A Quantitative Evaluation Method for Safety Situation in a Vehicle-Following Scenario Based on Multi-Sensor Information

Abstract

The present invention belongs to the technical field of intelligent connected vehicles, and particularly relates to a method for quantitatively evaluating the safety situation of a vehicle-following scenario based on multi-sensor information, including the following steps: Step 1: Construct a database of characteristic parameters for the vehicle-following driving scenario; Step 2: Observe, label, and quantify a specific vehicle-following scenario; Step 3: Determine the membership functions of each characteristic parameter respectively; Step 4: The membership functions of each characteristic parameter are processed by a fuzzy inference system to output a safety situation index. The beneficial effects are as follows: By online integrating multiple characteristic parameters of the vehicle-following driving scenario, the present invention effectively reflects the driving behavior characteristics of the driver and the motion state characteristics between the host vehicle and the preceding vehicle with the safety situation quantification value, improving the accuracy and rationality of the evaluation parameters for vehicle driving safety. In addition, when determining the membership function, the method driven by data effectively avoids the influence of insufficient human experience or evaluation errors on the accuracy of fuzzy inference.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent connected vehicles, and particularly relates to a method for quantitatively evaluating the safety situation of a vehicle following scenario based on multi-sensor information. Background Technique

[0002] With the commercial application of Intelligent and Connected Vehicle (ICV) technology, the riding comfort of vehicles has been effectively improved and the driver's burden has been reduced. Among them, the key to functions such as Adaptive Cruise Control (ACC), Autonomous Emergency Braking (AEB), and Forward Collision Warning (FCW) in ICV technology lies in analyzing the characteristic parameters of the safety situation of the current host vehicle driving scenario and actively controlling the host vehicle's powertrain or braking system, especially in the vehicle following scenario. Most current evaluation methods for safety situations evaluate the safety situation of the current driving scenario by calculating selected evaluation parameters and comparing them with default thresholds. The most widely used evaluation parameters include Time-to-Collision (TTC), Headway-Time (HWT), and Distance-to-Collision (DTC), etc. However, these evaluation parameters are too simplistic, only considering the motion state characteristics (such as relative distance, relative speed, etc.) between the host vehicle and the preceding vehicle, ignoring the influence of driving behavior characteristics, unable to describe the change process of the safety situation of the entire vehicle following scenario, and unable to meet the accuracy and rationality requirements for evaluating the safety situation of the vehicle following scenario, restricting the development of ICV technology. Summary of the Invention

[0003] The purpose of the present invention is to provide a method for quantitatively evaluating the safety situation of a vehicle following scenario based on multi-sensor information to solve the above problems.

[0004] The present invention achieves the above purpose through the following technical solutions:

[0005] A method for quantitatively evaluating the safety situation of a vehicle following scenario based on multi-sensor information includes the following steps:

[0006] Step 1: Construct a database of characteristic parameters for the vehicle following driving scenario;

[0007] Step 2: Observe, label, and quantify a specific vehicle following scenario; the specific method for observing, labeling, and quantifying a specific vehicle following scenario in Step 2 is as follows:

[0008] Cluster the above characteristic parameters into N data sets where N > 1, and represent each center with the center of each set; recruit m drivers with different driving experiences (m > 1) to observe and quantitatively label the driving scenarios represented by the set centers, and calculate the average opinion score of the observed quantification.

[0009] Step 3: Determine the membership functions of each characteristic parameter respectively; specifically, determining the membership functions of each characteristic parameter in Step 3 is as follows:

[0010] Cluster each characteristic parameter according to the fuzzy clustering method.

[0011] Approximately estimate the membership functions of each characteristic parameter using triangular membership functions.

[0012] Step 4: After the membership functions of each characteristic parameter are processed by the fuzzy inference system, a safety situation index is output; specifically, the method for observing the safety situation index after the membership functions of each characteristic parameter are processed by the fuzzy inference system in Step 4 is as follows:

[0013] ① When the driving scenario has no danger and the driver can follow the vehicle normally and smoothly, the safety situation score is 2 points.

[0014] ② When the driving scenario has potential danger and the driver needs to drive carefully, the safety situation score is 1 point.

[0015] ③ When the danger level of the driving scenario increases and the driver needs to concentrate and be ready to brake at any time, the safety situation score is 0 point.

