CN116229762A - Intersection right-turning collision early warning method based on intelligent luminous line - Google Patents

Intersection right-turning collision early warning method based on intelligent luminous line Download PDF

Info

Publication number
CN116229762A
CN116229762A CN202310233246.2A CN202310233246A CN116229762A CN 116229762 A CN116229762 A CN 116229762A CN 202310233246 A CN202310233246 A CN 202310233246A CN 116229762 A CN116229762 A CN 116229762A
Authority
CN
China
Prior art keywords
collision
vehicle
area
intersection
straight
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202310233246.2A
Other languages
Chinese (zh)
Inventor
解冰岩
宋风坤
张方伟
张兴林
孙正亮
林鹏
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shandong Hanxin Technology Co ltd
Original Assignee
Shandong Hanxin Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shandong Hanxin Technology Co ltd filed Critical Shandong Hanxin Technology Co ltd
Priority to CN202310233246.2A priority Critical patent/CN116229762A/en
Publication of CN116229762A publication Critical patent/CN116229762A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/16—Anti-collision systems
    • G08G1/167—Driving aids for lane monitoring, lane changing, e.g. blind spot detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
    • G08B21/18—Status alarms
    • G08B21/182—Level alarms, e.g. alarms responsive to variables exceeding a threshold
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0125—Traffic data processing
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/16—Anti-collision systems
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T10/00—Road transport of goods or passengers
    • Y02T10/10—Internal combustion engine [ICE] based vehicles
    • Y02T10/40—Engine management systems

Abstract

The invention discloses an intersection right-turn vehicle collision early warning method based on an intelligent luminous line, in particular to the field of vehicle-road cooperative intersection risk early warning, which comprises the following steps: the method comprises the steps of defining a detection area and a collision area of an intersection through data of road side equipment, vehicle-mounted equipment and a vehicle networking database center; the estimated time length of the collision between the straight-running vehicle and the right-turning vehicle is estimated rapidly through the trained neural network, so that the collision possibility is obtained; calculating estimated impact result severity values according to the kinetic energy of two vehicles, the number of passengers, the number of surrounding pedestrians and the number of non-motor vehicles; calculating to obtain estimated collision risk according to the collision possibility and the estimated collision result severity; and finally, mapping the collision risk value into the flicker frequency of the intelligent light-emitting lamp, and simultaneously carrying out signal flicker early warning on the straight-running vehicle and the right-turning vehicle. The invention can reduce the collision accident rate of the straight driving and the right-turning vehicles at the intersection and improve the traffic safety at the intersection.

