A self-vehicle driving risk quantification method and system based on human, vehicle and road angles

Abstract

The application discloses a self-vehicle driving risk quantification method and system based on human, vehicle and road angles, and comprises the following steps: obtaining driving information; obtaining self-vehicle lane deviation risk and self-vehicle overspeed risk based on the driving information, and then obtaining a first driving risk value of self-vehicle driving risk representing a road constraint angle; obtaining collision severity and collision urgency of self-vehicle and other vehicles based on the driving information, and then obtaining a second driving risk value of self-vehicle driving risk representing a collision angle of self-vehicle and other vehicles; obtaining a risk perception factor of self-vehicle driving based on the driving information, and then obtaining a third driving risk value of self-vehicle driving risk representing a driver perception angle. The method can help vehicles to comprehensively, accurately and intuitively know the risk degree of self-vehicle driving state, and timely adjust the state to ensure driving safety.

Description

Technical Field

[0001] This invention belongs to the field of traffic safety research technology, specifically relating to a method and system for quantifying vehicle driving risks based on the perspectives of people, vehicles, and roads. Background Technology

[0002] The quantification of driving risks is the foundation for improving active safety systems and developing vehicle intelligence. Vehicles, especially autonomous vehicles, employ a variety of driving strategies during operation, each facing different driving risks. Appropriate methods for quantifying driving risks can pre-assess the risks of various driving strategies before the vehicle makes a driving strategy decision, thereby providing a basis for the decision-making process and contributing to ensuring driving safety.

[0003] Existing methods for quantifying driving risks, whether from a single or multiple perspective, are mostly based on risk fields. They emphasize the interaction between various elements in the driving environment but neglect the autonomy of the vehicle, making them unsuitable for providing decision-making basis for driving strategies. Summary of the Invention

[0004] To address the aforementioned issues, this invention proposes a method and system for quantifying autonomous vehicle driving risks based on the perspectives of people, vehicles, and roads. This method quantifies driving risks from the perspective of autonomous vehicle driving in a dynamic and time-varying traffic environment, helping vehicles to promptly understand the risk level of their driving status and adjust their status in a timely manner to ensure driving safety.

[0005] On one hand, the present invention provides a method for quantifying vehicle driving risk based on the perspectives of people, vehicles, and roads, comprising the following steps: acquiring driving information, wherein the driving information includes vehicle data of the vehicle itself, vehicle data of other vehicles on the road in which the vehicle is traveling, and road data of the road in which the vehicle is traveling; based on the driving information, obtaining the vehicle's lane departure risk and speeding risk, and based on the vehicle's lane departure risk and speeding risk, obtaining a first driving risk value characterizing the vehicle's driving risk from the perspective of road constraints; based on the driving information, obtaining the severity and urgency of collision between the vehicle and other vehicles, and based on the severity and urgency of collision, obtaining a second driving risk value characterizing the vehicle's driving risk from the perspective of collision between the vehicle and other vehicles; based on the driving information, obtaining a risk perception factor of the vehicle while driving, and based on the risk perception factor, obtaining a third driving risk value characterizing the vehicle's driving risk from the perspective of driver perception.

[0006] Preferably, the driving information includes: lane line type (solid or dashed line), speed limit v limit The combined speed of the vehicle, v EV lateral speed v of the vehicle x,EVlongitudinal speed v of the vehicle y,EV Horizontal position of the vehicle (x) EV longitudinal position of the vehicle y EV Vehicle width W EV Vehicle length L EV Lateral speed v of the vehicle in front x,LV longitudinal speed v of the vehicle in front y,LV Horizontal position of the front vehicle (x) LV Longitudinal position y of the front vehicle LV Front vehicle width W LV Length of the front vehicle L LV Lateral speed v of the lane-changing vehicle x,LCV Longitudinal speed v of the lane-changing vehicle y,LCV Lateral position of the lane-changing vehicle (x) LCV Longitudinal position y of the lane-changing vehicle LCV Lane changing vehicle width W LCV Lane changer length L LCV Longitudinal position y of the lane-changing vehicle's entry point p The heading angle θ of the vehicle EV The heading angle θ of the vehicle in front LV The heading angle θ of the lane-changing vehicle LCV The lateral position x of the left lane line of the lane where the vehicle is located. leftboder The lateral position x of the right lane line of the lane where the vehicle is located. rightboder .

[0007] Preferably, the lateral positions of the vehicle, the preceding vehicle, and the lane-changing vehicle are all the lateral positions of the center of the front bumper of the vehicle, and the longitudinal positions of the vehicle, the preceding vehicle, and the lane-changing vehicle are all the longitudinal positions of the center of the front bumper of the vehicle; the values ​​of the lateral positions are the values ​​corresponding to the horizontal axis in the preset coordinate system, and the values ​​of the longitudinal positions are the values ​​corresponding to the vertical axis in the preset coordinate system.

[0008] Preferably, based on the driving information, the lane departure risk of the vehicle is obtained, specifically calculated using the following formula: R line =ke -Δd , where R line Δd represents the risk of lane departure by the vehicle; k represents the constraint coefficient, which is 1 when the lane line is solid and 1-p(lc) when the lane line is dashed, where p(lc) represents the probability of lane changing; Δd represents the minimum lateral distance between the vehicle's edge and the lane lines on both sides, i.e., Δd = min(d left ,d right ),in, Where, x EV Indicates the lateral position of the vehicle, x leftboder Indicates the lateral position of the left lane line of the lane where the vehicle is located, x rightboderIndicates the lateral position of the right lane line of the lane where the vehicle is located, W EV θ represents the width of the vehicle. EV This indicates the heading angle of the vehicle.

[0009] Preferably, based on the driving information, the risk of the vehicle speeding is obtained, specifically calculated using the following formula: Among them, R limitspeed This indicates the risk of a vehicle speeding, v EV v represents the combined velocity of the vehicle. limit This indicates the speed limit.

[0010] Preferably, based on the vehicle's lane departure risk and speeding risk, a first driving risk value characterizing the vehicle's driving risk of the road constraint angle is obtained, specifically calculated using the following formula: R road =0.5*R line +0.5*R limitspeed , where R road R represents the first driving risk value. line R represents the risk of lane departure for a vehicle. limitspeed This indicates the risk level of a vehicle speeding.

[0011] Preferably, based on the driving information, the collision urgency between the vehicle and other vehicles is obtained, specifically including the following steps: the driving information also includes vehicle data of other vehicles on the road where the vehicle is traveling; based on the driving information, the lateral collision urgency coefficient and the longitudinal collision urgency coefficient between the vehicle and other vehicles are obtained; based on the lateral collision urgency coefficient and the longitudinal collision urgency coefficient, the collision urgency experienced by the vehicle from other vehicles from the lateral and longitudinal directions is obtained through a piecewise function; based on the collision severity of the vehicle and other vehicles and the collision urgency, a second driving risk value representing the driving risk of the vehicle, which characterizes the collision angle between the vehicle and other vehicles, is obtained.

[0012] Preferably, the other vehicles include the vehicle in front of the vehicle in the lane where the vehicle is located, and the vehicle in the adjacent lane where the vehicle is located that intends to change lanes.

[0013] Preferably, based on the driving information, the lateral collision urgency coefficient between the vehicle and the lane-changing vehicle is obtained, specifically calculated using the following formula: r urgency,lat =|THW lon,LCV,P -THW lon,EV,P |, where r urgency,lat Indicates the lateral collision urgency coefficient; point P is the point of entry when the lane-changing vehicle cuts into the lane of the vehicle it is in from the lane of the lane-changing vehicle; THW lon,LCV,P This represents the time it takes for the lane-changing vehicle to reach the cut-off point P while maintaining a constant speed. THW lon,EV,P This represents the time it takes for the vehicle to maintain a constant speed and reach the entry point P. Among them, y P Indicates the longitudinal position of the lane-changing vehicle's entry point, y LCV Indicates the longitudinal position of the lane-changing vehicle, v y,LCV y represents the longitudinal speed of the lane-changing vehicle. EV Indicates the longitudinal position of the vehicle, v y,EV This indicates the longitudinal speed of the vehicle.

[0014] Preferably, based on the driving information, the longitudinal collision urgency coefficient between the vehicle and the vehicle in front is obtained, specifically calculated using the following formula: Where, r urgency,lon Indicates the longitudinal collision urgency factor; TTC EV,LV Indicates the time of collision between the vehicle and the vehicle in front; △y EV,LV △v represents the longitudinal distance between the rear of the vehicle in front and the front of the vehicle itself. EV,LV y represents the relative longitudinal velocity between the vehicle in front and the vehicle itself; LV Indicates the longitudinal position of the vehicle in front, W LV θ represents the width of the vehicle in front. LV L represents the heading angle of the vehicle in front. LV Indicates the length of the vehicle in front, y EV Indicates the longitudinal position of the vehicle, v y,EV v represents the longitudinal velocity of the vehicle. y,LV This indicates the longitudinal speed of the vehicle in front.

