Vehicle safety risk threshold identification method for heavy-duty truck

Abstract

The invention provides a vehicle safety risk threshold identification method for a heavy-duty truck, relates to the technical field of vehicle safety, and solves the problem of limited identification precision of a vehicle risk threshold in the prior art. According to the technical scheme, dynamic response data of a truck and environment variables under various external environments and different road conditions are obtained through a multi-source data acquisition system, and real-time key risk indexes of the heavy-duty truck are calculated; preprocessing the real-time key risk indexes to obtain cleaned real-time data; determining candidate thresholds corresponding to the real-time data through a quantile statistical method; the candidate threshold is fed back to the vehicle dynamics model for forward verification, the dynamic response of the vehicle under the threshold boundary condition is simulated until the simulation result passes forward simulation verification, and the candidate threshold is written into a dynamic threshold library. According to the technical scheme, accurate early warning can be achieved, and the accident rate is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle safety, and particularly to a method for identifying vehicle safety risk thresholds for heavy-duty trucks. Background Art

[0002] This section aims to provide background or context for the embodiments of the present invention recited in the claims. The descriptions herein may include concepts that can be explored, but not necessarily concepts that have been previously thought of or explored. Therefore, unless otherwise indicated herein, the content described in this section is not prior art for the specification and claims of this application, and is not admitted to be prior art merely by virtue of being included in this section.

[0003] With the rapid development of the road freight industry in China, heavy-duty trucks play an increasingly important role in logistics transportation. However, due to the characteristics of heavy-duty trucks such as large mass, high inertia, and long braking distance, they are extremely prone to serious accidents such as rollover and brake failure in complex road environments, posing a severe challenge to road traffic safety.

[0004] Statistics show that the proportion of freight vehicle liability accidents is 12.7% (among which heavy-duty trucks account for 63% of freight accidents), and the consequences of accidents are often more serious, not only causing heavy casualties and property losses, but also having a profound impact on the social and economic order.

[0005] At present, there are still significant deficiencies in the safety research for heavy-duty trucks. Traditional in-vehicle warning systems mainly rely on fixed thresholds for speeding or deviation warnings, and cannot adapt to the dynamically changing load distribution, complex road conditions, and environmental conditions, resulting in insufficient warning accuracy. Although existing navigation systems can provide basic speed limit suggestions, they lack personalized risk modeling for heavy-duty trucks and are difficult to meet the actual driving safety requirements. At the technical level, although the application of a single simulation platform can simulate some vehicle dynamics behaviors, the identification accuracy of risk thresholds is limited and it is difficult to support accurate warnings. In addition, existing research mainly focuses on single-scenario analysis, and there is still a lack of systematic solutions for comprehensive risk assessment in a multi-source information coupling environment. With the rapid development of vehicle networking technology, breakthroughs in high-precision positioning, real-time communication, and multi-source data fusion technologies provide new possibilities for improving the driving safety of heavy-duty trucks.

[0006] Therefore, in this context, there is an urgent need for a method for establishing a threshold library that can integrate vehicle parameters, road conditions, and environmental information. Summary of the Invention

[0007] To solve the problem in the prior art that the identification accuracy of vehicle risk thresholds is limited and it is difficult to support accurate warnings, the purpose of the present invention is to provide a method for identifying vehicle safety risk thresholds for heavy-duty trucks.

[0008] To solve the above technical problems, in a first aspect, according to some embodiments, the present invention provides a method for identifying safety risk thresholds for heavy-duty trucks, including:

[0009] Obtain the dynamic response data of the heavy-duty truck and environmental variables under various external environments and different road conditions through a multi-source data acquisition system. According to the dynamic response data and the environmental variables, calculate the real-time key risk indicators of the heavy-duty truck, and the key risk indicators correspond to specific risk scenarios;

[0010] Preprocess the real-time key risk indicators, and the preprocessing includes noise removal processing, missing data supplementation processing, and outlier removal processing to obtain the cleaned real-time data;

[0011] Determine the candidate thresholds corresponding to the real-time data through the quantile statistical method;

[0012] Input the candidate thresholds into the vehicle dynamics model of the heavy-duty truck for forward simulation verification, simulate the dynamic response of the heavy-duty truck under the threshold boundary conditions. If the simulation result passes the forward simulation verification, the candidate thresholds are written into the dynamic threshold library and associated with the scenario labels of the corresponding risk scenarios; if the simulation result deviates from the expectation, start the dynamic correction method, which specifically includes: adopting a parameter sensitivity classification correction mechanism to determine the sensitivity of the candidate thresholds. If the candidate thresholds are high-sensitivity parameters, linear correction is adopted; if the candidate thresholds are medium-low sensitivity parameters, the safety margin rule is used for adjustment. The corrected candidate thresholds are fed back to the vehicle dynamics model of the heavy-duty truck for forward verification again until the error between the simulation result and the candidate thresholds is less than or equal to 5% to pass the verification. The candidate thresholds are written into the dynamic threshold library and associated with the scenario labels of the corresponding risk scenarios.