Description

Intersection right-turning collision early warning method based on intelligent luminous line
Technical Field
The invention belongs to the technical field of intersection risk early warning implemented by a vehicle-road cooperation technology, and particularly relates to an intersection right-turning vehicle collision early warning method based on an intelligent luminous line.
Background
The road intersection is an important hub of the urban traffic system and carries important functions of collecting, steering and evacuating various motor vehicles, non-motor vehicles and pedestrians. Meanwhile, the intersection is also a convergence area of traffic behavior conflict points and traffic accident multiple points. According to statistics in the world, the proportion of road intersection accidents in traffic accidents on urban roads is up to 30% -80%. Traffic accident statistics data in recent years in China show that road intersection accidents still have gradually rising situations year by year.
Because the right-turn lanes of most of intersections in China are not provided with right-turn protection phases, when the traffic flow of the intersections is large, the right-turn vehicles have left-side view dead zones, drivers cannot timely see the straight-going vehicles which travel at a long distance in a lateral direction, the situation that the straight-going vehicles and the right-turn vehicles fight for lanes frequently occurs, and collision accidents are often caused.
On the other hand, at present, some cities in China apply luminous marking facilities at intersections, but the facilities mainly have the effect of increasing the sight distance accessibility of motor vehicles and pedestrians on roads at night and reminding motor vehicle drivers, non-motor vehicles and pedestrians of passing by, and the potential early warning capability of the facilities is not fully applied.
Disclosure of Invention
In order to reduce collision risk of right-turn vehicles and lateral straight vehicles at an intersection, the invention provides an intersection right-turn vehicle collision early warning method based on intelligent luminous lines based on a vehicle-road cooperation technology.
In order to achieve the above purpose, the present invention provides the following technical solutions: an intersection collision early warning method based on intelligent luminous marking facilities comprises the following steps:
s1, acquiring road side data and defining an area, wherein the area mainly comprises a detection area and a collision area.
S2, calculating the estimated time length and the collision possibility of the vehicle running to the collision area through the FNN neural network.
S3, estimating the severity value of the collision result through comprehensively calculating the vehicle condition and the surrounding condition.
S4, calculating collision risk, which is a function of the collision probability of the collision areas of the two vehicles and the estimated severity value of the collision result.
S5, converting the risk value into flicker frequency to perform flicker early warning of the luminous marking.
Wherein, the acquired data in S1 includes: the road side RSU, the vehicle-mounted OBU and the vehicle networking data center are used for acquiring the speed of all driving at the intersection, the real-time timing state of the traffic light, the actual engineering size of the intersection and the like.
Further, there are two demarcation ways of the detection area described in step S1:
when the straight signal lamp is green, the straight road detection range is length d a And a right-hand lane detection length of d b Is a region of (a) in the above-mentioned region(s).
When the straight-going signal lamp is yellow and the intersection is full red, namely, the lane right is about to alternate, the situation that the driver runs the yellow lamp to accelerate to cause accidents exists, and the detection range of the straight-going lane is enlarged to be the length d' a The right-turn detection range is still d b Is a region of (a) in the above-mentioned region(s).
D is as described above a ,d' a And d b The maximum limit value of (2) is the furthest distance in the flashing state of the visible marking to the vehicle, which varies due to the changing of the line of sight conditions caused by the conditions of the lane and the weather, time.
The collision area in the step S1 is a rectangular area, wherein the following manner is defined:
the width of the collision Area area_col is set to the straight-travel lane width w;
the length of the collision Area area_col is set as the braking distance dist_stop of the straight-running vehicle at the current speed, and the calculation formula is as follows:
Figure BDA0004121110680000021
wherein v is the real-time speed of the vehicle, t_reaction is the reaction time of the driver, g is the gravitational acceleration (9.8 m/s 2), and mu is the friction coefficient of the lane.
After the length and width of the collision area are determined, the coordinate range of the whole collision area is determined by determining the coordinates of any vertex of the area.
Specifically, the first nearest intersection point of the straight-going lane extension line and the left boundary of the right-turning vehicle running track is selected as the vertex A of the collision area, and the coordinate (A_xpos, A_ypos) is calculated as follows:
the A_ypos coordinate value is the y-direction coordinate of the extension line of the straight lane.
The calculation formula of the A_xpos coordinate value is as follows:
Figure BDA0004121110680000031
wherein:
l is the length of the straight-going vehicle;
psi is the vehicle turning steering angle;
p_xpos and p_ypos are coordinates of a virtual turning axis p;
in the step S2, a fully-connected neural network is established to calculate the estimated time length and the collision possibility of the vehicle running to the collision area. Wherein the calculation comprises the following steps:
and constructing parameters of which the input layers are vehicles and lanes, and outputting the parameters as a neural network for estimating the driving time.
The input layer comprises 6 input points, namely a current vehicle speed v, a current vehicle acceleration a, a distance s between the vehicle and a geometric center point o of a collision risk area from a vehicle number n between a stop line of an intersection and the current vehicle, a road weight remaining time g_dur of a current lane and a width w_lane of the lane.
Wherein the hidden layer is arranged as two layers, the number of neurons of the first layer is 7, and the number of neurons of the second layer is 4.
The output layer is a single neuron of the estimated travel time t_pred.
The activation functions of the hidden layer and the output layer are built, specifically, a ReLU activation function is adopted, and the formula is as follows.
f(x)＝max(0,x)
The loss function is built, and specifically, a HuberLoss loss function is adopted, and the formula is as follows:
Figure BDA0004121110680000041
wherein, delta is a super parameter, and HuberLoss approaches MSE when |y-f (x) | is less than or equal to delta; when |y-f (x) |is not less than δ, huberLoss becomes MAE-like. HuberLoss reduces the sensitivity problem to outliers and has everywhere conductive properties.
An optimizer is defined, and specifically, a small batch gradient descent method MBGD is adopted.
Each time one sample is selected, the Loss Function formula is abstracted as:
Figure BDA0004121110680000042
the formula for updating the weight parameters using the Loss function:
Figure BDA0004121110680000043
training a neural network through historical data of the Internet of vehicles to generate a direct-driving estimated time length model A and a right-turning vehicle estimated time length model B.
And calculating the possibility that two vehicles enter the collision area simultaneously by estimating the driving time.
Figure BDA0004121110680000051
Wherein t is a 、t b The estimated time for the straight-running vehicle and the right-turning vehicle to reach the collision area is respectively. Delta T is a preset super parameter representing the maximum time difference between two vehicles that can be accommodated.
The estimated severity value of the collision result in the step S3 is calculated by the following formula:
Figure BDA0004121110680000052
wherein f is an influence function representing the kinetic energy of the straight vehicle a and the kinetic energy of the right-turning vehicle b, k is an influence function of the number of passengers of the two vehicles, and g and z are influence functions of the number of passengers and the number of non-motor vehicles in a limited range around the collision area respectively.
In the step S4, the collision risk of the straight-driving and the right-turning vehicle at the moment i is calculated, and the calculation formula is as follows:
Figure BDA0004121110680000053
where i is the designated time, λ is the correction constant, P is the probability of collision, and S is the estimated outcome severity value of the collision.
In step S5, the risk value is converted into a flicker frequency to perform flicker early warning of the intelligent marking, and the calculation formula of the flicker frequency is as follows:
FRE blink ＝R i *μ
wherein R is i And (6) calculating the estimated risk at the ith moment in the step S4, wherein mu is a conversion coefficient.
Drawings
FIG. 1 is a diagram showing the calculation steps of an intelligent marking warning method
FIG. 2 is a schematic diagram of an intersection entity
FIG. 3 is a schematic view showing parameters of the collision zone
Fig. 4 is a schematic diagram of a neural network.
Detailed Description
It should be understood that the schematic drawings are not drawn to scale. A flowchart, as used in this disclosure, illustrates operations implemented according to some embodiments of the present invention. It should be appreciated that the operations of the flow diagrams may be implemented out of order and that steps without logical context may be performed in reverse order or concurrently. Moreover, one or more other operations may be added to or removed from the flow diagrams by those skilled in the art under the direction of the present disclosure.
Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of the invention. The appearances of such phrases in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. Those of skill in the art will explicitly and implicitly appreciate that the embodiments described herein may be combined with other embodiments.
The invention provides an intersection right-turning vehicle collision early warning method of an intelligent luminous line, which is described below.
Fig. 1 is a schematic flow chart of an embodiment of a method for early warning a collision of a right-turn vehicle at an intersection of an intelligent luminous line, which includes the steps of:
s1, acquiring road side data and defining an area, wherein the area mainly comprises a detection area and a collision area.
S2, calculating the estimated time length and the collision possibility of the vehicle running to the collision area through the FNN neural network.
S3, estimating the severity value of the collision result through comprehensively calculating the vehicle condition and the surrounding condition.
S4, calculating collision risk as a function of collision probability and estimated collision result severity value.
S5, converting the risk value into flicker frequency to perform flicker of the luminous marking.
The working principle and beneficial effects of the technical scheme are as follows: the method comprises the steps of defining a detection area and a collision area of an intersection through data of road side equipment, vehicle-mounted equipment and a vehicle networking database center; the estimated time length of the collision between the straight-running vehicle and the right-turning vehicle is estimated rapidly through the trained neural network, so that the collision possibility is obtained; calculating the estimated impact result severity after collision according to the kinetic energy of the two vehicles, the number of passengers, the number of surrounding pedestrians and the number of non-motor vehicles; calculating to obtain estimated collision risk according to the collision possibility and the estimated collision result severity; and finally, mapping the collision risk value into the flicker frequency of the intelligent light-emitting lamp, and simultaneously carrying out signal flicker early warning on the straight-running vehicle and the right-turning vehicle.
Specifically, the acquiring data in S1 includes: the road side RSU, the vehicle-mounted OBU and the vehicle networking data center are used for acquiring the speed of all driving at the intersection, the real-time timing state of the traffic light, the actual engineering size of the intersection and the like.
Fig. 2 is a schematic diagram of an intersection entity, including a detection area and a collision area described in S.
Specifically, the detection region described in S1 includes two cases:
when the straight signal lamp is green, the straight road detection range is length d a And a right-hand lane detection length of d b Is a region of (a) in the above-mentioned region(s).
When the straight-going signal lamp is yellow and the intersection is full red, namely, when the lane right is about to alternate, the situation that the driver runs the yellow lamp to accelerate to cause accidents exists, and the detection range of the straight-going lane is enlarged to d' a The right-turn detection range is still d b Is a region of (a) in the above-mentioned region(s).
D is as described above a 、d' a And d b The maximum limit value of (2) is the furthest distance in the visible flashing state to the vehicle, which varies due to the changing of the line of sight conditions caused by the conditions of the lane and the weather, time.
Specifically, the collision area of S1 is a rectangular area, as shown in fig. 3, and the following manner is defined:
the width of the collision Area area_col is set to the straight-travel lane width w;
the length of the collision Area area_col is set as the braking distance dist_stop of the straight-running vehicle at the current speed, and the calculation formula is as follows:
Figure BDA0004121110680000081
wherein v is the real-time speed of the vehicle, t_reaction is the reaction time of the driver, g is the gravitational acceleration (9.8 m/s 2), and mu is the friction coefficient of the lane.
After the length and the width of the collision area are determined, the coordinate of any vertex of the area needs to be determined, and then the coordinate range of the whole collision area can be determined.
Further, the first nearest collision point between the extension line of the straight-going lane and the running track of the right-turning vehicle is selected as a reference point A, and the coordinates (A_xpos, A_ypos) are calculated as follows:
the A_ypos coordinate value is the y-direction coordinate of the extension line of the straight lane.
The calculation formula of the A_xpos coordinate value is as follows:
Figure BDA0004121110680000082
wherein:
l is the length of the straight-going vehicle;
psi is the vehicle turning steering angle;
p_xpos and p_ypos are coordinates of a virtual turning axis p;
specifically, in the step S2, the step of establishing a fully connected neural network to calculate the estimated time length and the collision possibility of the vehicle traveling to the collision area includes the following steps:
the parameters of the vehicle and the lane are constructed as input layers, and the neural network with the estimated time length is output, and the structure diagram of the neural network is shown in fig. 4.
The input layer comprises 6 input points, namely a current vehicle speed v, a current acceleration a, the number of vehicles n between the current vehicle and a stop line of an intersection, a distance s between the vehicles and a geometric center point o of a collision risk area, a road right remaining time g_dur of a current lane and a width w_lane of the lane.
Specifically, the hidden layer is set to 2 layers, wherein the number of neurons of the first layer is 7, and the number of neurons of the second layer is 4.