[0015] Preferably, in the preset coordinate system, the horizontal axis extends from the origin of the preset coordinate system through the left lane line of the vehicle's lane to the right lane line of the vehicle's lane, and the vertical axis extends from the origin of the preset coordinate system in the direction of the vehicle's travel. Based on the lateral collision urgency coefficient and the longitudinal collision urgency coefficient, the degree of lateral and longitudinal collision urgency experienced by the vehicle from other vehicles is obtained through a piecewise function, specifically calculated using the following formula: Wherein, when a lane-changing vehicle intends to cut into its own lane from the right adjacent lane, ΔX represents the lateral distance between the lane-changing vehicle and the right lane line of its own lane, i.e., ΔX = x LCV -x rightboder When a lane-changing vehicle intends to cut into its own lane from the lane adjacent to its own lane, ΔX represents the lateral distance between the lane-changing vehicle and the left lane line of its own lane, i.e., ΔX = x. leftboder -x LCV ; where x LCV Indicates the lateral position of the lane-changing vehicle, x rightboder Indicates the lateral position of the right lane line of the lane where the vehicle is located, x leftboder ΔX represents the lateral position of the left lane line of the vehicle's current lane; ρ is the probability of the vehicle detecting a lane-changing vehicle cutting into its lane; ΔX≥0 indicates that the lane-changing vehicle is in the stage before cutting into the vehicle's current lane. Indicates that since the lane-changing car and from the car in front Larger values ​​in the value represent the degree of lateral and longitudinal collision urgency U experienced by the vehicle from other vehicles; when ΔX < 0, it indicates that the lane-changing vehicle is in the stage after cutting into the vehicle's lane. After the lane-changing vehicle cuts into the vehicle's lane, it is updated to the preceding vehicle, and the piecewise function... This indicates that the vehicle receives feedback from the vehicle in front after the update. From the updated predecessor The value U represents the degree of urgency of a collision with another vehicle from its lateral and longitudinal directions.

[0016] Preferably, based on the driving information, the severity of the collision between the vehicle and other vehicles is obtained. Specifically, the method is as follows: based on the driving information, a collision severity prediction model is used to obtain the collision severity.

[0017] Preferably, the severity S of the collision is divided into three levels: S=1 represents a minor collision, S=2 represents a moderate collision, and S=3 represents a severe collision.

[0018] Preferably, based on the severity of the collision between the vehicle and another vehicle and the urgency of the collision, a second driving risk value, representing the driving risk of the vehicle at the angle of collision with the other vehicle, is obtained and calculated using the following formula: R vehicle =U 1 / S , where R vehicle The value indicates the second driving risk, where U represents the urgency of the collision and S represents the severity of the collision.

[0019] Preferably, the self-driving vehicle is a manually driven vehicle, the risk perception factor of the self-driving vehicle is the actual instantaneous acceleration of the self-driving vehicle, and the third driving risk value is calculated using the following formula: Among them, R human denoted as α, where α is the driving risk value of the manually driven vehicle, and α is the actual instantaneous acceleration of the manually driven vehicle.

[0020] Preferably, the vehicle is an autonomous vehicle. The method for obtaining the risk perception factor of the vehicle while driving based on the driving information is as follows: constructing an input set and output set of a risk perception factor prediction model based on the NGSIM database; training a regression prediction model based on a Gaussian process using the input set and output set of the risk perception factor prediction model to construct the risk perception factor prediction model; the driving information also includes data information used to obtain the input set of the risk perception factor prediction model; and obtaining the risk perception factor of the vehicle while driving based on the driving information and the risk perception factor prediction model.

[0021] The risk perception factor prediction model includes a horizontal risk perception factor prediction model and a vertical risk perception factor prediction model. The risk perception factors include horizontal risk perception factors predicted by the horizontal risk perception factor prediction model and vertical risk perception factors predicted by the vertical risk perception factor prediction model.

[0022] Preferably, the method for constructing the input and output sets of the risk perception factor prediction model based on the NGSIM database is as follows: An initial impact feature set of vehicle driving risk is constructed based on the NGSIM database; the initial impact feature set is filtered according to feature influence and feature redundancy to obtain an impact feature set; the mean and standard deviation of the impact feature set within the time interval [t-ΔT,t] are obtained as inputs to the risk perception factor prediction model to construct the input set; the instantaneous acceleration at time t is obtained as the output of the model to construct the output set; ΔT is the driver's reaction lag time; wherein, the initial impact feature set includes an initial horizontal impact feature set and an initial vertical impact feature set, and the impact feature set includes a horizontal impact feature set for training the horizontal risk perception factor prediction model and a vertical impact feature set for training the vertical risk perception factor prediction model.

[0023] Preferably, the method for obtaining the influence feature set by feature screening of the initial influence feature set based on feature influence degree and feature redundancy is as follows: calculate the Pearson correlation coefficient between the initial features in the initial influence feature set and their corresponding instantaneous acceleration at time t, and screen out the initial features with a Pearson coefficient value ≥ 0.6 as the initial features with high feature influence degree; calculate the Pearson correlation coefficient between the screened initial features, and remove the initial features with lower feature influence degree from the two initial features with strong mutual correlation to obtain the influence feature set, where strong mutual correlation means that the Pearson correlation coefficient between the two initial features is ≥ 0.8.

[0024] Preferably, the driving information further includes data information for obtaining the input set of the risk perception factor prediction model. The method for obtaining the risk perception factor of the vehicle during driving based on the driving information and using the risk perception factor prediction model is as follows: based on the driving information, obtain the lateral influence feature set and the longitudinal influence feature set of the vehicle during driving; based on the lateral influence feature set of the vehicle during driving, obtain the lateral risk perception factor using the lateral risk perception factor prediction model; and based on the longitudinal influence feature set of the vehicle during driving, obtain the longitudinal risk perception factor using the longitudinal risk perception factor prediction model.

[0025] Preferably, based on the risk perception factor, a third driving risk value representing the vehicle's driving risk from the driver's perception angle is obtained, specifically calculated using the following formula: When a lane-changing vehicle cuts in from the right side of the vehicle, ΔX = x LCV -x rightboderWhen a lane-changing vehicle cuts in from the left side of the vehicle, ΔX = x leftboder -x LCV , where x LCV Indicates the lateral position of the lane-changing vehicle, x rightboder Indicates the lateral position of the right lane line of the lane where the vehicle is located, x leftboder Indicates the lateral position of the left lane line of the lane where the vehicle is located; R humanlike R represents the third driving risk value for the autonomous vehicle; ρ is the probability of the autonomous vehicle detecting a lane-changing vehicle cutting in; humanlike,lat This represents the lateral traffic risk value. Among them, a EV,LCV For horizontal risk perception factors; R humanlike,lon This represents the longitudinal driving risk value. Among them, a EV,LV This is a longitudinal risk perception factor.

[0026] Preferably, it further includes: obtaining a comprehensive driving risk value that characterizes the driving risk of a vehicle from multiple perspectives of people, vehicles, and roads based on the first driving risk value, the second driving risk value, and the third driving risk value.

[0027] Preferably, the comprehensive driving risk value is calculated using the following formula: R = R humanlike +R vehicle +R road Where R represents the comprehensive driving risk value, R humanlike R represents the third driving risk value of autonomous vehicles. vehicle R represents the second driving risk value. road This indicates the first driving risk value.

[0028] On the other hand, the present invention provides a vehicle driving risk quantification system based on the perspectives of people, vehicles, and roads, comprising:

[0029] The driving information acquisition module includes: a driving information acquisition unit, used to acquire driving information;

[0030] The first driving risk quantification module includes: a lane departure risk acquisition unit for obtaining lane departure risk of the vehicle based on the driving information; a speeding risk acquisition unit for obtaining speeding risk of the vehicle based on the driving information; and a first driving risk value acquisition unit for obtaining a first driving risk value representing the driving risk of the vehicle with respect to road constraint angles based on the lane departure risk and speeding risk.

[0031] The second driving risk quantification module includes: a severe collision severity acquisition unit, used to obtain the collision severity of the vehicle and other vehicles based on the driving information; a collision urgency acquisition unit, used to obtain the collision urgency of the vehicle and other vehicles based on the driving information; and a second driving risk value acquisition unit, used to obtain a second driving risk value representing the driving risk of the vehicle, which characterizes the collision angle between the vehicle and other vehicles, based on the collision severity and collision urgency.

[0032] The third driving risk quantification module includes: a risk perception factor acquisition unit, used to acquire risk perception factors of the vehicle while driving based on the driving information; and a third driving risk value acquisition unit, used to obtain a third driving risk value that characterizes the driving risk of the vehicle from the driver's perception angle based on the risk perception factors.

[0033] Preferably, it further includes: a comprehensive driving risk value acquisition unit, used to obtain a comprehensive driving risk value that characterizes the driving risk of a vehicle from multiple perspectives of people, vehicles and roads based on the first driving risk value, the second driving risk value and the third driving risk value.

[0034] Compared with the prior art, the present invention has the following beneficial effects:

[0035] (1) This invention quantifies driving risks from the perspective of the driver's perception, the collision angle between the vehicle and surrounding vehicles, and the road constraints of the vehicle. It can comprehensively quantify the multi-angle driving risks from the perspective of the vehicle in a dynamic and time-varying traffic environment. It can also use radar charts to help describe the driving risks faced by the vehicle from various angles. This helps the vehicle to quantify the driving environment from its own perspective, thereby providing a comprehensive and effective basis for the vehicle's future decisions.

[0036] (2) This invention addresses the risk of poor driving of a vehicle under the constraints of road traffic rules by obtaining the risk of lane departure and the risk of speeding of the vehicle in the horizontal and vertical directions, respectively, and then obtaining the first driving risk value that characterizes the driving risk of the vehicle in terms of the road constraint angle. This helps the vehicle to know the risk level of its driving status in a timely manner and adjust its status in a timely manner to ensure driving safety.

[0037] (3) To quantify the lane departure risk of a vehicle while driving, this invention proposes a lane line constraint coefficient for the vehicle by combining the driver's intention and the lane line type. Then, it establishes a quantitative expression for the lane departure risk of a vehicle in a structured road that considers lateral constraints. This expression can quantitatively assess the lateral departure risk of a vehicle in real time, which helps the vehicle to perceive and adjust its driving state in a timely manner. In response to the bad driving behavior of speeding, this invention proposes a quantitative expression for the speeding risk of a vehicle in a structured road that considers longitudinal constraints. This expression can represent the risk change of a vehicle from normal driving speed to speeding, which helps to provide early warning of speeding behavior.

