A real-time assessment method for vehicle operation risk based on risk field

Abstract

The present invention proposes a real-time assessment method for vehicle operation risk based on risk field. The method collects environmental information around the target vehicle through various devices, then divides the driving environment into three parts: static traffic environment, dynamic traffic control, and surrounding moving objects. The risk field theory is combined to calculate the risks of static traffic environment factors and dynamic traffic control to the target vehicle; the approach risk and collision risk are divided by calculating the conflict time, and the approach risk and collision risk caused by moving objects to the target vehicle are calculated. Finally, all risks are superimposed to obtain the risk set faced by the target vehicle at the current moment, providing a reliable basis for subsequent safety warnings and assisting driving decisions.

Description

Technical Field

[0001] The present invention relates to the fields of intelligent driving and driving risk assessment, and in particular to a real-time vehicle operation risk assessment method based on a risk field. Background Art

[0002] In the era of booming vehicle-road collaboration and assisted driving technologies, real-time risk assessment is a crucial component of intelligent driving systems. Accurately assessing dangerous conditions in the surrounding environment is crucial for safe driving. Being able to timely assess vehicle operating risks can effectively prevent accidents.

[0003] Currently, vehicle operation risk assessment methods can be divided into three categories: single-target-oriented, reachable-set-based, and potential field-based. Potential field-based methods have attracted widespread attention because they can incorporate multiple scenario elements and are applicable to complex scenarios. Potential field-based methods draw on potential field theory from physics, using repulsive fields to describe collision risk. Wang Jianqiang et al. used field theory to represent risk factors arising from drivers, vehicles, road conditions, and other traffic factors. They established a unified model of the driving safety field, which consists of three components: a potential field determined by stationary objects on the road (such as parked vehicles); a dynamic field determined by moving objects on the road (such as vehicles and pedestrians); and a behavioral field determined by the driver's individual characteristics. Huang et al. used LSTM to establish an intention recognition model for surrounding vehicles. They then used a driving safety field risk assessment model to output potential risks, thereby achieving real-time assessment of driving safety. With the advancement of driving safety field theory, the vehicle's own kinematic characteristics have also been incorporated. To address the issue of driving safety fields consistently neglecting the impact of road direction on driving safety, Tian et al. introduced an elliptical correction formula to the driving safety field, weighting and compressing the model's parameters. This further improved the driving safety field theory, but failed to consider the impact of the ego vehicle's motion state on driving risk. Wang Anjie et al. integrated the assessment of ego vehicle stability risk into their risk assessment model, establishing a predictive risk field model that integrates surrounding vehicle risk, road environment risk, and stability risk. This model accurately reflects the changing trends in collision risk, but fails to consider the risk of surrounding non-motor vehicles and is insufficient to reflect the actual level of risk on the road. Wang Jinxiang et al. employed a game theory-based combined weighted TOPSIS method to effectively integrate multiple factors related to intelligent vehicle driving safety, establishing a multi-level vehicle collision hazard situation assessment system that couples "human-vehicle-road-environment." However, the sample size used to construct the assessment decision information was limited, making it insufficient to reflect the actual level of collision risk.

[0004] Therefore, in response to the shortcomings of driving risk assessment, the present invention proposes a real-time assessment method for vehicle operation risk based on risk field. Different from other risk field risk assessment methods, this method divides the driving environment into three parts: static traffic environment, dynamic traffic control, and surrounding moving objects. The risk characteristics generated by them are considered separately, and the conflict time introduced by moving objects is used to classify the risks they cause to the target vehicle. The impact when they approach or have potential conflicts with the current trajectory of the target vehicle is comprehensively considered, and the risks generated by moving objects are comprehensively described, thereby improving the accuracy of risk assessment and the level of driving safety. Summary of the Invention

[0005] To address the shortcomings of existing driving risk assessment methods, this invention aims to classify and quantify the risks surrounding a target vehicle based on risk field theory, identify the risks faced by the target vehicle, and issue an early warning when there is a collision risk or the imminent risk exceeds a safety threshold. The specific implementation steps of this method are as follows:

[0006] Step 1: Collect information about the vehicle's surrounding environment, including lane markings, obstacles, traffic lights, and the movement of surrounding moving objects, using millimeter-wave radar, GPS, lidar, V2X network, and combined inertial navigation equipment.