[0013] Optionally, in some embodiments, the noise removal processing specifically includes:

[0014] Adopt the Savitzky-Golay smoothing processing method to eliminate high-frequency noise by using the local polynomial fitting method, and use Formula 1 as follows:

[0015]

[0016] where y kis the output value of the k-th data point after smoothing, representing the result after filtering; N is the normalization coefficient, specifically the total number of data points within the window, i.e., the window width; N = 2m + 1 = 15, representing a time window covering 150 ms, determined by the sensor sampling rate and the dynamic response frequency of the vehicle, with the current sensor sampling rate being 100 Hz; m is the window radius, defined as including m neighboring points before and after the current point k; a i is the weighting coefficient, determined by local polynomial fitting, and its value depends on the polynomial order and the window width, used to assign different weights to data points at different positions; x k+i is the value of the (k + i)-th point in the original input data, i.e., the original observed values of the current point k and its neighboring points.

[0017] Optionally, in some embodiments, the missing data supplementation process specifically includes:

[0018] When the missing data is less than 100 ms, linear interpolation is used, as shown in Formula 2, as follows:

[0019]

[0020] where x(t) is the interpolation result at time t, i.e., the estimated value for filling the missing point; x(t1) and x(t2) are the original data values of adjacent known points t1 and t2;

[0021] When the missing data is greater than or equal to 100 ms, cubic spline interpolation is used, as shown in Formula 3, as follows:

[0022] S(t) = a(t - t i ) 3 + b(t - t i ) 2 + c(t - t i ) + d(3)

[0023] where S(t) is the cubic spline interpolation result at time t; t i is the starting time point of the current segmented interval, and a, b, c, and d are the coefficients of the cubic polynomial; by piecewise polynomial fitting, the smoothness of the data curve is maintained.

[0024] Optionally, in some embodiments, the outlier removal process specifically includes:

[0025] The 3σ principle is used for outlier filtering, using Formula 4, as follows:

[0026] μ x - 3σ x ≤ x ≤ μ x + 3σ x (4)

[0027] Among them, μ x is the data mean value, and σ x is the standard deviation.

[0028] Optionally, in some embodiments, the key risk indicators at least include at least any one of the braking attenuation rate, kinetic energy conversion rate, tire slip rate, roll angle change rate, lateral load transfer rate, dynamic center of mass offset, yaw angular acceleration, crosswind equivalent mass coefficient, and risk response time margin. Specifically, it includes:

[0029] The braking attenuation rate corresponds to a specific risk scenario representing the risk of braking system performance decline, and is expressed by Formula Five as follows:

[0030]

[0031] Among them, v0 is the initial vehicle speed, μ is the road surface friction coefficient, D actual is the actual braking distance, D min is the minimum braking distance, and g = 9.8 m / s²;

[0032] The kinetic energy conversion rate corresponds to a specific risk scenario representing the risk of braking system performance decline, and is expressed by Formula Six as follows:

[0033]

[0034] Among them, E brake is the energy consumed during braking, is the initial kinetic energy of the vehicle, F brake is the braking force of the vehicle, v is the instantaneous vehicle speed, v0 is the initial vehicle speed, and m is the total mass of the vehicle;

[0035] The tire slip rate corresponds to a specific risk scenario representing the risk of brake failure and skidding, and is expressed by Formula Seven as follows:

[0036]

[0037] Among them, v is the vehicle driving speed, r is the tire radius, and ω is the tire angular velocity;

[0038] The roll angle change rate corresponds to a specific risk scenario representing the risk of vehicle attitude instability, and is expressed by Formula Eight as follows:

[0039]

[0040] Among them, θ t is the roll angle at time t, and Δt = 0.01 s;

[0041] The lateral load transfer ratio corresponds to a specific risk scenario of vehicle rollover risk scenario, which is expressed by Formula 9 as follows:

[0042]

[0043] Among them, F z,r F z,l is the vertical load of the left and right wheels;

[0044] The dynamic center of mass offset corresponds to a specific risk scenario of vehicle rollover risk scenario, which is expressed by Formula 10 as follows:

[0045]

[0046] Among them, h0 is the static center of mass height, a y is the lateral acceleration, B is the wheelbase, g is the acceleration due to gravity, and g = 9.8m / s 2 ;

[0047] The yaw angular acceleration corresponds to a specific risk scenario of vehicle steering out of control risk scenario, which is expressed by Formula 11 as follows:

[0048]

[0049] Among them, ψ t is the yaw angle at time t, Δt = 0.01s;

[0050] The crosswind equivalent mass coefficient corresponds to a specific risk scenario of crosswind-dominated vehicle lateral instability phenomenon risk scenario, which is expressed by Formula 12 as follows:

[0051]

[0052] Among them, F wind is the lateral wind force, h zx is the center of mass height, B is the wheelbase, ρ is the air density, C d is the air dynamic drag coefficient, A is the vehicle lateral projection area, v wind is the crosswind speed;

[0053] The risk response time margin corresponds to a specific risk scenario of quantifying the remaining time from the current state to reaching the danger threshold, providing a reaction window time for the driver or control system, which is expressed by Formula 17 as follows:

[0054]

[0055] Among them, X lim is a certain threshold, X real is the real-time measured value of the corresponding parameter, is the maximum value of the parameter change rate.