[0038] (4) This invention starts from the perspective of the vehicle's driving in a complex and ever-changing traffic environment, and takes into account the vehicle's motion state. It comprehensively considers the traffic conflicts between the vehicle and other vehicles during the driving process. It considers the collision problem between the vehicle and other vehicles from the perspective of collision between the vehicle and other vehicles. It considers the impact of lateral collision on driving risk as well as the impact of longitudinal collision on driving risk. It can accurately quantify the driving risk of the vehicle and provide a basis for decision-making for the driving strategy to be executed by the vehicle, which is conducive to ensuring driving safety.

[0039] (5) When obtaining the lateral collision urgency coefficient, this invention combines the lane-changing vehicle's intention to cut into the lane of the vehicle and the position of the cutting point to express the lateral collision urgency during the lane-changing vehicle's cutting process; when obtaining the longitudinal collision urgency coefficient, it considers the influence of the geometry and heading angle of the vehicle and the preceding vehicle on the longitudinal relative distance between the two vehicles, which can more accurately express the longitudinal collision urgency. This invention considers that when a lane-changing vehicle cuts in, during the ΔX≥0 stage, the vehicle may simultaneously be subjected to both longitudinal collision from the preceding vehicle and lateral collision from the lane-changing vehicle, and extracts the lateral collision urgency coefficient from the lane-changing vehicle's lateral collision urgency coefficient. and from the car in front Larger values ​​in the piecewise function represent the degree of lateral and longitudinal collision urgency U experienced by the vehicle from other vehicles, providing a more precise indication of the collision urgency. During the phase ΔX < 0, if a lane-changing vehicle successfully enters the vehicle's lane, it becomes the vehicle in front, and the vehicle may experience a longitudinal collision with this updated vehicle. Indicates that it comes from the vehicle preceding the updated vehicle. From the updated predecessor The value U represents the degree of urgency of a collision from another vehicle in both the lateral and longitudinal directions. It takes into account the change of role of the vehicle before and after changing lanes (whether it continues to be the vehicle changing lanes or becomes the vehicle in front), and can more accurately reflect the degree of collision urgency.

[0040] (6) This invention starts from the driver's perception perspective and uses the acceleration taken by the driver as the risk factor perceived by the driver. The driving risk quantification method based on this risk factor can express the impact of other vehicles in the environment on the driving safety of the vehicle from the driver's perception perspective. It is conducive to the vehicle to discover early safety hazards, can provide a basis for driving warning mechanism, and can also provide a theoretical reference for the intervention mechanism of driving assistance system.

[0041] (7) Based on the human driving database, this invention uses the data mining capabilities of machine learning to summarize the risk perception patterns of human drivers in the driving environment during the driving process, and establishes a risk perception factor prediction model, which helps autonomous vehicles quantify the risk level in the environment from a human-like perception perspective, so as to serve the decision-making process of autonomous vehicles.

[0042] (8) The autonomous vehicle driving risk quantification system of the present invention corresponds to the autonomous vehicle driving risk quantification method of the present invention. It can accurately, comprehensively and intuitively quantify autonomous vehicle driving risks, provide decision-making basis for the driving strategy to be executed by the vehicle, and help ensure driving safety. Attached Figure Description

[0043] Figure 1 This is a schematic diagram of the method for quantifying vehicle driving risks based on the perspectives of people, vehicles, and roads in Embodiment 1 of the present invention;

[0044] Figure 2 This is a schematic diagram of the vehicle's lane departure during the driving process in Embodiment 1 of the present invention;

[0045] Figure 3 This is a schematic diagram of the vehicle, the vehicle in front, and the vehicle changing lanes on the road where the vehicle is traveling in Embodiment 1 of the present invention, which may be involved in lateral or longitudinal collisions.

[0046] Figure 4 This is a schematic diagram of the state before and after a lane-changing vehicle cuts into the lane where the vehicle is located on the road in Embodiment 1 of the present invention.

[0047] Figure 5 This is a line graph showing the feature influence degree of the lateral initial influence feature in Embodiment 1 of the present invention;

[0048] Figure 6 This is a line graph showing the feature influence degree of the initial longitudinal influence feature in Embodiment 1 of the present invention.

[0049] Figure 7 This is a test result diagram of a test example of the horizontal risk perception factor prediction model in Embodiment 1 of the present invention;

[0050] Figure 8 This is a test result diagram of another test example of the horizontal risk perception factor prediction model in Embodiment 1 of the present invention;

[0051] Figure 9 This is a test result diagram of a test example of the longitudinal risk perception factor prediction model in Embodiment 1 of the present invention;

[0052] Figure 10 This is a test result diagram of another test example of the longitudinal risk perception factor prediction model in Embodiment 1 of the present invention;

[0053] Figure 11 These are schematic diagrams of three simulation scenarios in Embodiment 1 of the present invention;

[0054] Figure 12 This is a trend chart of the changes in the first driving risk value, the second driving risk value, and the third driving risk value obtained by the quantification method of the present invention under simulation scenario 1 in embodiment 1 of the present invention;

[0055] Figure 13 This is a trend chart of the changes in the first driving risk value, the second driving risk value, and the third driving risk value obtained by the quantification method of the present invention under simulation scenario 2 in embodiment 1 of the present invention;

[0056] Figure 14 This is a trend chart of the changes in the first driving risk value, the second driving risk value, and the third driving risk value obtained by the quantification method of the present invention under simulation scenario 3 in embodiment 1 of the present invention;

[0057] Figure 15 These are risk radar charts for three simulation scenarios in Embodiment 1 of the present invention. Detailed Implementation

[0058] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0059] Example 1

[0060] This embodiment provides a method for quantifying vehicle driving risks based on the perspectives of people, vehicles, and roads, including the following steps:

[0061] 1. Obtain driving information.

[0062] The driving information includes vehicle data of the vehicle and road data of the road on which the vehicle is traveling, in order to obtain a first driving risk value that characterizes the driving risk of the vehicle from the road constraint angle; further, in order to obtain a second driving risk value that characterizes the driving risk of the vehicle from the angle of collision with other vehicles, the driving information also includes vehicle data of other vehicles on the road on which the vehicle is traveling; further still, in order to obtain a third driving risk value that characterizes the driving risk of the vehicle from the driver's perception angle, the driving information also includes data information for obtaining the input set of the risk perception factor prediction model (the risk perception factor prediction model is described in detail in step 4).

[0063] In this embodiment, taking a highway driving environment as an example, the vehicle is traveling on a highway, and the driving environment is a normal traffic environment, without considering extreme operating conditions or extreme weather conditions. To facilitate the description of the relationships between various traffic participants in the highway driving environment, they are distinguished by codes. The vehicle whose driving risk to be quantified is referred to as the ego vehicle (EV, which could be a future autonomous vehicle). Other vehicles include the leading vehicle (LV, the first vehicle in front of the ego vehicle in its lane) and lane-changing vehicles (LCVs) in adjacent lanes that intend to change lanes. In addition, other vehicles around the ego vehicle on its driving road are dynamic traffic vehicles (SVs) around the ego vehicle.

[0064] In this embodiment, the method of obtaining driving information is not limited, and those skilled in the art can choose as needed. For example, they can use radar technology, navigation and positioning systems, cameras, sensors and other electronic devices to complete the perception and positioning of surrounding vehicles, and obtain driving information in the driving environment in real time through vehicle-to-vehicle communication, vehicle-road cooperation and other Internet of Vehicles technologies.

[0065] When the required driving information is future driving information, the driving information can be obtained based on the future driving trajectories of the vehicle and surrounding vehicles. The future driving trajectories of the vehicle and surrounding vehicles can be obtained according to the method described in the published literature "Yuan-YuanRen,Lan Zhao,Xue-Lian Zheng,et al.A Method for Predicting Diverse Lane-Changing Trajectories of Surrounding Vehicles Based on Early Detection of Lane Change[J].IEEE Access,2022,Vol.10:17451-17472.", which will not be elaborated here.

[0066] In other words, the required driving information can be obtained through existing technologies, such as the following: lane marking type (solid or dashed), speed limit (v). limit The combined speed of the vehicle, v EV lateral speed v of the vehicle x,EV longitudinal speed v of the vehicle y,EV Horizontal position of the vehicle (x) EV longitudinal position of the vehicle y EV Vehicle width W EV Vehicle length L EV Lateral speed v of the vehicle in front x,LVlongitudinal speed v of the vehicle in front y,LV Horizontal position of the front vehicle (x) LV Longitudinal position y of the front vehicle LV Front vehicle width W LV Length of the front vehicle L LV Lateral speed v of the lane-changing vehicle x,LCV Longitudinal speed v of the lane-changing vehicle y,LCV Lateral position of the lane-changing vehicle (x) LCV Longitudinal position y of the lane-changing vehicle LCV Lane changing vehicle width W LCV Lane changer length L LCV Longitudinal position y of the lane-changing vehicle's entry point p The heading angle θ of the vehicle EV The heading angle θ of the vehicle in front LV The heading angle θ of the lane-changing vehicle LCV The lateral position x of the left lane line of the lane where the vehicle is located. leftboder The lateral position x of the right lane line of the lane where the vehicle is located. rightboder Table 1 shows the above driving information symbols and their descriptions.