[0007] Step 2: Based on the lane lines, obstacles, traffic light positions and related data obtained in step 1, the risks posed to the target vehicle by static traffic environment factors and dynamic traffic control information are calculated respectively.

[0008] Step 21, calculate the risk of the road boundary line and lane line in the static traffic environment factors to the target vehicle using the following risk function:

[0009]

[0010] Among them, x i and y i Respectively represent the distance between the i-th road boundary line or lane dividing line and the target vehicle in the horizontal and vertical directions, is the target vehicle’s speed, A i Indicates the different field strength coefficients generated by road boundaries and different lane lines.

[0011] Step 22, calculate the risk posed by the traffic light to the target vehicle in the dynamic traffic control information using the following risk function:

[0012]

[0013] Among them, T0 is the remaining time of green light, v max is the maximum speed limit at the intersection, v0 is the design speed at the intersection, is the time required for a vehicle at position (x, y) to drive from entrance lane i to exit lane j, is the risk faced by a vehicle at position (x, y) at time t when it moves from entrance lane i to exit lane j, is the distance for a vehicle to safely exit from entrance lane i at exit lane j. The value of n is 0, 1, 2, etc., indicating multiple exit lanes.

[0014] Step 3: Based on the position and speed data of the target vehicle and surrounding moving objects obtained in step 1, calculate the conflict time, divide the risk into imminent risk and collision risk based on the conflict time, and calculate the imminent risk and collision risk caused by the moving objects to the target vehicle respectively.

[0015] In step 31, assuming that the non-motorized vehicle maintains its current speed and direction, and the motorized vehicle maintains its current speed and yaw rate, if the spatiotemporal trajectory of the target vehicle and the surrounding objects overlap, the overlap point is a potential collision point, and the collision time is calculated:

[0016]

[0017] Where T is the conflict time, D host and D n is the distance between the target vehicle and surrounding moving objects and the potential collision point, and is the speed of the target vehicle and surrounding moving objects.

[0018] Step 32: Determine the collision type between the target vehicle and surrounding moving objects. The determination rules are as follows:

[0019]

[0020] Where t is the critical time, a is the vehicle deceleration, v is the vehicle speed limit of the specific road section, and t r Driver reaction time

[0021] Step 33: Based on the determination result of step 32, if there is a collision risk, the collision risk caused by the moving object to the target vehicle is calculated using the following risk function:

[0022]

[0023] Among them, m host and m n are the weights of the target vehicle and surrounding vehicles, R veh and R nonveh The mass correction coefficients for motor vehicles and non-motor vehicles are respectively, and are the speeds of the target vehicle and surrounding vehicles, d cis the relative distance between the target vehicle and the moving object with collision risk, ρ and μ are unknown coefficients.

[0024] Step 34: Based on the determination result of step 32, if there is an imminent risk, the imminent risk caused by the moving object to the target vehicle is calculated using the following risk function:

[0025]

[0026] Among them, m host and m n are the weights of the target vehicle and surrounding vehicles, R veh and R nonveh The mass correction coefficients for motor vehicles and non-motor vehicles are respectively, and are the speeds of the target vehicle and surrounding vehicles, d a is the relative distance between the target vehicle and the moving object with approaching risk, θ is the angle between the target vehicle and the moving direction of the moving object, and ρ and μ are unknown coefficients.

[0027] In step 4, based on the risks caused by static traffic environment factors and dynamic traffic control information obtained in step 2, and the risks caused by moving objects obtained in step 3, all risks are added together to obtain the risk set faced by the target vehicle as follows:

[0028]

[0029] Among them, Risk i ,col is the collision risk faced by the target vehicle,Risk i , app is the imminent risk faced by the target vehicle, t0 represents the current moment, and Δt represents a short period of time in the future.

[0030] Step 5: Based on the risk set calculated in step 4, the real-time safety status of the target vehicle is determined according to the risk classification. If there is a collision risk, the vehicle is sorted by the collision risk value. If there is no collision risk, the vehicle is sorted by the imminent risk value.

