of vehicle safety technology, and in particular to a vehicle safety assurance method and system based on artificial intelligence. Background Technology

[0002] Despite the widespread adoption of intelligent driving assistance systems, traffic accidents caused by driver distraction or dangerous maneuvers remain frequent. In complex urban road or highway scenarios, it is necessary to identify abnormal driver behavior and predict potential risks in real time, triggering multi-level warnings (audio-visual alerts, seat vibration, automatic speed limits) within 0.5 to 3 seconds before an accident occurs to compensate for the response delay deficiencies of traditional passive safety systems.

[0003] Current mainstream solutions employ vision-based driver behavior monitoring systems. These systems utilize infrared cameras and image recognition algorithms deployed in the cockpit to capture the driver's facial features (such as eyelid closure frequency and head turning angle) and hand movements (such as the duration of time the driver is off the steering wheel). This data, combined with vehicle sensor data (steering wheel angle fluctuations and lane departure rates), forms a distraction index model. When the distraction index exceeds a threshold, the system triggers a warning via the in-vehicle voice system, and some systems can activate the emergency lane keeping function.

[0004] The existing solutions suffer from several drawbacks. They heavily rely on lighting conditions and camera field of view clarity, with recognition rates dropping by over 60% in backlight, at night, or when the driver is wearing sunglasses / masks. Furthermore, the image recognition algorithms are insufficient in distinguishing between subtle movements (such as briefly looking down to check navigation) and genuinely dangerous maneuvers (such as intentional serpentine driving), resulting in a false alarm rate as high as 35%. In addition, the coordination between the warning mechanism and vehicle dynamic control is weak, failing to dynamically adjust warning thresholds based on real-time road conditions (such as curve curvature and distance to the vehicle ahead), leading to delayed warnings or excessive interference with driving in complex scenarios. Summary of the Invention

[0005] This application provides a vehicle safety assurance method and system based on artificial intelligence to address the problems in the prior art.

[0006] Firstly, this application provides a vehicle safety assurance method based on artificial intelligence, including:

[0007] Multiple piezoelectric elements distributed on the vehicle chassis and tire inner wall are used to collect response signals generated by the vehicle's contact with the road surface during driving. The response signals include the correlation characteristics between the tread deformation amplitude and vibration frequency under different road conditions.

[0008] The system continuously records the rate of change of steering wheel angle, the increase in brake pedal travel, and the interval of accelerator pedal triggering through the in-cockpit operating device, and forms a multi-dimensional operating mode sequence.

[0009] The correlation characteristics between the tread deformation amplitude and vibration frequency in the response signal are coupled with the steering wheel angle change rate in the multi-dimensional operation mode sequence, and a composite safety parameter including the road adhesion coefficient and steering inertia deviation is generated based on the coupling result.

[0010] Based on the change range of the composite safety parameters during vehicle operation, the sampling frequency of the piezoelectric element on the inner wall of the tire and the signal gain of the piezoelectric element in the chassis are adjusted so that the matching relationship between the tread deformation amplitude and the steering wheel angle change rate based on the adjustment results is adapted to the current road conditions, and a parameter adaptation state is generated.

[0011] When the magnitude of the change exceeds a preset critical range in continuous steering or emergency braking scenarios, based on the parameter adaptation state, the linear growth slope of the power output torque in the vehicle power system is limited and the pulse response intensity of the vehicle brake hydraulic system is enhanced, thereby simultaneously correcting the coordinated control priority of lateral slip and longitudinal braking.

[0012] Optionally, the step of coupling the correlation characteristics between the tread deformation amplitude and vibration frequency in the response signal with the steering wheel angle change rate in the multi-dimensional operation mode sequence, and generating a composite safety parameter including the road adhesion coefficient and steering inertia deviation based on the coupling result, includes:

[0013] Based on the correlation characteristics between the tread deformation amplitude and the vibration frequency of the response signal, the pressure distribution characteristics of the contact area between the vehicle tire and the road surface are extracted. The pressure distribution characteristics are jointly characterized by the extreme value distribution law of the tread deformation amplitude and the periodic fluctuation trajectory of the vibration frequency.

[0014] The rate of change of steering wheel angle in the multi-dimensional operation mode sequence is divided into a positive increasing segment and a negative decreasing segment according to the direction of steering operation, and the maximum instantaneous slope value of the rate of change of steering wheel angle in each segment is extracted to form a steering offset feature.

[0015] The periodic fluctuation trajectory of the vibration frequency in the pressure distribution feature is time-series superimposed with the maximum instantaneous slope value of the steering offset feature, and a fused waveform is generated based on the superposition result. During the superposition process, the waveform amplitude of the pressure distribution feature is scaled proportionally as the instantaneous slope value of the steering offset feature increases.

[0016] The amplitude abrupt change interval of the periodic fluctuation trajectory in the fused waveform is divided into intervals. When the duration of the amplitude abrupt change interval exceeds a preset multiple of the single contact cycle between the tire and the road surface, the road adhesion coefficient corresponding to the extreme value distribution pattern in the amplitude abrupt change interval and the steering inertia deviation corresponding to the extreme value of the slope change of the fused waveform are extracted. The product of the road adhesion coefficient and the steering inertia deviation is used as a composite safety parameter.

[0017] Optionally, adjusting the sampling frequency of the piezoelectric element on the tire inner wall and the signal gain of the piezoelectric element in the chassis based on the change amplitude of the composite safety parameters during vehicle operation, so that the matching relationship between the tread deformation amplitude and the steering wheel angle change rate based on the adjustment result adapts to the current road conditions, and generating a parameter adaptation state, includes:

[0018] The variation range of the composite safety parameters is monitored. By distinguishing the proportion of high-frequency fluctuation range and low-frequency gradual change range in the variation range, the sampling frequency adjustment amount of the piezoelectric element on the inner wall of the tire is determined so as to adjust the sampling frequency. When the proportion of high-frequency fluctuation range increases, the sampling frequency is increased, and when the proportion of low-frequency gradual change range increases, the sampling frequency is decreased.

[0019] Based on the changing trend of the low-frequency gradual change range, determine whether the cumulative change of the tread deformation amplitude exceeds the deformation threshold range corresponding to the current road condition. If it exceeds, increase the signal gain of the chassis piezoelectric element step by step according to a preset ratio until the tread deformation amplitude returns to the deformation threshold range.

[0020] The adjusted sampling frequency and the enhanced signal gain are input into the matching model. The matching model is used to calculate the time-series correlation between the steering wheel angle change rate and the tread deformation amplitude to generate matching weight coefficients.

[0021] Based on the matching weight coefficient, the sampling interval uniformity of the piezoelectric element on the inner wall of the tire and the signal amplification smoothness of the piezoelectric element in the chassis are adjusted synchronously. According to the adjustment result, the matching relationship between the tread deformation amplitude and the steering wheel angle change rate is adapted to the current road conditions, and a parameter adaptation state is generated.

[0022] Optionally, when the change amplitude exceeds a preset critical range in continuous steering or emergency braking scenarios, based on the parameter adaptation state, the coordinated control priority of lateral slip and longitudinal braking is simultaneously corrected by limiting the linear growth slope of the power output torque in the vehicle power system and enhancing the pulse response intensity of the vehicle brake hydraulic system, including:

[0023] Detect whether the change amplitude exceeds a preset critical range. If it does, identify the current scenario as a continuous steering scenario or an emergency braking scenario, and generate a corresponding scenario identifier based on the parameter adaptation state.

[0024] If the scenario identifier is a continuous steering scenario identifier, the linear growth slope of the power output torque of the vehicle power system is divided into a positive increasing stage and a negative pullback stage according to the direction of steering angle change. In the positive increasing stage, the slope adjustment step size is gradually reduced, and in the negative pullback stage, the slope recovery rate is compressed according to a preset ratio.

[0025] If the scenario identifier is an emergency braking scenario identifier, the pulse interval reference value of the vehicle brake hydraulic system is set according to the instantaneous change rate of the brake pedal travel increment in the vehicle power system, and the change amplitude is converted into a pulse pressure amplification coefficient proportionally based on the pulse interval reference value.

[0026] The reduced slope adjustment step size, the compressed slope recovery rate, and the pulse pressure amplification coefficient are used as weighting factors input into the allocation model. The priority weights of the vehicle lateral slip control signal and the longitudinal braking control signal are then redistributed through the allocation model.

[0027] Optionally, the step of time-series superimposing the periodic fluctuation trajectory of the vibration frequency in the pressure distribution characteristics with the maximum instantaneous slope value of the steering offset characteristics, and generating a fused waveform based on the superimposition result, includes:

[0028] Based on the periodic fluctuation trajectory of the vibration frequency in the pressure distribution characteristics, the tire and road surface are divided into equal-length time periods according to the single contact period. The peak points and valley points within the equal-length time periods are extracted to form a fluctuation peak-valley sequence.

[0029] Based on the time point at which the maximum instantaneous slope value of the steering offset feature is located, the reference amplitude of the fluctuation peak-valley sequence within the time period corresponding to the time point is determined, and the ratio of the maximum instantaneous slope value to the reference amplitude is used as the amplitude scaling factor.

[0030] Within the equal time period, with the reference amplitude as the center value, the peak amplitude of the fluctuation peak-valley sequence is expanded forward according to the amplitude scaling factor, and the valley amplitude is compressed backward according to the amplitude scaling factor to generate a scaled waveform segment.

[0031] The scaled waveform segments are joined end-to-end in chronological order to form a continuous waveform segment. The amplitude jump regions at the junctions of the continuous waveform segments are smoothed by slope transition, and a fused waveform is formed based on the smoothing result.

[0032] Optionally, the step of inputting the reduced slope adjustment step size, the compressed slope recovery rate, and the pulse pressure amplification coefficient as weighting factors into the allocation model, and then redistributing the priority weights of the vehicle lateral slip control signal and the longitudinal braking control signal through the allocation model, includes:

[0033] The reduced slope adjustment step size is converted into the suppression coefficient of lateral slip control, the compressed slope recovery rate is converted into the compensation coefficient of lateral slip control, and the pulse pressure amplification coefficient is converted into the superposition coefficient of longitudinal braking control.

[0034] In the allocation model, independent horizontal weight channels and vertical weight channels are divided. The horizontal weight channel receives the difference between the suppression coefficient and the compensation coefficient, and the vertical weight channel receives the cumulative value of the superposition coefficient.

[0035] When the scene identifier is a continuous steering scene identifier, the difference is converted into a weight allocation ratio adjustment instruction for the lateral slip control signal based on the positive and negative characteristics of the steering angle change direction.

[0036] When the scenario identifier is an emergency braking scenario identifier, the accumulated value is mapped to a weight allocation ratio enhancement instruction for the longitudinal braking control signal based on the rate of change of the brake pedal travel increment.