[0056] Optionally, in some embodiments, determining the candidate threshold corresponding to the real-time data by the quantile statistical method specifically includes:

[0057] Output a parameter change sequence based on the real-time data, where the parameter change sequence refers to an index sequence calculated for each state of the vehicle, and the parameter change sequence is defined as X = {x1, x2... x n}, where x n represents the corresponding parameter, n represents the number of parameters, and n is a positive integer greater than or equal to 1;

[0058] wherein, the parameter change sequence is arranged in ascending order and satisfies the following conditions:

[0059] x1 ≤ x2 ≤... ≤ x n

[0060] The corresponding threshold X q satisfies the condition: X q = x k +(n·q - k)(x k+1 - x k );

[0061] wherein, the value of q ranges from 0 to 1, expressed as a percentile value, and k is the value obtained by rounding down n·q;

[0062] Compare the value of q with the theoretical value, and take the minimum value as the candidate threshold, which satisfies Formula Thirteen, as follows:

[0063] X candidate = min{X q , X the} (13)

[0064] wherein, X the is the threshold obtained by theoretical calculation.

[0065] Optionally, in some embodiments, the parameter sensitivity grading and correction mechanism is adopted to determine the sensitivity of the candidate threshold, specifically including:

[0066] Adopt Formula Fourteen to determine the sensitivity of the candidate threshold, as follows:

[0067]

[0068] S i is the sensitivity of parameter P i , X lim is the threshold in a specific environment, and P i is the input parameter that affects the threshold Xlim variables

[0069] Optionally, in some embodiments, when the candidate threshold is a highly sensitive parameter, linear correction is adopted, which specifically includes:

[0070] The linear correction is performed using Equation XV as follows:

[0071]

[0072] where X new is the corrected threshold, X candidate is the candidate threshold, D the is the theoretical braking distance, D ini is the initial braking distance, and α is the safety factor;

[0073] When the candidate threshold is a medium or low sensitive parameter, the safety margin rule is adopted for adjustment, which specifically includes:

[0074] The adjustment is performed using Equation XVI as follows:

[0075] μ safe = μ base ×(1 + β·S i ) (16)

[0076] where μ safe is the threshold modified by the safety margin rule, μ base is the theoretical or reference threshold, β is the margin coefficient, and S i is the parameter sensitivity coefficient.

[0077] In a second aspect, an embodiment of the present invention further provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of the method according to any one of the first aspects are implemented.

[0078] In a third aspect, an embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, characterized in that: when the computer program is executed by a processor, the steps of the method according to any one of the first aspects are implemented.

[0079] The above technical solution of the present invention has at least the following beneficial technical effects: The embodiment of the present invention provides a method for identifying the safety risk threshold of a heavy-duty truck. According to the cleaned data and based on the vehicle dynamics model and real-time data, the key risk indicators of the vehicle are calculated. The key risk indicators correspond to specific risk scenarios. The quantile statistical method is used to determine the candidate thresholds corresponding to the key risk indicators, and the candidate thresholds are fed back to the vehicle dynamics model for positive verification to simulate the dynamic response of the vehicle under the threshold boundary conditions. If the candidate thresholds do not meet the requirements, they are corrected, and finally qualified candidate thresholds are obtained. The present invention conducts personalized risk modeling for heavy-duty trucks and performs multi-scenario analysis, providing a systematic solution for comprehensive risk assessment in a multi-source information coupling environment. It integrates vehicle parameters, combines specific road conditions and relevant environmental information to establish a threshold library method. The identification accuracy of the vehicle risk threshold is high, enabling precise early warning. Through dynamic threshold identification, the personalization and precision of risk early warning are realized, thereby effectively reducing the accident rate and ensuring road traffic safety. Description of the Drawings

[0080] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or in the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0081] Figure 1 It is a process flow chart of a method for identifying the safety risk threshold of a heavy-duty truck vehicle provided by an embodiment of the present invention.

[0082] Figure 2 It is a process flow chart of threshold verification in a method for identifying the safety risk threshold provided by an embodiment of the present invention.

[0083] Figure 3 It is a schematic block diagram of an electronic device provided by an embodiment of the present invention.

[0084] Figure 4 It is a lateral load transfer rate data provided by an embodiment of the present invention.

[0085] Figure 5 It is a preprocessed lateral load transfer rate data provided by an embodiment of the present invention.