[0031]

[0032] Among them, Risk i , app is the imminent risk faced by the target vehicle, Risk i , col is the collision risk faced by the target vehicle, σ is the safety threshold. When the imminent risk at that moment exceeds the threshold or there is a collision risk, the point with the minimum imminent risk value in the traffic environment is selected as the recommended position for the driver to adjust the vehicle's motion state.

[0033] Step 6: Based on the determination result of the target vehicle's safety status in step 5, the vehicle's speed and direction are recommended through voice prompts and display prompts to remind the driver to make changes to the vehicle's movement status in advance to avoid dangerous events. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 This is a block diagram of the overall concept of the present invention.

[0035] Figure 2 Schematic diagram for calculating conflict time at an unsignalized intersection. DETAILED DESCRIPTION

[0036] The present invention is described in detail below with reference to the accompanying drawings and embodiments. It should be understood that this example is only used to illustrate the present invention and is not intended to limit the scope of the present invention. The overall idea of ​​the driving risk assessment method based on the risk field provided by the present invention is as follows: Figure 1 As shown, the specific implementation method includes the following steps:

[0037] Step 1: Collect information about the vehicle's surrounding environment, including lane markings, obstacles, traffic lights, and the movement of surrounding moving objects, using millimeter-wave radar, GPS, lidar, V2X network, and combined inertial navigation equipment.

[0038] Step 2: Based on the lane lines, obstacles, signal light positions and related data obtained in step 1, calculate the risks posed to the target vehicle by static traffic environment factors and dynamic traffic control information. The specific implementation steps of step 2 are as follows:

[0039] Step 21, calculate the risk of the road boundary line and lane line in the static traffic environment factors to the target vehicle using the following risk function:

[0040]

[0041] Among them, x i and y i Respectively represent the distance between the i-th road boundary line or lane dividing line and the vehicle in the horizontal and vertical directions, is the target vehicle’s speed, A i Indicates the different field strength coefficients generated by different road boundaries and lane lines. This characterizes the differences in the risks posed to the target vehicle by different types of environmental elements, such as the A i The value must be greater than the white dotted line A i Specifically determine A i When the value is set, a medium-sized car is selected as the standard vehicle, and its A i The value is 1, and a frontal collision test is conducted between standard vehicles at a speed of 40km / h to obtain vehicle deformation data as the basis for damage.i The value is the ratio of the deformation of the object and the standard vehicle caused by a head-on collision to the deformation of the standard vehicle and the standard vehicle caused by a collision; let A of the double yellow line be i The value is 0.3, the A of the lane line i Values, such as A of the white dotted line and the white solid line i The value is the ratio of the driving points and property damage caused by a vehicle crossing the line to the driving points and property damage caused by a standard vehicle crossing the double yellow line, with the degree of violation and the fine as the basis for damage.

[0042] Step 22: Calculate the risk corresponding to the dynamic traffic control information. The most important part of the dynamic traffic control information is the traffic lights. The risk function corresponding to the traffic lights is as follows:

[0043]

[0044] Among them, T0 is the remaining time of green light, v max is the maximum speed limit at the intersection, v0 is the design speed at the intersection, is the time required for a vehicle at position (x, y) to drive from entrance lane i to exit lane j, is the risk faced by a vehicle at position (x, y) at time t when it moves from entrance lane i to exit lane j, is the distance for a vehicle to safely exit from entrance lane i at exit lane j. The value of n is 0, 1, 2, etc., indicating multiple exit lanes.

[0045] Step 3: Based on the position and speed data of the target vehicle and surrounding moving objects obtained in Step 1, the conflict time is calculated to divide the risk into imminent risk and collision risk. The imminent risk and collision risk caused by the moving objects to the target vehicle are calculated separately. The specific implementation steps of Step 3 are as follows:

[0046] In step 31, assuming that the non-motorized vehicle maintains its current speed and direction, and the motorized vehicle maintains its current speed and yaw rate, if the spatiotemporal trajectory of the target vehicle and the surrounding objects overlap, the overlap point is a potential collision point, and the collision time is calculated:

[0047]

[0048] Where T is the conflict time, D host and D n is the distance between the target vehicle and surrounding moving objects and the potential collision point, and is the speed of the target vehicle and surrounding moving objects. Figure 2 , assuming that the vehicle traveling from south to north is the target vehicle, it detects surrounding vehicles when it reaches an unsignaled intersection, and determines the presence of a potential collision point based on the collected data, thereby calculating the conflict time.