[0037] Optionally, the step of continuously recording the steering wheel angle change rate, brake pedal travel increment, and accelerator pedal trigger interval via the in-cabin operating device, and forming a multi-dimensional operating mode sequence, includes:

[0038] A three-axis motion sensor group is installed on the steering wheel shaft, brake pedal linkage and the bottom of the accelerator pedal to capture the steering wheel angular displacement, brake pedal depressing stroke and accelerator pedal contact trigger status.

[0039] The absolute value of the angle change is calculated for the steering wheel angular displacement in a unit time window, and the steering wheel angular change rate is constructed based on the temporal characteristics of the absolute value of the angle change.

[0040] The brake pedal depressing stroke is divided into stepped incremental marker points according to a preset displacement interval, and the time difference between the stepped incremental marker points is recorded to generate the brake pedal stroke increment;

[0041] The contact trigger state is converted into a contact interval timing signal. The interval duration of the contact interval timing signal is calculated as the accelerator pedal trigger interval. The steering wheel angle change rate, the brake pedal travel increment, and the accelerator pedal trigger interval are arranged in parallel according to the operation occurrence time. A multi-dimensional operation mode sequence with time synchronization is generated based on the arrangement result.

[0042] Secondly, this application provides a vehicle safety assurance system based on artificial intelligence, including:

[0043] The acquisition module is used to acquire response signals generated by the vehicle's contact with the road surface during driving through multiple piezoelectric elements distributed on the vehicle chassis and tire inner wall. The response signals include the correlation characteristics between the tread deformation amplitude and vibration frequency under different road conditions.

[0044] The acquisition module is used to continuously record the rate of change of steering wheel angle, the increase in brake pedal travel, and the trigger interval of accelerator pedal through the in-cabin operating device, and form a multi-dimensional operating mode sequence.

[0045] The coupling module is used to couple the correlation characteristics between the tread deformation amplitude and vibration frequency in the response signal with the steering wheel angle change rate in the multi-dimensional operation mode sequence, and generate a composite safety parameter including the road adhesion coefficient and steering inertia deviation based on the coupling result.

[0046] The adjustment module is used to adjust the sampling frequency of the piezoelectric element on the inner wall of the tire and the signal gain of the piezoelectric element in the chassis according to the change range of the composite safety parameters during vehicle driving, so that the matching relationship between the tread deformation amplitude and the steering wheel angle change rate based on the adjustment result is adapted to the current road conditions, and a parameter adaptation state is generated.

[0047] The correction module is used to simultaneously correct the coordinated control priority of lateral slip and longitudinal braking of the vehicle when the change amplitude exceeds a preset critical range under continuous steering or emergency braking scenarios, based on the parameter adaptation state, by limiting the linear growth slope of the power output torque in the vehicle power system and enhancing the pulse response intensity of the vehicle brake hydraulic system.

[0048] Thirdly, embodiments of this application provide a computing device, including a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are invoked and executed by the processing component to implement an artificial intelligence-based vehicle safety assurance method as described in the first aspect above.

[0049] Fourthly, embodiments of this application provide a computer storage medium storing a computer program, which, when executed by a computer, implements an artificial intelligence-based vehicle safety assurance method as described in the first aspect.

[0050] In this embodiment, multiple piezoelectric elements distributed on the vehicle chassis and tire inner walls collect response signals generated during vehicle operation when in contact with the road surface. These response signals include correlation characteristics between tire tread deformation amplitude and vibration frequency under different road conditions. The steering wheel angle change rate, brake pedal travel increment, and accelerator pedal trigger interval are continuously recorded via an in-cab operating device, forming a multi-dimensional operating mode sequence. The correlation characteristics between tire tread deformation amplitude and vibration frequency in the response signals are coupled with the steering wheel angle change rate in the multi-dimensional operating mode sequence. Based on the coupling result, a composite safety factor including road adhesion coefficient and steering inertia deviation is generated. Full parameters; based on the variation range of the composite safety parameters during vehicle operation, the sampling frequency of the piezoelectric element on the inner wall of the tire and the signal gain of the piezoelectric element in the chassis are adjusted so that the matching relationship between the tread deformation amplitude and the steering wheel angle change rate based on the adjustment results is adapted to the current road conditions, generating a parameter adaptation state; when the variation range exceeds the preset critical range in continuous steering or emergency braking scenarios, based on the parameter adaptation state, the linear growth slope of the power output torque in the vehicle power system is limited and the pulse response intensity of the vehicle brake hydraulic system is enhanced, and the coordinated control priority of lateral slip and longitudinal braking of the vehicle is simultaneously corrected.

[0051] The technical solution of this application has the following beneficial effects:

[0052] By synchronously capturing the dynamic deformation and vibration characteristics of the tire tread using piezoelectric elements at multiple locations, a real-time sensing capability for the tire-road contact mechanical characteristics is established, providing high-frequency dynamic response data for road condition identification. Based on the quantitative analysis of the timing of steering wheel and brake / accelerator pedal operations, a multi-dimensional dynamic pattern profile of driving behavior is constructed, reflecting the driver's operational intentions and operational stability. Through cross-domain coupling of tire tread vibration characteristics and steering wheel turning rate, dual safety elements of vehicle dynamics (steering inertia) and environmental perception (road adhesion) are integrated to generate a global safety index that can be quantified for risk assessment. The piezoelectric sensing parameters (sampling frequency / gain) are dynamically optimized according to changes in safety parameters to achieve adaptive calibration of the matching relationship between tire tread state and steering wheel operation, improving the real-time matching degree between data acquisition and road conditions. In emergency scenarios, power output and braking response are synchronously intervened based on parameter adaptation status, and the lateral / longitudinal control weight allocation is reconstructed to solve the dynamic instability problem caused by the single-dimensional control of traditional systems.

[0053] Furthermore, the tire-road contact pressure distribution characteristics are characterized by extracting the extreme distribution of tread deformation amplitude and the vibration frequency fluctuation trajectory. Steering offset features are extracted by dividing the steering wheel angular rate into forward / reverse segments. The pressure distribution waveform is scaled proportionally to the instantaneous slope value of the steering offset to generate a fused waveform. The duration threshold of the waveform amplitude abrupt change interval is detected, and the road adhesion coefficient corresponding to the extreme tread distribution and the steering inertia deviation corresponding to the extreme waveform slope within that interval are extracted. Composite safety parameters are generated through multiplication. Through asymmetric waveform scaling and abrupt change interval detection mechanisms, the coupling relationship between road adhesion capability and steering dynamic deviation is accurately quantified, achieving physically interpretable generation of composite safety parameters. Waveform processing rules based on tire contact cycle constraints ensure that parameter generation is strictly synchronized with the actual vehicle motion state, providing a dual decision-making basis for real-time safety control, combining environmental perception and operational intent analysis.

[0054] These or other aspects of this application will become more apparent in the following description of the embodiments. Attached Figure Description

[0055] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0056] Figure 1 A flowchart of an artificial intelligence-based vehicle safety assurance method provided in this application is shown;

[0057] Figure 2 A schematic diagram of the structure of an artificial intelligence-based vehicle safety system provided in this application is shown;

[0058] Figure 3 A schematic diagram of the structure of a computing device provided in this application is shown. Detailed Implementation

[0059] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0060] In some of the processes described in the specification, claims, and accompanying drawings of this application, multiple operations appearing in a specific order are included. However, it should be clearly understood that these operations may not be executed in the order they appear herein, or may be executed in parallel. The operation numbers, such as 101, 102, etc., are merely used to distinguish different operations and do not themselves represent any execution order. Furthermore, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel. It should be noted that the descriptions such as "first," "second," etc., in this document are used to distinguish different messages, devices, modules, etc., and do not represent a chronological order, nor do they limit "first" and "second" to different types.

[0061] Researchers have discovered that existing vehicle safety systems face a mismatch between real-time tire tread dynamic response and driving operation under complex road conditions. This is particularly evident in scenarios involving continuous steering or emergency braking, where the disconnect between sensor data and temporal analysis of driving behavior leads to delayed or excessive safety control decisions, hindering multi-dimensional collaborative optimization. To address this, an AI-based vehicle safety assurance method is proposed. Specifically, a piezoelectric sensor network captures tire tread deformation and vibration characteristics in real time, simultaneously constructing a multi-dimensional temporal sequence of driving operations. This dynamically couples tire tread response characteristics with steering wheel speed to generate composite safety parameters, and adaptively adjusts the sensor acquisition strategy based on parameter changes. Finally, in emergency scenarios, the lateral and longitudinal control priorities are simultaneously corrected. This method achieves millisecond-level adaptive matching between vehicle dynamic response and road conditions through cross-domain collaborative analysis of piezoelectric signals and driving behavior, significantly improving stability control accuracy and braking safety under extreme driving conditions.

[0062] The technical solution of this application can be applied to scenarios involving driver distraction or dangerous operation warnings.

[0063] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0064] Figure 1 A flowchart of an artificial intelligence-based vehicle safety assurance method is provided as an embodiment of this application, such as... Figure 1 As shown, the method includes:

[0065] 101. By using multiple piezoelectric elements distributed on the vehicle chassis and inner wall of the tires, the response signals generated by the vehicle contacting the road surface during driving are collected. The response signals include the correlation characteristics between the tread deformation amplitude and vibration frequency under different road conditions.

[0066] In this step, the response signal refers to the electrical signal generated when the piezoelectric elements distributed on the vehicle chassis and inner wall of the tire sense the contact between the tire tread and the road surface. It includes the correlation data between the tread deformation amplitude (the physical quantity change of tire deformation under pressure) and the vibration frequency (the speed characteristics of the periodic vibration of the tire tread).

[0067] In this embodiment, firstly, the piezoelectric element captures the original electrical signals generated by tread deformation and vibration in real time; secondly, the extreme points of deformation amplitude (such as maximum and minimum values) are extracted from the original electrical signals using peak detection technology, and the dominant frequency component of vibration frequency is separated using spectrum analysis technology; thirdly, the deformation amplitude range (difference between maximum and minimum values) and the dominant frequency component are aligned according to a time window and encapsulated into a response signal containing road condition characteristics.

[0068] Suppose that the driver is distracted and fails to avoid a pothole in time. The piezoelectric element of the right front tire detects a sudden increase in deformation amplitude (0.5mm→1.8mm) and high-frequency vibration (40Hz→95Hz), triggering a response signal marked as "abnormal impact road condition".

[0069] 102. The steering wheel angle change rate, brake pedal travel increment and accelerator pedal trigger interval are continuously recorded through the cockpit operating device, and a multi-dimensional operating mode sequence is formed.

[0070] In this step, the multidimensional operation mode sequence is a time-series dataset composed of the steering wheel angle change rate (absolute value of steering wheel angle change per unit time), brake pedal travel increment (change in travel per single pedal press), and accelerator pedal trigger interval (time interval between two pedal presses).

