Road section entrance and exit vehicle safety induction method considering risk avoidance

Abstract

The invention discloses a road section entrance and exit vehicle safety guidance method considering risk avoidance, and belongs to the field of traffic guidance. Comprising the following steps: S1, carrying out global video shooting on a given entrance and exit road section through a monitor, and learning a track of a vehicle during turning in a video by using a YOLO algorithm; s2, extracting road information in the video, and constructing a two-dimensional coordinate system; s3, uploading information of vehicles entering the monitoring range to a traffic data service center through a vehicle network technology, and mapping the vehicle information into a two-dimensional coordinate system to realize spatial positioning; s4, according to the driving direction of the vehicle entering the monitoring range, the number of possible collision conflicts is judged, and safety induction is carried out; according to the steps, risk avoidance of the vehicle at the entrance and exit of the road section is achieved, and delay and exhaust emission caused by braking of the vehicle due to information blocking are reduced.

Description

Technical Field

[0001] The invention relates to the field of traffic guidance, and in particular to a road section entrance and exit vehicle safety induction method taking risk avoidance into consideration. Background Art

[0002] Road entrances and exits play an important role in urban traffic. They exist to meet the traffic needs of different areas in the city and ensure smooth and safe traffic flow. However, for road entrances and exits with low traffic flow and no traffic lights, drivers often pay more attention at the forks when passing through the road, which causes the vehicle to slow down, resulting in unnecessary delays and exhaust pollution. How to improve the driver's travel experience and reduce unnecessary delays is one of the important issues to solve the problems of poor road transportation efficiency, pollution emissions, and environmental protection. For large entrances and exits on roads with low traffic flow and no signals, drivers often pass at a higher speed, but if there are other vehicles trying to enter or exit the entrance and exit, the travel experience of both parties will be greatly reduced. However, if the vehicle position and speed information of each other is known in advance and the information is passed to the traveler, so that he can accelerate or decelerate at the right time and then pass through the entrance and exit smoothly, the driver's travel experience can be greatly improved, traffic delays caused by sudden stops can be reduced, and the occurrence of traffic accidents can be reduced.

[0003] In summary, transmitting vehicle information to the entrances and exits of large road sections with low traffic flow and no signals and guiding them to avoid risks is an effective guiding measure to reduce traffic accidents.

[0004] Existing traffic management systems, including traffic lights, intelligent transportation systems (ITS), vehicle-to-vehicle communication (V2V) technology, and navigation services, are mainly concentrated in urban areas with dense traffic flow or highways, relying on fixed infrastructure or the vehicle's own sensors to provide traffic information and assist decision-making. However, these systems often ignore the entrances and exits of large sections of roads with less traffic flow and lack of signal control, and the traffic management and risk avoidance measures in these areas are relatively insufficient. The present invention proposes an innovative vehicle information transmission and dynamic guidance system specifically for these scenarios ignored by existing technologies. By exchanging vehicle position and speed information in real time, the system can predict and guide the driver to adopt appropriate acceleration or deceleration strategies at the entrance and exit, thereby effectively reducing traffic delays and accident risks caused by sudden stops or avoidances, and significantly improving the driver's travel experience. This system does not require expensive infrastructure investment, has a high degree of adaptability and economy, which is not available in existing technologies. Summary of the invention

[0005] The present invention aims to improve the defects of current technologies and proposes a method for safety induction of large-scale road entrances and exits that takes risk avoidance into consideration. First, global monitoring is performed on the target entrance and exit to identify the trajectory of vehicles passing through the entrance and exit and use machine learning to familiarize them with the approximate trajectory of vehicle turns in the entrance and exit. Then, the divided areas are globally monitored and detected to identify the vehicle positions and current vehicle speeds in the monitoring area. Different risk avoidance measures are implemented according to the number of conflict points. By calculating the positional relationship, decision information is provided to the vehicle, such as non-interference passage, accelerated passage, and decelerated passage. This allows the vehicle to safely and smoothly avoid risks and pass through the target entrance and exit.

[0006] The present invention is implemented by adopting the following technical scheme: a road section entry and exit vehicle safety induction method considering risk avoidance, the specific steps are as follows:

[0007] S1. Use the monitor to take a global video of a given entrance and exit section, and use the YOLO algorithm to learn the trajectory of the vehicle turning in the video. The specific steps are as follows:

[0008] S1.1. Take a global video of a given import and export section;

[0009] S1.2, select the vehicle in the key frame of the video and mark the bounding box;

[0010] S1.3, divide the labeled data into training set, validation set and test set, where the training set accounts for 50-90%, the validation set accounts for 2-25%, and the test set accounts for 2-25%;

[0011] S1.4. Use the training set to train the YOLO model, evaluate the detection performance of the model on the test set, and finally verify the model in the validation set;

[0012] S1.5. Use the trained YOLO model to detect the vehicle in the video, and apply the tracking algorithm to track the vehicle in consecutive frames to obtain the consecutive frame tracking trajectory;

[0013] S1.6. Use the trained YOLO model to distinguish the turning trajectory and straight trajectory of the vehicle;

[0014] S2, extracting road information in the video and constructing a two-dimensional coordinate system, specifically, taking the main road as the x-axis, the branch road as the y-axis, and the intersection of the main road and the branch road as the origin to construct the two-dimensional coordinate system;