[0067] Table 1

[0068] <![CDATA[v EV ]]> Self-speed <![CDATA[v limit ]]> Speed ​​limit <![CDATA[v x,EV ]]> The lateral speed of the vehicle <![CDATA[v x,LV ]]> Lateral speed of the vehicle in front <![CDATA[v y,EV ]]> longitudinal speed of the vehicle <![CDATA[v y,LV ]]> longitudinal speed of the vehicle in front <![CDATA[x EV ]]> lateral position of the vehicle <![CDATA[x LV ]]> Lateral position of the front vehicle <![CDATA[y EV ]]> longitudinal position of the vehicle <![CDATA[y LV ]]> longitudinal position of the front vehicle <![CDATA[W EV ]]> Vehicle width <![CDATA[W LV ]]> Front vehicle width <![CDATA[L EV ]]> Self-length <![CDATA[L LV ]]> Front vehicle length <![CDATA[x LCV ]]> Lateral position of lane change vehicle <![CDATA[y p ]]> Longitudinal position of lane-changing vehicle entry point <![CDATA[y LCV ]]> Longitudinal position of lane changer <![CDATA[θ LCV ]]> Heading angle of the lane-changing vehicle <![CDATA[v x,LCV ]]> Lateral speed of lane-changing vehicle <![CDATA[θ EV ]]> The vehicle's heading angle <![CDATA[v y,LCV ]]> Longitudinal speed of lane-changing vehicle <![CDATA[θ LV ]]> The heading angle of the vehicle in front <![CDATA[W LCV ]]> Lane change vehicle width <![CDATA[x leftboder ]]> The lateral position of the left lane line of the lane where the vehicle is located <![CDATA[L LCV ]]> Lane changer length <![CDATA[x rightboder ]]> The right lane line of the lane where the vehicle is located is laterally positioned.

[0069] In this embodiment, the horizontal axis of the preset coordinate system extends from the origin of the preset coordinate system through the left lane line of the vehicle's lane to the right lane line of the vehicle's lane, and the vertical axis extends from the origin of the preset coordinate system in the direction of the vehicle's travel. Those skilled in the art will understand that the setting of the preset coordinate system is not limited, and technicians can make reasonable selections and adjustments according to actual conditions.

[0070] In this embodiment, the lateral positions of the vehicle, the preceding vehicle, and the lane-changing vehicle are all lateral positions of the center of the front bumper of their respective vehicles. Similarly, the longitudinal positions of the vehicle, the preceding vehicle, and the lane-changing vehicle are all longitudinal positions of the center of the front bumper of their respective vehicles. In this embodiment, both the lateral and longitudinal positions are coordinate points in a preset coordinate system. The value of the lateral position is the value corresponding to the horizontal axis of the preset coordinate system, and the value of the longitudinal position is the value corresponding to the vertical axis of the preset coordinate system. For example, the lateral and longitudinal positions of the vehicle represent the coordinate points of the center of the front bumper of the vehicle in the preset coordinate system; the lateral and longitudinal positions of the preceding vehicle represent the coordinate points of the center of the front bumper of the preceding vehicle in the preset coordinate system; and the lateral and longitudinal positions of the lane-changing vehicle represent the coordinate points of the center of the front bumper of the lane-changing vehicle in the preset coordinate system.

[0071] 2. Obtain the first driving risk value.

[0072] When vehicles travel on structured roads, they must comply with two fundamental traffic rules: lane boundary lines and speed limits. Lane boundary lines restrict the lateral range of movement, while speed limits restrict the longitudinal speed. Being constrained laterally by lane boundary lines means there is a risk of lane departure; being constrained by speed limits means there is a risk of speeding. Both lane departure and speeding are considered poor driving behaviors. Therefore, the risks of poor driving under traffic rules can be expressed in two main ways: the risk of lane departure due to lateral lane boundary line constraints, and the risk of speeding due to longitudinal speed limits.

[0073] 2.1 Based on the driving information, the lane departure risk of the vehicle is obtained.

[0074] like Figure 2 As shown, there is a risk of lane departure for vehicles. This embodiment aims to quantify the risk of lane departure while driving. By combining the driving intention of the vehicle with the lane line type, a constraint coefficient of the lane line on the vehicle is proposed. Then, a quantitative expression for the lane departure risk of vehicles considering lateral constraints in structured roads is established. This can quantitatively assess the lateral departure risk of vehicles in real time, which helps vehicles to perceive and adjust their driving status in a timely manner.

[0075] Specifically, it is calculated using the following formula:

[0076] R line =ke -Δd ,

[0077] in,

[0078] R line This indicates the risk of lane departure for the vehicle.

[0079] k represents the constraint coefficient. When the lane line is a solid line, k = 1. When the lane line is a dashed line, k = 1 - p(lc), where p(lc) represents the probability of a vehicle changing lanes.

[0080] △d represents the minimum lateral distance between the edge of the vehicle and the lane lines on both sides, i.e., △d = min(d left ,d right ),in, Where, x EV Indicates the lateral position of the vehicle, x leftboder Indicates the lateral position of the left lane line of the lane where the vehicle is located, x rightboder Indicates the lateral position of the right lane line of the lane where the vehicle is located, W EV θ represents the width of the vehicle. EV This indicates the heading angle of the vehicle.

[0081] The method for obtaining the vehicle lane-changing probability p(lc) in this embodiment corresponds to the "maximum log-likelihood value of the model output probability loglik" in the patent application with application number 201910836440.3 entitled "A method for recognizing a driver's lane-changing intention", i.e., ln(p(lc))=loglik. Taking the exponent of both sides of the formula, we can get p(lc)=e loglik This is the vehicle lane-changing probability p(lc) required in this embodiment, which will not be elaborated further here.

[0082] This embodiment uses the relative distance Δd between the vehicle and the lane line to express the risk of lane departure. The main types of lane lines (solid and dashed lines) have different levels of constraint on the vehicle. Therefore, this invention proposes a constraint coefficient k to express the degree of constraint imposed on the vehicle by the lane line type. If the lane boundary line is a solid line, the constraint strength of the lane boundary is the greatest; vehicles are not allowed to cross the lane line. Driving over a solid line is also a dangerous behavior. Therefore, the constraint risk of a solid lane boundary line is inversely proportional to the lateral distance between the vehicle and the lane line, and the vehicle's lane-changing intention does not need to be considered; the constraint coefficient k is set to 1. If the lane boundary line is a dashed line, it means that vehicles are allowed to cross the lane line to perform lane changing behavior. Therefore, the constraint strength of the dashed lane line on the vehicle changes with the vehicle's behavioral intention. If the vehicle is maintaining its lane, then the dashed line is equivalent to a solid line. If the vehicle intends to change lanes, then the constraint of the dashed line on the vehicle weakens as the probability of changing lanes increases. That is, the constraint coefficient k of the dashed lane line decreases as the probability of the vehicle changing lanes p(lc) increases. If the probability of the vehicle changing lanes is 1, then the dashed lane line has no constraint on the vehicle.

[0083] 2.2 Based on the driving information, the vehicle's speeding risk is obtained.

[0084] This embodiment proposes a quantitative expression for the risk of vehicle speeding in structured roads that considers longitudinal constraints, targeting the undesirable driving behavior of speeding. This expression can represent the risk change of a vehicle from normal driving speed to speeding, which helps to provide early warning of speeding behavior.

[0085] Specifically, it is calculated using the following formula:

[0086]

[0087] Among them, R limitspeed This indicates the risk of a vehicle speeding, v EV v represents the combined velocity of the vehicle. limit This indicates the speed limit.

[0088] For roads with speed limits, the maximum speed a vehicle can travel must not exceed the limit; otherwise, there is a risk of speeding. The risk value of the speed limit constraint is directly proportional to the speed difference between the vehicle's combined speed and the speed limit. Furthermore, there is a risk of speeding even before the vehicle exceeds the speed limit, because once the speed limit is exceeded, the risk has already been created. Therefore, the potential risk before speeding gradually increases as the speed difference between the vehicle's combined speed and the speed limit increases. A negative speed difference indicates that the vehicle is not speeding, and the risk value is less than 1; a speed difference of 0 indicates that the speed limit has been reached, and the risk value reaches 1; a speed difference greater than 0 indicates that speeding has occurred, and the risk value is greater than 1.

[0089] 2.3 Based on the vehicle lane departure risk and vehicle speeding risk, a first driving risk value representing the vehicle driving risk of the road constraint angle is obtained.

[0090] Vehicles are constrained by road structure and traffic rules, and face risks of vehicle deviation and speeding. Since these two factors have roughly the same constraining effect on vehicles, they are combined by assigning equal weights, specifically calculated using the following formula:

[0091] R road =0.5*R line +0.5*R limitspeed ,

[0092] Among them, R road R represents the first driving risk value. line R represents the risk of lane departure for a vehicle. limitspeed This indicates the risk level of a vehicle speeding.

[0093] 3. Obtain the second driving risk value.

[0094] 3.1 Based on the driving information, the severity of the collision between the vehicle and other vehicles is obtained.

[0095] In this embodiment, the collision severity is obtained through a collision severity prediction model, which is described in our patent application with application number 202210497607.X entitled "A Method for Predicting the Severity of a Potential Collision Accident" and will not be repeated here. The collision severity S in this embodiment corresponds to the collision severity in the patent application with application number 202210497607.X entitled "A Method for Predicting the Severity of a Potential Collision Accident": {severity1, severity2, severity3}, that is, the collision severity S is divided into three levels, S=1 represents a minor collision, S=2 represents a moderate collision, and S=3 represents a severe collision.

[0096] 3.2 Based on the driving information, the degree of urgency of collision between the vehicle and other vehicles is obtained.

[0097] In this embodiment, as Figure 3 As shown, during the driving of an EV, it may collide with the preceding vehicle (LV) and the lane-changing vehicle (LCV). The EV has a potential longitudinal collision with the preceding vehicle (LV) and a potential lateral collision with the lane-changing vehicle (LCV). Both directions are equally important to the EV's driving safety. Therefore, this embodiment obtains the lateral collision imminence coefficient and the longitudinal collision imminence coefficient between the EV and other vehicles from both the lateral and longitudinal directions.

[0098] (1) Based on the driving information, obtain the lateral collision urgency coefficient between the vehicle and the lane-changing vehicle.