[0071] In this embodiment, firstly, the angular displacement is recorded by the angular velocity sensor of the steering wheel shaft, and the absolute value of the angle difference between adjacent sampling points is calculated according to a 50ms time window to generate a sequence of angular change rate. Secondly, the travel interval of the brake pedal displacement sensor data is divided according to a preset threshold (e.g., 10mm), and the time difference between adjacent intervals is recorded to generate the brake pedal travel increment. Thirdly, the accelerator pedal contact state is timed to obtain the accelerator pedal trigger interval. Finally, the timing data of the three types of actions are aligned according to the timestamp to construct a multi-dimensional operation mode sequence.

[0072] For example, continuing the previous example, the driver panicked on a bumpy road, and the rate of change of the steering wheel angle soared from 5° / s to 22° / s in 0.5 seconds. The increase in the brake pedal travel reached 80mm / s, and the interval between accelerator pedal triggering was irregular (0.3 seconds → 1.2 seconds → 0.4 seconds). The constructed multi-dimensional operation mode sequence presents dangerous operation characteristics of "high-frequency steering + sudden braking + accidental acceleration".

[0073] 103. The correlation characteristics between the tread deformation amplitude and vibration frequency in the response signal are coupled with the steering wheel angle change rate in the multi-dimensional operation mode sequence, and a composite safety parameter including the road adhesion coefficient and steering inertia deviation is generated based on the coupling result.

[0074] In this step, the composite safety parameter is a comprehensive index generated by the dynamic coupling of tread deformation / vibration characteristics and steering wheel speed, including road adhesion coefficient (tire-road friction characteristics) and steering inertia deviation (the difference between actual steering torque and theoretical value).

[0075] In this embodiment, firstly, the tread vibration frequency fluctuation trajectory is time-aligned with the steering wheel angle change rate, and phase deviation caused by operation delay is eliminated; secondly, the sliding window weighting method is used to multiply the time alignment result with the mean of the steering wheel angle change rate to generate an instantaneous coupling coefficient; thirdly, noise interference in the instantaneous coupling coefficient is removed by dynamic threshold filtering, and combined with the pre-calibrated mapping relationship table of road surface adhesion and steering inertia, composite safety parameters are output.

[0076] For example, continuing the previous example, firstly, the vibration trajectory amplitude (15%) after time alignment is multiplied with the average steering wheel speed (22° / s) to generate an instantaneous coupling coefficient of 3.3 (0.15×22); then, noise interference is removed by dynamic threshold filtering (coefficients <1.0 are filtered out), and combined with the pre-calibrated mapping table, the coupling coefficient of 3.3 is converted into a road surface adhesion coefficient of 0.25 (low adhesion) and a steering inertia deviation of +18% (steering torque overshoot), and finally, the composite safety parameter value is output as 0.25×18=4.5, triggering a dual risk warning of "low adhesion road surface + oversteering".

[0077] 104. Based on the change range of the composite safety parameters during vehicle operation, adjust the sampling frequency of the piezoelectric element on the inner wall of the tire and the signal gain of the piezoelectric element on the chassis, so that the matching relationship between the tread deformation amplitude and the steering wheel angle change rate based on the adjustment result is adapted to the current road conditions, and generate a parameter adaptation state.

[0078] In this step, the parameter adaptation state is a quantitative state characterizing whether the matching degree between tire tread deformation and steering wheel operation is adapted to the current road conditions after adjusting the tire piezoelectric sampling frequency (data acquisition density) and chassis piezoelectric signal gain (signal amplification factor).

[0079] In this embodiment, firstly, the variation amplitude of the composite safety parameter is decomposed in the frequency domain to separate high-frequency fluctuations (>5Hz) and low-frequency slow changes (<1Hz) components; if the high-frequency proportion exceeds 60%, the tire piezoelectric sampling frequency is increased to capture transient deformation details; if the low-frequency trend continues to exceed the threshold (e.g., an increase of >25% within 30 seconds), the gain of the chassis piezoelectric signal is increased exponentially (e.g., 1→4 times), and finally, the parameter adaptation state is generated.

[0080] For example, continuing the previous example, if the vehicle detects that the composite safety parameter is continuously exceeding the limit at a low frequency (adhesion coefficient 0.25 for 15 seconds) during driving, the chassis piezoelectric gain is automatically increased to 3 times to enhance the ability to capture weak vibration signals on wet and slippery roads and generate a parameter adaptation state of "low frequency gain mode".

[0081] 105. When the change amplitude exceeds the preset critical range in continuous steering or emergency braking scenarios, based on the parameter adaptation state, the linear growth slope of the power output torque in the vehicle power system is limited and the pulse response intensity of the vehicle brake hydraulic system is enhanced, thereby simultaneously correcting the coordinated control priority of lateral slip and longitudinal braking of the vehicle.

[0082] In this step, the priority of coordinated control is the weight allocation relationship between lateral slip control (anti-skid) and longitudinal braking control (deceleration), with the one with higher weight responding first.

[0083] In this embodiment, firstly, pattern recognition is used to determine whether the changes in composite safety parameters match the scenarios of continuous steering (sine wave) or emergency braking (step change). During continuous steering, the increase in power torque is limited by the slope, and the lateral control weight is increased based on the gain data in the parameter adaptation state. During emergency braking, the braking force is increased by a multiple of the pulse pressure amplification coefficient, and the longitudinal control weight is increased. Finally, multi-dimensional collaborative intervention is achieved through dynamic weight allocation.

[0084] For example, continuing the previous example, in the oversteer scenario on a low-adhesion road surface, the power output torque slope of the vehicle's power system is limited to 0.5% / s, the lateral control weight is increased to 75%, and the braking pulse intensity is increased by 1.8 times, forcibly correcting the vehicle's lateral slippage and shortening the braking distance by 1.2 meters.

[0085] Steps 101-105 utilize deep fusion of piezoelectric sensing data and driving operation timing to generate real-time composite safety parameters reflecting road conditions and operational risks. Based on these parameters, the sensing acquisition strategy is dynamically optimized to achieve precise matching between tire tread status and steering wheel control. In distracted or dangerous operating scenarios, the execution priority of lateral and longitudinal control is dynamically adjusted through dual intervention of power output slope limitation and braking pulse intensity enhancement. This significantly improves vehicle handling stability and active safety in complex scenarios such as low-traction surfaces, continuous steering, or emergency braking, resolving the core defects of traditional solutions such as sensor response delay and disconnected control decisions.

[0086] To address the challenges of traditional vehicle safety systems in accurately quantifying the coupling relationship between tire tread dynamic response and steering operation under distracted driving scenarios, and their inability to adaptively adjust safety decision-making logic based on real-time road conditions, some embodiments involve coupling the correlation characteristics between tire tread deformation amplitude and vibration frequency in the response signal with the steering wheel angle change rate in the multi-dimensional operation mode sequence. Based on the coupling result, a composite safety parameter including the road adhesion coefficient and steering inertia deviation is generated, including:

[0087] 201. Based on the correlation characteristics between the tread deformation amplitude and the vibration frequency of the response signal, extract the pressure distribution characteristics of the contact area between the vehicle tire and the road surface. The pressure distribution characteristics are jointly characterized by the extreme value distribution law of the tread deformation amplitude and the periodic fluctuation trajectory of the vibration frequency.

[0088] In step 201, the pressure distribution characteristics are characterized by the extreme value distribution law of the tread deformation amplitude and the periodic fluctuation trajectory of the vibration frequency, which together represent the dynamic pressure change in the contact area between the tire and the road surface, and are used to reflect the uniformity of the tread force and the feedback characteristics of the road surface to the tire.

[0089] In this embodiment, the tread deformation amplitude data is first divided into spatial grids, and the number of deformation extreme points in each grid cell is counted to generate an extreme value density distribution map. Secondly, the vibration frequency sequence is periodically trajectory fitted to extract the main frequency fluctuation amplitude (e.g., ±8%) and phase offset angle (e.g., 30° hysteresis). Finally, the extreme value density distribution map is superimposed with the main frequency fluctuation amplitude and phase offset angle to generate a three-dimensional pressure distribution heat map, in which the high-density extreme value area corresponds to the tread pressure concentration area.

[0090] 202. Divide the rate of change of steering wheel angle in the multi-dimensional operation mode sequence into a positive increasing segment and a negative decreasing segment according to the direction of steering operation, and extract the maximum instantaneous slope value of the rate of change of steering wheel angle in each segment to form a steering offset feature.

[0091] In step 202, the steering offset feature is obtained by dividing the steering direction according to the rate of change of steering wheel angle and extracting the maximum instantaneous slope in each segment to quantify the dynamic offset trend of steering operation.

[0092] In this embodiment, the sequence of steering wheel angle change rates is first directionally marked (clockwise is positive, counterclockwise is negative), and the sequence is divided into increasing segments (positive slope continuous segment) and decreasing segments (negative slope continuous segment) according to continuous same-direction operation; secondly, an extreme value detection algorithm is used in each segment to capture the maximum instantaneous slope value (e.g., +12° / s²) and its corresponding timestamp; finally, the maximum slope values ​​are arranged in chronological order to generate steering offset features with directional markings.

[0093] 203. The periodic fluctuation trajectory of the vibration frequency in the pressure distribution feature is time-series superimposed with the maximum instantaneous slope value of the steering offset feature, and a fused waveform is generated based on the superposition result, wherein the waveform amplitude of the pressure distribution feature is scaled proportionally as the instantaneous slope value of the steering offset feature increases during the superposition process;

[0094] In step 203, the fused waveform is a composite waveform generated by superimposing the vibration frequency fluctuation trajectory in the pressure distribution characteristics and the maximum instantaneous slope value of the steering offset characteristics along the time axis. Its amplitude scales linearly as the steering offset slope value increases, and it is used to characterize the dynamic correlation between tread vibration and steering strength.

[0095] In this embodiment, the time axis of the vibration frequency fluctuation trajectory (e.g., a 50Hz sine wave) is first aligned with the pulse sequence of the steering offset characteristics to ensure that the wave peak matches the pulse time point. The scaling ratio is calculated based on the maximum slope value of each pulse point (e.g., slope value / reference value 10° / s² → scaling factor 1.2), and the amplitude of the fluctuation trajectory within the corresponding time window is dynamically adjusted (e.g., the original amplitude ±5% is adjusted to ±6%). Finally, the scaled waveform amplitudes are spliced ​​together in time sequence to generate a fused waveform with dynamically changing amplitude.