[0049] Step 32: Determine the collision type between the target vehicle and surrounding moving objects. The determination rules are as follows:

[0050]

[0051] Where t is the critical time, a is the vehicle deceleration, v is the vehicle speed limit of the specific road section, and t r Driver reaction time

[0052] Step 33: Based on the result of step 32, if there is a collision risk, calculate the collision risk caused by the moving objects to the target vehicle. The moving objects are mainly motor vehicles and non-motor vehicles around the target vehicle, including the front and rear vehicles in the vehicle lane and the front and rear vehicles in the left and right lanes. The following risk function is used:

[0053]

[0054] Among them, m host and m n are the weights of the target vehicle and surrounding vehicles, R veh and R nonveh The mass correction coefficients for motor vehicles and non-motor vehicles are respectively, and are the speeds of the target vehicle and surrounding vehicles, d c is the relative distance between the target vehicle and the moving object with collision risk, and ρ and μ are unknown coefficients. When determining the values ​​of ρ and μ, traffic accident data is used to represent the impact of speed on driving risk. Existing research has shown that the value of ρ is 1.566×10 -14 , the μ value is 6.687.

[0055] In step 35, based on the result of step 32, if there is an approaching risk, the approaching risk posed by the moving objects to the target vehicle is calculated. The moving objects are mainly motor vehicles and non-motor vehicles around the target vehicle, including the vehicles in front and behind the target vehicle lane and the vehicles in front and behind the left and right lanes. The following risk function is used:

[0056]

[0057] Among them, m host and m n are the weights of the target vehicle and surrounding vehicles, R veh and R nonveh The mass correction coefficients for motor vehicles and non-motor vehicles are respectively, and are the speeds of the target vehicle and surrounding vehicles, d ais the relative distance between the target vehicle and the moving object with approaching risk, θ is the angle between the target vehicle and the moving direction of the moving object, and ρ and μ are unknown coefficients.

[0058] In step 4, based on the risks caused by static traffic environment factors and dynamic traffic control information obtained in step 2, and the risks caused by moving objects obtained in step 3, all risks are added together to obtain the risk set faced by the target vehicle as follows:

[0059]

[0060] Among them, Risk i ,col is the collision risk faced by the target vehicle,Risk i , app is the imminent risk faced by the target vehicle, t0 represents the current moment, and Δt represents a short period of time in the future.

[0061] Step 5: Based on the risk value calculated in step 4, the real-time safety status of the target vehicle is determined according to the risk classification. If there is a collision risk, the vehicle is sorted by the collision risk value; if there is no collision risk, the vehicle is sorted by the imminent risk value.

[0062]

[0063] Among them, Risk i , app is the imminent risk faced by the target vehicle, Risk i , col is the collision risk faced by the target vehicle, σ is the safety threshold. When the imminent risk at that moment exceeds the threshold or there is a collision risk, the point with the minimum imminent risk value in the traffic environment is selected to tell the driver to adjust the vehicle's motion state.

[0064] Step 6: Based on the determination result of the target vehicle's safety status in step 5, the vehicle's speed and direction are recommended through voice prompts and display prompts to remind the driver to make changes to the vehicle's movement status in advance to avoid dangerous events.

[0065] The above steps describe the implementation process of the present invention in detail, but the present invention is not limited to the specific details of the above embodiments. Anything within the scope of the present invention should not be excluded from the scope of protection of the present invention.