[0099] Specifically, it is calculated using the following formula:

[0100] r urgency,lat =|THW lon,LCV,P -THW lon,EV,P |,

[0101] in,

[0102] r urgency,lat Indicates the lateral collision urgency factor;

[0103] Point P is the entry point when a lane-changing vehicle cuts into the lane it is in from its own lane.

[0104] THW lon,LCV,P This represents the time it takes for the lane-changing vehicle to reach the cut-off point P while maintaining a constant speed.

[0105] THW lon,EV,P This represents the time it takes for the vehicle to maintain a constant speed and reach the entry point P.

[0106] Among them, y P Indicates the longitudinal position of the lane-changing vehicle's entry point, y LCV Indicates the longitudinal position of the lane-changing vehicle, v y,LCV y represents the longitudinal speed of the lane-changing vehicle. EV Indicates the longitudinal position of the vehicle, v y,EV This indicates the longitudinal speed of the vehicle.

[0107] In this embodiment, the entry point P can be obtained from the lane-changing trajectory prediction model of other vehicles in the published literature "Yuan-Yuan Ren, Lan Zhao, Xue-LianZheng, et al. AMethod for Predicting Diverse Lane-Changing Trajectories of Surrounding Vehicles Based on Early Detection of Lane Change[J].IEEE Access,2022,Vol.10:17451-17472.", which will not be elaborated here.

[0108] In this embodiment, THW lon,LCV,P and THW lon,EV,P The larger the difference, the lower the urgency and risk of a collision between the EV and the LCV. When the difference is 0, it means that the two vehicles arrive at the entry point P at the same time, and the urgency of the collision is the greatest.

[0109] (2) Based on the driving information, obtain the longitudinal collision urgency coefficient between the vehicle and the vehicle in front.

[0110] Specifically, it is calculated using the following formula:

[0111]

[0112] in,

[0113] r urgency,lon Indicates the longitudinal collision urgency factor;

[0114] TTC EV,LV Indicates the time of collision between the vehicle and the vehicle in front;

[0115] △y EV,LV △v represents the longitudinal distance between the rear of the vehicle in front and the front of the vehicle itself. EV,LV Indicates the relative longitudinal speed between the vehicle in front and the vehicle behind;

[0116] y LV Indicates the longitudinal position of the vehicle in front, W LV θ represents the width of the vehicle in front. LV L represents the heading angle of the vehicle in front. LV Indicates the length of the vehicle in front, y EV Indicates the longitudinal position of the vehicle, v y,EV v represents the longitudinal velocity of the vehicle. y,LV This indicates the longitudinal speed of the vehicle in front.

[0117] (3) Based on the lateral collision urgency coefficient and the longitudinal collision urgency coefficient, the degree of lateral and longitudinal collision urgency experienced by the vehicle from other vehicles is obtained by using a piecewise function.

[0118] In this embodiment, the relative lateral position of the two vehicles is used to distinguish the type of collision between the EV and another vehicle. For example... Figure 4 As shown, taking the lane-changing vehicle LCV with the intention to cut into the lane of the vehicle EV from the right adjacent lane of the vehicle EV as an example, during the process of the lane-changing vehicle LCV cutting into the lane, the influence between the vehicle EV and the lane-changing vehicle LCV gradually changes from lateral to longitudinal. The potential collision between the vehicle EV and the lane-changing vehicle LCV also changes from a lateral collision (before the lane-changing vehicle LCV cuts in, the vehicle EV is simultaneously subjected to a potential lateral collision from the lane-changing vehicle LCV and a potential longitudinal collision from the preceding vehicle LV) to a longitudinal collision (after the lane-changing vehicle LCV cuts in, the lane-changing vehicle LCV's identity changes to the preceding vehicle, and the vehicle EV is only subjected to a potential longitudinal collision from the preceding vehicle LV, which is now the identity of the lane-changing vehicle LCV). Therefore, this embodiment proposes a piecewise function related to the lateral position of the lane-changing vehicle LCV to express the degree of urgency of the collisions from other vehicles to the vehicle EV under different relative motion states.

[0119] Specifically, it is calculated using the following formula:

[0120]

[0121] in,

[0122] When a lane-changing vehicle intends to cut into its own lane from the right-adjacent lane, ΔX represents the lateral distance between the lane-changing vehicle and the right lane line of its own lane, i.e., ΔX = x. LCV -x rightboder When a lane-changing vehicle intends to cut into its own lane from the lane adjacent to its own lane, ΔX represents the lateral distance between the lane-changing vehicle and the left lane line of its own lane, i.e., ΔX = x. leftboder -x LCV ; where x LCV Indicates the lateral position of the lane-changing vehicle, x rightboder Indicates the lateral position of the right lane line of the lane where the vehicle is located, x leftboder Indicates the lateral position of the left lane line of the lane where the vehicle is located;

[0123] ρ represents the probability of the lane-changing vehicle merging into the lane detected by the vehicle; ΔX≥0 indicates that the lane-changing vehicle is in the stage before merging into the lane of the vehicle. Indicates that since the lane-changing car and from the car in front Larger values ​​in the value represent the degree of lateral and longitudinal collision urgency U experienced by the vehicle from other vehicles; when ΔX < 0, it indicates that the lane-changing vehicle is in the stage after cutting into the vehicle's lane. After the lane-changing vehicle cuts into the vehicle's lane, it is updated to the preceding vehicle, and the piecewise function... This indicates that the vehicle receives feedback from the vehicle in front after the update. From the updated predecessor The value U represents the degree of urgency of a collision with another vehicle from its lateral and longitudinal directions.

[0124] In this embodiment, the cut-in probability ρ of the lane-changing vehicle (LCV) detected by the EV can be obtained from the lane-changing intent identification model of other vehicles in the published literature "Yuan-Yuan Ren, Lan Zhao, Xue-Lian Zheng, et al. A Method for Predicting Diverse Lane-Changing Trajectories of Surrounding Vehicles Based on Early Detection of Lane Change[J].IEEE Access,2022,Vol.10:17451-17472.", which will not be elaborated here.

[0125] In this embodiment, it should be noted that the piecewise function is based on ΔX. Under the preset coordinate system of this embodiment, the collision urgency U is the piecewise function mentioned above. However, when the preset coordinate system is different, the piecewise function's segmentation conditions with respect to ΔX change adaptively.

[0126] 3.3 Based on the severity of the collision between the vehicle and other vehicles and the urgency of the collision, a second driving risk value is obtained, which characterizes the driving risk of the vehicle and the collision angle between the vehicle and other vehicles.

[0127] Calculated using the following formula:

[0128] R vehicle =U 1 / S ,

[0129] Among them, R vehicle The value indicates the second driving risk, where U represents the urgency of the collision and S represents the severity of the collision.

[0130] This embodiment uses the power function R. vehicle =U 1 / S The vehicle's driving risk in relation to the angle of collision between itself and other vehicles is expressed as a value within the range of 0-1. This quantification method uses 0 and 1 as upper and lower limits. For various driving strategies, the driving information is different, and the quantified second driving risk value corresponding to each driving strategy is also different. The closer the second driving risk value is to 1, the greater the vehicle's driving risk in relation to the angle of collision between itself and other vehicles corresponding to that driving strategy. The closer the driving risk value is to 0, the smaller the vehicle's driving risk in relation to the angle of collision between itself and other vehicles corresponding to that driving strategy. This is more conducive to the decision-making basis of vehicle driving strategy.

[0131] 4. Obtain the third driving risk value.

[0132] 4.1 Based on the driving information, obtain the risk perception factor of the vehicle while driving.

[0133] Human drivers constantly adjust their acceleration to regulate their perceived real-time risk level during driving; this is a unique risk perception ability of human drivers. Therefore, if the vehicle is manually driven, its instantaneous acceleration can be directly used as a risk perception factor. The instantaneous acceleration of a manually driven vehicle can be acquired in real time during driving (i.e., the actual instantaneous acceleration during driving), for example, through the vehicle's acceleration sensor. Accordingly, based on this risk perception factor, a third driving risk value, representing the vehicle's driving risk from the driver's perception perspective, is obtained and can be directly calculated using the formula... Calculate, where R human denoted as the third driving risk value for manually driven vehicles, and 'a' as the actual instantaneous acceleration of the manually driven vehicle.

[0134] In this embodiment, as a preferred solution, the vehicle is an autonomous vehicle. The risk perception factor of the autonomous vehicle is the instantaneous acceleration predicted by the risk perception factor prediction model from a human-like perspective. Since the driver's perceived risk mainly comes from vehicles in front of the vehicle in the lane and vehicles in adjacent lanes, but considering that the interaction between the two and the vehicle is different, the resulting risks are also different. Therefore, this embodiment establishes risk perception factor prediction models from both the lateral and longitudinal aspects to predict the lateral risk perception factor and the longitudinal risk perception factor of the autonomous vehicle, respectively, to characterize the risk perception factor of the vehicle.

[0135] Among them, the lateral risk perception factor refers to the instantaneous acceleration of the vehicle when a lane-changing vehicle cuts in, at which time the lane-changing vehicle is cutting between the vehicle and the vehicle in front, and the instantaneous acceleration of the vehicle is mainly affected by the lane-changing vehicle; the longitudinal risk perception factor refers to the instantaneous acceleration of the vehicle when no lane-changing vehicle cuts in, at which time the instantaneous acceleration of the vehicle is mainly affected by the vehicle in front.

[0136] Specifically, the method for obtaining risk perception factors while driving includes the following steps:

[0137] S1. Construct the input and output sets of the risk perception factor prediction model based on the NGSIM database.

[0138] S11. Construct an initial impact feature set of vehicle driving risks based on the NGSIM database.