[0015] S3. Through vehicle network technology, the vehicle information entering the monitoring range is uploaded to the traffic data service center, and the vehicle information is mapped into a two-dimensional coordinate system to achieve spatial positioning. The specific steps are as follows:

[0016] S3.1. Use vehicle networking technology to monitor vehicles entering the monitoring area;

[0017] S3.2, collecting the vehicle's speed and two-dimensional coordinates;

[0018] S3.3, uploading the collected driving speed and two-dimensional coordinates to the traffic data service center;

[0019] S3.4, converting the driving speed and two-dimensional coordinates uploaded to the traffic data service center into a unified format;

[0020] S3.6, mapping the vehicle speed and the two-dimensional coordinates to a two-dimensional coordinate system to form a spatial positioning;

[0021] S4. According to the driving direction of the vehicle entering the monitoring range, determine the number of possible collisions and conflicts, and conduct safety guidance. The specific steps are as follows:

[0022] S4.1. Determine the number of possible collisions on the road based on the trained YOLO model.

[0023] S4.2. According to the number of possible collision conflicts, three collision conflict situations are divided, including: no collision conflict, one collision conflict, and two collision conflicts;

[0024] S4.3. In the event of a single collision or double collision, the vehicle will be guided to safety through the LED screens on the main road and branch roads.

[0025] Furthermore, in the case of no collision conflict, that is, there is no intersection between the main road and the branch road, no intervention is required, and "Please pass with the current driving state" is displayed on the LED screens beside the main road and the branch road. No collision conflict specifically includes two driving states, as follows:

[0026] When the position coordinates of the vehicle head in the main road are (x 0 ,y 0 ), the main road traffic length is X 1 When , the expression of the first driving state without collision conflict is as follows:

[0027] S 1 <|X 0 |+X 2

[0028] In the formula, S 1 The distance traveled by vehicles on the main road, and V 01 V is the speed of the main road vehicle when it enters the monitoring range. 02 is the speed of the branch road vehicle when it enters the monitoring range, L 0 is the trajectory of branch road vehicles passing through the entrance and exit without intervention, t 2is the time it takes for vehicles on the branch road to pass through the entrance and exit, and X 0 X is the lateral movement distance of the vehicle head on the main road, 2 is the distance that the branch road vehicle moves laterally, and Among them, V 02Y is the longitudinal velocity of the branch road vehicle, and V 02X is the lateral moving speed of the branch road vehicle, and

[0029] The specific expression of the second driving state without collision conflict is as follows:

[0030] S 1 -X 1 -(|x 0 |+X 2 )>0

[0031] In the formula, S 1 is the distance traveled by vehicles on the main road, X 1 The main road has long traffic, X 2 is the lateral movement distance of the branch road vehicle;

[0032] Furthermore, if a collision occurs, that is, there is a single vehicle or multiple vehicles on the main road, and there is a single vehicle on the branch road that wants to turn through the entrance or exit, it is necessary to judge and intervene in the three driving states in this situation;

[0033] The first driving state of a collision conflict is that the branch road vehicle will collide with the main road traffic without any operation. The specific expression is as follows:

[0034] S 1 -X 1 -(|x 0 |+X 2 )<0, and |x 0 |+X 2 <S 1

[0035] When the first driving state is predicted to occur, the vehicle will be prompted to stop and give way to the main road traffic, that is, the vehicle will be prompted to stop and give way to the branch road vehicles, and the LED screen of the branch road will display "Please slow down in advance and stop at the entrance and exit turns, stop and give way to the main road traffic".

[0036] The second driving state after a collision occurs is that the main road traffic rear-ends the branch road vehicle without any operation. The specific expression is as follows:

[0037] S 1 -X 1 -(|x 0 |+X2 )=0

[0038] When the second driving state is predicted to occur, a deceleration strategy is adopted, that is, the branch road vehicle rear-ends the main road traffic, and a deceleration prompt information is provided to the branch road vehicle on the branch road LED screen. The deceleration prompt information is the speed that the vehicle should reduce. The specific calculation formula is as follows:

[0039]

[0040] Where, t s is the time to pass the import and export after deceleration, a 2x is the lateral deceleration of the branch road vehicle;

[0041] The third driving state of a collision conflict is that the main road traffic rear-ends the branch road vehicle without any operation. The specific expression is as follows:

[0042] S 1 -X 1 -(|x 0 |+X 2 )<0

[0043] When the third driving state is predicted to occur, an acceleration strategy prompt is adopted, that is, the main road traffic rear-ends the branch road vehicle, and an acceleration prompt information is provided to the branch road vehicle on the branch road LED screen. The acceleration prompt information is the speed at which the vehicle should accelerate. The specific calculation formula is as follows:

[0044] V 2 =V 02 +a 2 ×t j ,V 2 <V

[0045] Where V 2 is the vehicle speed after acceleration, t j is the time it takes for vehicles on the branch road to pass through the entrance and exit under the acceleration strategy, a 2 is the acceleration of the branch road vehicle, V is the speed limit of the road section, and in,

[0046] Furthermore, the two collision conflicts occur, that is, there are three driving routes of vehicles in the entrance and exit, and their driving routes intersect at two intersections, which is judged as two conflict points. In this case, it is necessary to judge and intervene in the three driving states in this situation;

[0047] When two collisions occur, take the main road as the x-axis and the branch road as the y-axis, and assume:

[0048] The vehicle coming from the negative y axis direction turns right, the vehicle coming from the positive x axis direction turns left, and the vehicle coming from the negative x axis goes straight. The vehicle position information starting from the negative x axis is (x 1 ,y 1 ), with a velocity of v 1 , the lateral velocity is v 1x , the longitudinal velocity is v 1y , marked as vehicle 1;

[0049] The vehicle position information starting from the positive x-axis is (x 2 ,y 2 ), with a velocity of v 2 , the lateral velocity is v 2x , the longitudinal velocity is v 2y , marked as vehicle 2;

[0050] The vehicle position information starting from the negative y axis is (x 3 ,y 3 ), with a velocity of v 3 , the lateral velocity is v 2x , the longitudinal velocity is v 3y , marked as vehicle 3;

[0051] Specifically, when two collision conflicts occur in the first driving state, vehicle 1 first collides with vehicle 2, and then vehicle 1 collides with vehicle 3;

[0052] The first processing is performed, and the existence time T 1 So that:

[0053] v 1x ×T 1 +v 2x ×T 1 =|x 1 |+|x 2 |

[0054] |y 1 |-v 1y ×T 1 =|y 2 |-v 2y ×T 1

[0055] v 1 2 =v 1x 2 +v 1y 2

[0056] v 2 2 =v 2x 2 +v 2y2

[0057] The trajectory of vehicle 1 is from the negative x-axis to the positive x-axis, so v 1 =v 1x ;

[0058] Then we can get the following formula:

[0059] (|y 1 |-|y 2 |)×(v 1 +v 2x )=(|x 1 |+|x 2 |)×v 2y

[0060] At this time, let v 2x =x,v 2y =y, we get:

[0061] (v 2 2 -y 2 -2|y 2 ×y 1 |+2y 1 2 -2|y 1 |×v 1 )×x 2 +(2×|y 2 ×y 1 |×v 1 -2×y 1 2 ×v 1 +2×v 2 ×v 1 -2×|x 1 × 2 |-2×|x 2 |×v 2 )×x+[v 2 2 (x 1 2 +2×|x 1 × 2 |+x 2 2 )-y 2 2 ×v 1 2 -y 1 2 ×v 1 2 +2×y 2 2 ×|y 1 |×v1 ]=0

[0062] make:

[0063] A=(v 2 2 -y 2 -2|y 2 ×y 1 |+2y 1 2 -2|y 1 |×v 1 );

[0064] B=(2×|y 2 ×y 1 |×v 1 -2×y 1 2 ×v 1 +2×v 2 ×v 1 -2×|x 1 × 2 |-2×|x 2 |×v 2 );

[0065] C=[v 2 2 (x 1 2 +2×|x 1 × 2 |+x 2 2 )-y 2 2 ×v 1 2 -y 1 2 ×v 1 2 +2×y 2 2 ×|y 1 |×v 1 ]

[0066] get:

[0067] You can get:

[0068]

[0069] And because: 1x ×T 1 +v 2x ×T 1 =|x 1 |+|x 2 |, where

[0070] Bring in v 2x The collision conflict time T of the first driving state can be solved, and when the collision conflict time T is met, information is transmitted to the vehicle;

[0071] The information transmission to the vehicles is to transmit the deceleration information to vehicle No. 1, and the deceleration information is: if no collision is required, there is an acceleration a 1 , so that Where S is the driving distance at which the conflict occurs, S = v 1 T, S' is the driving distance without conflict;

[0072] At the same time, vehicle 1 has reached the conflict point, while vehicle 1 that is slowing down has not reached the conflict point. At this time, the information is transmitted to the main road LED screen to remind vehicle 1 that "there is a left-turning vehicle ahead, please slow down", and the branch road LED screen displays information to remind vehicle 2 to "maintain the current state and pass at a constant speed";

[0073] The second processing is performed, and the speed information obtained through the vehicle networking technology is that the speed of vehicle No. 1 passing the first conflict point through the deceleration measures is v 1 ', the coordinates of the first conflict are obtained through the previously obtained intersection information, vehicle position information and coordinate system (v 1 T 1 +x 1 ,y 1 ), then for the second conflict, there is a time T 2 >0, such that:

[0074] v 1 'T 2 +v 1 T 1 +x 1 =v 3x T 1

[0075] v 3y T 2 =|y 1 |+|y 2 |

[0076] v 3 2 =v 3x 2 +v 3y 2

[0077] Time T 2 Simplifying the formula for the unknowns yields:

[0078] v3 2 T 2 2 -v 1 2 -2v 1 '(v 1 T 1 +x 1 )T 2 -(v 1 T 1 +x 1 ) 2 -(|y 1 |+|y 2 |) 2 =0

[0079] The solution is:

[0080]

[0081] Then when the calculation meets the condition, information is transmitted to the vehicle;

[0082] The information transmission to the vehicles is carried out through the vehicle networking technology, providing vehicle 1 with the information "please pass at a constant speed at the current speed", and providing vehicle 3 with the information "there is a vehicle passing on the left main road, please slow down and pass the entrance and exit";

[0083] If the above conditions are not met, the information will be transmitted to the vehicle involved, that is, "pass the entrance and exit with the current driving status";