[0016] Preferably, specifically constructing the characteristic parameter database of the following vehicle-following driving scenario in Step 1:

[0017] The acquisition module acquires the driving data of the vehicle-following scenario, and the data acquired by the acquisition module includes the opening degree of the main vehicle's acceleration pedal, the opening degree of the main vehicle's braking pedal, the speed of the main vehicle, the relative distance and relative speed between the main vehicle and the vehicle in front.

[0018] Extract the characteristic parameters of the vehicle-following scenario according to the acquired driving data.

[0019] Construct a characteristic parameter database of the vehicle-following driving scenario according to the above characteristic parameters.

[0020] Preferably, the acquisition module includes an acceleration pedal opening sensor, a braking pedal opening sensor, a vehicle speed sensor, and a millimeter wave radar sensor.

[0021] Preferably, the characteristic parameters of the vehicle-following scenario include the driver's desired acceleration a F 、HWT and TTC -1 .

[0022] Preferably, the mean opinion score is the weighted average of the security situation quantification values.

[0023] Preferably, the acceleration a F , HWT and TTC -1 The characteristic parameters are calculated as follows;

[0024] The characteristic parameter a F The calculation method is as follows:

[0025]

[0026] The characteristic parameters HWT and TTC -1 The calculation methods are as follows:

[0027]

[0028] Preferably, the calculation method of the average score is as follows:

[0029]

[0030] The beneficial effects are as follows: By online fusing multiple characteristic parameters of the following vehicle driving scenarios, the present invention effectively reflects the driving behavior characteristics of the driver and the motion state characteristics between the host vehicle and the preceding vehicle with the security situation quantification value, improving the accuracy and rationality of the evaluation parameters for vehicle driving safety. In addition, when determining the membership function, the method driven by data effectively avoids the influence of insufficient human experience or evaluation errors on the accuracy of fuzzy inference. The application scenarios of the present invention include but are not limited to:

[0031] 1. The present invention can be used in the ADAS system of intelligent connected vehicles, such as functions like ACC, AEB, FCW, etc., to quantitatively evaluate the security situation of the current following vehicle scenario of the host vehicle and improve vehicle driving safety.

[0032] 2. The present invention can also be used in the vehicle integrated energy management system (EMS) and power battery management system (BMS) of new energy vehicles for the integrated energy optimization of new energy vehicles and the charging and discharging management of power batteries, and is especially applicable to the multi-objective optimization control of vehicle power performance, economy, and driving safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0034] Figure 1It is a flow chart of a safety situation quantitative evaluation method for a vehicle - following scenario based on multi - sensor information;

[0035] Figure 2 It is a schematic diagram of a safety situation quantitative evaluation method for a vehicle - following scenario based on multi - sensor information;

[0036] Figure 3 It is a schematic diagram of sensors involved in the quantitative evaluation of the safety situation in the vehicle - following scenario;

[0037] Figure 4 It is an example of the division of the characteristic parameter data set;

[0038] Figure 5 It is a schematic diagram for determining the membership function of the triangular membership function. Specific implementation mode

[0039] The technical solution of the present invention will be further explained below in conjunction with the accompanying drawings:

[0040] As Figures 1-5 shown;

[0041] Example 1:

[0042] A safety situation quantitative evaluation method for a vehicle - following scenario based on multi - sensor information includes the following steps:

[0043] Step 1: Construct a characteristic parameter database for the vehicle - following driving scenario;

[0044] Step 2: Observe and label and quantify a specific vehicle - following scenario;

[0045] Step 3: Determine the membership function of each characteristic parameter respectively;

[0046] Step 4: The membership functions of each characteristic parameter are processed by a fuzzy inference system to output a safety situation index.

[0047] In Step 1, the accelerator pedal opening sensor, the brake pedal opening sensor, the vehicle speed sensor, and the millimeter - wave radar sensor in the acquisition module are all electrically connected to the fuzzy inference module. The accelerator pedal opening sensor, the brake pedal opening sensor, the vehicle speed sensor, and the millimeter - wave radar sensor respectively collect corresponding driving data on the main vehicle accelerator pedal opening, the main vehicle brake pedal opening, the main vehicle speed, the relative distance and relative speed between the main vehicle and the vehicle ahead in the vehicle - following scenario. According to the collected driving data, the acceleration a F 、HWT and TTC -1 of the vehicle - following driving scenario are extracted through the calculation methods of each data. The calculation method of the characteristic parameter a F is as follows;

[0048] Characteristic parameter a FThe calculation method is as follows:

[0049]

