Intelligent warning method and system for driving risk of signal intersection considering influence of large vehicle

Abstract

The application discloses a signal intersection driving risk intelligent early warning method and system considering the influence of large vehicles, and the method comprises the following steps: selecting a high-risk signal intersection; constructing a high-risk signal intersection risk field model, and calibrating the static parameters of the model by using historical trajectory data; monitoring the in-and-out vehicles of the high-risk signal intersection, and when a large vehicle is identified, real-time target data of the large vehicle and surrounding vehicles are extracted, the dynamic parameters of the model are calibrated based on the shielding condition of the large vehicle, the blind area of the large vehicle and the target data; the target data is input into the high-risk signal intersection risk field model in real time to calculate the high-risk signal intersection risk field, and a high-risk discrimination index is calculated according to the high-risk signal intersection risk field; if the high-risk discrimination index is greater than a preset threshold, safety warning is performed. The application fully considers the high accident risk caused by the blind area of the large vehicle and the shielding to the surrounding vehicles, and can accurately judge the safety level of the intersection.

Description

Technical Field

[0001] This invention relates to traffic safety technology, and more particularly to an intelligent early warning method and system for traffic risks at signalized intersections that takes into account the impact of large vehicles. Background Technology

[0002] Signalized intersections, as the most important nodes in urban traffic systems, are "bottleneck areas" in road traffic. They connect roads in various directions, allowing traffic flows to freely change direction and intersect within them. However, this function also leads to a large concentration of vehicles and pedestrians at intersections, creating traffic bottlenecks and further causing frequent traffic conflicts. Therefore, conducting scientific and reasonable safety analysis of the behavioral characteristics of vehicles passing through intersections, exploring the influencing factors of collisions, identifying dangerous scenarios for vehicles at signalized intersections in real time, and using early warning systems for safety alerts are of paramount value in improving intersection safety and reducing the intersection accident rate. Existing methods for safety analysis of intersections mainly employ the following approaches:

[0003] 1. Based on raw accident data or simulated data, traffic flow parameters are used to evaluate the traffic safety status of signalized intersections. For example, patent CN116469272A determines vehicle conflict relationships based on the expected driving trajectories of vehicles within the intersection; it projects vehicles into virtual lanes to form a virtual convoy, calculates the minimum safe distance between the vehicle and conflicting vehicles in the virtual lanes, calculates the comprehensive driving risk of the vehicle, and outputs the driving risk level evaluation result. However, traffic accidents have a high degree of randomness and cannot reflect the impact of short-term changes in traffic flow on intersection safety.

[0004] 2. Studies on the traffic safety impact of large vehicles at intersections typically focus on the individual vehicle, neglecting the influence of the signalized intersection environment and the mutual interference between large vehicles and other surrounding vehicles. For example, patent CN115171431 proposes a multi-view blind spot warning method for large vehicles at intersections. This method uses multiple cameras deployed within the intersection to detect large vehicles and pedestrians from multiple perspectives, achieving blind spot warnings for large vehicles under these conditions. However, few studies on the traffic safety impact of large vehicles consider their influence on intersection safety when passing through them. Research on the motion and safety characteristics of large vehicles passing through intersections, and studies considering the impact of large vehicles on intersection traffic safety, are scarce.

[0005] 3. This paper addresses the problem of quantitative analysis of complex signalized intersections using existing field theory models. In complex traffic environments, various factors affecting vehicle traffic are considered. From a field theory perspective, the risk level to motor vehicle driving safety is evaluated, and a dynamic obstacle avoidance model based on a risk field is constructed. This model effectively describes complex driving environments and can effectively assess vehicle risk intensity and optimize driving behavior in real time. However, it cannot explain or express vehicle following behavior and does not clearly define the risk range of motor vehicles.

[0006] 4. While evaluating intersection safety based on risk field theory, this method fails to consider the impact of large vehicles. For example, patent CN114724376A proposes an intersection safety evaluation method based on risk field theory. This method, using intersection environmental information and vehicle operating status data, constructs a risk field for the vehicle operating environment based on risk field theory, characterizing the risk borne by vehicles at various locations within the intersection for safety evaluation. However, the movement and safety characteristics of large vehicles passing through the intersection will interact with surrounding vehicles, and the method fails to consider the guidance and pressure exerted by large vehicles on surrounding vehicles.

[0007] In summary, current research on large vehicles at intersections primarily focuses on analyzing intersection driving data and studying individual large vehicles. However, existing technologies lack depth in data analysis, neglecting the impact of signalized intersection environments. When studying interference between large vehicles and other surrounding vehicles, both static traffic facilities and dynamic traffic flow affect their behavior. Furthermore, when assessing intersection safety based on the intersection environment and vehicle conditions, the impact of large vehicles on surrounding vehicles is ignored. Summary of the Invention

[0008] The technical problem to be solved by the present invention is to provide an intelligent early warning method and system for traffic risks at signalized intersections that takes into account the impact of large vehicles, so as to improve the safety of intersections and reduce the accident rate at intersections.

[0009] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0010] A method for intelligent early warning of traffic risks at signalized intersections considering the impact of large vehicles, comprising the following steps:

[0011] S01. Select a high-risk signal intersection;

[0012] S02. Construct a risk field model for high-risk signalized intersections and calibrate the static parameters of the risk field model for high-risk signalized intersections using historical trajectory data of the high-risk signalized intersections;

[0013] S03. Monitor vehicles entering and exiting high-risk signal intersections. When a large vehicle is detected, proceed to step S04.