[0139] The specific method for constructing the lateral initial influence feature set is as follows: ① Obtain a vehicle lane-changing sample set based on the NGSIM database. The specific method can be found in the invention patent application document No. 202210323142.6, entitled "A Method for Extracting Lane-Changing Process and Analyzing Lane-Changing Characteristics Based on Unsupervised Technology"; ② Select a vehicle-entering sample set from the vehicle lane-changing sample set. The specific criteria for defining the entry behavior are: when the front of the vehicle touches the lane line, the time difference (THW) between the vehicle and the lane-changing vehicle is ≤2s; during the entire lane-changing process (within 10s before and after crossing the lane line as determined in patent application document 202210323142.6), the vehicle exhibits braking behavior, i.e., the minimum acceleration of the vehicle during the lane-changing process is <-0.92m / s². 2 ③ Based on the sample set of vehicles entering the lane, a lateral initial impact feature set is constructed to characterize the influence of the lane-changing vehicle's LCV on the vehicle's EV. The lateral initial impact feature set constructed in this embodiment is shown in Table 2:

[0140] Table 2 Horizontal Initial Influence Feature Set

[0141]

[0142] The method for constructing the longitudinal initial impact feature set is as follows: ① Obtain vehicle following samples based on the NGSIM dataset. For specific methods, please refer to the published literature "Zhu Ting, Yang Hongtai, Zhong Xinzhi, Zou Yajie. Analysis of dangerous following behavior characteristics based on intelligent driver model [J]. Transportation and Transport, 2021, 37(3):87-91", which will not be repeated here. ② Construct the longitudinal initial impact feature set representing the impact of the preceding vehicle's LV on the vehicle's EV based on the following samples. The longitudinal initial impact feature set constructed in this embodiment is shown in Table 3:

[0143] Table 3. Initial Vertical Influence Feature Set

[0144]

[0145]

[0146] S12. Based on the feature influence degree and feature redundancy, the initial influence feature set is filtered to obtain the influence feature set.

[0147] To reduce computational complexity, it is necessary to find features that have a significant impact on the vehicle. Therefore, this invention performs feature filtering on the initial impact feature set based on feature impact and feature redundancy to obtain features that have a significant impact on the vehicle, i.e., the impact feature set.

[0148] The characteristic influence degree of this invention refers to the correlation between initial influence features and risk perception factors. This invention obtains the characteristic influence degree by calculating the Pearson correlation coefficient between the initial features in the initial influence feature set and their corresponding instantaneous acceleration at time t. The characteristic influence degree of the horizontal initial influence feature in this embodiment is as follows: Figure 5 As shown, the feature influence degree of the initial vertical influence feature is as follows: Figure 6 As shown.

[0149] This invention uses feature influence calculation to select features with a Pearson coefficient value ≥ 0.6 as features strongly correlated with the risk perception factor for prediction. Based on this, in this embodiment, the initial horizontal feature selected according to the feature influence is v. LCV v y,LV a x,LCV , dx-1, dy-1, THW-1, d-1; initial vertical feature is v LV v y,LV , THW-2, dx-2, dy-2, d_θ-2, dv-2, dvy-2, d-2;

[0150] To reduce the cross-correlation between influencing features and thus decrease the complexity of the risk factor prediction model, this invention performs feature redundancy analysis on the initially selected features. By calculating the Pearson correlation coefficient between the initial features, this invention removes the initial feature with lower influence from two highly correlated initial features, thus obtaining the influencing feature set. The cross-correlation results of the horizontal initial features in this embodiment are shown in Table 4, and the cross-correlation results of the vertical initial features are shown in Table 5.

[0151] Table 4 Results of initial cross-correlation of lateral features

[0152] f1 1 0.99 0.61 -0.41 -0.24 0.21 -0.24 f4 / 1 0.60 -0.39 -0.22 0.22 -0.23 f5 / / 1 -0.52 -0.73 -0.51 -0.73 f8 / / / 1 0.37 0.75 0.37 f9 / / / / 1 0.48 1.00 f13 / / / / / 1 0.57 f16 / / / / / / 1

[0153] Table 5. Results of initial longitudinal feature cross-correlation.

[0154]

[0155]

[0156] In this invention, two initial features with a Pearson correlation coefficient greater than or equal to 0.8 are selected as features with strong cross-correlation. Based on the cross-correlation results of the initial features, features f4 and f9 in the initial lateral features are excluded in this embodiment. The constructed lateral influence feature set includes: the absolute distance d-1 between the lane-changing vehicle and the vehicle, the relative lateral distance dx-1 between the lane-changing vehicle and the vehicle, the headway THW-1 between the lane-changing vehicle and the vehicle, and the lateral acceleration a of the lane-changing vehicle. x,LCV Lane change vehicle speed v LCV, that is, Fi-1={d-1, dx-1, THW-1, a x,LCV v LCV After excluding features f1', f11', and f16' from the initial longitudinal features, the constructed longitudinal influence feature set includes: the longitudinal speed v of the preceding vehicle. y,LV The relative headway between the preceding and following vehicles is THW-2, the relative heading angle between the preceding and following vehicles is d_θ-2, the relative lateral distance between the preceding and following vehicles is dx-2, the relative longitudinal distance between the preceding and following vehicles is dy-2, and the relative resultant velocity between the preceding and following vehicles is dv-2, i.e., Fi-2={v y,LV , THW-2, d_θ-2, dx-2, dy-2, dv-2}.

[0157] S13. Obtain the mean and standard deviation of the influencing feature set in the time interval [t-ΔT,t] as input to the risk perception factor prediction model, construct the input set, obtain the instantaneous acceleration at time t as the output of the model, construct the output set, where ΔT is the driver's reaction lag time.

[0158] The inputs and outputs of the horizontal risk perception factor prediction model are obtained from the horizontal impact feature set, while the inputs and outputs of the vertical risk perception factor prediction model are obtained from the vertical impact feature set.

[0159] Assuming the moment when the vehicle reacts to acceleration or deceleration when faced with a vehicle cutting in is t, considering the driver's reaction time and the braking system's reaction time, we can conclude that there will be a lag of △T between the impact of the lane-changing vehicle or the vehicle in front on the vehicle and the driver's actual reaction. In this embodiment, △T is set to 1s. This value is determined with reference to the conclusions on driver reaction time in existing research.

[0160] S2. Use the input and output sets of the risk perception factor prediction model to train the regression prediction model based on Gaussian process, and construct the risk perception factor prediction model.

[0161] Gaussian Process Regression (GPR) is a regression prediction model based on Gaussian processes. Because Gaussian processes support the modeling of univariate or multivariate normally distributed random variables and have a high degree of flexibility in their model structure, compared with traditional machine learning methods such as ANN and SVM, the GPR model reduces the number of parameters that need to be identified, simplifies the parameter optimization process, and enhances the convergence of the model. It has gradually developed into a mainstream machine learning model. Therefore, this invention constructs a risk perception factor prediction model based on the GPR algorithm to achieve risk perception factor prediction.

[0162] The input and output sets of the risk perception factor prediction model constructed in step S1 are divided into training and validation sets in an 8:2 ratio.

[0163] S21. Using the GPR toolkit in MATLAB software, construct a risk perception factor prediction model for autonomous vehicles based on the training set.

[0164] S22. The trained risk perception factor prediction model is validated using the validation set. Table 6 shows the indicators used to evaluate the model of this invention.

[0165] Table 6

[0166]

[0167] The results show that the ACU of the horizontal risk perception factor prediction model in this embodiment is 0.9815 on the validation set, and the ACU of the vertical risk perception factor prediction model is 0.9690 on the validation set, indicating that both the horizontal and vertical risk perception factor prediction models of this invention have excellent prediction performance.

[0168] Furthermore, to further evaluate the advantages of the prediction model of this invention, this embodiment compares its advantages with other prediction models. Since neural network prediction algorithms require a large sample size, we use Support Vector Machines (SVMs), which do not have strict requirements on sample size, as a comparison model. Using Python's built-in SVM toolbox, we built a Support Vector Machine Regression (SVR) prediction model. Based on the same validation set, and using RMSE and MAE as evaluation metrics, the model performance comparison is shown in Table 7. It can be concluded that the prediction error of the GPR-based model is much smaller than that of the SVR model.

[0169] Table 7 Model Performance Comparison

[0170] SVR 0.6868 0.5433 GPR 0.1475 0.2986 Error reduction 78.52％ 45.04％

[0171] S23. Test the constructed risk perception factor prediction model.

[0172] The predictive performance of the horizontal risk perception factor prediction model is tested using samples outside the training set. Two test samples are randomly selected, and the prediction results are as follows. Figure 7 , Figure 8 As shown, where, Figure 7 The test sample had a MAE value of 0.0115 and an RMSE value of 0.0139. Figure 8 The test sample in the sample has a MAE value of 0.0237 and an RMSE value of 0.0322.

[0173] The predictive performance of the longitudinal risk perception factor prediction model was tested using samples outside the training set. Two test samples were randomly selected, and the prediction results are as follows. Figure 9 , Figure 10 As shown, where, Figure 9 The test sample had a MAE value of 0.0922 and an RMSE value of 0.1113. Figure 10 The test sample in the model has an MAE value of 0.0711 and an RMSE value of 0.0969. These test results indicate that the risk perception factor prediction model of this invention has superior predictive performance.

[0174] S3. Based on the driving information, obtain the risk perception factor of the vehicle while driving using the risk perception factor prediction model.

[0175] S31. Based on the driving information, obtain the lateral influence feature set and longitudinal influence feature set of the vehicle during driving.

[0176] S32. Based on the lateral impact feature set of the vehicle during driving, the lateral risk perception factor is obtained using the lateral risk perception factor prediction model. Based on the longitudinal impact feature set of the vehicle during driving, the longitudinal risk perception factor is obtained using the longitudinal risk perception factor prediction model.

[0177] The risk perception factor prediction model includes a horizontal risk perception factor prediction model and a vertical risk perception factor prediction model. The risk perception factors include horizontal risk perception factors predicted by the horizontal risk perception factor prediction model and vertical risk perception factors predicted by the vertical risk perception factor prediction model.