[0096] 204. Divide the amplitude abrupt change interval of the periodic fluctuation trajectory in the fused waveform into intervals. When the duration of the amplitude abrupt change interval exceeds a preset multiple of the single contact cycle between the tire and the road surface, extract the road adhesion coefficient corresponding to the extreme value distribution pattern within the amplitude abrupt change interval, and the steering inertia deviation corresponding to the extreme value of the slope change of the fused waveform. Use the product of the road adhesion coefficient and the steering inertia deviation as a composite safety parameter.

[0097] In step 204, the composite safety parameter is the product of the road adhesion coefficient (reflecting the friction characteristics between the tire and the road surface) and the steering inertia deviation (reflecting the steering torque imbalance) extracted from the amplitude abrupt change range of the fused waveform, and is used to assess the overall safety risk level of the vehicle.

[0098] In this embodiment, firstly, continuous amplitude abrupt change intervals in the fused waveform exceeding a preset threshold (e.g., ±7%) are detected; secondly, the ratio of the duration of the continuous amplitude abrupt change interval to the tire's single contact cycle (e.g., 0.1 seconds) is calculated, and if the ratio is greater than 2.5 times, it is determined to be a valid interval; then, the median of the extreme value density distribution map within the interval (e.g., 4 points / cm²) is extracted and mapped to the road surface adhesion coefficient (4→0.4), and the maximum slope of the waveform (e.g., +18% / s) is extracted and mapped to the steering inertia deviation (18→+15%); finally, the product of the road surface adhesion coefficient and the steering inertia deviation is output as a composite safety parameter.

[0099] Here is a specific example:

[0100] Assume that when a driver is driving on a curve in the rain, operating a mobile phone causes the steering wheel to rotate slightly back and forth (steering offset features show a maximum slope of +9° / s² in the positive increasing segment and -7° / s² in the reverse segment), while the tires pass over a local water accumulation area (tread deformation amplitude extreme density drops to 2 points / cm², vibration frequency fluctuation trajectory amplitude ±10%). Step 201 extracts pressure distribution features to generate a three-dimensional pressure distribution heatmap with extreme density of 2 points / cm² and fluctuation trajectory amplitude ±10%; Step 202 divides the steering segment and extracts the maximum slope pulse sequence (+9° / s² and -7° / s²); Step 203 adjusts the fluctuation trajectory amplitude to ±9% by a slope scaling factor of 0.9 (9 / 10) to generate a fused waveform; Step 204 detects an amplitude abrupt change interval lasting 0.3 seconds (3 times the contact cycle), maps the road adhesion coefficient of 0.3 and steering inertia deviation of +12%, and outputs a composite parameter of 3.6 (0.3×12). When the parameter value exceeds the threshold of 3.0, a tiered warning is triggered: the first-level warning is a steering wheel vibration alert; the second-level intervention is to automatically limit the torque increase slope to 0.5% / s to avoid oversteering.

[0101] Steps 201-204 construct an amplitude-adaptive fusion waveform by dynamically fusing tread pressure distribution and steering offset, accurately capturing the coupling risks of low-traction road surfaces and abnormal steering operations. In scenarios where the driver is distracted, interpretable composite safety parameters are generated based on the duration of waveform abrupt change intervals and parameter mapping rules, achieving seamless integration from tread mechanical state perception to active safety control. This solves the problems of misjudgment and response delay caused by single-dimensional data analysis in traditional solutions. Through parameter-driven hierarchical warnings and dynamic control weight allocation, the reliability of safety decisions and the real-time performance of control are significantly improved in complex scenarios such as slippery roads and continuous steering.

[0102] To address the issue of mismatch between tire tread dynamic response and steering wheel operation caused by fixed sensor parameters in traditional vehicle safety systems during distracted driving scenarios, and the inability to optimize sensing accuracy in real time according to road condition changes, some embodiments involve adjusting the sampling frequency of the piezoelectric element on the tire inner wall and the signal gain of the chassis piezoelectric element based on the change amplitude of the composite safety parameters during vehicle operation. This ensures that the matching relationship between the tire tread deformation amplitude and the steering wheel angle change rate, based on the adjustment result, adapts to the current road conditions, generating a parameter adaptation state, including:

[0103] 301. Monitor the change range of the composite safety parameter, and determine the sampling frequency adjustment amount of the piezoelectric element on the inner wall of the tire by distinguishing the proportion of high-frequency fluctuation range and low-frequency gradual change range in the change range, so as to adjust the sampling frequency. When the proportion of high-frequency fluctuation range increases, the sampling frequency is increased, and when the proportion of low-frequency gradual change range increases, the sampling frequency is decreased.

[0104] In step 301, the high-frequency fluctuation range refers to the rapid fluctuation component with a frequency higher than 5Hz in the range of changes in the composite safety parameters, reflecting transient tread deformation or sudden steering operations. The low-frequency gradual change range refers to the slow change component with a frequency lower than 1Hz in the range of changes in the composite safety parameters, reflecting continuous changes in road conditions.

[0105] In this embodiment, the composite safety parameters are first decomposed using multi-scale wavelet decomposition to separate high-frequency (5-20Hz) and low-frequency (0.1-1Hz) components. Then, the energy percentage of the high-frequency components is calculated (e.g., high-frequency energy / total energy × 100%). If the percentage is greater than 55%, the tire piezoelectric sampling frequency is increased to capture transient deformation details. If the low-frequency percentage is greater than 70%, the sampling frequency is reduced to reduce redundant data.

[0106] 302. Based on the changing trend direction of the low-frequency gradual change range, determine whether the cumulative change of the tread deformation amplitude exceeds the deformation threshold range corresponding to the current road condition. If it exceeds, gradually increase the signal gain of the chassis piezoelectric element according to a preset ratio until the tread deformation amplitude returns to the deformation threshold range.

[0107] In step 302, the cumulative change is the total cumulative change in the tread deformation amplitude within the low-frequency gradual change range, which is used to determine whether the tread condition continues to deviate from the normal threshold range.

[0108] In this embodiment, the tread deformation amplitude in the low-frequency gradual change range is first integrated. If the cumulative change exceeds the deformation threshold within 30 seconds (e.g., the cumulative increase in deformation amplitude > 2.0 mm), the gain enhancement process is started. The gain of the chassis piezoelectric signal is adjusted step by step according to a preset ratio (e.g., the gain increases by 0.5 times every 5 seconds) until the tread deformation amplitude returns to the threshold range (e.g., 1.0 ± 0.3 mm). Finally, the gain value is locked and marked as a stable state.

[0109] 303. Input the adjusted sampling frequency and the enhanced signal gain into the matching model, and use the matching model to perform a weighted calculation on the temporal correlation between the steering wheel angle change rate and the tread deformation amplitude to generate matching weight coefficients;

[0110] In step 303, the weighted calculation quantifies the degree of dynamic matching between the steering wheel angle change rate and the tire tread deformation amplitude by analyzing the temporal synchronicity between the two.

[0111] In this embodiment, the adjusted sampling frequency (e.g., 2kHz) and the enhanced signal gain (e.g., 2.0 times) are first input into the matching model; second, dynamic time warping (DTW) is performed on the steering wheel angular rate and the tread deformation amplitude in the matching model to calculate the alignment path distance between the two waveforms; and third, matching weight coefficients are generated in reverse based on the alignment path distance (e.g., distance < 0.1 → weight 0.9, distance > 0.5 → weight 0.3), with a higher weight indicating a better matching degree.

[0112] 304. Based on the matching weight coefficient, the sampling interval uniformity of the piezoelectric element on the inner wall of the tire and the signal amplification smoothness of the piezoelectric element in the chassis are adjusted synchronously. According to the adjustment result, the matching relationship between the tread deformation amplitude and the steering wheel angle change rate is adapted to the current road conditions, and a parameter adaptation state is generated.

[0113] In step 304, the sampling interval uniformity is the variance of the time interval between adjacent sampling points of the tire piezoelectric element, reflecting the temporal stability of data acquisition. The signal amplification smoothness is the smoothness index of the waveform amplitude after the chassis piezoelectric signal gain is adjusted, used to evaluate the degree of signal distortion.

[0114] In this embodiment, the sampling interval adjustment amount (weight × reference interval, reference 100ms → adjusted 70ms) is first calculated based on the matching weight coefficient (e.g., 0.7) to reduce the sampling interval variance to a preset range (e.g., <5ms²). Then, the cutoff frequency of the signal gain smoothing filter is adjusted according to the weight coefficient (e.g., a weight of 0.7 corresponds to a 15Hz low-pass filter) to suppress high-frequency noise, so that the matching relationship between the tread deformation amplitude and the steering wheel angle change rate is adapted to the current road conditions, and finally, the parameter adaptation state is generated (e.g., "high-frequency optimization mode: sampling interval 70ms ± 2ms, smoothness index 0.85").

[0115] Here is a specific example:

[0116] Assume a driver on a rainy highway frequently makes minor steering wheel adjustments due to navigation operation (steering rate fluctuation standard deviation > 6° / s), while the vehicle enters a localized area of ​​standing water (tire tread deformation amplitude low-frequency gradual change cumulative amount exceeds the threshold of 1.8mm). The system executes the following coordinated control: Step 301 detects that the high-frequency proportion of composite parameters is 60%, and increases the tire sampling frequency to 1.8kHz; Step 302 calculates the low-frequency cumulative change amount of 2.1mm (exceeding the threshold of 0.3mm), and progressively increases the chassis gain to 2.2 times; Step 303, the matching model calculates the waveform alignment distance between the steering wheel and the tire tread to 0.25, generating a matching weight coefficient of 0.75; Step 304: adjust the sampling interval to 75ms (variance < 3ms²) based on the weight of 0.75, and set the smoothing filter cutoff frequency to 12Hz, finally generating the parameter adaptation state "gain enhancement mode", triggering the following controls: steering wheel vibration to indicate distraction risk; limit the torque growth slope to 0.6% / s to suppress oversteering.

[0117] Steps 301-304 achieve precise matching between tire tread deformation and steering wheel operation by frequency domain decomposition of composite parameter variation amplitude and dynamic adjustment of sensor parameters. In scenarios of driver distraction or dangerous operation, the sampling frequency is optimized based on the proportion of high-frequency / low-frequency components to capture key transient features. The reliability of data under low signal-to-noise ratio road conditions is improved through adaptive signal gain enhancement. The matching weight coefficient is generated by weighted calculation based on temporal correlation. Finally, a stable parameter adaptation state is formed by adjusting the sampling interval and signal smoothness as dual parameters.

[0118] To address the issue of lateral slip and longitudinal braking coordination failure in traditional vehicle safety systems during distracted or dangerous driving scenarios due to limited scenario recognition and rigid control weight allocation, some embodiments, when the change amplitude exceeds a preset critical range in continuous steering or emergency braking scenarios, based on the parameter adaptation state, simultaneously correct the coordinated control priority of lateral slip and longitudinal braking by limiting the linear growth slope of the power output torque in the vehicle powertrain system and enhancing the pulse response intensity of the vehicle's brake hydraulic system, including:

[0119] 401. Detect whether the change amplitude exceeds a preset critical range. If it does, identify the current scenario as a continuous steering scenario or an emergency braking scenario, and generate a corresponding scenario identifier based on the parameter adaptation state.