[0050] Wherein, δ is the rotational mass conversion coefficient, m is the vehicle mass, T a is the output torque of the drive shaft, T b is the braking torque, r w is the wheel rolling radius, g is the acceleration due to gravity, α is the road gradient, f is the rolling resistance coefficient, C D is the air resistance coefficient, A is the frontal area, v F is the speed of the host vehicle, f a () is the relationship between the output torque of the drive shaft and the accelerator pedal opening, x is the accelerator pedal opening, f b () is the relationship between the braking torque and the brake pedal opening, y is the brake pedal opening. The above formula (1) can be abbreviated as

[0051] a d = f(x, y, v F ) (2)

[0052] The calculation methods of the characteristic parameters HWT and TTC -1 are as follows:

[0053]

[0054]

[0055] Wherein, Δd is the relative distance between the preceding vehicle and the host vehicle, v R is the relative speed between the preceding vehicle and the host vehicle.

[0056] a F is classified into emergency acceleration, gentle acceleration, constant speed, gentle deceleration and emergency deceleration; the category of HWT is classified into small, medium and large; TTC -1 is classified into small, medium and large; then, according to the calculated parameters, a database of characteristic parameters of the following-following driving scenario is constructed through the fuzzy inference system in the fuzzy inference module.

[0057] After that, in step two, the characteristic parameters of the above following-following driving scenario are clustered. The clustered characteristic parameters are divided into N, N>1 data sets. The characteristic parameters with different categories are divided into different data set centers, and each set center represents each center;

[0058] Recruit \(m\) drivers with different driving experiences (\(m>1\)), collect various parameters of different drivers' following - vehicle scenarios, perform fuzzy - system processing on different parameter data to construct a database, then conduct clustering and classification on the database to form different data - set centers, observe and quantitatively label the driving scenarios represented by the set centers, and calculate the average opinion score of the observation and quantification. The average opinion score is the weighted average of the safety - situation quantification values.

[0059] The calculation method of the average score is as follows:

[0060]

[0061] where \(S\) ki is the score of the \(k\) - th group of the \(i\) - th observer, and \(\omega\) i is the weight of the \(i\) - th observer's driving experience, which is expressed as: when the observer has more than 3 years of driving experience, \(\omega\) i = 0.5; when the observer has 2 - 3 years of driving experience, \(\omega\) i = 0.3; when the observer has 1 - 2 years of driving experience, \(\omega\) i = 0.2.

[0062] Then, in step three, cluster each characteristic parameter according to the fuzzy - clustering method, and then approximately estimate the membership - function of each characteristic parameter using the triangular membership function and transfer it to the fuzzy module; the method for determining the membership function is as follows:

[0063] Assume that a certain characteristic parameter is divided into \(K\) classes in the fuzzy - clustering process, that is, it is divided into \(K\) fuzzy sets \(C_1,C_2,\cdots,C\) K . As Figure 3 shown, if the \(h\) - th (\(1 < h < K\)) fuzzy set \(C\) h contains \(N\) h elements, which are \(\{p_1,p_2,\cdots,p\}\) Nh respectively, and the 3 vertices of the corresponding triangular membership function are \((a\) h , 0), \((b\) h , 1), \((c\) h , 0), where

[0064]

[0065] The membership function of the fuzzy set \(C\) h is:

[0066]

[0067] The membership functions of the fuzzy sets \(C_1\) and \(C\) K are respectively:

[0068]

[0069]

[0070] After that, the membership functions of each characteristic parameter in Step 4 are processed by the fuzzy inference system to output the safety situation index. The specific method for observing the safety situation index is as follows;

[0071] ① When there is no danger in the driving scenario and the driver can follow the vehicle normally and smoothly, the safety situation score is 2 points;

[0072] ② When there is a potential danger in the driving scenario and the driver needs to drive carefully, the safety situation score is 1 point;

[0073] ③ When the danger level in the driving scenario increases and the driver needs to concentrate on driving and be ready to brake at any time, the safety situation score is 0 point.

[0074] Embodiment 2:

[0075] A method for quantitatively evaluating the safety situation of following-vehicle driving with multi-sensor information, as Figure 1 shown, includes the following steps:

[0076] S1. Construct a database of characteristic parameters for the following-vehicle driving scenario.

[0077] S11: Collect the driving data of the vehicle following scenario. The collected data includes the opening degree of the main vehicle's acceleration pedal, the opening degree of the main vehicle's brake pedal, the speed of the main vehicle, the relative distance and relative speed between the main vehicle and the vehicle in front. The sensors involved in the above driving data, as Figure 3 shown, include an acceleration pedal opening degree sensor, a brake pedal opening degree sensor, a vehicle speed sensor, and a millimeter-wave radar sensor.