[0084] Specifically, when two collisions occur in the second driving state, if two collisions are to occur, vehicle 3 first collides with vehicle 1 and then collides with vehicle 2;

[0085] The first processing is performed, and the existence time T 3 So that:

[0086] v 3x T 3 +v 1x T 3 =|x 1 |+|x 3 |

[0087] |y 1 |+|y 3 |=v 3x T 3

[0088] v 3 2 =v 3x 2 +v 3y2

[0089] Since vehicle 1 only moves laterally, we have v 1x =v 1 , similarly there is v 3x =v 3 , and we get:

[0090]

[0091] and That is, the existence time T 3 The first conflict occurs. When the calculation meets the condition, the information is transmitted to the target vehicle through the vehicle networking technology. The message "the vehicle coming from the branch road at the intersection ahead, please slow down and pass" is transmitted to vehicle 1, and the message "please pass the entrance and exit in the current state" is transmitted to vehicle 3, so that the main road gives way to the vehicles on the branch road to pass the entrance and exit;

[0092] The second processing is carried out, and the coordinates of the first conflict point are measured as (x 3 -v 3x T 3 ,y 3 +v 3y T 3 ), let x 4 =x 3 -v 3x T 3 ,y 4 =y 3 +v 3y T 3 , then:

[0093] v 2 T 4 -v 3x T 4 =|x 2 |-|x 4 |

[0094] In the formula, And|y 4 |+v 3y T 4 =|y 2 |;

[0095] When the calculated result satisfies the above formula, it is predicted that a collision will occur if the two vehicles pass through the entrance and exit at the current speed and state;

[0096] The deceleration information is transmitted to vehicle 2 through the Internet of Vehicles technology, that is, "There is a vehicle passing through the entrance and exit of the branch road ahead, please slow down", and the information is also transmitted to vehicle 3, that is, "Please pass through the entrance and exit at the current speed and driving status", thereby completing risk avoidance;

[0097] If the above conditions are not met, the information will be transmitted to the vehicle involved, that is, "pass the entrance and exit with the current driving status";

[0098] Specifically, when two collisions occur, the third driving state is processed for the first time, and there is a time T 5 , so that:

[0099] (v 1y +v 2y )T 5 =|y 1 |+|y 2 |, and (v 1x -v 2x )T 5 =|x 1 |-|x 2 |

[0100]

[0101] v 1x 2 +v 1y 2 =v 1 2

[0102] v 2x 2 +v 2y 2 =v 2 2

[0103] Then, when the measured parameters meet the above requirements when entering the monitoring range, the first conflict occurs;

[0104] When the speed and distance parameters meet the above requirements, a deceleration strategy is prompted for vehicle 2, that is, a deceleration prompt is given to the branch road vehicle, and the vehicle networking technology transmits the deceleration information to the branch road vehicle, that is, "There is a vehicle turning left on the main road ahead, please slow down", and prompts the main road vehicle and vehicle 1 with the message "Please pass the entrance and exit at a constant speed in the current state";

[0105] The second processing is carried out, and the coordinates of the first conflict point are measured as (x 1 ',y 1 '), then when the speed and distance parameters satisfy the following formula, a second conflict will occur. The specific expression is as follows:

[0106] (v 3x +v 1x )T 6 =|y 1 |+|y 2 |

[0107] In the formula,

[0108] You can get:

[0109]

[0110] v 1x 2 +v 1y 2 =v 1 2

[0111] v 3x 2 +v 3y 2 =v 3 2

[0112] Then, when the measured parameters meet the above requirements when entering the monitoring range, the first conflict occurs;

[0113] When the speed and distance parameters meet the above requirements, the vehicle networking technology will prompt vehicle 3 with a deceleration strategy, that is, it will prompt the vehicle turning right on the main road to decelerate. The deceleration information will be transmitted to the LED display board through the vehicle networking technology, that is, "There is a vehicle turning left on the main road ahead, please slow down", and the vehicle on the main road will be prompted with a message, that is, "Please pass the entrance and exit at a constant speed in the current state."

[0114] Beneficial effects of the present invention

[0115] It combines the current hot topics of machine learning and applies it to the field of transportation; the decision information proposed by the present invention is based on the target entrance and exit, and its previous learning data comes from a large number of driving trajectories in the target entrance and exit, so the decision information proposed by the present invention is convincing; if the speed and position parameters of the driver meet the relevant conflict conditions during driving, the intervention information will be transmitted to the vehicles involved through the Internet of Vehicles technology to achieve risk avoidance.

[0116] Compared with the prior art, the present invention provides an innovative solution for traffic monitoring and safety management by integrating the YOLO algorithm, vehicle networking technology and coordinate system parameter calculation. Utilizing the efficient target detection capability of the YOLO algorithm, the system can capture and analyze the trajectory of vehicles passing through the entrance and exit of a road section in real time. Combined with vehicle networking technology, the system further obtains the speed and position information of the vehicle, laying a solid foundation for the precise parameter calculation in the coordinate system. This combination not only improves the comprehensiveness and accuracy of data collection, but also makes the analysis of vehicle behavior more intuitive and accurate through the establishment of a coordinate system.

[0117] On this basis, the system can predict and calculate possible conflicts, such as vehicles merging or potential collision risks. More importantly, the present invention is not limited to passive monitoring, but can actively intervene in safety based on the conflict prediction results. This intervention can provide real-time warning information to the driver, thereby effectively reducing the risk of accidents.