[0086] Figure 6 It is a dynamic center of mass offset data provided by an embodiment of the present invention.

[0087] Figure 7 It is a preprocessed dynamic center of mass offset data provided by an embodiment of the present invention. Detailed implementation manners

[0088] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0089] In addition, in the following description, the description of known structures and technologies is omitted to avoid unnecessarily confusing the concepts of the present invention.

[0090] If the description in the embodiments of the present application involves "first", "second", etc., the description of "first", "second", etc. is only for descriptive purposes, and cannot be understood as indicating or implying their relative importance or implicitly indicating the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one of such features; the technical solutions between various embodiments may be combined with each other based on what can be achieved by those of ordinary skill in the art.

[0091] It should be noted that the sequence numbers of the order mentioned in the present application do not necessarily represent strict execution in the actual specific implementation process. The sequence numbers are used to distinguish each step for convenience of description and to prevent confusion.

[0092] In addition, the technical features involved in different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0093] Currently, in the prior art, there are problems such as being unable to adapt to dynamic load distributions, complex road conditions, and environmental conditions, resulting in insufficient warning accuracy. Although existing navigation systems can provide basic speed limit suggestions, they lack personalized risk modeling for heavy-duty trucks and are difficult to meet the actual driving safety requirements.

[0094] The embodiments of the present invention provide a method for identifying adaptive thresholds for safety risks of heavy-duty trucks based on multi-dimensional parameter coupling, which is used to solve the problems that current heavy-duty trucks only rely on fixed thresholds for warning, have insufficient safety, and lack analysis of personalized driving safety scenarios for heavy-duty trucks. The present invention focuses on a more comprehensive method for establishing a threshold library.

[0095] Before analyzing the data, it is necessary to clarify the data source and generation logic to ensure the reliability of the analysis basis and the self-consistency of the technical solution. The data is derived from the multi-parameter coupling modeling and dynamic parameter control process: by integrating vehicle body parameters (load, wheelbase, tire pressure, etc.), road geometric conditions (gradient, curvature, superelevation, etc.) and environmental variables (road surface adhesion coefficient, crosswind speed, etc.), a digital model covering typical risk scenarios (ice and snow sharp turns, long downhill braking failure, etc.) is constructed; based on the control variable method, a single risk parameter is dynamically adjusted, and the vehicle dynamic response data (roll angle, lateral acceleration, braking distance, etc.) is collected in real time, and the environmental variables and vehicle state changes are recorded synchronously to form a multi-dimensional data set. This process provides high-density and high-confidence data input for subsequent analysis by isolating interference variables and capturing critical states, and at the same time clarifies the causal relationship between parameters and risk events, supporting the scientific nature of feature extraction and threshold identification. Now enter the main description of the content of the present invention.

[0096] An embodiment of the present invention provides a method for identifying the vehicle safety risk threshold of a heavy-duty truck, as Figure 1 shown, including:

[0097] S1. Data feature extraction: Obtain the vehicle dynamic response data of the heavy-duty truck and environmental variables under various external environments and different road conditions through a multi-source data acquisition system, and calculate the real-time key risk indicators of the heavy-duty truck according to the dynamic response data and the environmental variables, and the key risk indicators correspond to specific risk scenarios;

[0098] S2. Data preprocessing: Preprocess the real-time key risk indicators, and the preprocessing includes noise removal processing, missing data supplement processing, and outlier removal processing to obtain the cleaned real-time data;

[0099] S3. Determine the candidate threshold: Determine the candidate threshold corresponding to the real-time data by the quantile statistical method;

[0100] S4. Threshold Verification: Input the candidate threshold into the vehicle dynamics model of the heavy-duty truck for forward simulation verification, simulate the dynamic response of the heavy-duty truck under the threshold boundary conditions. If the simulation result passes the forward simulation verification, the candidate threshold is written into the dynamic threshold library and associated with the scenario label of the corresponding risk scenario; if the simulation result deviates from the expectation, start the dynamic correction method, which specifically includes: adopting a parameter sensitivity grading correction mechanism to determine the sensitivity of the candidate threshold. If the candidate threshold is a highly sensitive parameter, linear correction is used; if the candidate threshold is a medium or low sensitive parameter, the safety margin rule is used for adjustment. The corrected candidate threshold is fed back to the vehicle dynamics model of the heavy-duty truck for forward verification again until the error between the simulation result and the candidate threshold is less than or equal to 5% to pass the verification. Then, the candidate threshold is written into the dynamic threshold library and associated with the scenario label of the corresponding risk scenario.

[0101] For the specific threshold verification process, refer to Figure 2 as shown.

[0102] The following is a detailed description.

[0103] S1. In the data feature extraction stage, key risk indicators need to be calculated from the dynamic response data and environmental variable data to support the scientific nature of threshold identification.