[0178] 4.2 Based on the aforementioned risk perception factor, a third driving risk value is obtained, which characterizes the driving risk of the vehicle from the driver's perception angle.

[0179] Specifically, it is calculated using the following formula:

[0180]

[0181] in,

[0182] When the lane-changing vehicle cuts in from the right side of the vehicle, ΔX = x LCV -x rightboder When a lane-changing vehicle cuts in from the left side of the vehicle, ΔX = x leftboder -x LCV , where x LCV Indicates the lateral position of the lane-changing vehicle, x rightboder Indicates the lateral position of the right lane line of the lane where the vehicle is located, x leftboder Indicates the lateral position of the left lane line of the lane where the vehicle is located;

[0183] R humanlike The third driving risk value for autonomous vehicles;

[0184] ρ is the probability of a lane-changing vehicle cutting in, detected by the autonomous vehicle;

[0185] R humanlike,lat This represents the lateral traffic risk value. Among them, a EV,LCV For horizontal risk perception factors;

[0186] R humanlike,lon This represents the longitudinal driving risk value. Among them, a EV,LV This is a longitudinal risk perception factor.

[0187] In this embodiment, the probability ρ of the lane-changing vehicle detected by the vehicular vehicle can be obtained from the lane-changing intent identification model of other vehicles in the published literature "Yuan-YuanRen,Lan Zhao,Xue-Lian Zheng,et al.A Method for Predicting Diverse Lane-Changing Trajectories of Surrounding Vehicles Based on Early Detection of Lane Change[J].IEEE Access,2022,Vol.10:17451-17472.", which will not be elaborated here.

[0188] In this embodiment, the method for obtaining the third driving risk value adopts anthropomorphism and quantifies the driving risk of autonomous vehicles from the perspective of driver perception. The risk quantification method of this embodiment can quantify the driving risk that may be faced when a certain strategy is executed from the perspective of driver perception, and can be used as the basis for judging the strategy selection when making decisions.

[0189] 5. Obtain the comprehensive driving risk value.

[0190] Specifically, based on the first driving risk value, the second driving risk value, and the third driving risk value, a comprehensive driving risk value is obtained that represents the driving risk of a vehicle from multiple perspectives of people, vehicles, and roads.

[0191] In this embodiment, the formula for calculating the comprehensive driving risk value is as follows:

[0192] R = R humanlike +R vehicle +R road ,

[0193] Where R represents the comprehensive driving risk value, R humanlike R represents the third driving risk value of autonomous vehicles. vehicle R represents the second driving risk value. road This indicates the first driving risk value.

[0194] To verify the feasibility of the vehicle driving risk quantification method in this embodiment, this embodiment designs the following... Figure 11 The three simulation scenarios are shown.

[0195] Simulation Scenario 1: Lane-changing vehicle (LCV) cuts in without collision with the EV. Specific scenario description: The LCV initially travels at a constant speed, then decelerates uniformly before cutting in, and maintains a constant speed after successfully cutting in; the EV initially travels at a constant speed, then decelerates uniformly after the LCV cuts in, and maintains a constant speed after stabilizing and following the deceleration.

[0196] Simulation Scenario 2: Collision between a lane-changing LCV and an EV. Specific scenario description: The LCV slowly accelerates before cutting in, colliding with the target EV behind it during the cut-in process; the EV begins to decelerate before the LCV cuts in, colliding with the accelerating LCV that is changing lanes.

[0197] Simulation Scenario 3: The LCV (Lane Changer Vehicle) cuts into the lane and collides with the EV (Electric Vehicle). Specific scenario description: The LCV initially moves at a constant speed, then begins to decelerate uniformly before cutting in, maintaining a constant speed until the cutting process is complete; the EV initially moves at a constant speed, then decelerates as the LCV cuts in, subsequently maintaining a constant speed. Due to insufficient deceleration time, the EV rear-ends the LCV.

[0198] According to the autonomous vehicle driving risk quantification method of this embodiment, autonomous vehicle driving risk quantification is performed on simulation scenarios 1-3 respectively. The autonomous vehicle driving risk curves plotted from the specific autonomous vehicle driving risk quantification equations of simulation scenarios 1-3 are as follows: Figure 12-14 As shown.

[0199] like Figure 12 As shown, in simulation scenario 1, the driving risk values ​​at each angle (i.e., the first driving risk value, the second driving risk value, and the third driving risk value) did not reach 1. The driving risk at the human perception angle (i.e., the third driving risk value) increased as the LCV cut in. After the LCV cut in, the risk at the human perception angle dropped to below 0.5 and remained stable, which is consistent with the EV's driving state of first decelerating uniformly and then maintaining a constant speed after detecting the LCV's cut in during the simulation. From the point of significant lateral movement before the LCV cut in, due to the potential lateral collision threat it posed to the EV, the driving risk at the potential collision angle faced by the EV (i.e., the second driving risk value) showed an upward trend. Only after the LCV cut in and the two vehicles were in a longitudinal following state did the driving risk at the potential collision angle of the EV gradually drop to below 0.2, without significant fluctuations.

[0200] like Figure 13As shown, in simulation scenario 2, except for the driving risk at the road constraint angle (i.e., the first driving risk value), which remains stable at around 0.2, the driving risk at the human perception angle (i.e., the third driving risk value) and the driving risk at the potential collision angle (i.e., the second driving risk value) both gradually increase as the vehicle approaches a collision. Because the continuous acceleration of the LCV before its entry forces the EV to brake continuously, the EV perceives the continuously increasing risk and maintains risk balance through braking. Therefore, the driving risk at the human perception angle of the EV increases continuously with the entry of the LCV, approaching 1 at the moment of collision. Similarly, because the LCV's acceleration entry ultimately leads to a collision, the driving risk at the potential collision angle of the EV gradually increases as it approaches a collision, reaching its maximum at the moment of collision. In this simulation scenario, since the EV does not engage in lane departure or speeding, the driving risk at the road constraint angle remains consistently around 0.2 without significant fluctuations.

[0201] like Figure 14 As shown in simulation scenario 3, the driving risk of the potential collision angle (i.e., the second driving risk value) generally increases during the approach collision process, reaching its maximum at the moment of collision. Unlike the potential collision risk of the vehicle, the maximum value of the driving risk of the human perception angle (i.e., the third driving risk value) occurs when the LCV crosses the line. At the moment of vehicle collision, the human perception risk is less than 0.5. The reason for this is that the EV behaves relatively aggressively during the LCV's entry and does not take reasonable avoidance actions in time. Specifically, the EV only decelerates briefly to avoid the LCV when it enters the line and then maintains a constant speed instead of continuously decelerating to maintain a safe distance, ultimately leading to a rear-end collision between the EV and the LCV after the LCV successfully enters the line.

[0202] In the simulation scenarios 1-3 above, since the EV does not engage in lane departure or speeding, the driving risk of the road constraint angle (i.e., the first driving risk value) remains at around 0.2 without significant fluctuations.

[0203] To more intuitively illustrate the driving risks of the vehicle from various angles, a radar chart is used to describe the driving risks faced by the vehicle from various angles (i.e., the first driving risk value, the second driving risk value, and the third driving risk value) at the moment of entry without collision in simulation scenario 1, the moment of entry collision in simulation scenario 2, and the moment of entry collision after collision in simulation scenario 3. Figure 15 As shown.

[0204] As can be seen, the vehicle driving risk quantification method in this embodiment can vividly, intuitively and accurately reflect the driving risks faced by vehicles in different traffic scenarios, providing a basis for decision-making on the driving strategies to be executed by the vehicle, which is conducive to ensuring driving safety.

[0205] This embodiment also provides a vehicle driving risk quantification system based on the perspectives of people, vehicles, and roads, including:

[0206] (1) Driving information acquisition module. Specifically, it includes a driving information acquisition unit, which is used to acquire driving information.

[0207] (2) First Traffic Risk Quantification Module. Specifically includes:

[0208] The vehicle lane departure risk acquisition unit is used to obtain the vehicle lane departure risk based on the driving information.

[0209] The vehicle speeding risk acquisition unit is used to obtain the vehicle speeding risk based on the driving information.

[0210] The first driving risk value acquisition unit is used to obtain a first driving risk value that characterizes the driving risk of the vehicle based on the vehicle's lane departure risk and speeding risk.

[0211] (3) Second Traffic Risk Quantification Module. Specifically includes:

[0212] The severe collision severity acquisition unit is used to obtain the collision severity between the vehicle and other vehicles based on the driving information.

[0213] The collision urgency acquisition unit is used to obtain the collision urgency between the vehicle and other vehicles based on the driving information.

[0214] The second driving risk value acquisition unit is used to obtain a second driving risk value, which characterizes the driving risk of the vehicle based on the collision severity and collision urgency.

[0215] (4) Third-level vehicle risk quantification module. Specifically, it includes:

[0216] The risk perception factor acquisition unit is used to acquire the risk perception factor of the vehicle while driving based on the driving information.

[0217] The third driving risk value acquisition unit is used to obtain a third driving risk value that characterizes the driving risk of the vehicle from the driver's perception angle based on the risk perception factor.

[0218] (5) Comprehensive driving risk quantification module. Specifically, it includes: a comprehensive driving risk value acquisition unit, which is used to obtain a comprehensive driving risk value that represents the driving risk of the vehicle from multiple perspectives of people, vehicles and roads based on the first driving risk value, the second driving risk value and the third driving risk value.