[0014] S04. Extract target data of large vehicles and surrounding vehicles in real time, obtain the occlusion status of large vehicles and calculate the blind spots of large vehicles. Based on the occlusion status of large vehicles, the blind spots of large vehicles and target data, calibrate the dynamic parameters of the risk field model of high-risk signal intersection.

[0015] S05. Input the target data into the high-risk signal intersection risk field model in real time to calculate the high-risk signal intersection risk field, and calculate the high-risk judgment index based on the high-risk signal intersection risk field. If the high-risk judgment index is greater than the preset threshold, a safety warning is issued; otherwise, proceed to step S04 until the large vehicle leaves the high-risk signal intersection.

[0016] Furthermore, step S01 includes: if the current signalized intersection meets one or more of the preset traffic flow conditions, accident rate conditions, and visibility conditions, the current signalized intersection is added to the set of high-risk signalized intersections.

[0017] Furthermore, in step S02, when constructing the risk field model for a high-risk signalized intersection, the following steps are included:

[0018] Based on the coordinate data of the vehicle trajectory, the desired trajectories for straight-going and turning are fitted to obtain the desired trajectory field;

[0019] Based on the varying degrees of impact of different traffic light displays on vehicles as they pass through signalized intersections, a traffic light constraint field is constructed.

[0020] Based on the shading effect of large vehicles on small cars in the same lane in the basic road section area and intersection weaving area, a large vehicle shading field is constructed.

[0021] Based on the impact of large vehicles on small vehicles located in the blind spots of large vehicles in the basic road section area and the intersection weaving area, a blind spot constraint field is constructed.

[0022] Based on the spring-damping model, construct the motion field for large vehicles and small cars;

[0023] The field strength of the high-risk signal intersection is obtained by superimposing the field strength of the desired trajectory field, the field strength of the traffic light constraint field, the field strength of the large vehicle obstruction field, the field strength of the blind spot constraint field, and the field strength of the motion field.

[0024] Furthermore, in step S01, after selecting high-risk signalized intersections, the process also includes establishing a static traffic scene database for high-risk signalized intersections, including: acquiring static data for different high-risk signalized intersections, wherein the static data includes lane data for each approach lane, signal light height, and maximum permissible speed at the signalized intersection; the field strength expression for the desired trajectory field is as follows:

[0025]

[0026]

[0027] In the formula, k lat The horizontal constraint coefficient (dimensionless); l lane Lane width (unit: m), k des It is the target lane attraction coefficient (dimensionless), d exp This represents the perpendicular distance (in meters) from the center line of the trajectory band to the spatial point (x, y). The perpendicular point is (x ij ,y ij The trajectory length (in meters) between the target endpoint of exit channel j and exit channel j. Let be the dimensionless field strength value of the constraint field on the expected trajectory of the vehicle affected by exit lane n.

[0028] The expression for the field strength of the traffic light constraint field is as follows:

[0029]

[0030]

[0031] In the formula, V represents the dimensionless signal light constraint field strength value (where T is the remaining green light time, in seconds); v max v1 represents the maximum permissible speed at the intersection (in m / s); v2 represents the average speed at the intersection (in m / s). k is the time (in seconds) required for a vehicle to reach the exit lane of the intersection. sig It is the target signal control constraint coefficient (dimensionless).

[0032] Furthermore, in step S02, when calibrating the static parameters of the risk field model of the high-risk signal intersection using historical trajectory data of the high-risk signal intersection, the following steps are taken: acquiring historical trajectory data of the high-risk signal intersection, preprocessing the historical trajectory data and fitting the expected trajectory equations for each turn of each approach lane of the high-risk signal intersection, calculating the time required for a vehicle to reach the exit lane of the intersection, the length of the trajectory strip and the center line of the trajectory strip, and calibrating the lateral constraint coefficient, the target lane attraction coefficient and the target signal control constraint coefficient based on the time required for a vehicle to reach the exit lane of the intersection, the length of the trajectory strip and the center line of the trajectory strip.

[0033] Furthermore, the field strength expression for the large vehicle shading field is as follows:

[0034]

[0035]

[0036] in,

[0037]

[0038] d she =Δl1tanα

[0039] l she =Δl2tanβ

[0040] Let be the dimensionless value of the shading constraint field intensity of the spatial point (x, y) in the basic road segment region at time t; α be the angle between the driver's horizontal line of sight of the small car and the traffic light (unit: °); h2 be the height of the traffic light (unit: m); h3 be the height of the large vehicle (unit: m); h1 be the height of the small car (unit: m); Δx be the horizontal distance between the small car and the traffic light (unit: m); d she Δl1 is the obstruction height (in meters); Δl1 is the distance between large vehicles and small vehicles on the basic road section (in meters); k sher The shading coefficient (dimensionless) for the basic road section area; l she Δl2 is the obstruction distance (in meters); Δl2 is the average headway between large vehicles and small vehicles when large vehicles pass through the weaving zone (in meters); k shei The occlusion coefficient (dimensionless) is the weaving zone occupancy coefficient at the intersection.