[0104] The present invention determines the key risk indicators to be calculated and the corresponding related risks as follows:

[0105] (1) Braking attenuation rate, and the scenario label of the corresponding risk scenario is: Braking system effectiveness decline. The formula is as follows:

[0106]

[0107] v0 is the initial vehicle speed, μ is the road surface friction coefficient, D actual is the actual braking distance, D min is the minimum braking distance, g = 9.8m / s

[0108] (2) Kinetic energy conversion rate, and the scenario label of the corresponding risk scenario is: Braking system effectiveness decline. The formula is as follows:

[0109]

[0110] E brake is the energy consumed during braking, is the initial kinetic energy of the vehicle, F brake is the braking force of the vehicle, v is the instantaneous vehicle speed, v0 is the initial vehicle speed, and m is the total mass of the vehicle.

[0111] (3) Tire slip ratio, and the scenario labels for the corresponding risk scenarios are: brake failure, skidding. The formula is as follows:

[0112]

[0113] where v is the vehicle driving speed, r is the tire radius, and ω is the tire angular velocity.

[0114] (4) Roll angle rate of change, and the scenario label for the corresponding risk scenario is: vehicle attitude instability. The formula is as follows:

[0115]

[0116] θ t is the roll angle at time t, and Δt = 0.01 s.

[0117] (5) Lateral load transfer ratio, and the scenario label for the corresponding risk scenario is: vehicle rollover. The formula is as follows:

[0118]

[0119] F z,r F z,l are the vertical loads on the left and right wheels. As Figure 4 shown, Figure 4 is the lateral load transfer ratio calculated based on the real-time left and right wheel vertical loads in the dynamic response data obtained from the multi-source data acquisition system.

[0120] (6) Dynamic center of mass offset, and the scenario label for the corresponding risk scenario is: vehicle rollover. The formula is as follows:

[0121]

[0122] h0 is the static center of mass height, a y is the lateral acceleration, B is the wheelbase, and g is the acceleration due to gravity, generally taking g = 9.8 m / s 2 .

[0123] Specifically, as Figure 6 shown, Figure 6 is the dynamic center of mass offset data calculated based on the above parameters in the dynamic response data obtained from the multi-source data acquisition system.

[0124] (7) Yaw angular acceleration, and the scenario label for the corresponding risk scenario is: steering out of control. The formula is as follows:

[0125]

[0126] ψ t is the yaw angle at time t, and Δt = 0.01 s.

[0127] (8) Crosswind equivalent mass coefficient. The scenario label for the corresponding risk scenario is: Crosswind-dominated vehicle lateral instability phenomenon. The formula is as follows:

[0128]

[0129] F wind is the lateral wind force, h zx is the height of the center of mass, B is the wheelbase, ρ is the air density, C d is the air drag coefficient, A is the lateral projected area of the vehicle, v wind is the crosswind speed.

[0130] (9) Risk response time margin. The scenario label for the corresponding risk scenario is: Quantify the remaining time from the current state to reach the danger threshold, providing a reaction window for the driver or control system. The formula is as follows:

[0131]

[0132] X lim is a certain threshold, X real is the real-time measured value of the corresponding parameter, is the maximum value of the change rate of this parameter.

[0133] S2. Data preprocessing stage mainly includes noise removal, missing data supplementation, outlier removal, etc., which can be simply understood as smoothing processing to ensure the accuracy of subsequent calculations.

[0134] First, use the Savitzky-Golay smoothing algorithm to eliminate high-frequency noise by local polynomial fitting. The formula is:

[0135]

[0136] where, y k is the output value of the k-th data point after smoothing, representing the result after filtering; N is the normalization coefficient, usually the total number of data points within the window, that is, the window width: N = 2m + 1 = 15 (covering a 150 ms time window, determined by the sensor sampling rate and the vehicle dynamic response frequency. Generally, the frequency of a typical frequency band is about 0 - 10 Hz. Here, it is considered that the sensor sampling rate is 100 Hz); m is the window radius, defining m neighboring points before and after the current point k; a i is the weighting coefficient, determined by local polynomial fitting (such as the least squares method), and its value depends on the polynomial order and the window width, used to assign different weights to data points at different positions; x k+i is the value of the (k + i)-th point in the original input data, that is, the original observed values of the current point k and its neighboring points.

[0137] Next, process the missing data:

[0138] ① When the short - term data is missing (< 100ms), use the linear interpolation method as follows:

[0139]

[0140] x(t) is the interpolation result at time t, that is, the estimated value for filling the missing point; x(t1) and x(t2) are the original data values of adjacent known points t1 and t2.

[0141] ② When the long - term data is missing (≥ 100ms), use the cubic spline interpolation method as follows:

[0142] S(t) = a(t - t i ) 3 + b(t - t i ) 2 + c(t - t i ) + d

[0143] S(t) is the cubic spline interpolation result at time t; t i is the starting time point of the current segmented interval (for example, the left - hand endpoint of the i - th interval); a, b, c, and d are the coefficients of the cubic polynomial respectively.