[0120] In step 401, the parameter adaptation state is dynamic state data that characterizes the matching relationship between the piezoelectric element sampling frequency and the tread deformation and steering wheel operation after signal gain adjustment, including quantitative indicators such as sampling interval uniformity and signal smoothness index.

[0121] In this embodiment, the change amplitude of the composite safety parameter is first detected by a threshold. If it exceeds the preset critical range (e.g., >5.0), the sampling interval variance (e.g., 3ms²) and signal smoothness (e.g., 0.8) in the parameter adaptation state are extracted. Then, based on the combined features of variance and smoothness (e.g., low variance + high smoothness), a preset scene template is matched. If the continuous steering template (sine wave mode) is matched, it is marked as "continuous steering scene identifier". If the emergency braking template (step change mode) is matched, it is marked as "emergency braking scene identifier".

[0122] 402. If the scene identifier is a continuous steering scene identifier, the linear growth slope of the power output torque of the vehicle power system is divided into a positive increasing stage and a reverse pullback stage according to the direction of steering angle change. In the positive increasing stage, the slope adjustment step size is gradually reduced, and in the reverse pullback stage, the slope recovery rate is compressed according to a preset ratio.

[0123] In step 402, the forward increasing phase is the steering operation phase where the steering wheel angle continuously increases. The reverse return phase is the straightening operation phase where the steering wheel angle decreases.

[0124] In this embodiment, the increasing and decreasing phases are first divided according to the direction of the change in the steering angle. For the positive increasing phase, a slope decay strategy is adopted: the initial slope decreases by 0.1% / s for every 5° of steering angle until it reaches a minimum of 0.5% / s. For the negative decreasing phase, a rate compression strategy is adopted: the slope recovery rate is reduced by the percentage of the decreasing angle (e.g., a compression ratio of 0.7 corresponds to a 20° decrease) (e.g., from 0.8% / s to 0.56% / s).

[0125] 403. If the scenario identifier is an emergency braking scenario identifier, set the pulse interval reference value of the vehicle brake hydraulic system according to the instantaneous change rate of the brake pedal travel increment in the vehicle power system, and convert the change amplitude into a pulse pressure amplification coefficient proportionally based on the pulse interval reference value.

[0126] In step 403, the pulse interval reference value is a basic value of the brake pulse time interval that is dynamically set according to the instantaneous change rate of the brake pedal travel increment, and is used to control the frequency of brake pressure application.

[0127] In this embodiment, the instantaneous rate of change of the brake pedal travel increment is first calculated (e.g., travel increases by 15mm within 100ms → rate 150mm / s), and the pulse interval reference value is set according to the rate and interval mapping table (e.g., 150mm / s → interval 50ms); then the change amplitude of the composite safety parameter (e.g., 6.0) is converted into a pulse pressure amplification coefficient (e.g., reference pressure 10MPa → adjusted 12MPa) according to a linear ratio based on the pulse interval reference value (change amplitude / reference value 5.0 → 1.2 times).

[0128] 404. The reduced slope adjustment step size, the compressed slope recovery rate, and the pulse pressure amplification coefficient are used as weighting factors input into the allocation model. The priority weights of the vehicle lateral slip control signal and the longitudinal braking control signal are then redistributed through the allocation model.

[0129] In step 404, the weighting factors include the reduced torque slope adjustment step size, the compressed slope recovery rate, and the pulse pressure amplification coefficient, which are used to calculate the control weights in the dynamic allocation model.

[0130] In this embodiment, the weighting factor is first input into the allocation model to assign weights to the lateral slip control signal (e.g., torque slope step size × 0.6 + slope recovery rate × 0.4) and to the longitudinal braking control signal (e.g., pulse amplification coefficient × 1.0). Then, the control priority is adjusted based on the weight ratio (e.g., lateral weight 0.52 vs. longitudinal weight 0.48). If the lateral weight is greater than the longitudinal weight, the anti-skid control response speed is increased to 1.5 times, and vice versa, the braking pressure application frequency is increased to 2.0 times.

[0131] Here is a specific example:

[0132] The driver was distracted on a wet curve, causing the steering wheel to turn continuously to the left (the steering angle increased from 0° to 35° and then returned to 10°), while simultaneously accidentally pressing the brake pedal (the travel increment rate was 180mm / s). The combined safety parameter change range reached 7.2 (exceeding the critical value of 5.0). In step 401, the parameter adaptation status shows a sampling interval variance of 2ms² and a smoothness of 0.85, and a "continuous steering scene identifier" is generated by matching the continuous steering template; in step 402, the torque slope in the positive incremental phase (0°→35°) gradually decreases from 1.2% / s to 0.5% / s, and the slope recovery rate in the reverse callback phase (35°→10°) is compressed to 0.4% / s; in step 403, the brake pedal increment rate of 180mm / s is mapped to a pulse interval reference value of 40ms, and the change amplitude of 7.2 is converted into a pulse amplification coefficient of 1.44 times (reference 10MPa→14.4MPa); in step 404, the model is assigned a lateral weight of 0.58 (0.5×0.6+0.4×0.4) and a longitudinal weight of 0.42 (14.4×0.03), the lateral control priority is increased to 1.5 times the response speed, the steering wheel reverse torque assists in returning to center, and the torque output is limited to 80% of the current value.

[0133] Steps 401-404 achieve scenario-based collaborative optimization of lateral slip and longitudinal braking control through parameter adaptation-driven scenario-based accurate identification and dynamic weight allocation. In distraction-prone or dangerous operation scenarios, based on the torque slope decay during continuous steering and the pulse parameter conversion during emergency braking, combined with multi-factor weight calculation of the allocation model, the control priority is dynamically adjusted to solve the oversteering or underbraking problems caused by fixed weight allocation in traditional solutions. Through deep binding of scenario identifiers and control parameters, vehicle stability and operational tolerance are significantly improved in complex scenarios such as slippery curves and sudden braking.

[0134] To address the waveform distortion issue caused by inaccurate dynamic coupling analysis of tire tread vibration characteristics and steering operation in traditional vehicle safety systems during distracted driving scenarios, some embodiments involve time-series superposition of the periodic fluctuation trajectory of the vibration frequency in the pressure distribution characteristics with the maximum instantaneous slope value of the steering offset characteristics, and generating a fused waveform based on the superposition result, including:

[0135] 501. Based on the periodic fluctuation trajectory of the vibration frequency in the pressure distribution characteristics, divide the tire and road surface into equal-length time periods according to the single contact period, and extract the peak points and valley points within the equal-length time periods to form a fluctuation peak-valley sequence.

[0136] In step 501, the wave peak-valley sequence is formed by dividing the periodic wave trajectory of the tread vibration frequency into multiple equal-length time periods according to the complete cycle of a single tire contact with the road surface, and extracting the peak point (highest amplitude point) and valley point (lowest amplitude point) of the waveform in each time period to form serialized data.

[0137] In this embodiment, the vibration frequency fluctuation trajectory is first divided into multiple equal-length time periods based on the single contact cycle between the tire and the road surface; then, the local highest and lowest points of the waveform curve are searched within each time period and marked as peak points and valley points, respectively; finally, the peak points and valley points of all time periods are arranged in chronological order to form a fluctuation peak-valley sequence containing time sequence information.

[0138] 502. Based on the time point at which the maximum instantaneous slope value of the steering offset feature is located, determine the reference amplitude of the fluctuation peak-valley sequence within the time period corresponding to the time point, and use the ratio of the maximum instantaneous slope value to the reference amplitude as the amplitude scaling factor;

[0139] In step 502, the amplitude scaling factor is a dynamic adjustment coefficient generated based on the ratio of the steering operation intensity (maximum instantaneous slope value) to the reference amplitude.

[0140] In this embodiment, firstly, based on the occurrence time of the maximum instantaneous slope value in the steering offset feature, the tire contact cycle time period corresponding to the occurrence time point is locked (e.g., time period 3: 0.2~0.3 seconds); secondly, the peak and valley data recorded in the fluctuation peak and valley sequence within this time period are extracted, and the median amplitude of the two is calculated as a benchmark reference; finally, by performing a proportional operation between the maximum instantaneous slope value and the benchmark reference, an amplitude scaling factor for dynamically adjusting the waveform amplitude is generated.

[0141] 503. Within the equal time period, with the reference amplitude as the center value, the peak amplitude of the fluctuation peak-valley sequence is expanded forward according to the amplitude scaling factor, and the valley amplitude is compressed backward according to the amplitude scaling factor to generate a scaled waveform segment.

[0142] In step 503, forward expansion amplifies the peak amplitude by a scaling factor, enhancing the explicit expression of tread vibration characteristics. Reverse compression reduces the valley amplitude by a scaling factor, suppressing non-critical vibration noise.

[0143] In this embodiment, firstly, using the reference amplitude as the center reference line, the peak amplitude of the wave peak-valley sequence is positively amplified by a scaling factor (e.g., peak value within time period 3 +8% → +24%) to enhance the high-pressure area characteristics of the tire tread under steering force; secondly, the valley amplitude is negatively reduced by a scaling factor (e.g., valley value within time period 3 -5% → -1.67%) to weaken the interference of non-critical vibration noise; finally, a scaled waveform segment with asymmetric amplitude characteristics is generated, where peak expansion reflects the instantaneous impact effect of steering operation, and valley compression suppresses the negative impact of environmental noise.

[0144] 504. Connect the scaled waveform segments end to end in time sequence to form a continuous waveform segment. Perform slope transition smoothing on the amplitude jump area at the junction of the continuous waveform segments. Form a fused waveform based on the smoothing result.

[0145] In step 504, the slope transition smoothing process is to gradually transition the amplitude abrupt region at the junction of adjacent waveform segments to eliminate the distortion caused by waveform splicing.

[0146] In this embodiment, the scaled waveform segments of each time period are first spliced ​​together sequentially along the time axis to form a preliminary continuous waveform. Then, the amplitude jump difference at the junction of adjacent waveform segments of the preliminary continuous waveform is detected (e.g., the difference between the amplitude at the end of the previous segment (+24%) and the amplitude at the beginning of the next segment (+5%) is 19%). A linear gradual transition curve is inserted within the jump interval to limit the amplitude change rate between the end of the previous segment and the beginning of the next segment within a preset range. Finally, a smooth and continuous fused waveform is output, eliminating waveform distortion caused by segmentation processing.