[0041] Furthermore, the field strength expression for the blind zone constraint field is as follows:

[0042]

[0043]

[0044]

[0045] in,

[0046]

[0047]

[0048] d bli =W car -Δd

[0049]

[0050]

[0051] Let be the blind zone constraint field strength value (dimensionless) of spatial point (x, y) in the basic road segment region at time t; Δl3 is the distance (in meters) from the rearview mirror of the large vehicle to the rear of the small vehicle; the lengths of the large vehicle and the small vehicle are l. tru With l car (Unit: m), the width of large vehicles and the width of small vehicles are W. tru With W car (Unit: m); Δd is the distance (unit: m) of the small car entering the field of view of the large vehicle's rearview mirror in the y-coordinate; d bli Blind spot distance for large vehicles (unit: m); d w Ω represents the actual distance between the large vehicle and the small car (in meters); Ω is the angle between the rear of the small car and the rearview mirror of the large vehicle (in degrees); when P = 0, the small car is in front of the large vehicle; when P = 1, the small car is behind the large vehicle; k wblir k is the dimensionless constraint coefficient for the blind spot in front of large vehicles; blir The blind spot constraint coefficient (dimensionless) for large vehicles; l pra The actual distance between small cars and large vehicles (unit: m); Δl5 is the distance from the rearview mirror of the large vehicle to the rear of the large vehicle (unit: m); l bli φ is the distance to the blind spot behind the large vehicle (in meters); φ is the angle between the center point of the small car and the rearview mirror of the large vehicle (in degrees), used to determine whether the small car has entered the blind spot of the large vehicle.

[0052] Furthermore, the expression for the field strength of the sports field is as follows:

[0053]

[0054]

[0055] Among them, L car Safety distance for small vehicles (unit: m); L truSafety distance for large vehicles (unit: m); speed of the vehicles in front and behind is v. p and v ego (Unit: m / s); i indicates that the model is a large vehicle and the model is a small vehicle, respectively; c i k is the virtual damping coefficient (dimensionless) in the model. i U represents the virtual spring stiffness coefficient (dimensionless) in the model; mov(x,y,t) The intensity value of the motion field when the vehicle is moving (dimensionless).

[0056] Furthermore, step S04 includes the following steps:

[0057] Based on the vehicle's coordinate data, identify the small car that is closest to the large vehicle and travels in the same direction within a first length range centered on the large vehicle.

[0058] The identified large vehicles and their corresponding small vehicles are analyzed, and their length, width, height, speed, and distance between the two vehicles are extracted as target data.

[0059] The front blind spot of a large vehicle is defined as a semicircle with the front of the vehicle as the center and a radius equal to the second length. The rear blind spot of a large vehicle is defined as the range between the target angle intervals.

[0060] Measured trajectory data of high-risk signal intersections are obtained, and the safe distance, actual distance, vehicle acceleration and vehicle speed are calculated based on the vehicle information in the measured trajectory data. Based on the calculated parameters and the corresponding actual data, the PSO algorithm is used to iteratively optimize the vehicle-type parameters in the spring-damping model, select the optimal parameter values, and use them to calibrate the occlusion coefficient, blind spot constraint coefficient, virtual damping coefficient and virtual spring stiffness coefficient for large vehicles.

[0061] Furthermore, the high-risk identification index in step S05 is the Relative Driving Safety Index (RDIS), whose expression is as follows:

[0062]

[0063] Among them, U (x,y,t) This indicates the risk field intensity value of a high-risk signalized intersection where vehicles are traveling. To assess the standard driving risk field intensity of a vehicle in a specific hazardous scenario, when RIDS is greater than 1, the vehicle is determined to be in a hazardous state; otherwise, the vehicle is in a safe state.

[0064] The present invention also proposes a high-risk scenario driving risk intelligent early warning system, including a storage medium of interconnected microprocessors, the microprocessors being programmed or configured to execute any of the above-described signalized intersection driving risk intelligent early warning methods considering the impact of large vehicles.

[0065] Compared with the prior art, the advantages of the present invention are as follows:

[0066] 1. This invention utilizes risk field theory to quantitatively express environmental and motion factors at signalized intersections. Based on historical trajectory data, the static parameters of the risk field model are calibrated. Simultaneously, considering the influence of large vehicles, the dynamic parameters of the risk field model are calibrated based on the obstruction and blind spots of large vehicles. This ensures that at any given moment, a vehicle at any position within the signalized intersection has its corresponding risk field intensity value, identifying high-risk scenarios to reduce the intersection accident rate and thus improve intersection safety.

[0067] 2. This invention takes large vehicles and surrounding smaller vehicles as the research object, studies the motion and safety characteristics of large vehicles at signalized intersections, combines environmental factors and influencing factors of moving vehicles to construct a risk field model, and calculates high-risk discrimination indexes based on this model to assess the collision risk of smaller vehicles around large vehicles, reduce accidents, and improve intersection safety. Attached Figure Description

[0068] Figure 1 This is a flowchart of a method according to an embodiment of the present invention.

[0069] Figure 2 This is a schematic diagram of the high-level signal intersection selected in an embodiment of the present invention.

[0070] Figure 3 This is a diagram illustrating how large vehicles obstruct traffic lights for surrounding vehicles.

[0071] Figure 4 This is a diagram illustrating blind spots for large vehicles.

[0072] Figure 5 This is a flowchart of the intelligent early warning system for high-risk driving scenarios according to an embodiment of the present invention. Detailed Implementation

[0073] The present invention will be further described below with reference to the accompanying drawings and specific preferred embodiments, but this does not limit the scope of protection of the present invention.

[0074] Example 1

[0075] To improve the safety of signalized intersections and reduce the accident rate, this embodiment proposes an intelligent early warning method for traffic risks at signalized intersections that considers the impact of large vehicles, such as... Figure 1 As shown, it includes the following steps:

[0076] S01. Select a high-risk signal intersection;

[0077] S02. Construct a risk field model for high-risk signalized intersections and calibrate the static parameters of the risk field model for high-risk signalized intersections using historical trajectory data of the high-risk signalized intersections;

[0078] S03. Monitor vehicles entering and exiting high-risk signal intersections. When a large vehicle is detected, proceed to step S04.