[0118] In addition, the advantages of the present invention are its high adaptability and flexibility. The present invention can provide reliable monitoring and early warning services. At the same time, the system's real-time data processing and dynamic response capabilities provide strong decision-making support for traffic management departments, which helps to improve the efficiency and effectiveness of traffic management.

[0119] Ultimately, the implementation of the present invention can significantly improve the traffic safety of road sections, reduce the accident rate, improve traffic smoothness, enhance the public's awareness of traffic safety, and provide a solid foundation for the development of intelligent transportation systems and the integration of autonomous driving technology. Through this comprehensive safety management method, the present invention can not only ensure the safety of every driver, but also bring broader economic and social benefits to society. BRIEF DESCRIPTION OF THE DRAWINGS

[0120] Figure 1 is a flow chart of the present invention;

[0121] Figure 2 A diagram illustrating the situation at a conflict point;

[0122] Figure 3 This is a diagram illustrating the first type of situation at two conflict points;

[0123] Figure 4 This is a diagram illustrating the second type of situation at two conflict points;

[0124] Figure 5 This is a diagram illustrating the third type of situation at two conflict points. DETAILED DESCRIPTION

[0125] The present invention is further described in detail below in conjunction with a specific embodiment. A road section entry and exit vehicle safety induction method considering risk avoidance has the following specific steps:

[0126] S1. Use the monitor to take a global video of a given entrance and exit section, and use the YOLO algorithm to learn the trajectory of the vehicle turning in the video. The specific steps are as follows:

[0127] S1.1. Take a global video of a given import and export section;

[0128] S1.2, select the vehicle in the key frame of the video and mark the bounding box;

[0129] S1.3, the labeled data is divided into a training set, a validation set, and a test set. In this embodiment, the training set accounts for 70%, the validation set accounts for 15%, and the test set accounts for 15%;

[0130] S1.4. Use the training set to train the YOLO model, evaluate the detection performance of the model on the test set, and finally verify the model in the validation set;

[0131] S1.5. Use the trained YOLO model to detect the vehicle in the video, and apply the tracking algorithm to track the vehicle in consecutive frames to obtain the consecutive frame tracking trajectory;

[0132] S1.6. Use the trained YOLO model to distinguish the turning trajectory and straight trajectory of the vehicle;

[0133] S2. Extract road information in the video and construct a two-dimensional coordinate system. Specifically, the two-dimensional coordinate system is constructed by taking the main road as the x-axis, the branch road as the y-axis, and the intersection of the main road and the branch road as the origin. In this embodiment, the monitoring range is 100m upstream of the x-axis and 100m downstream of the x-axis; 70m upstream of the y-axis and 70m downstream of the y-axis.

[0134] S3. Through vehicle network technology, the vehicle information entering the monitoring range is uploaded to the traffic data service center, and the vehicle information is mapped into a two-dimensional coordinate system to achieve spatial positioning. The specific steps are as follows:

[0135] S3.1. Use vehicle networking technology to monitor vehicles entering the monitoring area;

[0136] S3.2, collecting the vehicle's speed and two-dimensional coordinates;

[0137] S3.3, uploading the collected driving speed and two-dimensional coordinates to the traffic data service center;

[0138] S3.4, converting the driving speed and two-dimensional coordinates uploaded to the traffic data service center into a unified format;

[0139] S3.6, mapping the vehicle speed and the two-dimensional coordinates to a two-dimensional coordinate system to form a spatial positioning;

[0140] S4. According to the driving direction of the vehicle entering the monitoring range, determine the number of possible collisions and conflicts, and conduct safety guidance. The specific steps are as follows:

[0141] S4.1. Determine the number of possible collisions on the road based on the trained YOLO model.

[0142] S4.2. According to the number of possible collision conflicts, three collision conflict situations are divided, including: no collision conflict, one collision conflict, and two collision conflicts;

[0143] S4.3. In the event of a single collision or double collision, the vehicle will be guided to safety through the LED screens on the main road and branch roads.

[0144] Furthermore, in the case of no collision conflict, that is, there is no intersection between the main road and the branch road, no intervention is required, and "Please pass with the current driving state" is displayed on the LED screens beside the main road and the branch road. No collision conflict specifically includes two driving states, as follows:

[0145] When the position coordinates of the vehicle head in the main road are (x 0 ,y 0 ), the main road traffic length is X 1 When , the expression of the first driving state without collision conflict is as follows:

[0146] S 1 <|X 0 |+X 2

[0147] In the formula, S 1 The distance traveled by vehicles on the main road, and V 01 V is the speed of the main road vehicle when it enters the monitoring range. 02 is the speed of the branch road vehicle when it enters the monitoring range, L 0 is the trajectory of branch road vehicles passing through the entrance and exit without intervention, t 2 is the time it takes for vehicles on the branch road to pass through the entrance and exit, and X 0 X is the lateral movement distance of the vehicle head on the main road, 2 is the distance that the branch road vehicle moves laterally, and Among them, V 02Y is the longitudinal velocity of the branch vehicle, and V 02X is the lateral moving speed of the branch road vehicle, and

[0148] The specific expression of the second driving state without collision conflict is as follows:

[0149] S 1 -X 1 -(|x 0 |+X 2 )>0

[0150] In the formula, S 1 is the distance traveled by vehicles on the main road, X1 The main road has long traffic, X 2 is the lateral movement distance of the branch road vehicle;

[0151] Furthermore, if a collision occurs, that is, there is a single vehicle or multiple vehicles on the main road, and there is a single vehicle on the branch road that wants to turn through the entrance or exit, it is necessary to judge and intervene in the three driving states in this situation;

[0152] The first driving state of a collision conflict is that the branch road vehicle will collide with the main road traffic without any operation. The specific expression is as follows:

[0153] S 1 -X 1 -(|x 0 |+X 2 )<0, and |x 0 |+X 2 <S 1

[0154] When the first driving state is predicted to occur, the vehicle will be prompted to stop and give way to the main road traffic, that is, the vehicle will be prompted to stop and give way to the branch road vehicles, and the LED screen of the branch road will display "Please slow down in advance and stop at the entrance and exit turns, stop and give way to the main road traffic".

[0155] The second driving state after a collision occurs is that the main road traffic rear-ends the branch road vehicle without any operation. The specific expression is as follows:

[0156] S 1 -X 1 -(|x 0 |+X 2 )=0

[0157] When the second driving state is predicted to occur, a deceleration strategy is adopted, that is, the branch road vehicle rear-ends the main road traffic, and a deceleration prompt information is provided to the branch road vehicle on the branch road LED screen. The deceleration prompt information is the speed that the vehicle should reduce. The specific calculation formula is as follows:

[0158]

[0159] Where, t s is the time to pass the import and export after deceleration, a 2x is the lateral deceleration of the branch road vehicle;

[0160] The third driving state of a collision conflict is that the main road traffic rear-ends the branch road vehicle without any operation. The specific expression is as follows:

[0161] S 1 -X 1-(|x 0 |+X 2 )<0

[0162] When the third driving state is predicted to occur, an acceleration strategy prompt is adopted, that is, the main road traffic rear-ends the branch road vehicle, and an acceleration prompt information is provided to the branch road vehicle on the branch road LED screen. The acceleration prompt information is the speed at which the vehicle should accelerate. The specific calculation formula is as follows:

[0163] V 2 =V 02 +a 2 ×t j ,V 2 <V

[0164] Where V 2 is the vehicle speed after acceleration, t j is the time it takes for vehicles on the branch road to pass through the entrance and exit under the acceleration strategy, a 2 is the acceleration of the branch road vehicle, V is the speed limit of the road section, and in,

[0165] Furthermore, the two collision conflicts occur, that is, there are three driving routes of vehicles in the entrance and exit, and their driving routes intersect at two intersections, which is judged as two conflict points. In this case, it is necessary to judge and intervene in the three driving states in this situation;

[0166] When two collisions occur, take the main road as the x-axis and the branch road as the y-axis, and assume:

[0167] The vehicle coming from the negative y axis direction turns right, the vehicle coming from the positive x axis direction turns left, and the vehicle coming from the negative x axis goes straight. The vehicle position information starting from the negative x axis is (x 1 ,y 1 ), with a velocity of v 1 , the lateral velocity is v 1x , the longitudinal velocity is v 1y , marked as vehicle 1;

[0168] The vehicle position information starting from the positive x-axis is (x 2 ,y 2 ), with a velocity of v 2 , the lateral velocity is v 2x , the longitudinal velocity is v 2y , marked as vehicle 2;

[0169] The vehicle position information starting from the negative y axis is (x 3 ,y 3 ), with a velocity of v 3 , the lateral velocity is v 2x , the longitudinal velocity is v 3y, marked as vehicle 3;

[0170] Specifically, when two collision conflicts occur in the first driving state, vehicle 1 first collides with vehicle 2, and then vehicle 1 collides with vehicle 3;

[0171] The first processing is performed, and the existence time T 1 So that:

[0172] v 1x ×T 1 +v 2x ×T 1 =|x 1 |+|x 2 |

[0173] |y 1 |-v 1y ×T 1 =|y 2 |-v 2y ×T 1

[0174] v 1 2 =v 1x 2 +v 1y 2

[0175] v 2 2 =v 2x 2 +v 2y 2

[0176] The trajectory of vehicle 1 is from the negative x-axis to the positive x-axis, so v 1 =v 1x ;

[0177] Then we can get the following formula:

[0178] (|y 1 |-|y 2 |)×(v 1 +v 2x )=(|x 1 |+|x 2 |)×v 2y

[0179] At this time, let v 2x =x,v 2y =y, we get:

[0180] (v 2 2 -y 2 -2|y2 × y 1 | + 2y 1 2 -2|y 1 |× v 1 )× x 2 +(2×|y 2 × y 1 |× v 1 -2× y 1 2 × v 1 +2× v 2 × v 1 -2×|x 1 × x 2 |-2×|x 2 |× v 2 )× x + [v 2 2 (x 1 2 +2×|x 1 × x 2 |+x 2 2 )-y 2 2 × v 1 2 -y 1 2 × v 1 2 +2× y 2 2 ×|y 1 |× v 1 =0

[0181] Let:

[0182] A = (v 2 2 -y 2 -2|y 2 × y 1 |+2y 1 2 -2|y 1 |× v 1 );

[0183] B = (2×|y 2 × y 1 |× v 1 -2× y 1 2 × v 1 +2× v 2 × v 1 -2×|x 1 × x 2 |-2×|x 2|×v 2 );