[0147] Here is a specific example:

[0148] Suppose that in a snow-covered bend, the driver is distracted and causes the steering wheel to turn sharply to the right (the steering offset characteristic shows a maximum instantaneous slope of 24° / s², corresponding to a time point of 0.32 seconds), while the tires experience low-frequency, high-amplitude vibrations due to the compaction of snow (vibration frequency fluctuation trajectory amplitude ±15%). First, the vibration trajectory is divided into time periods of 0.30 to 0.45 seconds based on the tire's single contact cycle of 0.15 seconds. The peak-valley sequence within this time period is extracted (peak value +15%, valley value -12%). Then, the time period corresponding to the maximum instantaneous slope value is locked, and the baseline amplitude of 13.5% (median of peak-valley amplitude) is calculated, generating an amplitude scaling factor of 1.78 (steering intensity modulation coefficient). Next, the peak value is forward expanded to +26.7%, and the valley value is reverse compressed to -6.74%, forming an asymmetric waveform segment (amplitude range -6.74% to +26.7%). Finally, when splicing the waveform segments, an amplitude jump difference of 18.7% between adjacent intervals is detected, and a 0.15-second linear gradual transition is inserted (amplitude decrease of 1.25% every 0.01 seconds) to generate a continuous and smooth fused waveform. When the amplitude abrupt change in the fused waveform lasts for 0.3 seconds (up to twice the contact cycle), graded control is triggered: the steering wheel return torque is increased by 40% to suppress oversteering, the pre-pressure braking is simultaneously applied to 12MPa, and a red warning of "low adhesion steering risk" is projected through the HUD to achieve dynamic coordination of lateral and longitudinal control.

[0149] Steps 501-504 accurately reflect the dynamic impact of steering intensity on tire vibration characteristics through time-segmented asymmetric waveform scaling and dynamic smoothing transition processing, avoiding waveform distortion caused by global uniform scaling in traditional solutions. In distracted driving scenarios, the accuracy of coupling feature analysis of low-adhesion road surfaces and sudden steering operations is significantly improved by time-domain segmentation of peak-valley sequences and local amplitude adjustment of the maximum steering slope. Combined with smoothing processing to eliminate waveform splicing noise, high-fidelity fused waveform data is provided for safety control decisions, solving the problem of false safety warning triggering caused by phase deviation and amplitude distortion in traditional methods.

[0150] To address the dynamic imbalance problem caused by the rigid allocation of lateral and longitudinal control weights in traditional vehicle safety systems during distracted or dangerous operation scenarios, some embodiments incorporate the reduced slope adjustment step size, the compressed slope recovery rate, and the pulse pressure amplification coefficient as weighting factors input into the allocation model. The allocation model then redistributes the priority weights of the vehicle's lateral slip control signal and longitudinal braking control signal, including:

[0151] 601. The reduced slope adjustment step size is converted into the suppression coefficient of the lateral slip control, the compressed slope recovery rate is converted into the compensation coefficient of the lateral slip control, and the pulse pressure amplification coefficient is converted into the superposition coefficient of the longitudinal braking control.

[0152] In step 601, the suppression coefficient is derived from the slope adjustment step size of the power output torque and is used to limit the intervention intensity of lateral slip control. The compensation coefficient is derived from the slope recovery rate and is used to correct the control lag during the steering return process; the superposition coefficient is derived from the brake pulse pressure amplification coefficient and is used to enhance the execution force of longitudinal braking control.

[0153] In this embodiment, the slope adjustment step size (e.g., 0.5% / s) is first scaled proportionally to the suppression coefficient of lateral slip control (step size × 0.8 → 0.4), the slope recovery rate (e.g., 0.6% / s) is compressed to the compensation coefficient of lateral slip control (rate × 0.6 → 0.36), and the pulse amplification coefficient (e.g., 1.5 times) is linearly mapped to the superposition coefficient of longitudinal braking control (1.5 → 0.75) according to a preset mapping rule.

[0154] 602. In the allocation model, the horizontal weight channel and the vertical weight channel are divided into independently operated horizontal weight channels. The horizontal weight channel receives the difference between the suppression coefficient and the compensation coefficient, and the vertical weight channel receives the accumulated value of the superposition coefficient.

[0155] In step 602, the lateral weighting channel is an independent processing channel for the lateral slip control signal, and its input is the difference between the suppression coefficient and the compensation coefficient. The longitudinal weighting channel is an independent processing channel for the longitudinal braking control signal, and its input is the accumulated value of the superposition coefficients.

[0156] In this embodiment, firstly, a horizontal and vertical dual channel is established in the allocation model. The horizontal channel inputs the real-time difference between the suppression coefficient and the compensation coefficient, and the vertical channel inputs the sliding window accumulation value of the superposition coefficient (such as the average value of the superposition coefficient in the last 5 seconds). Secondly, the difference (0.04) and the accumulation value (0.75) are mapped to the weight range of 0~1 through normalization processing. Finally, the weight calculation logic of the corresponding channel is dynamically activated based on the scene identifier.

[0157] 603. When the scene identifier is a continuous steering scene identifier, based on the positive and negative characteristics of the steering angle change direction, the difference is converted into a weight allocation ratio adjustment instruction for the lateral slip control signal.

[0158] In step 603, the weight allocation ratio adjustment instruction adjusts the execution weight ratio of the lateral slip control signal according to the positive and negative characteristics of the direction of steering angle change (e.g., left turn is positive and right turn is negative).

[0159] In this embodiment, the positive and negative markers of the direction of steering angle change are first identified, where "+" indicates positive steering and "-" indicates reverse correction. If it is a positive direction, the difference result of the lateral weight channel (e.g., 0.04) is multiplied by the positive gain coefficient (e.g., 1.5) to generate a weight ratio adjustment value (0.06). If it is a negative direction, the difference result is multiplied by the reverse attenuation coefficient (e.g., 0.8) to generate an adjustment value (0.032). Finally, the adjustment value is superimposed on the baseline weight (e.g., 0.5→0.56 or 0.532) to generate a weight allocation ratio adjustment instruction for the lateral slip control signal.

[0160] 604. When the scenario identifier is an emergency braking scenario identifier, the accumulated value is mapped to a weight allocation ratio enhancement instruction for the longitudinal braking control signal according to the rate of change of the brake pedal travel increment.

[0161] In step 604, the weight allocation ratio enhancement instruction dynamically increases the execution weight ratio of the longitudinal braking control signal based on the rate of change of the brake pedal travel increment.

[0162] In this embodiment, firstly, the instantaneous rate of change of the brake pedal travel increment is calculated by recording the travel difference between adjacent sampling points and dividing it by the sampling time window length (e.g., when the travel increases by 15mm within a 50ms time window, the instantaneous rate of change is 15mm / 0.05s=300mm / s); secondly, the accumulated value of the longitudinal weight channel is adjusted according to the rate and weight mapping table (e.g., 180mm / s → weight increase of 0.3) (e.g., 0.75 → 0.75+0.3=1.05); finally, the accumulated value is limited (e.g., upper limit 1.0), and the weight allocation ratio enhancement command of the longitudinal braking control signal is output according to the limited accumulated value.

[0163] Here is a specific example:

[0164] Suppose a driver, while on a slippery curve, continuously turns the steering wheel to the left due to checking their mobile phone (steering angle change direction marked as +, suppression coefficient 0.1, compensation coefficient 0.06), and simultaneously accidentally depresses the brake pedal (stroke increment rate 180mm / s, superposition coefficient 0.45). The system executes the following coordinated control: Step 601 converts and generates suppression coefficient 0.1, compensation coefficient 0.06, and superposition coefficient 0.45; in step 602, the lateral channel receives the difference value 0.04 (0.1−0.06), and the longitudinal channel receives the accumulated value 0.75 (0.45+historical value 0.3); in step 603, based on the positive steering mark (+), the difference value 0.04×gain 1.5→weight adjustment value 0.06, and the lateral control weight is increased from 0.5 to 0.56; in step 604, based on the stroke difference between adjacent sampling points recorded by the brake pedal displacement sensor (e.g., stroke from 0~50ms from 0.1 to 0.06), the system adjusts the weight of the brake pedal. The difference between the 20mm and 35mm increases (15mm difference), divided by the time window length of 0.05 seconds, and the instantaneous change rate is calculated to be 300mm / s, triggering a longitudinal weight increase of 0.3. The final control decision is as follows: Lateral control: weight 0.56 → steering wheel return torque increased by 28% (10N·m → 12.8N·m), suppressing sideslip; Longitudinal control: weight 1.0 → brake pulse pressure increased to 1.5 times the baseline value (8MPa → 12MPa), shortening the braking distance by 1.8 meters; Warning linkage: triggers the instrument panel to flash warning "oversteering + braking conflict", and an audible prompt "please concentrate".

[0165] Steps 601-604 achieve scenario-based priority allocation of lateral slip and longitudinal braking control through dynamic conversion and weighted channel separation calculation of steering and braking control parameters. In distraction or dangerous operation scenarios, the control weights are adjusted in real time based on steering direction characteristics and braking intensity, resolving the conflict between steering torque and braking force caused by fixed weights in traditional systems. Lateral control optimization is driven by the difference between suppression coefficient and compensation coefficient, and longitudinal control enhancement is driven by the accumulation of superposition coefficients, significantly improving vehicle stability and operational safety in wet and slippery road surfaces, continuous steering, or emergency braking scenarios, achieving dynamic self-adaptation and collaborative intervention of multi-dimensional control.

[0166] To address the warning delay issue caused by the disconnect between single-sensor data acquisition and multi-dimensional operational behavior analysis in distracted driving scenarios, some embodiments involve continuously recording the steering wheel angle change rate, brake pedal travel increment, and accelerator pedal trigger interval via an in-cabin operating device, forming a multi-dimensional operational mode sequence, including:

[0167] 701. Install a three-axis motion sensor group on the steering wheel shaft, brake pedal linkage and the bottom of the accelerator pedal to capture the steering wheel angular displacement, brake pedal depressing stroke and accelerator pedal contact trigger status.

[0168] In step 701, the three-axis motion sensor group is a hardware device consisting of a steering wheel shaft angle sensor, a brake pedal displacement sensor, and an accelerator pedal contact sensor, used to synchronously capture the steering wheel angular displacement, brake pedal depressing stroke, and accelerator pedal contact triggering state operated by the driver.

[0169] In this embodiment, a high-precision angle encoder is first embedded in the steering wheel shaft to record the steering wheel rotation angle in real time; a linear displacement sensor is installed at the brake pedal linkage to detect the pedal pressing stroke; and a pressure-sensitive contact switch is deployed at the bottom of the accelerator pedal to capture the contact trigger state. The three sets of sensors achieve timing alignment of data acquisition through a synchronous clock signal.