[0079] S04. Extract target data of large vehicles and surrounding vehicles in real time, obtain the occlusion status of large vehicles and calculate the blind spots of large vehicles. Based on the occlusion status of large vehicles, the blind spots of large vehicles and target data, calibrate the dynamic parameters of the risk field model of high-risk signal intersection.

[0080] S05. Input the target data into the high-risk signal intersection risk field model in real time to calculate the high-risk signal intersection risk field, and calculate the high-risk judgment index based on the high-risk signal intersection risk field. If the high-risk judgment index is greater than the preset threshold, a safety warning is issued; otherwise, proceed to step S04 until the large vehicle leaves the high-risk signal intersection.

[0081] Through the above steps, this embodiment uses risk field theory to quantitatively express the environmental and motion factors of signalized intersections. Based on historical trajectory data, the static parameters of the risk field model are calibrated. Simultaneously, considering the influence of large vehicles, the dynamic parameters of the risk field model are calibrated based on the obstruction and blind spots of large vehicles. This ensures that at any given moment, a vehicle at any position within the signalized intersection has a corresponding risk field intensity value, identifying high-risk scenarios to reduce the intersection accident rate and thus improve intersection safety. Furthermore, this embodiment takes large vehicles and surrounding smaller vehicles as the research object, studying the motion and safety characteristics of large vehicles at signalized intersections. By combining environmental factors and the influencing factors of moving vehicles, a risk field model is constructed. Based on this model, a high-risk discrimination index is calculated to assess the collision risk of smaller vehicles surrounding large vehicles, reducing accidents and improving intersection safety.

[0082] The following provides a detailed explanation of each step.

[0083] In step S01 of this embodiment, high-risk signalized intersections are selected based on different factors such as traffic flow, accident rate, and visibility, and a set of high-risk signalized intersections A = {a1, a2, a3, ..., a...} is established. n}, where a n This is represented as the nth high-risk signalized intersection in the set. Therefore, the steps for selecting a high-risk signalized intersection include:

[0084] If the current signalized intersection meets one or more of the preset traffic flow conditions, accident rate conditions, and visibility conditions, the current signalized intersection will be added to the set of high-risk signalized intersections. Specifically, the traffic flow condition can be that the average daily traffic flow is greater than the preset traffic flow threshold, the accident rate condition can be that the average daily accident rate is greater than the preset accident rate threshold, and the visibility condition can be that the visibility of the road is less than the preset visibility threshold. Figure 2 This example describes a high-risk signalized intersection selected in step S01 of this embodiment. Traffic flow is moderate, and during the green light period, left-turning and straight-through vehicles are allowed to proceed together at each approach lane. This intersection is a typical cross-shaped intersection. Left-turning vehicles have their own dedicated lanes. The north, south, and west approach lanes each have 3 straight-through lanes, 2 left-turn lanes, and 1 right-turn lane, while the east approach lane has 2 straight-through lanes, 2 left-turn lanes, and 1 straight-through and right-turn lane. The traffic conditions for this high-risk signalized intersection are shown in Table 1.

[0085] Table 1. Intersection Traffic Conditions

[0086]

[0087] In order to record the relevant data of the selected high-risk signalized intersections for model construction in subsequent steps, step S01 of this embodiment, after selecting the high-risk signalized intersections, also includes the step of establishing a static traffic scene database for high-risk signalized intersections, including: obtaining static data of different high-risk signalized intersections, the static data including lane data of each approach lane, traffic light height, and maximum allowable speed of the signalized intersection, and then establishing a static traffic scene database for high-risk signalized intersections based on the static data of different high-risk signalized intersections.

[0088] In step S02 of this embodiment, the risk field of a high-risk signalized intersection includes the desired trajectory field, the traffic light constraint field, the large vehicle obstruction field, the blind spot constraint field, and the motion field. Therefore, the construction of the risk field model of a high-risk signalized intersection includes the following steps:

[0089] S021) Constructing the desired trajectory field: Based on the coordinate data of the vehicle trajectory, fit the desired trajectories for straight-line and turning directions to obtain the desired trajectory field. The field strength of the desired trajectory field is the superposition of the lateral constraint field strength and the target lane attraction field strength. The expression for the field strength of the desired trajectory field is as follows:

[0090]

[0091]

[0092] In the above formula, k lat The horizontal constraint coefficient (dimensionless); l lane Lane width (unit: m), k desIt is the target lane attraction coefficient (dimensionless), d exp This represents the perpendicular distance (in meters) from the center line of the trajectory band to the spatial point (x, y). The perpendicular point is (x ij ,y ij The trajectory length (in meters) between the target endpoint of exit channel j and exit channel j. Let be the dimensionless field strength value of the constraint field on the expected trajectory of the vehicle affected by exit lane n.

[0093] S022) Constructing a traffic light constraint field: Based on the varying degrees of impact of different traffic light displays on vehicles passing through signalized intersections, a traffic light constraint field is constructed. The field strength expression for the traffic light constraint field is as follows:

[0094]

[0095]

[0096] In the above formula, V represents the dimensionless signal light constraint field strength value (where T is the remaining green light time, in seconds); v max v1 represents the maximum permissible speed at the intersection (in m / s); v2 represents the average speed at the intersection (in m / s). k is the time (in seconds) required for a vehicle to reach the exit lane of the intersection. sig It is the target signal control constraint coefficient (dimensionless);

[0097] S023) Construct a large vehicle shielding area: such as Figure 3 As shown, large vehicles will have a shading effect on small cars in the same lane in the basic road section area and the intersection weaving area. Based on the shading effect of large vehicles on small cars in the same lane in the basic road section area and the intersection weaving area, a large vehicle shading field is constructed, and its field strength is calculated according to different areas.