[0144] By piece - wise polynomial fitting, the smoothness of the data curve can be effectively maintained.

[0145] Next, implement the 3σ principle for outlier filtering as follows:

[0146] μ x - 3σ x ≤ x ≤ μ x + 3σ x

[0147] μ x is the data mean, and σ x is the standard deviation.

[0148] Specifically, as shown in Figures 4 to 7 , Figure 4 is the original data of the lateral load transfer rate LTR, one of the key risk indicators calculated based on real - time dynamic response data and environmental variable data. It can be seen that the original data has a lot of noise and the curve is not smooth. After pre - processing, the obtained smooth data is as shown in Figure 5 ; Figure 6 is the original data of the dynamic center - of - mass offset, one of the key risk indicators calculated based on real - time dynamic response data and environmental variable data. After pre - processing, the obtained smooth data is Figure 7 .

[0149] So far, the data preprocessing process is completed.

[0150] S3. Determine the candidate threshold. The candidate threshold corresponding to the key risk indicator is determined by the quantile statistics method. The specific process is as follows:

[0151] Through the above steps, a parameter sequence containing the key risk indicator is initially obtained. It is necessary to determine the required data from this parameter sequence, that is, the "candidate threshold". The present invention determines the "candidate threshold" by the quantile statistics method. The following is the specific process:

[0152] First, a set of change sequences of a parameter is obtained according to the output of the vehicle dynamics digital model, which is defined as:

[0153] X = {x1, x2... x n}

[0154] This set of change sequences is arranged in ascending order and can be:

[0155] x1 ≤ x2 ≤... ≤ x n

[0156] Then the corresponding threshold X q satisfies:

[0157] X q = x k +(n·q - k)(x k+1 - x k )

[0158] The value of q ranges from 0 to 1, representing the percentage value. For example, q = 0.95 represents the 95% quantile. k is the value obtained by rounding down n·q (in the present invention, 0.95 is taken to cover 5% of the high-risk working conditions. Those skilled in the art can set it according to the actual situation. The present invention adopts 0.95 because the effect is better at 0.95. If it is too high, it is easily affected by outliers; if it is too low, the candidate threshold will deviate greatly from the expectation).

[0159] For further detailed description, when the key risk indicator is calculated and preprocessed through step S2, a curve graph is obtained, as Figure 5 shown, Figure 5 which is the change curve of the lateral load transfer rate (LTR), one of the key risk indicators obtained by real-time calculation. This curve is reflected in the set X = {x1, x2... x n}, that is, LTR = {LTR1, LTR2, LTR3………LTR n}, LTR nFor each point on the corresponding curve, the so-called 95% percentile means taking the top 5% of the LTR data as the candidate threshold. That is, in the current LTR data, take the top 5% of the maximum values in the set of change sequences. It can be seen that not all calculated LTR values are candidate thresholds.

[0160] Finally, compare with the theoretical value and take the minimum value to locate the "candidate threshold":

[0161] X candidate = min{X q , X the}

[0162] So far, the determination of the candidate threshold is completed.

[0163] S4. Threshold verification process. The threshold verification and dynamic correction module refers to ensuring the effectiveness and adaptability of the risk threshold through a "simulation - theory" double - closed - loop verification mechanism. In the forward simulation verification stage, the "candidate threshold" is used for the kinetic model simulation to simulate the dynamic response of the vehicle under extreme working conditions and verify whether the threshold completely covers the theoretical risk scenario. If the simulation result passes the forward simulation verification and is consistent with the theoretical expected result, it is directly written into the threshold library; if there is a deviation between the simulation result and the theoretical expectation (such as the braking distance not reaching the expected critical value), dynamic correction is carried out.

[0164] First, declare the method for confirming the sensitivity, that is, the degree of influence of each parameter on the threshold.

[0165]

[0166] S i is the sensitivity of parameter P i , X lim is the threshold in a specific environment, P i is the input parameter, a variable that affects the threshold X lim .

[0167] Through the above formula, the sensitivity of a certain characteristic threshold to related parameters can be obtained.

[0168] Next, classification and correction will be carried out according to the sensitivity level:

[0169] ① When the high - sensitivity parameter (S i ≥ 0.8), the following mathematical expression is used:

[0170]

[0171] Among them, X new is the corrected threshold, X candidate is the candidate threshold, D the is the theoretical braking distance, D iniis the initial braking distance, and α is the safety factor.

[0172] ② When the sensitivity parameter (S i < 0.8), this step is corrected according to the safety margin rule. The aim is to bring a buffer to the risk threshold by adjusting the theoretical threshold upward to avoid the uncertainty caused by environmental fluctuations and unknown interferences. The formula is as follows:

[0173] μ safe = μ base × (1 + β · S i )

[0174] μ safe is the threshold value after being modified by the safety margin rule, μ base is the theoretical or benchmark threshold value, β is the margin coefficient, and S i is the parameter sensitivity coefficient.