The output layer is a single neuron of the estimated travel time t_pred.
And constructing an activation function of the hidden layer and the output layer. A ReLU activation function is used, the formula of which is as follows.
f(x)＝max(0,x)
The loss function is built, and a HuberLoss loss function is adopted, wherein the formula is as follows:
Figure BDA0004121110680000091
wherein, delta is a super parameter, and HuberLoss approaches MSE when |y-f (x) | is less than or equal to delta; when |y-f (x) |is not less than δ, huberLoss becomes MAE-like. HuberLoss reduces the sensitivity problem to outliers and has everywhere conductive properties.
An optimizer is defined, and specifically, a small batch gradient descent method MBGD is adopted.
Each time one sample is selected, the Loss Function formula is abstracted as:
Figure BDA0004121110680000092
Figure BDA0004121110680000101
the formula for updating the weight parameters using the Loss function:
Figure BDA0004121110680000102
training a neural network through historical driving data of the Internet of vehicles to generate a direct driving estimated time length model A and a right turning vehicle estimated time length model B.
And calculating the possibility that two vehicles enter the collision area simultaneously by estimating the driving time.
Figure BDA0004121110680000103
Wherein t is a 、t b The estimated time for the straight-running vehicle and the right-turning vehicle to reach the collision area is respectively. Delta T is a preset super parameter representing the maximum time difference between two vehicles that can be accommodated.
The working principle and beneficial effects of the technical scheme are as follows: the traditional prediction method mostly calculates the possibility of collision of two vehicles according to a joint probability formula through the current speed and distance. The method is simple in calculation, but because the method does not consider the comprehensive states of the vehicle, the road network and the signal control, the calculated running time is low in accuracy, and the estimated time is estimated more accurately through the comprehensive data of the intersection condition. In addition, the neural network model after being trained through the historical data can be calculated at a high speed, so that the real-time performance of budget is guaranteed.
Specifically, in the step S3, the severity of the collision result is estimated, and the calculation formula is as follows:
Figure BDA0004121110680000104
Figure BDA0004121110680000111
wherein f is an influence function representing the kinetic energy of the straight vehicle a and the kinetic energy of the right-turning vehicle b, k is an influence function of the number of passengers of the two vehicles, and g and z are influence functions of the number of passengers and the number of non-motor vehicles in a limited range around the collision area respectively. v and m are the vehicle speed and the mass of the vehicle, respectively. pas and ful define total numbers for the total number of passengers and passengers, respectively. Alpha and beta are the influence coefficients of the surrounding non-motor vehicles and the surrounding pedestrians, respectively.
The working principle and beneficial effects of the technical scheme are as follows: in a real collision scenario, in addition to serious consequences for both collision vehicles, secondary injuries to surrounding entities such as pedestrians or non-machines are often caused. The method estimates the severity of the consequences by comprehensively considering the kinetic energy conditions of the two vehicles and the quantity of passenger capacity, surrounding pedestrians and non-motor vehicles, and is more comprehensive and realistic than considering the consequences of the vehicles alone.
Specifically, in the step S4, the collision risk of the straight-running vehicle and the right-turning vehicle at the moment i is calculated, and the formula is as follows:
Figure BDA0004121110680000112
where λ is the correction constant, P is the probability of collision, and S is the estimated outcome severity value.
Specifically, in the step S5, the risk value is converted into the flicker frequency to perform the flicker of the luminous marking, and the calculation formula of the flicker frequency is as follows:
FRE blink ＝R i *μ
wherein R is i And (6) calculating the estimated risk at the ith moment in the step S4, wherein mu is a conversion coefficient.
The working principle and beneficial effects of the technical scheme are as follows: according to the method, the collision possibility and the estimated collision result of the vehicle are combined, the collision risk value is obtained through calculation, and then the value is mapped into the flicker frequency of the intelligent luminous line through a certain mapping relation, namely, the flicker frequency of the intelligent luminous line is larger and larger along with the increase of the estimated risk value, so that the effect of reminding a direct vehicle and a right-turning vehicle is achieved, and the occurrence of collision accidents is reduced.
It will be apparent that the foregoing is merely a preferred embodiment of the present invention and is not intended to limit the invention, and that various modifications and variations of the present invention will be apparent to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (9)