[0098] In this embodiment, the field strength of the large vehicle occupancy field in the basic road section area is calculated using the following formula:

[0099]

[0100] d she =Δk1tanα (6)

[0101]

[0102] In the above formula, Let be the dimensionless value of the shading constraint field intensity of the spatial point (x, y) in the basic road segment region at time t; α be the angle between the driver's horizontal line of sight of the small car and the traffic light (unit: °); h2 be the height of the traffic light (unit: m); h3 be the height of the large vehicle (unit: m); h1 be the height of the small car (unit: m); Δx be the horizontal distance between the small car and the traffic light (unit: m); d she Δl1 is the obstruction height (in meters); Δl1 is the distance between large vehicles and small vehicles on the basic road section (in meters); k sher The occlusion coefficient (dimensionless) for the basic road section area;

[0103] In this embodiment, the field strength of the large vehicle shading field in the intersection weaving area is calculated using the following formula:

[0104] l she =Δl2tanβ (8)

[0105]

[0106] In the above formula, l she Δl2 is the obstruction distance (in meters); Δl2 is the average headway between large vehicles and small vehicles when large vehicles pass through the weaving zone (in meters); k shei The shading coefficient (dimensionless) is the weaving zone shading coefficient at the intersection.

[0107] In summary, the expression for the field strength of the field obstructed by a large vehicle is as follows:

[0108]

[0109] S024) Construct the blind zone constraint field: such as Figure 4 As shown, large vehicles have a certain impact on cars located in their blind spots in both the basic road section area and the intersection weaving area. Therefore, based on the impact of large vehicles on cars located in the blind spots of large vehicles in the basic road section area and the intersection weaving area, a blind spot constraint field is constructed; and the field strength is calculated according to the blind spots of different areas.

[0110] In this embodiment, the blind zone constraint field strength of the basic road segment area is calculated using the following formula:

[0111]

[0112]

[0113] d bli =W car -Δd (13)

[0114]

[0115] In the above formula, Let be the blind zone constraint field strength value (dimensionless) of spatial point (x, y) in the basic road segment region at time t; Δl3 is the distance (in meters) from the rearview mirror of the large vehicle to the rear of the small vehicle; the lengths of the large vehicle and the small vehicle are l. tru With l car (Unit: m), the width of large vehicles and the width of small vehicles are W. tru With W car (Unit: m); Δd is the distance (unit: m) of the small car entering the rearview mirror field of view of the large vehicle in the y-coordinate; d bli Blind spot distance for large vehicles (unit: m); d w Ω represents the actual distance between the large vehicle and the small car (in meters); Ω is the angle between the rear of the small car and the rearview mirror of the large vehicle (in degrees); when P = 0, the small car is in front of the large vehicle; when P = 1, the small car is behind the large vehicle; k wblir k is the dimensionless constraint coefficient for the blind spot in front of large vehicles; blir The constraint coefficient for the rear blind spot of large vehicles (dimensionless);

[0116] In this embodiment, the blind zone constraint field strength of the intersection weaving area is calculated using the following formula:

[0117]

[0118]

[0119]

[0120]

[0121] In the above formula, l pra The actual distance between small cars and large vehicles (unit: m); Δl5 is the distance from the rearview mirror of the large vehicle to the rear of the large vehicle (unit: m); l bli φ is the distance to the blind spot behind the large vehicle (in meters); φ is the angle between the center point of the small car and the rearview mirror of the large vehicle (in degrees), used to determine whether the small car has entered the blind spot of the large vehicle.

[0122] In summary, the expression for the field strength of the blind zone constraint field is as follows:

[0123]

[0124] S025) Constructing the motion field: Based on the spring-damped model, construct the motion field for large vehicles and small vehicles; the expression for the field strength is as follows:

[0125]

[0126]

[0127] Among them, L car The safe following distance for small vehicles is expressed as follows:

[0128]

[0129] L tru The safe distance for large vehicles is expressed as follows:

[0130]

[0131] In the above formula, the speeds of the front and rear vehicles are v. p and v ego (Unit: m / s); i indicates that the model is a large vehicle and the model is a small vehicle, respectively; c i k is the virtual damping coefficient (dimensionless) in the model. i U represents the virtual spring stiffness coefficient (dimensionless) in the model; mov(x,y,t) δ represents the intensity of the motion field during vehicle movement (dimensionless); δ is the angle between the safe distance in a certain direction and the safe forward distance of the vehicle (unit: °); the length and width of the small car are l and l respectively. car l dcar (Unit: m), the length and width of the large vehicle are l tru l dtru (Unit: m), β represents the driver's reaction time (unit: s), W0 represents the distance between the left and right vehicles parallel to the k-th vehicle (unit: m), v k(t) Let a represent the instantaneous velocity (in m / s) of the k-th vehicle at second t. k This represents the maximum deceleration of the kth vehicle (unit: m / s). 2 ).

[0132] S026) The field strength of the high-risk signal intersection is obtained by superimposing the field strength of the desired trajectory field, the traffic light constraint field, the large vehicle obstruction field, the blind spot constraint field, and the motion field. The calculation formula is as follows:

[0133] U (x,y,t) =U exp(x,y,t) +U sig(x,y,t) +U she(x,y,t) +U bli(x,y,t) +U mov(x,y,t) (twenty four)

[0134] In the formula, U (x,y,t) This indicates the risk field intensity value of a high-risk signalized intersection where vehicles travel within it.