[0184] C=[v 2 2 (x 1 2 +2×|x 1 × 2 |+x 2 2 )-y 2 2 ×v 1 2 -y 1 2 ×v 1 2 +2×y 2 2 ×|y 1 |×v 1 ]

[0185] get:

[0186] You can get:

[0187]

[0188] And because: 1x ×T 1 +v 2x ×T 1 =|x 1 |+|x 2 |, where

[0189] Bring in v 2x The collision conflict time T of the first driving state can be solved, and when the collision conflict time T is met, information is transmitted to the vehicle;

[0190] The information transmission to the vehicles is to transmit the deceleration information to vehicle No. 1, and the deceleration information is: if no collision is required, there is an acceleration a 1 , so that Where S is the driving distance at which the conflict occurs, S = v 1 T, S' is the driving distance without conflict;

[0191] At the same time, vehicle 1 has reached the conflict point, while vehicle 1 that is slowing down has not reached the conflict point. At this time, the information is transmitted to the main road LED screen to remind vehicle 1 that "there is a left-turning vehicle ahead, please slow down", and the branch road LED screen displays information to remind vehicle 2 to "maintain the current state and pass at a constant speed";

[0192] The second processing is performed, and the speed information obtained through the vehicle networking technology is that the speed of vehicle No. 1 passing the first conflict point through the deceleration measures is v 1 ', the coordinates of the first conflict are obtained through the previously obtained intersection information, vehicle position information and coordinate system (v 1 T 1 +x 1 ,y 1 ), then for the second conflict, there is a time T 2 >0, such that:

[0193] v 1 'T 2 +v 1 T 1 +x 1 =v 3x T 1

[0194] v 3y T 2 =|y 1 |+|y 2 |

[0195] v 3 2 =v 3x 2 +v 3y 2

[0196] Time T 2 Simplifying the formula for the unknowns yields:

[0197] v 3 2 T 2 2 -v 1 2 -2v 1 '(v 1 T 1 +x 1 )T 2 -(v 1 T 1 +x 1 ) 2 -(|y 1 |+|y 2 |) 2 =0

[0198] The solution is:

[0199]

[0200] Then when the calculation meets the condition, information is transmitted to the vehicle;

[0201] The information transmission to the vehicles is carried out through the vehicle networking technology, providing vehicle 1 with the information "please pass at a constant speed at the current speed", and providing vehicle 3 with the information "there is a vehicle passing on the left main road, please slow down and pass the entrance and exit";

[0202] If the above conditions are not met, the information will be transmitted to the vehicle involved, that is, "pass the entrance and exit with the current driving status";

[0203] Specifically, when two collisions occur in the second driving state, if two collisions are to occur, vehicle 3 first collides with vehicle 1 and then collides with vehicle 2;

[0204] The first processing is performed, and the existence time T 3 So that:

[0205] v 3x T 3 +v 1x T 3 =|x 1 |+|x 3 |

[0206] |y 1 |+|y 3 |=v 3x T 3

[0207] v 3 2 =v 3x 2 +v 3y 2

[0208] Since vehicle 1 only moves laterally, we have v 1x =v 1 , similarly there is v 3x =v 3 , and we get:

[0209]

[0210] and That is, the existence time T 3 The first conflict occurs. When the calculation meets the condition, the information is transmitted to the target vehicle through the vehicle networking technology. The message "the vehicle coming from the branch road at the intersection ahead, please slow down and pass" is transmitted to vehicle 1, and the message "please pass the entrance and exit in the current state" is transmitted to vehicle 3, so that the main road gives way to the vehicles on the branch road to pass the entrance and exit;

[0211] The second processing is carried out, and the coordinates of the first conflict point are measured as (x 3 -v 3x T 3,y 3 +v 3y T 3 ), let x 4 =x 3 -v 3x T 3 ,y 4 =y 3 +v 3y T 3 , then:

[0212] v 2 T 4 -v 3x T 4 =|x 2 |-|x 4 |

[0213] In the formula, And|y 4 |+v 3y T 4 =|y 2 |;

[0214] When the calculated result satisfies the above formula, it is predicted that a collision will occur if the two vehicles pass through the entrance and exit at the current speed and state;

[0215] The deceleration information is transmitted to vehicle 2 through the Internet of Vehicles technology, that is, "There is a vehicle passing through the entrance and exit of the branch road ahead, please slow down", and the information is also transmitted to vehicle 3, that is, "Please pass through the entrance and exit at the current speed and driving status", thereby completing risk avoidance;

[0216] If the above conditions are not met, the information will be transmitted to the vehicle involved, that is, "pass the entrance and exit with the current driving status";

[0217] Specifically, when two collisions occur, the third driving state is processed for the first time, and there is a time T 5 , so that:

[0218] (v 1y +v 2y )T 5 =|y 1 |+|y 2 |, and (v 1x -v 2x )T 5 =|x 1 |-|x 2 |

[0219]

[0220] v 1x 2 +v 1y2 =v 1 2

[0221] v 2x 2 +v 2y 2 =v 2 2

[0222] Then, when the measured parameters meet the above requirements when entering the monitoring range, the first conflict occurs;

[0223] When the speed and distance parameters meet the above requirements, a deceleration strategy is prompted for vehicle 2, that is, a deceleration prompt is given to the branch road vehicle, and the vehicle networking technology transmits the deceleration information to the branch road vehicle, that is, "There is a vehicle turning left on the main road ahead, please slow down", and prompts the main road vehicle and vehicle 1 with the message "Please pass the entrance and exit at a constant speed in the current state";

[0224] The second processing is carried out, and the coordinates of the first conflict point are measured as (x 1 ',y 1 '), then when the speed and distance parameters satisfy the following formula, a second conflict will occur. The specific expression is as follows:

[0225] (v 3x +v 1x )T 6 =|y 1 |+|y 2 |

[0226] In the formula,

[0227] You can get:

[0228]

[0229] v 1x 2 +v 1y 2 =v 1 2

[0230] v 3x 2 +v 3y 2 =v 3 2

[0231] Then, when the measured parameters meet the above requirements when entering the monitoring range, the first conflict occurs;

[0232] When the speed and distance parameters meet the above requirements, the vehicle networking technology will prompt vehicle 3 with a deceleration strategy, that is, it will prompt the vehicle turning right on the main road to decelerate. The deceleration information will be transmitted to the LED display board through the vehicle networking technology, that is, "There is a vehicle turning left on the main road ahead, please slow down", and the vehicle on the main road will be prompted with a message, that is, "Please pass the entrance and exit at a constant speed in the current state."

Claims

1. A road section entry and exit vehicle safety induction method considering risk aversion, characterized in that: The following steps are involved: S1. Use the monitor to take a global video of a given entrance and exit section, and use the YOLO algorithm to learn the trajectory of the vehicle turning in the video; S2, extracting road information from the video and constructing a two-dimensional coordinate system; S3. Through vehicle network technology, the vehicle information entering the monitoring range is uploaded to the traffic data service center, and the vehicle information is mapped into a two-dimensional coordinate system to achieve spatial positioning; S4. Determine the number of possible collisions based on the driving directions of vehicles entering the monitoring range, and provide safety guidance.

2. A road section entry and exit vehicle safety induction method considering risk avoidance according to claim 1, characterized in that: The specific steps of taking a global video of a given entrance and exit section through a monitor and using the YOLO algorithm to learn the trajectory of the vehicle turning in the video are as follows: S1.

1. Take a global video of a given import and export section; S1.2, select the vehicle in the key frame of the video and mark the bounding box; S1.3, divide the labeled data into training set, validation set and test set; S1.

4. Use the training set to train the YOLO model, evaluate the detection performance of the model on the test set, and finally verify the model in the validation set; S1.

5. Use the trained YOLO model to detect the vehicle in the video, and apply the tracking algorithm to track the vehicle in consecutive frames to obtain the consecutive frame tracking trajectory; S1.

6. Use the trained YOLO model to distinguish the turning trajectory and straight trajectory of the vehicle.

3. A road section entry and exit vehicle safety induction method considering risk avoidance according to claim 1, characterized in that: The specific method of extracting road information in the video and constructing a two-dimensional coordinate system is: taking the main road as the x-axis, the branch road as the y-axis, and the intersection of the main road and the branch road as the origin to construct the two-dimensional coordinate system.

4. A road section entry and exit vehicle safety induction method considering risk avoidance according to claim 1, characterized in that: The specific steps of uploading the vehicle information entering the monitoring range to the traffic data service center through the vehicle network technology and mapping the vehicle information to the coordinate system to achieve spatial positioning are as follows: S3.

1. Use vehicle networking technology to monitor vehicles entering the monitoring area; S3.2, collecting the vehicle's speed and two-dimensional coordinates; S3.3, uploading the collected driving speed and two-dimensional coordinates to the traffic data service center; S3.4, converting the driving speed and two-dimensional coordinates uploaded to the traffic data service center into a unified format; S3.

6. Map the vehicle speed and two-dimensional coordinates to a two-dimensional coordinate system to form spatial positioning.

5. A road section entry and exit vehicle safety induction method considering risk avoidance according to claim 1, characterized in that: The specific steps of determining the number of possible collisions and conflicts according to the driving direction of the vehicle entering the monitoring range and conducting safety guidance are as follows: S4.

1. Determine the number of possible collisions on the road based on the trained YOLO model. S4.

2. According to the number of possible collision conflicts, three collision conflict situations are divided, including: no collision conflict, one collision conflict, and two collision conflicts; S4.

3. In the event of a single collision or double collision, the vehicle will be guided to safety through the LED screens on the main road and branch roads.

6. A road section entry and exit vehicle safety induction method considering risk avoidance according to claim 5, characterized in that: The non-collision conflict is divided into two driving states. When a non-collision conflict occurs, no intervention is required for the vehicle, and information is displayed on the LED screens on the main road and the branch road to provide prompts.

7. A road section entry and exit vehicle safety induction method considering risk avoidance according to claim 5, characterized in that: The occurrence of a collision conflict is divided into three driving states. When a collision conflict occurs, the vehicle needs to be intervened, and three kinds of information are displayed on the LED screens on the main road and the branch road for prompting.

8. A road section entry and exit vehicle safety induction method considering risk avoidance according to claim 5, characterized in that: The occurrence of two collision conflicts is divided into three driving states. When two collision conflicts occur, vehicle intervention is required, and three types of information are displayed on the LED screens on the main road and the branch road for prompting.

Images