[0078] S12: Extract the characteristic parameters of the following-vehicle scenario according to the collected driving data. The characteristic parameters include the driver's desired acceleration a F , HWT, and TTC -1 . The calculation method of the characteristic parameter a F is as follows:

[0079]

[0080] where δ is the rotating mass conversion coefficient, m is the vehicle mass, T a is the output torque of the drive shaft, T b is the braking torque, r w is the wheel rolling radius, g is the gravitational acceleration, α is the road gradient, f is the rolling resistance coefficient, C D is the air resistance coefficient, A is the frontal area, v F is the speed of the main vehicle, f a ( ) is the relationship between the output torque of the drive shaft and the opening degree of the acceleration pedal, x is the opening degree of the acceleration pedal, fb () is the relationship between the braking torque and the braking pedal opening, and y is the braking pedal opening. The above formula (1) can be abbreviated as

[0081] a d = f(x, y, v F ) (2)

[0082] The calculation methods of the characteristic parameters HWT and TTC are as follows: -1

[0083]

[0084]

[0085] where Δd is the relative distance between the leading vehicle and the host vehicle, and v R is the relative speed between the leading vehicle and the host vehicle.

[0086] S13: Construct a following driving scenario characteristic parameter database according to the above characteristic parameters.

[0087] S2. Quantitatively label a specific following scenario.

[0088] S21: Cluster the above characteristic parameters, divide them into N data sets, and represent each set by the center of the set. The data sets are divided as shown in Figure 3 , or can be divided in other ways. The categories of a F are divided into emergency acceleration, gentle acceleration, constant speed, gentle deceleration, and emergency deceleration; the categories of HWT are divided into small, medium, and large; the categories of TTC -1 are divided into small, medium, and large.

[0089] S22: Recruit m drivers with different driving experiences to observe and quantitatively label the following scenarios represented by the center of the set, and calculate the average score of the observation and quantification. The average opinion score is the weighted average of the safety situation quantification values.

[0090] The calculation method of the above average score is as follows:

[0091]

[0092] where S ki is the score of the kth group of the ith observer, and ω i is the weight of the driving experience of the ith observer, expressed as: when the observer has more than 3 years of driving experience, ω i = 0.5; when the observer has 2 - 3 years of driving experience, ω i = 0.3; when the observer has 1 - 2 years of driving experience, ω i = 0.2. ​

[0093] S3. Determine the membership functions of each characteristic parameter respectively.

[0094] S31: Cluster each characteristic parameter according to the fuzzy clustering method.

[0095] S32: Approximately estimate the membership functions of each characteristic parameter by using triangular membership functions. The method for determining the membership functions is as follows:

[0096] Assume that a certain characteristic parameter is divided into K classes during the fuzzy clustering process, that is, it is divided into K fuzzy sets C1, C2,..., C K . As Figure 3 shown, if the hth (1 < h < K) fuzzy set C h contains N h elements, which are {p1, p2,..., p Nh} respectively, and the 3 vertices of the corresponding triangular membership function are (a h , 0), (b h , 1), (c h , 0), where

[0097]

[0098] the membership function of the fuzzy set C h is:

[0099]

[0100] The membership functions of the fuzzy sets C1 and C K are respectively:

[0101]

[0102]

[0103] S4. Output the security situation index after processing the membership functions of each characteristic parameter by the fuzzy inference system. The specific observation method of the security situation index is as follows;

[0104] If there is no danger in the current driving scenario and the driver can follow the vehicle normally and smoothly, the security situation score is 2 points; if the current driving scenario has potential danger and the driver needs to drive carefully, the security situation score is 1 point; if the danger level in the current driving scenario increases and the driver needs to concentrate on driving and be ready to brake at any time, the security situation score is 0 point.