[0135] In this embodiment, the static parameters of the high-risk signal intersection risk field model specifically refer to the lateral constraint coefficient k.lat Target lane attraction coefficient k des and target signal control constraint coefficient k sig Correspondingly, in step S02, when calibrating the static parameters of the risk field model of the high-risk signal intersection using historical trajectory data of the high-risk signal intersection, the following steps are taken: acquiring historical trajectory data of the high-risk signal intersection; preprocessing the historical trajectory data and fitting the expected trajectory equations of each approach lane and each turn of the intersection using the polyfit built-in function of the numpy library in Python, wherein the straight-ahead direction is fitted as a linear function, and the left-turn and right-turn directions are fitted as cubic polynomial functions; calculating the time required for the vehicle to reach the exit lane of the intersection, the length of the trajectory strip, and the centerline of the trajectory strip; and calibrating the lateral constraint coefficient, the target lane attraction coefficient, and the target signal control constraint coefficient based on the time required for the vehicle to reach the exit lane of the intersection, the length of the trajectory strip, and the centerline of the trajectory strip.

[0136] In step S02 of this embodiment, after constructing the risk field model of the high-risk signalized intersection, the Relative Traffic Safety Index (RDIS) (dimensionless) is used as the high-risk discrimination index based on the established risk field of the high-risk signalized intersection. Its expression is as follows:

[0137]

[0138] in, The standard driving risk field intensity (RIDS) for the vehicle under evaluation in a specific hazardous scenario is defined as the field intensity generated when the angle δ between the safe distance in a certain direction and the vehicle's forward safe distance is 0. The specific scenario involves the vehicle under evaluation traveling in the same direction as the vehicle ahead in the same lane, without interference from other surrounding vehicles, and with identical braking performance, road surface slippage, and other influencing factors. When RIDS is greater than 1, the vehicle is considered to be in a hazardous state; otherwise, it is considered to be in a safe state.

[0139] Before step S03 in this embodiment, the following steps are included: installing a surveillance camera next to the entrance road of the high-risk intersection and ensuring that the camera can fully cover the entrance road of the intersection; in step S03, when monitoring vehicles entering and exiting the high-risk signal intersection, the following steps are included: the camera collects video data of the intersection in real time, and uses computer vision technology to perform target detection and tracking on the video data in real time to identify vehicles on the entrance and exit roads of the intersection.

[0140] In step S04 of this embodiment, the dynamic parameters of the risk field model for high-risk signalized intersections include the occlusion coefficient k of the basic road segment area. sher Shielding coefficient k of the intersection weaving area shei Blind spot constraint coefficient k for large vehicles wblir Blind spot constraint coefficient k for large vehicles blirand virtual damping coefficient c i and virtual spring stiffness coefficient k i Step S04 includes the following steps:

[0141] S041) Based on the vehicle's coordinate data, identify the small vehicle that is closest to the large vehicle and travels in the same direction within a first length range centered on the large vehicle. In this embodiment, the first length is 10m.

[0142] S042) Analyze the identified large vehicles and their corresponding small vehicles, and extract the length, width, height, speed, and distance between the two vehicles as target data respectively;

[0143] S043) A semicircle with the front of the large vehicle as the center and a radius of the second length is the front blind spot of the large vehicle; the range between the target angle intervals behind the large vehicle is the rear blind spot of the large vehicle. In this embodiment, the second length is 1.5m and the target angle interval is between 19° and 45°.

[0144] S044) Using the measured trajectory data of the intersection in this example, the corresponding vehicle information is input into the calculation program. The calculated safe distance, actual distance, vehicle acceleration and vehicle speed are used. Based on the calculated parameters and actual data, the PSO algorithm is used to iteratively optimize the vehicle-type parameters in the spring-damping model. The parameters are adjusted and optimized in each iteration, and the optimal parameter values ​​are selected. The occlusion coefficient, blind spot constraint coefficient, virtual damping coefficient and virtual spring stiffness coefficient corresponding to large vehicles are calibrated with the optimal parameter values.

[0145] In step S05 of this embodiment, the high-risk signalized intersection risk field model calculates the risk field strength and corresponding Relative Traffic Safety Index (RDIS) of the high-risk signalized intersection in real time based on the target data of large vehicles and surrounding cars extracted in real time. When RIDS is greater than 1, it is determined that the large vehicle and surrounding cars are in a dangerous state, the judgment result is output and a safety warning is implemented; otherwise, the large vehicle and surrounding cars are in a safe state, and the judgment continues until the large vehicle leaves the signalized intersection.

[0146] The method of this embodiment is applied to Figure 2 During the afternoon rush hour at the intersection, risk fields were established and high-risk scenarios were identified under the influence of large vehicles. The results are shown in Table 2.

[0147] Table 2 RDIS Indicator Identification

[0148]

[0149] Table 3 ETTC Indicator Identification

[0150]

[0151] Using the collision time parameter ETTC Figure 2 Table 3 is obtained from the safety assessment of the intersection. Comparing Table 2 and Table 3, it can be seen that the RDIS index can identify high-risk scenarios more quickly than the ETTC index, while correctly identifying high-risk scenarios.

[0152] Example 2

[0153] This embodiment proposes a high-risk scenario driving risk intelligent early warning system, including a storage medium for interconnected microprocessors. The microprocessors are programmed or configured to execute the signalized intersection driving risk intelligent early warning method considering the impact of large vehicles described in Embodiment 1. Figure 5 As shown, the workflow of using the high-risk scenario driving risk intelligent early warning system of this embodiment to perform intelligent early warning of driving risks at signalized intersections considering the impact of large vehicles is as follows:

[0154] S101. Select high-risk signalized intersections, retrieve static traffic information from the traffic basic information database, and establish a static traffic scenario database for high-risk signalized intersections;

[0155] S102. Construct a risk field model for high-risk signal intersections, calibrate the static parameters of the risk field model for high-risk signal intersections using historical trajectory data of high-risk signal intersections, and determine the relationship between high-risk discrimination indicators and the risk field of high-risk signal intersections.