[0175] The modified threshold value is verified by forward simulation again. If a closed-loop feedback is formed, the threshold value library is updated.

[0176] An embodiment of the present invention also provides an electronic device 300, as Figure 3 shown, including a memory 301, a processor 302, and a computer program stored on the memory 301 and executable on the processor. When the processor 302 executes the program, the steps of the method described in any one of the above embodiments are implemented.

[0177] An embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method described in any one of the above embodiments are implemented.

[0178] An embodiment of the present invention also provides a computer program product, including a computer program. The computer program is stored in a computer-readable storage medium; when a processor of an electronic device reads the computer program from the computer-readable storage medium, the processor executes the computer program, so that the electronic device executes the steps of the method described in any one of the above embodiments.

[0179] Obviously, the described embodiments are part of the embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0180] It should be understood that the above specific embodiments of the present invention are only used for illustrative explanation or interpretation of the principles of the present invention, and do not constitute a limitation on the present invention. Therefore, any modifications, equivalent replacements, improvements, etc. made without departing from the spirit and scope of the present invention shall be included within the protection scope of the present invention. In addition, the appended claims of the present invention are intended to cover all variations and modifications that fall within the scope and boundaries of the appended claims, or equivalent forms of such scope and boundaries.

Claims

1. A method for identifying vehicle safety risk thresholds for heavy-duty trucks, characterized in that, Including: Obtain the dynamic response data of the heavy-duty truck vehicle and environmental variables under various external environments and different road conditions through a multi-source data acquisition system. According to the dynamic response data and the environmental variables, calculate the real-time key risk indicators of the heavy-duty truck vehicle, and the key risk indicators correspond to specific risk scenarios; Preprocess the real-time key risk indicators. The preprocessing includes noise removal processing, missing data supplementation processing, and outlier removal processing to obtain the cleaned real-time data; Determine the candidate threshold corresponding to the real-time data through the quantile statistical method; Input the candidate threshold into the vehicle dynamics model of the heavy-duty truck for forward simulation verification, simulate the dynamic response of the heavy-duty truck vehicle under the threshold boundary conditions. If the simulation result passes the forward simulation verification, the candidate threshold is written into the dynamic threshold library and associated with the scenario label of the corresponding risk scenario; If the simulation result deviates from the expectation, start the dynamic correction method, which specifically includes: adopting a parameter sensitivity classification correction mechanism to determine the sensitivity of the candidate threshold. If the candidate threshold is a high-sensitivity parameter, linear correction is adopted. If the candidate threshold is a medium-low sensitivity parameter, the safety margin rule is used for adjustment. The corrected candidate threshold is fed back to the vehicle dynamics model of the heavy-duty truck for forward verification again until the error between the simulation result and the candidate threshold is less than or equal to 5% to pass the verification. The candidate threshold is written into the dynamic threshold library and associated with the scenario label of the corresponding risk scenario.

2. The method according to claim 1, wherein The noise removal processing specifically includes: Adopt the Savitzky-Golay smoothing processing method to eliminate high-frequency noise by using the local polynomial fitting method, and use Formula 1 as follows: where y k is the output value of the k-th data point after smoothing, representing the result after filtering; N is the normalization coefficient, specifically the total number of data points within the window, i.e., the window width; N = 2m + 1 = 15, representing a time window covering 150 ms, determined by the sensor sampling rate and the dynamic response frequency of the vehicle, and currently the sensor sampling rate is 100 Hz; m is the window radius, defined as including m neighboring points before and after the current point k; a i is the weighting coefficient, determined by local polynomial fitting, and its value depends on the polynomial order and the window width, and is used to assign different weights to data points at different positions; x k+i is the value of the (k + i)-th point in the original input data, i.e., the original observation values of the current point k and its neighboring points.

3. The method according to claim 1, wherein The missing data supplementation processing specifically includes: If the missing data is less than 100ms, adopt the linear interpolation method, as shown in Formula 2 as follows: Where x(t) is the interpolation result at time t, that is, the estimated value of the missing point; x(t1) and x(t2) are the original data values of adjacent known points t1 and t2; If the missing data is greater than or equal to 100ms, adopt the cubic spline interpolation method, as shown in Formula 3 as follows: S(t) = a(t - t i ) 3 + b(t - t i ) 2 + c(t - t i ) + d(3) Among them, S(t) is the cubic spline interpolation result at time t; t i is the starting time point of the current segmented interval, and a, b, c, and d are the coefficients of the cubic polynomial respectively; through piecewise polynomial fitting, the smoothness of the data curve is maintained.

4. The method according to claim 1, wherein The outlier removal processing specifically includes: Adopt the 3σ principle for outlier filtering, and use Formula 4 as follows: μ x -3σ x ≤x≤μ x +3σ x (4) Among them, μ x is the data mean, and σ x is the standard deviation.