1. An intersection right turning collision early warning method based on an intelligent luminous line is characterized in that: the method comprises the following steps:
s1, acquiring road side data and defining an area, wherein the area mainly comprises a detection area and a collision area;
s2, calculating estimated time length and collision possibility of the vehicle running to a collision area through the FNN neural network;
s3, estimating a severity value of a collision result by comprehensively calculating the vehicle condition and the surrounding non-mechanical and pedestrian conditions;
s4, calculating collision risks of the direct vehicle and the right-turn vehicle at the appointed moment;
s5, converting the risk value into flicker frequency to perform flicker of the luminous marked line.
2. The method for pre-warning a right turn vehicle at an intersection based on an intelligent luminous line according to claim 1, wherein the step of acquiring road side data is as described in S1, and is characterized in that: the acquired data content comprises:
the road side RSU, the vehicle-mounted OBU and the vehicle networking data center are used for acquiring parameters of all driving at the intersection, real-time timing state of traffic lights, actual engineering dimensions of the intersection and the like.
3. The method for pre-warning a right turn vehicle collision at an intersection based on an intelligent luminous line according to claim 1, wherein the delimited area in S1 comprises a detection area, and is characterized in that: comprising the following steps:
when the signal lamp corresponding to the lateral straight driving is green light, the detection range of the straight driving path is length d a And a right-hand lane detection length of d b Is a region of (2);
when the straight-going signal lamp is yellow and the intersection is full red, namely, when the lane right is about to alternate, the situation that the driver runs the yellow lamp to accelerate to cause accidents exists, and the detection range of the straight-going lane is enlarged to be the length d' a The right-turn detection range is still d b Is a region of (a) in the above-mentioned region(s). Calculating the phase of the geometrical center position of all parking subareas in the cluster and the distance of the destinationPairing distance arrays;
d is as described above a ,d' a And d b The maximum limit value of (2) is the furthest distance in the visible flashing state to the vehicle, which varies due to the changing of the line of sight conditions caused by the conditions of the lane and the weather, time.
4. The method for pre-warning an intersection right-turn car collision based on an intelligent luminous line according to claim 1, wherein the delimited area in S1 comprises a collision area, and is characterized in that: comprising the following steps:
the width of the collision Area area_col is set to the straight-travel lane width w;
the length of the collision Area area_col is set as the braking distance dist_stop of the straight-running vehicle at the current speed, and the calculation formula is as follows:
Figure FDA0004121110670000021
wherein v is the real-time speed of the vehicle, t_reaction is the reaction time of the driver, g is the gravitational acceleration (9.8 m/s 2), and mu is the friction coefficient of the lane;
after the length and the width of the collision area are determined, the coordinate of any vertex of the area needs to be determined, and then the coordinate range of the whole collision area can be determined. The first nearest collision point of the extension line of the straight-going lane and the running track of the right-turning vehicle is selected as a reference point A, and the coordinates (A_xpos, A_ypos) are calculated as follows:
the A_ypos coordinate value is the y-direction coordinate of the extension line of the straight lane.
The calculation formula of the A_xpos coordinate value is as follows:
Figure FDA0004121110670000022
wherein: l is the length of the straight-going vehicle; psi is the vehicle turning steering angle; p_xpos and p_ypos are coordinates of the virtual turning axis p.
5. The method for pre-warning the collision of the right-turn vehicle at the intersection based on the intelligent luminous line according to claim 1, wherein the estimated time period for the vehicle to travel to the collision area is calculated through the FNN neural network as described in S2, and is characterized in that: comprising the following steps:
constructing parameters of which the input layers are vehicles and lanes, and outputting a neural network which is a predicted duration;
the input layer comprises 5 input points, namely a current speed v, a current acceleration a, a distance s between the vehicle and a geometric center point o of a collision risk area, a road right remaining time g_dur of a current lane and a width w_lane of the lane;
the output layer is a single neuron for estimating the driving time t_pred;
constructing an activation function of the hidden layer and the output layer;
constructing a loss function and defining an optimizer;
training a neural network through historical data of the Internet of vehicles to generate a direct-driving estimated time length model A and a right-turning vehicle estimated time length model B.
6. The method for pre-warning the collision of the right-turn vehicles at the intersection based on the intelligent luminous line according to claim 1, wherein the method for calculating the collision possibility is characterized in that: the method comprises the following calculation formula:
Figure FDA0004121110670000031
wherein t is a 、t b The estimated time for the straight-running vehicle and the right-turning vehicle to reach the collision area is respectively. Delta T is a preset super parameter representing the maximum time difference between two vehicles that can be accommodated.
7. The method for pre-warning the collision of the right-turn vehicle at the intersection based on the intelligent luminous line as set forth in claim 1, wherein the estimated severity of the collision is as set forth in S3, and is characterized in that: the method comprises the following calculation formula:
Figure FDA0004121110670000032
wherein f is an influence function representing the kinetic energy of the straight vehicle a and the kinetic energy of the right-turning vehicle b, k is an influence function of the number of passengers of the two vehicles, and g and z are influence functions of the number of passengers and the number of non-motor vehicles in a limited range around the collision area respectively.
8. The method for pre-warning the collision of the right-turn vehicle at the intersection based on the intelligent luminous line according to claim 1, wherein the step of calculating the collision risk of the direct-drive vehicle and the right-turn vehicle at the designated moment as described in the step S4 is characterized in that: the method comprises the following calculation formula:
Figure FDA0004121110670000041
where i is the designated time, λ is the correction constant, P is the probability of collision, and S is the estimated outcome severity value of the collision.
9. The method for early warning of a right turn vehicle collision at an intersection based on an intelligent luminous line according to claim 1, wherein the step of converting the risk value into a flicker frequency to perform flicker of the luminous line in S5 is characterized in that: the calculation formula of the flicker frequency is:
FRE blink ＝R i *μ
wherein R is i And (6) calculating the estimated risk at the ith moment in the step S4, wherein mu is a conversion coefficient.
CN202310233246.2A 2023-03-13 2023-03-13 Intersection right-turning collision early warning method based on intelligent luminous line Pending CN116229762A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310233246.2A CN116229762A (en) 2023-03-13 2023-03-13 Intersection right-turning collision early warning method based on intelligent luminous line