[0156] S103. By monitoring vehicles entering and exiting high-risk signal intersections through roadside surveillance cameras, when a large vehicle is detected entering the monitoring area, the high-risk scenario driving risk intelligent early warning system is triggered to execute the following steps;

[0157] S104. Identify large vehicles and surrounding vehicles, extract vehicle status information of large vehicles and surrounding vehicles in real time as target data, obtain the occlusion of large vehicles and calculate the blind spot of large vehicles, and calibrate the dynamic parameters of the risk field model of high-risk signal intersection based on the occlusion of large vehicles, the blind spot of large vehicles and target data.

[0158] S05. Input the real-time extracted target data into the high-risk signal intersection risk field model, calculate the high-risk signal intersection risk field in real time, calculate the high-risk discrimination index based on the high-risk signal intersection risk field, and then determine whether a high-risk event has occurred based on the high-risk discrimination index. Output the discrimination result to the high-risk scenario driving risk intelligent early warning system. The high-risk scenario driving risk intelligent early warning system implements safety warnings until the large vehicle leaves the high-risk signal intersection.

[0159] In summary, the intelligent early warning method and system for traffic risks at signalized intersections that considers the impact of large vehicles proposed in this invention have the following advantages:

[0160] 1. Using risk field theory, the environmental and motion factors of signalized intersections are quantitatively expressed. The influence of large vehicles is considered in the potential energy field, and the large vehicle shading field and blind spot constraint field are constructed. In the kinetic energy field, the safe distance of vehicles in each direction is determined according to vehicle type. The kinetic energy field parameters are calibrated separately considering the influence of vehicle type, so that at any time, the vehicle at any position in the signalized intersection has its corresponding risk field intensity value, high-risk scenarios are identified, so as to reduce the intersection accident rate and improve the intersection safety.

[0161] 2. Taking large vehicles and surrounding cars as the research object, this study investigates the motion and safety characteristics of large vehicles at signalized intersections. By combining environmental factors and the influencing factors of moving vehicles, a driving risk field is constructed. Based on this, a high-risk discrimination index is calculated to assess the collision risk of surrounding cars, reduce accidents, and improve intersection safety.

[0162] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the invention. Therefore, any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention should fall within the protection scope of the present invention.

Claims

1. A method for intelligent early warning of traffic risks at signalized intersections considering the impact of large vehicles, characterized by the following steps: include: S01. Select a high-risk signal intersection; S02. Construct a risk field model for high-risk signalized intersections. Use historical trajectory data of high-risk signalized intersections to calibrate the static parameters of the risk field model. The construction of the risk field model for high-risk signalized intersections includes: Based on the coordinate data of the vehicle trajectory, the desired trajectories for straight-going and turning are fitted to obtain the desired trajectory field; Based on the varying degrees of impact of different traffic light displays on vehicles as they pass through signalized intersections, a traffic light constraint field is constructed. Based on the shading effect of large vehicles on small cars in the same lane in the basic road section area and intersection weaving area, a large vehicle shading field is constructed. Based on the impact of large vehicles on small vehicles located in the blind spots of large vehicles in the basic road section area and the intersection weaving area, a blind spot constraint field is constructed. Based on the spring-damping model, construct the motion field for large vehicles and small cars; The field strength of the high-risk signal intersection is obtained by superimposing the field strength of the desired trajectory field, the field strength of the traffic light constraint field, the field strength of the large vehicle obstruction field, the field strength of the blind spot constraint field, and the field strength of the motion field. S03. Monitor vehicles entering and exiting high-risk signal intersections. When a large vehicle is detected, proceed to step S04. S04. Extract target data of large vehicles and surrounding vehicles in real time, obtain the occlusion status of large vehicles and calculate the blind spots of large vehicles. Based on the occlusion status of large vehicles, the blind spots of large vehicles and target data, calibrate the dynamic parameters of the risk field model of high-risk signal intersection. S05. Input the target data into the high-risk signal intersection risk field model in real time to calculate the high-risk signal intersection risk field, and calculate the high-risk judgment index based on the high-risk signal intersection risk field. If the high-risk judgment index is greater than the preset threshold, a safety warning is issued; otherwise, proceed to step S04 until the large vehicle leaves the high-risk signal intersection.

2. The intelligent early warning method for traffic risks at signalized intersections considering the impact of large vehicles, as described in claim 1, is characterized in that... Step S01 includes: if the current signalized intersection meets one or more of the preset traffic flow conditions, accident rate conditions, and visibility conditions, add the current signalized intersection to the set of high-risk signalized intersections.

3. The intelligent early warning method for traffic risks at signalized intersections considering the impact of large vehicles as described in claim 1, characterized in that, In step S01, after selecting high-risk signalized intersections, the process further includes establishing a static traffic scene database for high-risk signalized intersections, including: acquiring static data for different high-risk signalized intersections, whereby the static data includes lane data for each approach lane, signal light height, and the maximum permissible speed at the signalized intersection; the field strength expression for the desired trajectory field is as follows: In the formula, This refers to the lateral constraint coefficient; Lane width, It is the target lane attraction coefficient. This represents the perpendicular distance from the vehicle to the center line of the trajectory zone at the spatial point (x, y). The perpendicular point is ( , The length of the trajectory between the target endpoint of exit channel j and exit channel j; Let be the field strength value constraining the desired trajectory of the vehicle affected by exit lane n; The expression for the field strength of the traffic light constraint field is as follows: In the formula, This represents the signal light constraint field strength value for vehicles under signal control constraints. The remaining time for the green light; The maximum permitted speed at the intersection; The average speed of vehicles at the intersection; This is the time required for a vehicle to reach the exit lane of the intersection. It is the target signal control constraint coefficient.