[0170] 702. Calculate the absolute value of the angle change for the steering wheel angular displacement in a unit time window, and construct the steering wheel angular change rate based on the temporal characteristics of the absolute value of the angle change;

[0171] In step 702, the rate of change of steering wheel angle is a time sequence generated by the absolute value of the change in steering wheel angle displacement within a unit time window, reflecting the dynamic intensity of the steering operation.

[0172] In this embodiment, the steering wheel angle displacement data stream is first segmented according to a fixed time window; then, the angle difference between adjacent sampling points is calculated, and the absolute value of the angle difference is taken to eliminate the influence of the steering direction (e.g., the steering angle increases from 0° to 15° within 0.1 seconds → the difference is 15°); then, the difference sequence is filtered by moving average to generate a smoothed steering wheel angle change rate; finally, the rate peak point and abnormal fluctuation interval are marked.

[0173] 703. Divide the brake pedal depressing stroke into stepped incremental marker points according to a preset displacement interval, and record the time difference between the stepped incremental marker points to generate the brake pedal stroke increment;

[0174] In step 703, the stepped incremental marker points are used to divide the brake pedal depressing stroke into discrete stepped intervals according to a preset displacement interval, and mark the start and end points of each interval.

[0175] In this embodiment, the displacement interval of the brake pedal is first set (e.g., each 10mm depressor is a step); then, the time difference between the pedal moving from the starting point of the current step to the end point of the next step (a displacement interval of 1 unit) is recorded (e.g., it takes 0.3 seconds to go from 20mm to 30mm); then, the brake pedal travel increment composed of the time difference is generated (e.g., increment 10mm / 0.3s≈33.3mm / s); finally, noise interference (e.g., negative increments caused by displacement rebound) is removed.

[0176] 704. Convert the contact trigger state into a contact interval timing signal, calculate the interval duration of the contact interval timing signal as the accelerator pedal trigger interval, arrange the steering wheel angle change rate, the brake pedal travel increment and the accelerator pedal trigger interval in parallel according to the operation occurrence time order, and generate a multi-dimensional operation mode sequence with time synchronization based on the arrangement result.

[0177] In step 704, the contact interval timing signal is a timing signal generated by recording the time difference between two adjacent contact trigger actions of the accelerator pedal, reflecting the urgency of the driver's pedal action.

[0178] In this embodiment, the accelerator pedal contact trigger state is first converted into a timestamp sequence (e.g., contact closing time points t1=1.5s, t2=1.8s); then the time interval between adjacent contacts in the timestamp sequence is calculated as the accelerator pedal trigger interval (Δt=t2-t1=0.3s); next, the steering wheel angle change rate, brake pedal travel increment, and accelerator pedal trigger interval are aligned according to a unified time axis; finally, missing values ​​are filled (e.g., zeros are added for periods without operation) and a multi-dimensional operation mode sequence is generated.

[0179] Here is a specific example:

[0180] While driving on the highway, the driver frequently made minor adjustments to the steering wheel due to using a mobile phone (the steering wheel angle displacement fluctuated from 0° to ±12° within 0.5 seconds), and at the same time accidentally touched the brake pedal (the downward travel suddenly increased from 0mm to 45mm, divided into 0.2 seconds for 0-10mm, 0.1 seconds for 10-20mm, 0.05 seconds for 20-30mm, and 0.15 seconds for 30-45mm), and irregularly pressed the accelerator pedal (the contact trigger interval was irregular: 0.4 seconds → 0.2 seconds → 0.7 seconds). In step 701, the triaxial sensor synchronously captures the steering wheel angular displacement (±12°), brake pedal travel (45mm), and acceleration contact status (interval 0.4s / 0.2s / 0.7s); in step 702, the steering wheel angular change rate is calculated in a 100ms window (e.g., 120° / s from 0 to 100ms, -90° / s from 100 to 200ms); in step 703, the brake travel is divided into four steps (0-10 / 10-20 / 20-30 / 30-45mm) in 10mm intervals, correspondingly generating brake pedal travel increments (50mm / s→100mm / s→200mm / s→100mm / s); in step 704, the acceleration contact interval is converted into a timing signal (0.4s→0.2s→0.7s), and a multi-dimensional operation mode sequence is generated. When the standard deviation of steering wheel speed fluctuation is detected to be >8° / s, the brake increment is suddenly increased by 200mm / s, and the accelerator pedal trigger interval is <0.3 seconds, the following warnings are triggered: audible and visual warnings: the instrument panel flashes "dangerous operation"; control preload: steering wheel damping is increased by 20%, and the braking system is pre-pressurized to 10MPa.

[0181] Steps 701-704 construct a holographic profile of the driver's operating behavior through multi-dimensional synchronous data acquisition and time-series fusion analysis using a triaxial sensor array, accurately identifying distracted or dangerous operating characteristics. Based on the spatiotemporal correlation detection of multi-dimensional operating pattern sequences, a millisecond-level response from raw data to risk warning is achieved, solving the problem of missed judgments caused by single-dimensional data analysis or inaccurate timing in traditional systems, and significantly improving the active safety control capabilities and warning reliability in complex driving scenarios.

[0182] Figure 2 This application provides a schematic diagram of the structure of an artificial intelligence-based vehicle safety system, as shown in the embodiment. Figure 2 As shown, the system includes:

[0183] The acquisition module 21 is used to acquire response signals generated by the vehicle contacting the road surface during driving through multiple piezoelectric elements distributed on the vehicle chassis and the inner wall of the tire. The response signals include the correlation characteristics between the tread deformation amplitude and the vibration frequency under different road conditions.

[0184] The acquisition module 22 is used to continuously record the rate of change of steering wheel angle, the increase of brake pedal travel and the trigger interval of accelerator pedal through the in-cabin operating device, and form a multi-dimensional operating mode sequence.

[0185] The coupling module 23 is used to couple the correlation characteristics between the tread deformation amplitude and vibration frequency in the response signal with the steering wheel angle change rate in the multi-dimensional operation mode sequence, and generate a composite safety parameter including the road adhesion coefficient and steering inertia deviation based on the coupling result.

[0186] The adjustment module 24 is used to adjust the sampling frequency of the piezoelectric element on the inner wall of the tire and the signal gain of the piezoelectric element on the chassis according to the change range of the composite safety parameters during vehicle driving, so that the matching relationship between the tread deformation amplitude and the steering wheel angle change rate based on the adjustment result is adapted to the current road conditions, and a parameter adaptation state is generated.

[0187] The correction module 25 is used to simultaneously correct the coordinated control priority of lateral slip and longitudinal braking of the vehicle when the change amplitude exceeds a preset critical range in continuous steering or emergency braking scenarios, based on the parameter adaptation state, by limiting the linear growth slope of the power output torque in the vehicle power system and enhancing the pulse response intensity of the vehicle brake hydraulic system.

[0188] Figure 2 The aforementioned AI-based vehicle safety system can execute... Figure 1 The implementation principle and technical effects of the artificial intelligence-based vehicle safety assurance method described in the illustrated embodiment will not be repeated here. The specific methods by which each module and unit of the artificial intelligence-based vehicle safety assurance system in the above embodiments perform operations have been described in detail in the embodiments related to this method, and will not be elaborated upon here.

[0189] In one possible design, Figure 2 The vehicle safety system based on artificial intelligence shown in the embodiment can be implemented as a computing device, such as... Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0190] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are invoked and executed by the processing component 32.

[0191] The processing component 32 is used for the above Figure 1 The embodiment describes a vehicle safety assurance method based on artificial intelligence.

[0192] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above-described method. Alternatively, the processing component may be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above-described method.

[0193] Storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0194] Of course, computing devices may also include other components, such as input / output interfaces, display components, communication components, etc.

[0195] Input / output interfaces provide interfaces between processing components and peripheral interface modules, which can be output devices, input devices, etc.

[0196] The communication components are configured to facilitate wired or wireless communication between computing devices and other devices.

[0197] The computing device can be a physical device or an elastic computing host provided by a cloud computing platform. In this case, the computing device can refer to a cloud server, and the aforementioned processing components, storage components, etc., can be basic server resources rented or purchased from the cloud computing platform.

[0198] This application also provides a computer storage medium storing a computer program, which, when executed by a computer, can perform the above-described functions. Figure 1 The illustrated embodiment presents an artificial intelligence-based vehicle safety assurance method.

[0199] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0200] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0201] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0202] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. An artificial intelligence-based vehicle safety assurance method, characterized by, The application relates to a vehicle safety parameter coupling method, comprising: Collecting response signals generated by the contact between vehicle tires and road surfaces during vehicle driving through a plurality of piezoelectric elements distributed on the vehicle chassis and the inner wall of the tires, wherein the response signals contain the correlation characteristics of the deformation amplitude and vibration frequency of the tire tread under different road conditions; Recording the steering wheel rotation rate, brake pedal stroke increment and accelerator pedal trigger interval through the operation device in the cockpit, and forming a multi-dimensional operation mode sequence; Coupling the correlation characteristics of the deformation amplitude and vibration frequency of the tire tread in the response signals with the steering wheel rotation rate in the multi-dimensional operation mode sequence, and generating a composite safety parameter containing the road adhesion coefficient and steering inertia deviation according to the coupling result; Adjusting the sampling frequency of the piezoelectric element on the inner wall of the tire and the signal gain of the piezoelectric element on the chassis according to the change amplitude of the composite safety parameter during vehicle driving, so that the matching relationship between the deformation amplitude of the tire tread and the steering wheel rotation rate is adapted to the current road condition based on the adjustment result, and a parameter adaptation state is generated; When the change amplitude exceeds the preset critical range in the continuous steering or emergency braking scene, the parameter adaptation state is used to limit the linear growth slope of the power output torque in the vehicle power system and enhance the pulse response strength of the vehicle brake hydraulic system, so as to synchronously correct the cooperative control priority of vehicle lateral slip and longitudinal braking; The coupling of the correlation characteristics of the deformation amplitude and vibration frequency of the tire tread in the response signals with the steering wheel rotation rate in the multi-dimensional operation mode sequence, and the generation of a composite safety parameter containing the road adhesion coefficient and steering inertia deviation according to the coupling result, comprises: According to the correlation characteristics of the deformation amplitude and vibration frequency of the tire tread in the response signals, the pressure distribution characteristics of the contact area between the vehicle tire and the road surface are extracted, and the pressure distribution characteristics are represented by the extreme value distribution law of the deformation amplitude and the periodic fluctuation trajectory of the vibration frequency; The steering wheel rotation rate in the multi-dimensional operation mode sequence is divided into a positive incremental section and a reverse decreasing section according to the steering operation direction, and the maximum instantaneous slope value of the steering wheel rotation rate in each section is extracted to form a steering offset feature; The periodic fluctuation trajectory of the pressure distribution characteristics and the maximum instantaneous slope value of the steering offset feature are time-superposed, and a fusion waveform is generated according to the superposition result, wherein the waveform amplitude of the pressure distribution characteristics is scaled in proportion to the instantaneous slope value of the steering offset feature during the superposition process; The amplitude mutation interval of the periodic fluctuation trajectory in the fusion waveform is divided into intervals, and when the duration of the amplitude mutation interval exceeds the preset multiple of the single contact period between the tire and the road surface, the road adhesion coefficient corresponding to the extreme value distribution law in the amplitude mutation interval and the steering inertia deviation corresponding to the slope change extreme value of the fusion waveform are extracted, and the product of the road adhesion coefficient and the steering inertia deviation is taken as the composite safety parameter.