5. The method according to claim 1, characterized in that, The key risk indicators at least include at least any one of the braking attenuation rate, kinetic energy conversion rate, tire slip rate, roll angle change rate, lateral load transfer rate, dynamic center of mass offset, yaw angular acceleration, crosswind equivalent mass coefficient, risk response time margin, specifically including: The braking attenuation rate, corresponding to a specific risk scenario, represents the risk situation of the braking system efficiency decline, and is represented by Formula 5 as follows: Among them, v0 is the initial vehicle speed, μ is the road surface friction coefficient, and D actual is the actual braking distance, and D min is the minimum braking distance, g = 9.8 m / s²; The kinetic energy conversion rate, corresponding to a specific risk scenario, represents the risk situation of the braking system efficiency decline, and is represented by Formula 6 as follows: Among them, E brake is the energy consumed during the braking process, and E v0 is the initial kinetic energy of the vehicle. F brake is the braking force of the vehicle, v is the instantaneous speed of the vehicle, v0 is the initial speed of the vehicle, and m is the total mass of the vehicle; The tire slip rate, corresponding to a specific risk scenario, represents the risk situations of braking failure and slipping, and is represented by Formula 7 as follows: Where v is the vehicle driving speed, r is the tire radius, and ω is the tire angular velocity; The roll rate corresponds to a specific risk scenario of vehicle attitude instability and is expressed by Formula VIII as follows: where θ t is the roll angle at time t, and Δt = 0.01 s; The lateral load transfer rate corresponds to a specific risk scenario of vehicle rollover and is expressed by Formula IX as follows: Among them, F z,r F z,l is the vertical load of the left and right wheels; The dynamic center of mass offset corresponds to a specific risk scenario of vehicle rollover and is expressed by Formula X as follows: Among them, h0 is the static centroid height, a y is the lateral acceleration, B is the track width, g is the acceleration due to gravity, and g = 9.8 m / s 2 ; The yaw angular acceleration corresponds to a specific risk scenario of vehicle steering out of control and is expressed by Formula XI as follows: where, ψ t is the yaw angle at time t, and Δt = 0.01 s; The crosswind equivalent mass coefficient corresponds to a specific risk scenario of vehicle lateral instability dominated by crosswind and is expressed by Formula XII as follows: Among them, F wind is the lateral wind force, h zx is the height of the centroid, B is the track width, ρ is the air density, C d is the aerodynamic drag coefficient, A is the lateral projected area of the vehicle, v wind is the crosswind speed; The risk response time margin corresponds to a specific risk scenario of quantifying the remaining time from the current state to reaching the danger threshold, providing a reaction window time for the driver or control system, and is expressed by Formula XVII as follows: Among them, X lim is a certain threshold value, and X real is the real-time measured value of the corresponding parameter, and is the maximum value of the change rate of this parameter.

6. The method according to claim 1, wherein Determining the candidate threshold corresponding to the real-time data by the quantile statistical method specifically includes: A parameter change sequence is obtained based on the real-time data output. The parameter change sequence refers to an index sequence calculated for each state of the vehicle respectively, and the parameter change sequence is defined as X = {x1, x2... x n}, where x n represents the corresponding parameter, n represents the number of parameters, and n is a positive integer greater than or equal to 1; Wherein, the parameter change sequence is arranged in ascending order and satisfies the following conditions: x1≤x2≤...≤x n The corresponding threshold value X q Satisfy the condition: X q = x k +(n·q - k)(x k+1 - x k ); Wherein, the value of q ranges from 0 to 1, and k is the value obtained by rounding down n·q; The value of q is compared with the theoretical value, and the minimum value is taken as the candidate threshold, satisfying Formula XIII as follows X candidate = min{X q , X the} (13) Among them, X the is the threshold value obtained by theoretical calculation.

7. The method according to claim 1, wherein Adopting a parameter sensitivity grading and correction mechanism to determine the sensitivity of the candidate threshold specifically includes: Using Formula XIV to determine the sensitivity of the candidate threshold as follows: S i is the sensitivity of parameter P i , X lim is the threshold value under a specific environment, P i is the input parameter that affects the threshold value X lim variable.

8. The method according to claim 1, characterized in that If the candidate threshold is a highly sensitive parameter, linear correction is adopted, specifically including: Performing the linear correction using Formula XV as follows: Among them, X new is the corrected threshold, X candidate is the candidate threshold, D the is the theoretical braking distance, D ini is the initial braking distance, and α is the safety factor; If the candidate threshold is a medium-low sensitive parameter, the safety margin rule is adopted for adjustment, specifically including: Performing the adjustment using Formula XVI as follows: μ safe = μ base × (1 + β·S i ) (16) Among them, μ safe is the threshold value after the safety margin rule is modified, μ base is the theoretical or benchmark threshold value, β is the margin coefficient, and S i is the parameter sensitivity coefficient.

9. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the program, the steps of the method according to any one of claims 1-8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, the steps of the method according to any one of claims 1-8 are implemented.

Images