4. The intelligent early warning method for traffic risks at signalized intersections considering the impact of large vehicles, as described in claim 3, is characterized in that... In step S02, when calibrating the static parameters of the risk field model of the high-risk signal intersection using historical trajectory data of the high-risk signal intersection, the following steps are taken: acquiring historical trajectory data of the high-risk signal intersection, preprocessing the historical trajectory data and fitting the expected trajectory equations of each approach lane and each turn of the high-risk signal intersection, calculating the time required for the vehicle to reach the exit lane of the intersection, the length of the trajectory strip and the center line of the trajectory strip, and calibrating the lateral constraint coefficient, the target lane attraction coefficient and the target signal control constraint coefficient based on the time required for the vehicle to reach the exit lane of the intersection, the length of the trajectory strip and the center line of the trajectory strip.

5. The intelligent early warning method for traffic risks at signalized intersections considering the impact of large vehicles as described in claim 1, characterized in that, The field strength expression for the field obstructed by the large vehicle is as follows: in, Let (x, y) be the shading constraint field intensity value of the spatial point (x, y) in the basic road segment region at time t. The angle between the driver's horizontal line of sight and the traffic light in a small car; The height of the traffic light; For the height of large vehicles; For small car height; This refers to the horizontal distance between a small vehicle and the traffic light. To block the height; The distance between large vehicles and small vehicles on a basic road section; The shading coefficient for the basic road section area; To block the distance; This refers to the average headway between large vehicles and smaller vehicles when large vehicles pass through the weaving zone. The shading coefficient of the intersection weaving area; The expression for the field strength of the blind zone constraint field is as follows: in, Let (x, y) be the blind zone constraint field strength value of the spatial point (x, y) in the basic road section area at time t; The distance from the rearview mirror of a large vehicle to the rear of a small car; the length of both large and small vehicles is... and The width of large vehicles and the width of small vehicles are and ; The distance on the y-coordinate between the small car and the large vehicle's rearview mirror field of view. This refers to the blind spot distance in front of large vehicles. The actual distance between large vehicles and small cars; Let P be the angle between the rear of the small car and the rearview mirror of the large vehicle; when P=0, the small car is in front of the large vehicle, and when P=1, the small car is behind the large vehicle. The constraint coefficient for the blind spot in front of large vehicles; The blind spot constraint coefficient for large vehicles; The actual distance between small cars and large vehicles; The distance between the rearview mirror of a large vehicle and the rear of the vehicle. This refers to the blind spot distance behind large vehicles. The angle between the center point of the small car and the rearview mirror of the large vehicle is used to determine whether the small car has entered the blind spot of the large vehicle.

6. The intelligent early warning method for traffic risks at signalized intersections considering the impact of large vehicles, as described in claim 5, is characterized in that... The expression for the field strength of the sports field is as follows: in, Maintain a safe distance for small vehicles; For large vehicles, the safe distance is maintained; the speeds of the vehicles in front and behind are... and ; i indicates that the model is a large vehicle in front and a small vehicle in front, respectively; This refers to the virtual damping coefficient in the model. This refers to the virtual spring stiffness coefficient in the model; This represents the intensity value of the sports field where vehicles travel.

7. The intelligent early warning method for traffic risks at signalized intersections considering the impact of large vehicles according to claim 6, characterized in that, Step S04 includes the following steps: Based on the vehicle's coordinate data, identify the small car that is closest to the large vehicle and travels in the same direction within a first length range centered on the large vehicle. The identified large vehicles and their corresponding small vehicles are analyzed, and their length, width, height, speed, and distance between the two vehicles are extracted as target data. The front blind spot of a large vehicle is defined as a semicircle with the front of the vehicle as the center and a radius equal to the second length. The rear blind spot of a large vehicle is defined as the range between the target angle intervals. Measured trajectory data of high-risk signal intersections are obtained, and the safe distance, actual distance, vehicle acceleration and vehicle speed are calculated based on the vehicle information in the measured trajectory data. Based on the calculated parameters and the corresponding actual data, the PSO algorithm is used to iteratively optimize the vehicle-type parameters in the spring-damping model, select the optimal parameter values, and use them to calibrate the occlusion coefficient, blind spot constraint coefficient, virtual damping coefficient and virtual spring stiffness coefficient for large vehicles.

8. The intelligent early warning method for traffic risks at signalized intersections considering the impact of large vehicles according to claim 1, characterized in that, The high-risk identification index in step S05 is the Relative Driving Safety Index (RDIS), whose expression is shown below: in, This indicates the risk field intensity value of a high-risk signalized intersection where vehicles are traveling. To assess the standard driving risk field intensity of a vehicle in a specific hazardous scenario, when RIDS is greater than 1, the vehicle is determined to be in a hazardous state; otherwise, the vehicle is in a safe state.

9. A high-risk scenario driving risk intelligent early warning system, characterized in that, The storage medium includes interconnected microprocessors, which are programmed or configured to execute the intelligent warning method for traffic risks at signalized intersections considering the impact of large vehicles as described in any one of claims 1 to 8.

Images