[0219] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A method for quantifying vehicle driving risk based on the perspectives of people, vehicles, and roads, characterized in that, Includes the following steps: Obtain driving information, which includes vehicle data of the vehicle itself, vehicle data of other vehicles on the road where the vehicle is traveling, and road data of the road where the vehicle is traveling. Based on the driving information, the lane departure risk and speeding risk of the vehicle are obtained. Based on the lane departure risk and speeding risk of the vehicle, the first driving risk value representing the driving risk of the vehicle, which characterizes the road constraint angle, is obtained. Based on the driving information, the severity and urgency of the collision between the vehicle and other vehicles are obtained. Based on the severity and urgency of the collision, a second driving risk value representing the driving risk of the vehicle, which characterizes the angle of collision between the vehicle and other vehicles, is obtained. Based on the driving information, a risk perception factor is obtained when the vehicle is driving. Based on the risk perception factor, a third driving risk value is obtained, which represents the driving risk of the vehicle from the driver's perception angle. It also includes: based on the first driving risk value, the second driving risk value, and the third driving risk value, obtaining a comprehensive driving risk value that characterizes the driving risk of a vehicle from multiple perspectives of people, vehicles, and roads; Based on the driving information, the lane departure risk of the vehicle is obtained, specifically calculated using the following formula: ， in, This indicates the risk of lane departure for the vehicle. This represents the constraint coefficient; when the lane line is a solid line, When the lane lines are dashed, ,in, This represents the probability of a vehicle changing lanes. This represents the minimum lateral distance between the edge of the vehicle and the lane lines on both sides, i.e. ,in, , ,in, Indicates the lateral position of the vehicle. This indicates the lateral position of the left lane line of the lane where the vehicle is located. This indicates the lateral position of the right lane line of the lane where the vehicle is located. Indicates the width of the vehicle. Indicates the heading angle of the vehicle; Based on the driving information, the risk of the vehicle speeding is calculated using the following formula: ， in, This indicates the risk level of speeding by the vehicle. Indicates the combined speed of the vehicle. Indicates the speed limit; Based on the driving information, the degree of urgency of a collision between the vehicle and other vehicles is determined, specifically including the following steps: Based on the driving information, the lateral collision urgency coefficient and longitudinal collision urgency coefficient of the vehicle and other vehicles are obtained; Based on the lateral and longitudinal collision urgency coefficients, the degree of lateral and longitudinal collision urgency experienced by the vehicle from other vehicles is obtained through a piecewise function. Based on the severity of the collision between the vehicle and other vehicles and the urgency of the collision, a second driving risk value is obtained, which characterizes the driving risk of the vehicle and the collision angle between the vehicle and other vehicles. The other vehicles mentioned include the vehicle in front of the vehicle in the lane in which the vehicle is located, and the vehicle in the adjacent lane in which the vehicle is located and intends to change lanes. Based on the driving information, the lateral collision urgency coefficient between the vehicle and the lane-changing vehicle is obtained, specifically calculated using the following formula: ， in, Indicates the lateral collision urgency factor; The point is the entry point when a lane-changing vehicle cuts into the lane it is in. This indicates that the vehicle changing lanes maintains a constant speed until it reaches the entry point. The time of time, that is , This indicates that the vehicle maintains a constant speed until it reaches the entry point. The time of time, that is , in, Indicates the longitudinal position of the lane-changing vehicle's entry point. Indicates the longitudinal position of the lane-changing vehicle. Indicates the longitudinal speed of the lane-changing vehicle. Indicates the longitudinal position of the vehicle. Indicates the longitudinal speed of the vehicle; Based on the driving information, the longitudinal collision urgency coefficient between the vehicle and the vehicle in front is obtained, specifically calculated using the following formula: ， in, Indicates the longitudinal collision urgency factor; Indicates the time of collision between the vehicle and the vehicle in front; It indicates the longitudinal distance between the rear of the vehicle in front and the front of the vehicle itself. Indicates the relative longitudinal speed between the vehicle in front and the vehicle behind; Indicates the longitudinal position of the vehicle in front. Indicates the width of the vehicle in front. Indicates the heading angle of the vehicle in front. Indicates the length of the vehicle in front. Indicates the longitudinal position of the vehicle. Indicates the longitudinal speed of the vehicle. This indicates the longitudinal speed of the vehicle in front.

2. The method for quantifying vehicle driving risk based on the perspectives of people, vehicles, and roads according to claim 1, characterized in that, In the preset coordinate system, the horizontal axis extends from the origin of the preset coordinate system through the left lane line of the lane where the vehicle is located to the right lane line of the lane where the vehicle is located, and the vertical axis extends from the origin of the preset coordinate system in the direction of the vehicle's travel. Based on the aforementioned lateral and longitudinal collision urgency coefficients, the degree of lateral and longitudinal collision urgency experienced by the vehicle from other vehicles is obtained through a piecewise function, specifically calculated using the following formula: ， in, When a vehicle intends to change lanes, it must move from the lane immediately adjacent to its current lane to the lane it is currently in. This indicates the lateral distance between the vehicle changing lanes and the right lane line of the lane it is currently in. When a vehicle intends to change lanes, it must move from the lane adjacent to its own lane to its own lane. This indicates the lateral distance between the vehicle changing lanes and the left lane line of the lane it is currently in. ;in, Indicates the lateral position of the lane-changing vehicle. This indicates the lateral position of the right lane line of the lane where the vehicle is located. Indicates the lateral position of the left lane line of the lane where the vehicle is located; ρ This represents the probability of a lane-changing vehicle cutting in, detected by the autonomous vehicle. This indicates that the vehicle changing lanes is in the stage before it merges into its own lane. Indicates that since the lane-changing car and from the car in front Larger values ​​in the range represent the degree of urgency of a collision with another vehicle from both the lateral and longitudinal directions. U ; The time indicates the stage after the lane-changing vehicle has entered its own lane. After the lane-changing vehicle enters its own lane, it is updated to the preceding vehicle. This is the stage in the piecewise function. This indicates that the vehicle receives feedback from the vehicle in front after the update. From the updated preceding vehicle The value represents the degree of urgency of a collision with another vehicle from its lateral and longitudinal directions. U .

3. The method for quantifying vehicle driving risk based on the perspectives of people, vehicles, and roads according to claim 1, characterized in that, Based on the driving information, the severity of the collision between the vehicle and other vehicles is obtained. Specifically, the collision severity is obtained by using a collision severity prediction model based on the driving information. Based on the severity of the collision between the vehicle and other vehicles and the urgency of the collision, a second driving risk value, representing the driving risk of the vehicle at the angle of collision with other vehicles, is obtained and calculated using the following formula: ， in, This indicates the second driving risk value. U Indicates the urgency of the collision. S Indicates the severity of the collision.

4. The method for quantifying vehicle driving risk based on the perspectives of people, vehicles, and roads according to claim 1, characterized in that, The vehicle is a manually driven vehicle, the risk perception factor of the vehicle is the actual instantaneous acceleration of the vehicle, and the third driving risk value is calculated using the following formula: in, For manually driven vehicles, This refers to the actual instantaneous acceleration of a manually driven vehicle.

5. The method for quantifying vehicle driving risk based on the perspectives of people, vehicles, and roads according to claim 1, characterized in that, The vehicle in question is an autonomous vehicle. The specific method for obtaining the risk perception factor of the vehicle while driving based on the driving information is as follows: The input and output sets of a risk perception factor prediction model are constructed based on the NGSIM database. This input and output set is then used to train a Gaussian process-based regression prediction model, thus constructing the risk perception factor prediction model. The driving information also includes data information used to obtain the input set of the risk perception factor prediction model. Based on the driving information, the risk perception factor of the vehicle during driving is obtained using the risk perception factor prediction model. The risk perception factor prediction model includes a horizontal risk perception factor prediction model and a vertical risk perception factor prediction model. The risk perception factors include horizontal risk perception factors predicted by the horizontal risk perception factor prediction model and vertical risk perception factors predicted by the vertical risk perception factor prediction model. Based on the aforementioned risk perception factor, a third driving risk value, representing the vehicle's driving risk from the driver's perception perspective, is obtained, specifically calculated using the following formula: in, When a lane-changing vehicle cuts in from the right side of the vehicle. When a lane-changing vehicle cuts in from the left side of the vehicle, ,in, Indicates the lateral position of the lane-changing vehicle. This indicates the lateral position of the right lane line of the lane where the vehicle is located. This indicates the lateral position of the left lane line of the lane where the vehicle is located. The third driving risk value for autonomous vehicles. ρ This represents the probability of a lane-changing vehicle cutting in, detected by the autonomous vehicle. This represents the lateral traffic risk value. ,in, This is a horizontal risk perception factor. This represents the longitudinal driving risk value. ,in, This is a longitudinal risk perception factor.

6. A vehicle driving risk quantification system based on the perspectives of people, vehicles, and roads, applied to the vehicle driving risk quantification method based on the perspectives of people, vehicles, and roads described in claim 1, characterized in that, include: The driving information acquisition module includes: a driving information acquisition unit, used to acquire driving information; The first driving risk quantification module includes: a lane departure risk acquisition unit for obtaining lane departure risk of the vehicle based on the driving information; a speeding risk acquisition unit for obtaining speeding risk of the vehicle based on the driving information; and a first driving risk value acquisition unit for obtaining a first driving risk value representing the driving risk of the vehicle with respect to road constraint angles based on the lane departure risk and speeding risk. The second driving risk quantification module includes: a severe collision severity acquisition unit, used to obtain the collision severity of the vehicle and other vehicles based on the driving information; a collision urgency acquisition unit, used to obtain the collision urgency of the vehicle and other vehicles based on the driving information; and a second driving risk value acquisition unit, used to obtain a second driving risk value representing the driving risk of the vehicle, which characterizes the collision angle between the vehicle and other vehicles, based on the collision severity and collision urgency. The third driving risk quantification module includes: a risk perception factor acquisition unit, used to obtain the risk perception factor of the vehicle when driving based on the driving information; and a third driving risk value acquisition unit, used to obtain the third driving risk value representing the driving risk of the vehicle from the driver's perception angle based on the risk perception factor.

Images