2. The method of claim 1, wherein, The sampling frequency of the tire inner wall piezoelectric element and the signal gain of the chassis piezoelectric element are adjusted according to the change range of the composite safety parameter during vehicle driving, so that the matching relationship between the tire tread deformation amplitude and the steering wheel angle change rate based on the adjustment result adapts to the current road conditions, and a parameter adaptation state is generated, including: Monitoring the change range of the composite safety parameter, determining the sampling frequency adjustment amount of the tire inner wall piezoelectric element by distinguishing the proportion of the high-frequency fluctuation interval and the low-frequency slowly changing interval in the change range, and adjusting the sampling frequency, wherein the sampling frequency is increased when the proportion of the high-frequency fluctuation interval increases, and the sampling frequency is decreased when the proportion of the low-frequency slowly changing interval increases; According to the change trend direction of the low-frequency slowly changing interval, it is judged whether the cumulative change amount of the tire tread deformation amplitude exceeds the deformation threshold range corresponding to the current road conditions, and if it exceeds, the signal gain of the chassis piezoelectric element is gradually increased by a preset proportion until the tire tread deformation amplitude returns to the deformation threshold range; The adjusted sampling frequency and the enhanced signal gain are input into the matching model, and the time sequence correlation between the steering wheel angle change rate and the tire tread deformation amplitude is weighted calculated by the matching model, and a matching weight coefficient is generated; Based on the matching weight coefficient, the sampling interval uniformity of the tire inner wall piezoelectric element and the signal amplification smoothness of the chassis piezoelectric element are synchronously adjusted, and according to the adjustment result, the matching relationship between the tire tread deformation amplitude and the steering wheel angle change rate adapts to the current road conditions, and a parameter adaptation state is generated.

3. The method of claim 1, wherein, When the change range exceeds the preset critical range in continuous steering or emergency braking scenarios, based on the parameter adaptation state, by limiting the linear growth slope of the power output torque in the vehicle power system and enhancing the pulse response strength of the vehicle brake hydraulic system, the cooperative control priority of vehicle lateral slip and longitudinal braking is synchronously corrected, including: Detecting whether the change range exceeds the preset critical range, and identifying the current scenario as a continuous steering scenario or an emergency braking scenario when it exceeds, and generating a corresponding scenario identifier based on the parameter adaptation state; If the scenario identifier is a continuous steering scenario identifier, the linear growth slope of the power output torque of the vehicle power system is divided into a positive increasing stage and a reverse callback stage according to the change direction of the steering angle, the slope adjustment step is gradually reduced in the positive increasing stage, and the slope recovery rate is compressed by a preset proportion in the reverse callback stage; If the scenario identifier is an emergency braking scenario identifier, set the pulse interval reference value of the vehicle brake hydraulic system according to the instantaneous change rate of the brake pedal stroke increment in the vehicle power system, and convert the change range into a pulse pressure amplitude coefficient based on the pulse interval reference value; The reduced slope adjustment step, the compressed slope recovery rate and the pulse pressure amplitude coefficient are input into the distribution model as weight factors, and the priority weight of the vehicle lateral slip control signal and the longitudinal braking control signal is redistributed by the distribution model.

4. The method of claim 1, wherein, The vibration frequency periodic fluctuation trajectory in the pressure distribution feature is time-superimposed with the maximum instantaneous slope value of the steering offset feature, and a fusion waveform is generated according to the superimposition result, comprising: According to the vibration frequency periodic fluctuation trajectory in the pressure distribution feature, the equal-length time periods are divided according to the single contact period of the tire and the road, and the peak points and the valley points in the equal-length time periods are extracted to form a fluctuation peak-valley sequence; According to the time point at which the maximum instantaneous slope value of the steering offset feature is located, the reference amplitude value of the fluctuation peak-valley sequence in the corresponding time period of the time point is determined, and the ratio of the maximum instantaneous slope value to the reference amplitude value is taken as an amplitude scaling factor; In the equal-length time period, the peak point amplitude of the fluctuation peak-valley sequence is positively expanded by the amplitude scaling factor with the reference amplitude value as the center value, and the valley point amplitude is inversely compressed by the amplitude scaling factor, to generate a scaled waveform segment; The scaled waveform segment is connected in time sequence to form a continuous waveform segment, the amplitude jump region of the continuous waveform segment is subjected to slope transition smoothing processing, and a fusion waveform is formed according to the smoothing processing result.

5. The method of claim 3, wherein, The reduced slope adjustment step, the compressed slope recovery rate, and the pulse pressure increase coefficient are input as weight factors into a distribution model, and the priority weight of the vehicle lateral slip control signal and the longitudinal brake control signal is redistributed through the distribution model, comprising: The reduced slope adjustment step is converted into a lateral slip control inhibition coefficient, the compressed slope recovery rate is converted into a lateral slip control compensation coefficient, and the pulse pressure increase coefficient is converted into a longitudinal brake control superposition coefficient; In the distribution model, an independent operation lateral weight channel and a longitudinal weight channel are divided, the lateral weight channel receives the difference value of the inhibition coefficient and the compensation coefficient, and the longitudinal weight channel receives the accumulated value of the superposition coefficient; When the scene identifier is a continuous steering scene identifier, based on the positive and negative characteristics of the steering angle change direction, the difference value is converted into a weight distribution proportion adjustment instruction of the lateral slip control signal; When the scene identifier is an emergency braking scene identifier, according to the change rate of the brake pedal stroke increment, the accumulated value is mapped into a weight distribution proportion enhancement instruction of the longitudinal brake control signal.

6. The method of claim 1, wherein, The steering wheel rotation angle change rate, brake pedal stroke increment, and acceleration pedal trigger interval are continuously recorded through the in-cabin operating device, and a multi-dimensional operation mode sequence is formed, comprising: A three-axis action sensor group is installed on the steering wheel shaft, brake pedal link, and acceleration pedal bottom to capture the steering wheel rotation angle displacement, brake pedal depression stroke, and acceleration pedal contact trigger state; The steering wheel rotation angle displacement is calculated according to the angle change amount absolute value per unit time window, and the steering wheel rotation angle change rate is constructed according to the time sequence characteristics of the angle change amount absolute value; The brake pedal depression stroke is divided into stepped increment marker points according to a preset displacement interval, and the time difference between the stepped increment marker points is recorded to generate a brake pedal stroke increment; The contact trigger state is converted into a contact interval timing signal, the interval duration of the contact interval timing signal is calculated as an accelerator pedal trigger interval, the steering wheel rotation angle change rate, the brake pedal stroke increment and the accelerator pedal trigger interval are arranged in parallel in chronological order of operation occurrence, and a multi-dimensional operation mode sequence with time sequence synchronization is generated according to the arrangement result.

7. An artificial intelligence-based vehicle safety assurance system, characterized by, It comprises: The acquisition module is used for collecting the response signal generated by the contact between the vehicle and the road surface during driving through a plurality of piezoelectric elements distributed on the vehicle chassis and the inner wall of the tire, and the response signal contains the correlation characteristics of the deformation amplitude and vibration frequency of the tire under different road conditions; The acquisition module is used for continuously recording the steering wheel rotation angle change rate, brake pedal stroke increment and accelerator pedal trigger interval through the operating device in the cockpit, and forming a multi-dimensional operation mode sequence; The coupling module is used for coupling the correlation characteristics of the deformation amplitude and vibration frequency of the tire in the response signal with the steering wheel rotation angle change rate in the multi-dimensional operation mode sequence, and generating a composite safety parameter containing the road surface adhesion coefficient and the steering inertia deviation according to the coupling result; The adjustment module is used for adjusting the sampling frequency of the piezoelectric element on the inner wall of the tire and the signal gain of the piezoelectric element on the chassis according to the change amplitude of the composite safety parameter during vehicle driving, so that the matching relationship between the deformation amplitude of the tire and the steering wheel rotation angle change rate is adapted to the current road condition based on the adjustment result, and a parameter adaptation state is generated; The correction module is used for synchronously correcting the cooperative control priority of vehicle lateral slip and longitudinal braking by limiting the linear growth slope of power output torque in the vehicle power system and enhancing the pulse response strength of the vehicle brake hydraulic system based on the parameter adaptation state when the change amplitude exceeds the preset critical range in the continuous steering or emergency braking scene; The coupling module is used for coupling the correlation characteristics of the deformation amplitude and vibration frequency of the tire in the response signal with the steering wheel rotation angle change rate in the multi-dimensional operation mode sequence, and generating a composite safety parameter containing the road surface adhesion coefficient and the steering inertia deviation according to the coupling result; According to the correlation characteristics of the deformation amplitude and vibration frequency of the tire in the response signal, the pressure distribution characteristics of the contact area between the vehicle tire and the road surface are extracted, which are represented by the extreme value distribution law of the deformation amplitude and the periodic fluctuation trajectory of the vibration frequency; The steering wheel rotation angle change rate in the multi-dimensional operation mode sequence is divided into a positive incremental section and a reverse decreasing section according to the steering operation direction, and the maximum instantaneous slope value of the steering wheel rotation angle change rate in each section is extracted to form a steering offset feature; The periodic fluctuation trajectory of the vibration frequency in the pressure distribution characteristics and the maximum instantaneous slope value of the steering offset feature are time-superposed, and a fusion waveform is generated according to the superposition result, wherein the waveform amplitude of the pressure distribution characteristics is scaled in proportion to the instantaneous slope value of the steering offset feature during the superposition process. The amplitude mutation interval of the periodic fluctuation trajectory in the fusion waveform is divided, when the duration of the amplitude mutation interval exceeds the preset multiple of the single contact period of the tire and the road surface, the road surface adhesion coefficient corresponding to the extreme value distribution law in the amplitude mutation interval is extracted, and the steering inertia deviation corresponding to the slope change extreme value of the fusion waveform, and the product of the road surface adhesion coefficient and the steering inertia deviation is taken as a composite safety parameter.

8. A computing device, comprising: The method comprises a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to realize the vehicle safety guarantee method based on artificial intelligence in any one of claims 1-6.

9. A computer storage medium, characterized in that A computer program is stored, and when the computer program is executed by a computer, the vehicle safety guarantee method based on artificial intelligence in any one of claims 1-6 is realized.

Images