Claims

1. A real-time assessment method for vehicle operation risk based on risk field, characterized in that: The following steps are involved: Step 1: Collect information about the vehicle's surrounding environment, including lane markings, obstacles, traffic lights, and the movement of surrounding moving objects, using millimeter-wave radar, GPS, lidar, V2X network, and combined inertial navigation equipment. Step 2: Based on the lane lines, obstacles, traffic light positions and related data obtained in step 1, the risks posed to the target vehicle by static traffic environment factors and dynamic traffic control information are calculated respectively; Step 21, calculate the risk of the road boundary line and lane line in the static traffic environment factors to the target vehicle using the following risk function: Among them, A i Indicates the different field strength coefficients generated by different road boundary lines and lane lines, x i and y i Respectively represent the distance between the i-th road boundary line or lane dividing line and the target vehicle in the horizontal and vertical directions, is the speed of the target vehicle; Step 22, calculate the risk posed by the traffic light to the target vehicle in the dynamic traffic control information using the following risk function: Among them, T0 is the remaining time of green light, v max is the maximum speed limit at the intersection, v0 is the design speed at the intersection, is the time required for a vehicle at position (x, y) to drive from entrance lane i to exit lane j, is the risk faced by a vehicle at position (x, y) at time t when it moves from entrance lane i to exit lane j, is the distance that the vehicle needs to travel from the i-th entry lane to the j-th exit lane safely. The value of n is 0, 1, 2, etc., indicating multiple exit lanes. Step 3: Based on the position and velocity data of the target vehicle and surrounding moving objects obtained in Step 1, the conflict time is calculated. The risk is divided into imminent risk and collision risk based on the conflict time, and the imminent risk and collision risk caused by the moving objects to the target vehicle are calculated respectively. In step 31, assuming that the non-motorized vehicle maintains its current speed and direction, and the motorized vehicle maintains its current speed and yaw rate, if the spatiotemporal trajectory of the target vehicle and the surrounding objects overlap, the overlap point is a potential collision point, and the collision time is calculated: Where T is the conflict time, D host and D n is the distance between the target vehicle and surrounding moving objects and the potential collision point, and is the speed of the target vehicle and surrounding moving objects; Step 32: Determine the collision type between the target vehicle and surrounding moving objects. The determination rules are as follows: Where t is the critical time, a is the vehicle deceleration, v is the vehicle speed limit of the specific road section, and t r Driver reaction time; Step 33: Based on the determination result of step 32, if there is a collision risk, the collision risk caused by the moving object to the target vehicle is calculated using the following risk function: Among them, m host and m n are the weights of the target vehicle and surrounding vehicles, R veh and R nonveh The mass correction coefficients for motor vehicles and non-motor vehicles are respectively, and are the speeds of the target vehicle and surrounding vehicles, d c is the relative distance between the target vehicle and the moving object with collision risk, ρ and μ are unknown coefficients; Step 34: Based on the determination result of step 32, if there is an imminent risk, the imminent risk caused by the moving object to the target vehicle is calculated using the following risk function: Among them, m host and m n are the weights of the target vehicle and surrounding vehicles, R veh and R nonveh The mass correction coefficients for motor vehicles and non-motor vehicles are respectively, and are the speeds of the target vehicle and surrounding vehicles, d a is the relative distance between the target vehicle and the moving object with an approaching risk, θ is the angle between the target vehicle and the moving direction of the moving object, and ρ and μ are unknown coefficients; In step 4, based on the risks caused by static traffic environment factors and dynamic traffic control information obtained in step 2, and the risks caused by moving objects obtained in step 3, all risks are added together to obtain the risk set faced by the target vehicle as follows: Among them, Risk i ,col is the collision risk faced by the target vehicle,Risk i , app is the imminent risk faced by the target vehicle, t0 represents the current moment, and Δt represents a short period of time in the future; Step 5: Based on the risk set calculated in step 4, the real-time safety status of the target vehicle is determined according to the risk classification. If there is a collision risk, the vehicle is sorted by the collision risk value. If there is no collision risk, the vehicle is sorted by the imminent risk value. Among them, Risk i , app is the imminent risk faced by the target vehicle, Risk i , col is the collision risk faced by the target vehicle, σ is the safety threshold. When the imminent risk at that moment exceeds the threshold or there is a collision risk, the point with the minimum imminent risk value in the traffic environment is selected as the recommended position for the driver to adjust the vehicle's motion state.

2. The method according to claim 1, characterized in that This real-time vehicle operation risk assessment method calculates the vehicle's risk based on risk field theory. It divides the driving environment into three parts: static traffic environment, dynamic traffic control, and surrounding moving objects. The risk characteristics generated by each part are considered separately. The conflict time introduced by moving objects is used to classify the risks they pose to the target vehicle. The impact of their approach or potential conflict with the target vehicle's current trajectory is comprehensively considered to comprehensively describe the risks generated by moving objects, thereby improving the accuracy of risk assessment and enhancing driving safety.

Images