[0105] Working principle: The accelerator pedal opening sensor, brake pedal opening sensor, vehicle speed sensor, and millimeter-wave radar sensor in the acquisition module respectively obtain the corresponding driving data of the host vehicle's accelerator pedal opening, host vehicle's brake pedal opening, host vehicle's speed, relative distance, and relative speed between the host vehicle and the preceding vehicle in the following vehicle scenario. Different parameter data are used to calculate the acceleration a F , HWT, and TTC -1 of the following vehicle scenario characteristic parameters. Then, a following vehicle driving scenario characteristic parameter database is constructed based on the above-mentioned following vehicle scenario characteristic parameters. Next, each following vehicle scenario characteristic parameter in the database is observed, marked, and quantified. By clustering each following vehicle scenario characteristic parameter, N data sets are divided. At the same time, m drivers with different driving experiences are solicited to observe and quantify the driving scenarios represented by the set centers, and the observed and quantified average opinion scores are calculated through relevant formulas. After that, each characteristic parameter is clustered according to the fuzzy clustering method, and the membership function of each characteristic parameter is approximately estimated using the triangular membership function. Then, it is processed and discriminated through the fuzzy processing system in the fuzzy processing module, and the safety situation index is output. If there is no danger in the current driving scenario, the driver can follow the vehicle normally and smoothly, and the safety situation score is 2 points; if the current driving scenario has potential danger and the driver needs to drive carefully, the safety situation score is 1 point; if the danger level of the current driving scenario increases and the driver needs to concentrate and be ready to brake at any time, the safety situation score is 0 points.

[0106] The above shows and describes the basic principle, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification only illustrates the principle of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements fall within the scope of the present invention claimed.

Claims

1. A safety situation quantitative evaluation method for following vehicle scenarios based on multi-sensor information, characterized in that It includes the following steps: Step 1: Construct a database of characteristic parameters for the following - vehicle driving scenario; Step 2: Observe, annotate, and quantify a specific following - vehicle scenario; The specific methods for observing, annotating, and quantifying a specific following - vehicle scenario in Step 2 are as follows: Cluster the above - mentioned characteristic parameters, divide them into N, N>1 data sets, and represent each set with its center; Recruit m, m>1 drivers with different driving experiences to observe, quantify, and annotate the driving scenarios represented by the set centers, and calculate the average opinion score of the observation and quantification; Step 3: Determine the membership functions of each characteristic parameter respectively; The specific method for determining the membership functions of each characteristic parameter in Step 3 is as follows: Cluster each characteristic parameter according to the fuzzy clustering method; Approximately estimate the membership functions of each characteristic parameter using triangular membership functions; Step 4: After the membership functions of each characteristic parameter are processed by a fuzzy inference system, a safety situation index is output; After the membership functions of each characteristic parameter are processed by a fuzzy inference system in Step 4, a safety situation index is output. The specific observation method for the safety situation index is as follows: ① When there is no danger in the driving scenario and the driver can follow the vehicle normally and smoothly, the safety situation score is 2 points; ② When the driving scenario has potential danger and the driver needs to drive carefully, the safety situation score is 1 point; ③ When the danger level of the driving scenario increases and the driver needs to concentrate on driving and be ready to brake at any time, the safety situation score is 0 point.

2. The safety situation quantitative evaluation method for a vehicle following scenario based on multi-sensor information according to claim 1, wherein, The specific methods for constructing the database of characteristic parameters for the following - vehicle driving scenario in Step 1 are as follows: The acquisition module acquires the driving data of the vehicle following - vehicle scenario. The data acquired by the acquisition module includes the opening degree of the main vehicle's acceleration pedal, the opening degree of the main vehicle's brake pedal, the speed of the main vehicle, the relative distance and relative speed between the main vehicle and the vehicle in front; Extract the characteristic parameters of the following - vehicle scenario according to the acquired driving data; Construct a database of characteristic parameters for the following - vehicle driving scenario according to the above - mentioned characteristic parameters.

3. The safety situation quantitative evaluation method for a vehicle-following scenario based on multi-sensor information according to claim 2, wherein: The acquisition module includes an acceleration - pedal - opening sensor, a brake - pedal - opening sensor, a vehicle - speed sensor, and a millimeter - wave radar sensor.

4. A safety situation quantitative evaluation method for a vehicle following scenario based on multi-sensor information according to claim 2, characterized in that: The following vehicle scene characteristic parameters include the driver's expected acceleration a F , HWT and TTC -1 .

5. The safety situation quantitative evaluation method for following vehicle scenarios based on multi-sensor information according to claim 1, characterized in that: The average opinion score is the weighted average of the safety situation quantification values.

6. The safety situation quantitative evaluation method for following vehicle scenarios based on multi-sensor information according to claim 4, characterized in that: The acceleration a F , HWT, and TTC -1 The characteristic parameters are calculated as follows; The characteristic parameter a F The calculation method is as follows: The characteristic parameters HWT and TTC -1 are calculated as follows: 。 7. A safety situation quantitative evaluation method for following vehicle scenarios based on multi-sensor information according to claim 1, characterized in that: The calculation method of the average score is as follows:

Images