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310233246.2A CN116229762A (en) 2023-03-13 2023-03-13 Intersection right-turning collision early warning method based on intelligent luminous line

Publications (1)

Publication Number Publication Date
CN116229762A true CN116229762A (en) 2023-06-06

Family

ID=86569342

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310233246.2A Pending CN116229762A (en) 2023-03-13 2023-03-13 Intersection right-turning collision early warning method based on intelligent luminous line

Country Status (1)

Country Link
CN (1) CN116229762A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117710909A (en) * 2024-02-02 2024-03-15 多彩贵州数字科技股份有限公司 Rural road intelligent monitoring system based on target detection and instance segmentation

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117710909A (en) * 2024-02-02 2024-03-15 多彩贵州数字科技股份有限公司 Rural road intelligent monitoring system based on target detection and instance segmentation
CN117710909B (en) * 2024-02-02 2024-04-12 多彩贵州数字科技股份有限公司 Rural road intelligent monitoring system based on target detection and instance segmentation

Similar Documents

Publication Publication Date Title
CN110488802B (en) Decision-making method for dynamic behaviors of automatic driving vehicle in internet environment
CN108109415B (en) Internet automatic driving control system and method with anti-congestion function
CN109727469B (en) Comprehensive risk degree evaluation method for automatically driven vehicles under multiple lanes
CN111523822B (en) Driving risk unified quantification method comprehensively considering all factors of person-vehicle-road
CN111583693B (en) Intelligent traffic cooperative operation system for urban road and intelligent vehicle control method
CN113050455A (en) Digital twin test system for intelligent networked automobile and control method
CN110155046A (en) Automatic emergency brake hierarchical control method and system
CN111295628B (en) Method and system for controlling a vehicle in an autonomous driving mode
JP2019537414A5 (en)
CN107919022B (en) Intelligent traffic light signal control method for dynamic time length distribution
CN113723699B (en) Method and system for warning correction handle control of expressway safety vehicle speed in severe weather
CN104417561A (en) Context-aware threat response arbitration
CN110544377A (en) intersection pedestrian collision avoidance method based on vehicle-road cooperation
CN111028504A (en) Urban expressway intelligent traffic control method and system
CN106530691A (en) Hybrid vehicle model multilane cellular automaton model considering vehicle occupancy space
JP2018055296A (en) Information processor, information processing method and information processing program
CN104504902B (en) Vehicle uses right of way, right of way computational methods and the road charging method based on right of way
CN116229762A (en) Intersection right-turning collision early warning method based on intelligent luminous line
CN113536538B (en) Pedestrian and vehicle interaction simulation method and system based on traffic conflict
CN112885116B (en) Highway rain and fog scene vehicle and road collaborative induction system
CN114005290A (en) Intersection left-turn bus lane-borrowing passing method and device
CN111009125B (en) Urban intersection pedestrian and vehicle collision probability calculation method
CN113119945A (en) Automobile advanced auxiliary driving system based on environment model
CN206383950U (en) The vehicle ring sense information acquisition system of driving behavior analysis is judged for multiple target
CN115641736A (en) Intelligent induction system and method for special highway sections

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination