A decision-making method and system for intelligent connected vehicles at intersections without traffic lights.
By acquiring information about the intersection environment and traffic participants, predicting trajectories and generating conflict zones, and adjusting vehicle speeds, the system solves the decision-making challenges faced by connected vehicles and other traffic participants at intersections without traffic lights, thus achieving safe and efficient traffic management.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- NANJING UNIV OF AERONAUTICS & ASTRONAUTICS
- Filing Date
- 2023-07-04
- Publication Date
- 2026-06-30
AI Technical Summary
Existing technologies cannot effectively handle vehicles and other traffic participants, such as pedestrians and bicycles, that do not support direct vehicle-to-everything (V2X) communication in traffic management at intersections without traffic lights, leading to frequent traffic accidents. Furthermore, existing research has failed to fully consider the impact of road conditions and uncontrollable factors.
By acquiring intersection environmental information and traffic participant data, using semantic segmentation and convolutional neural networks to identify road conditions, combining deep learning neural networks to predict traffic participant trajectories, generating conflict zones and calculating occupancy time, adjusting vehicle speed to avoid collisions, and employing a smart connected vehicle decision-making system consisting of perception units, server units, computing units, and control units, a decision-making system is established.
It improves traffic safety and efficiency at intersections without traffic lights, reduces the variation in the time spent in conflict zones, mitigates the impact of traffic oscillations, and ensures vehicles pass through intersections safely and efficiently.
Smart Images
Figure CN116985790B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of intelligent connected vehicle control, and in particular to a decision-making method and system for intelligent connected vehicles at intersections without traffic lights. Background Technology
[0002] In my country, intersections are the most frequent locations for traffic accidents. The main causes of these accidents include drivers failing to yield, disobeying traffic signals, and speeding. Vehicles failing to yield or slow down when passing through intersections are particularly problematic.
[0003] To improve intersection safety while maintaining traffic efficiency, traffic lights, roundabouts, and All-Way Stop Control (AWSC) are commonly used for management. However, some intersections, such as at-grade intersections of national / provincial highways and rural roads, or intersections in the far suburbs of cities, cannot be equipped with traffic lights. This is because traffic volume on main roads is much higher than on side roads, and in the latter cases, traffic volume is lower. In some urban intersections, traffic lights switch to a flashing yellow light at midnight, a method of eliminating traffic lights to improve intersection efficiency. Traffic light-less intersections are widespread in our lives.
[0004] In recent years, with the development of wireless communication in the field of autonomous driving and vehicle-to-everything (V2X) communication, LTE-V2X (Vehicle-to-Everything) communication systems have enabled communication and information exchange between vehicles (V2V), between vehicles and infrastructure (V2I), and between vehicles and pedestrians (V2P). Many scholars both domestically and internationally have applied V2X technology to intersection decision-making and management, eliminating traffic lights and using V2X to obtain vehicle information for optimal decision-making, ensuring safe and efficient traffic flow. Most of these studies assume that all vehicles support V2X and are capable of autonomous driving or controllable operation. However, in real-world driving scenarios, the vast majority of vehicles do not support connectivity and intelligence, and driver behavior is often unpredictable. Furthermore, these studies typically assume that only vehicles are traffic participants, failing to incorporate pedestrians, bicycles, and other moving targets such as two-wheeled, three-wheeled, and electric motorcycles—objects that cannot be ignored in real-world driving scenarios—into the intersection decision-making mechanism. Summary of the Invention
[0005] To address the aforementioned shortcomings, this invention provides a decision-making method and system for intelligent connected vehicles at intersections without traffic lights. It predicts the trajectories of traffic participants at the intersection, generates conflict zones through speed weights, and incorporates road conditions and uncontrollable factors into the decision-making management of the intersection, enabling intelligent connected vehicles to safely and efficiently pass through intersections without traffic lights.
[0006] This invention provides a decision-making method for intelligent connected vehicles at intersections without traffic lights, specifically including the following steps:
[0007] Step 1), obtain road environment information at the intersection;
[0008] Step 2), obtain information on traffic participants at the intersection;
[0009] Step 3): Based on the information obtained in steps 1) and 2), predict the future trajectory of the traffic participants;
[0010] Step 4): Based on the future trajectory predicted in Step 3), generate intersection driving conflict points between the main vehicle and other traffic participants, and generate intersection driving conflict areas.
[0011] Step 5), calculate the occupancy time of the conflict zone, including the time when the main vehicle and other traffic participants at the intersection who are clashing with the main vehicle are about to enter and leave the conflict zone.
[0012] Step 6) Determine whether a collision event has occurred based on the time the conflict area occupies, and adjust the speed of the controlled vehicle to allow the vehicle to pass through the intersection safely and efficiently.
[0013] Preferably, step 1) road environment information specifically refers to the weather conditions of the intersection at the current moment. Based on computer vision methods such as semantic segmentation and convolutional neural networks, the weather conditions are identified, and the road surface conditions corresponding to the current weather are queried. There are three types of road surface conditions: dry, wet and slippery, and icy and snowy. The adhesion coefficient for dry road surface is 0.8 to 0.9, the adhesion coefficient for wet and slippery road surface is 0.5 to 0.7, and the adhesion coefficient for icy and snowy road surface is 0.1 to 0.2.
[0014] Preferably, in step 2), the acquisition of traffic participant information at the intersection mainly relies on cameras installed at intersections without traffic lights. The cameras can be set up in a single diagonal arrangement, a double diagonal arrangement, or other forms of multiple camera setups. This is to avoid the problem of low detection accuracy caused by factors such as occlusion between objects, lighting, and viewing angle when using a single camera. The traffic participant information includes the type of traffic participant, location coordinates, speed, acceleration, and direction of travel. The types of traffic participants include cars, pedestrians, two-wheeled vehicles (including motorized and non-motorized vehicles), three-wheeled vehicles, and non-motorized four-wheeled vehicles.
[0015] Preferably, step 3) uses the traffic participant information detected in step 2) to predict the future trajectories of different types of traffic participants at the intersection in real time. Multi-sensor technology is used to perceive traffic participants in advance, obtain their historical location and dynamic information, and store and collect it to generate historical driving or movement trajectories. The geometric features of the intersection obtained in advance are used as input to a pre-trained deep learning neural network for predicting the future trajectories of traffic participants, so as to predict the future trajectories of different types of traffic participants.
[0016] Preferably, step 4) generates the intersection driving conflict area, and the specific steps are as follows:
[0017] 41) Generate the intersection conflict point between the main vehicle and other traffic participants. Generate the intersection conflict point P based on the intersection of the main vehicle's planned trajectory and the predicted future trajectories of other traffic participants obtained in step 3). i =(x i ,y i ), where c and N are the number of intersection driving conflict points generated by the trajectories of the main vehicle and other traffic participants, respectively. It should be noted that the trajectory includes information such as position, speed, and acceleration;
[0018] 42) Generate the intersection conflict zone between the main vehicle and other traffic participants. Using the intersection conflict point P from step 41) as an example. i With R as the center, i Generate the intersection driving conflict zone δ based on the radius. i , Where K is the gain coefficient, v ego The speed of the main vehicle, v i The speeds of traffic participants whose future trajectories clash with the main vehicle's trajectory at the intersection are represented by A and B, respectively, which serve as weighting coefficients for the speeds of the main vehicle and other traffic participants, satisfying the condition A + B = 1. In this embodiment, considering that the main vehicle is controllable while other traffic participants are not, they are considered nominally uncontrollable. Therefore, the relationship A < B exists, meaning that the speed of other traffic participants crossing the intersection has a greater impact on the size of the intersection conflict area.
[0019] 43) Calculate the distance between adjacent conflict zones and choose to merge them or not. Define the distance L between the i-th and (i+1)-th intersection conflict zones as |P i P i+1 |. If L≤R i +R i+1 If +s0, then two adjacent conflicting regions are considered as the same conflicting region. Further judgment is needed if L≤R. i +R i+1Then, two adjacent but intersecting conflict regions are merged into a single conflict region, with δ i =δ i ∪δ i+1 If R i +R i+1 ≤L≤R i +R i+1 +s0, then a connected region Δδ is generated. i δ i =δ i ∪δ i+1 ∪Δδ i If L > R i +R i+1 If +s0, no action is taken. Where R... i and R i+1 s0 and s1 are the radii of the driving conflict zones at the i-th and i+1-th intersections, respectively, and s0 is the safety distance, which is mainly determined by the size of the main vehicle.
[0020] 44) Obtain the final set of conflict areas based on the processed conflict areas. After step 43), the original number of conflict areas may decrease. Re-traverse all the intersection driving conflict areas obtained after step 43) to obtain the set of intersection driving conflict areas for the main vehicle. Where j = 1, 2, ..., M, M ≤ N, and M is the final number of intersection driving conflict zones.
[0021] Preferably, the occupancy time of the conflict area in step 5) can be represented by a time interval:
[0022] 51) For traffic participants at intersections who conflict with the main vehicle, T i =[ta i ,tl i ], ta i =S i / v i +t0, tl i =ta i +S′ i / v i , where, i=1,2,...,N, ta i and tl i S represents the time when the traffic participant at the intersection where the conflict with the main vehicle is about to enter and leave the conflict area, respectively. i S′ i These represent the distance from the current location to the point of entering the conflict zone and the distance from the point of entering the conflict zone to the point of leaving the conflict zone, respectively, in the direction of travel or movement speed of the traffic participant. i t0 represents the speed at which traffic participants at the intersection travel or move, and t0 represents the absolute time corresponding to the current moment.
[0023] 52) For the main vehicle, T_ego j =[ta_ego j ,tl_ego j ], ta_ego j =S_ego j / v ego +t0,ta_ego j =ta_ego j +S′_ego j / v ego Where j = 1, 2, ..., M, ta_ego j and tl_ego j Let S and S′ represent the times when the main vehicle is about to enter and leave the j-th conflict zone, respectively. Let S and S′ represent the distances from the main vehicle's current position to when it is about to enter the j-th conflict zone and the distances from when it is about to enter the j-th conflict zone to when it is about to leave the j-th conflict zone, respectively, along the direction of its travel speed. ego The speed of the main vehicle.
[0024] Preferably, since the traffic participant detection in step 2) usually uses a box to represent the general outline of the detection target, the definition of the traffic participant about to enter the conflict area in step 5) is that the foremost point of the box corresponding to the object enters the conflict area for the first time in the direction of movement and intersects the boundary representing the conflict area; the definition of the traffic participant about to leave the conflict area is that the last point of the box corresponding to the object leaves the conflict area in the direction of movement and intersects the boundary representing the conflict area.
[0025] As a preferred method, step 6) determines whether a collision event has occurred by calculating the conflict area δ between the main vehicle and the intersection. j Does the time occupied by the traffic participant in the conflict zone overlap with the time occupied by N other traffic participants passing through the conflict zone? If The conflict zone δ between the main vehicle and the intersection j If the occupancy time of the first traffic participant overlaps with the occupancy time of N other traffic participants passing through the conflict zone, it indicates that a collision event will occur; if If the time occupied by the main vehicle and other traffic participants in the conflict zone does not overlap, it indicates that a collision event will not occur.
[0026] Preferably, step 6) adjusting the speed of the main vehicle to ensure its safe and efficient passage through the intersection mainly includes two stages:
[0027] 61) Before entering the intersection, the main vehicle moves at a speed of v ego If driving in the straight-ahead buffer zone, At this point, adjust the speed v of the main vehicle.ego ′=v ego -Δv d ,until Among them, v ego ' represents the speed of the master vehicle in the next decision cycle, Δv d To reduce the speed step length, otherwise the main vehicle will continue to travel at the originally planned speed;
[0028] 62) After entering the intersection, the main vehicle moves at a speed of v ego When going straight or turning at an intersection, the resulting intersection driving conflict area δ j ,like Adjust the speed v of the main vehicle ego ′=v ego +Δv a ,until Where, Δv a To accelerate the step length, otherwise the main vehicle will continue to travel at the originally planned speed;
[0029] As a preferred option, step 6) adjusting the speed of the main vehicle should be based on the road conditions where the main vehicle is located. The lower the road surface adhesion coefficient, the lower the planned speed should be. This should meet the requirements of the main vehicle's power and handling stability, as well as comply with the traffic speed limit rules of the intersection, to ensure that the main vehicle passes through the intersection without traffic lights safely and without collision.
[0030] The present invention also provides an intelligent connected vehicle decision-making system for intersections without traffic lights, comprising four parts: a sensing unit, a server unit, a computing unit, and a control unit;
[0031] The perception unit uses appropriate sensors and algorithms based on computer vision technology to obtain road conditions and achieve accurate detection of traffic participants. It collects information on the type of traffic participants, their location coordinates, speed, acceleration, and driving direction, and transmits the above information to the computing unit in real time.
[0032] The server unit includes a roadside server and a vehicle-side server. The server realizes bidirectional transmission of vehicle-road information through wireless communication technology. The roadside server can send a takeover request to the vehicle-side server. After obtaining permission from the vehicle-side server, it can send control commands. The vehicle-side server can send the planned trajectory and vehicle information to the roadside server and can receive control commands sent by the roadside server.
[0033] The computing unit consists of a prediction module and a decision module. The prediction module predicts the future trajectories of traffic participants at the intersection based on the information collected by the perception unit. The decision module generates conflict zones and calculates the occupancy time based on the predicted future trajectories and the vehicle planning trajectories received by the roadside server, optimizes the speed of vehicle planning, and generates control commands for the vehicles.
[0034] The control unit receives upper-level control commands and issues specific lower-level control commands to the throttle, braking, and steering systems to perform lateral and longitudinal control of the vehicle, ensuring that the vehicle travels along the planned trajectory.
[0035] Preferably, the roadside server in the sensing unit and server unit belongs to the roadside device, and the vehicle-side server and control unit belong to the vehicle-side device. The computing unit can be either a component of the roadside device or a component of the vehicle-side device. It should be noted that the aforementioned roadside device and vehicle-side device together constitute the intelligent connected vehicle decision-making device for traffic light-free intersections provided by the present invention.
[0036] Preferably, for the main vehicle, if it supports intelligent connectivity but not autonomous driving, and there is no computing unit on the vehicle that can provide sufficient computing power, then the computing unit is part of the roadside device, indicating that the main vehicle's prediction and decision-making process depends on the computing unit in the roadside device; if it supports intelligent connectivity and autonomous driving, and there is a computing unit on the vehicle with sufficient computing power, then the computing unit is part of the vehicle-side device, indicating that the main vehicle's prediction and decision-making process depends on the computing unit in the vehicle-side device.
[0037] Preferably, when the computing unit is part of the vehicle-mounted device, it can receive information processed by the sensing unit located on the roadside device, and can also receive information collected by the vehicle-mounted sensing unit, such as the environmental sensing sensors carried by the vehicle itself, such as lidar, millimeter-wave radar, and cameras, and perform sensing fusion to improve the accuracy and precision of sensing, and help the decision-making module make more scientific and reasonable decisions.
[0038] The present invention has the following beneficial effects:
[0039] (1) This invention addresses the widespread existence of intersections without traffic lights in current road scenarios by incorporating vehicles that do not support autonomous driving or are uncontrollable, as well as traffic participants such as pedestrians, bicycles, and motorcycles that cannot be ignored in actual driving scenarios into the decision-making management of intersections, which greatly improves the safety and reliability of intersection decision-making.
[0040] (2) The present invention calculates the radius of the conflict area based on the speed weight of the two, which can effectively reduce the change in the time interval of the conflict area caused by the sudden and rapid acceleration and deceleration of the conflict object when driving, and greatly reduce the impact of traffic oscillation on the decision of the controlled main vehicle at the intersection; secondly, by merging adjacent conflict areas without sufficient safety distance, it can avoid the dangerous situation of the main vehicle being stuck between adjacent conflict areas when passing through the intersection while the front or rear of the vehicle is in the conflict area.
[0041] (3) This invention comprehensively considers the impact of road conditions and uncontrollable factors on vehicle decision-making at intersections without traffic lights. By predicting the trajectory of traffic participants at the intersection in real time, generating conflict areas and calculating the occupation time, it adjusts the driving speed of the controlled main vehicle, which can ensure that it passes through intersections without traffic lights safely and efficiently without collision.
[0042] (4) The device for intelligent connected vehicles at intersections without traffic lights of the present invention can switch the hardware architecture according to whether the main vehicle supports autonomous driving, effectively reducing the computing burden of the computing unit in the intersection device. At the same time, for vehicles that support autonomous driving, it can give full play to the advantages of their on-board sensors, improve the accuracy and precision of perception, and further improve the safety of passing through the intersection. Attached Figure Description
[0043] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below.
[0044] Figure 1 This is a scene diagram of a traffic light-less intersection used in an embodiment of the present invention;
[0045] Figure 2 This is a flowchart illustrating a decision-making method for intelligent connected vehicles at intersections without traffic lights, according to an embodiment of the present invention.
[0046] Figure 3 This is a schematic diagram of the structure of an intelligent connected vehicle decision-making device for intersections without traffic lights, according to an embodiment of the present invention.
[0047] Figure 4 This is a schematic diagram of a specific example of the intelligent connected vehicle decision-making device for unsignal intersections according to the present invention, when the computing unit is a roadside device;
[0048] Figure 5 This is a schematic diagram of a specific example of a decision-making device for intelligent connected vehicles at intersections without traffic lights, according to the present invention, when the computing unit is a vehicle-side device. Detailed Implementation
[0049] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0050] This invention provides a decision-making method for intelligent connected vehicles at intersections without traffic lights, such as... Figures 1 to 5 As shown, it includes the following steps:
[0051] Step 1) Obtain road environment information at the intersection.
[0052] Road environment information specifically refers to the weather conditions of the intersection at the current moment. Based on computer vision methods such as semantic segmentation and convolutional neural networks, the weather conditions are identified, and the corresponding road surface conditions are queried. There are three types of road surface conditions: dry, wet and slippery, and icy and snowy. The adhesion coefficient for dry road surface is 0.8 to 0.9, the adhesion coefficient for wet and slippery road surface is 0.5 to 0.7, and the adhesion coefficient for icy and snowy road surface is 0.1 to 0.2.
[0053] Step 2) Obtain information on traffic participants at the intersection.
[0054] The acquisition of traffic participant information at intersections mainly relies on cameras installed at intersections without traffic lights. These cameras can be set up in a single diagonal arrangement, a double diagonal arrangement, or other forms of multiple camera setups. The purpose is to avoid the problem of low detection accuracy caused by factors such as occlusion between objects, lighting, and viewing angle when using a single camera. The traffic participant information includes the type of traffic participant, location coordinates, speed, acceleration, and direction of travel. The types of traffic participants include vehicles, pedestrians, and others (two-wheeled vehicles, three-wheeled vehicles, etc.).
[0055] Step 3) Predict the future trajectories of traffic participants.
[0056] Based on the traffic participant information detected in step 2), the future trajectories of different types of traffic participants at the intersection are predicted in real time. This requires prior perception of traffic participants, acquisition of their historical location information, generation of historical driving or movement trajectories, and combination with the geometric features of the intersection obtained in advance. On this basis, a deep learning neural network is used to predict multimodal trajectories.
[0057] Step 4) Generate the intersection traffic conflict zone. This includes the following steps:
[0058] Step 41) Generate the intersection conflict point P based on the intersection of the main vehicle's planned trajectory and the predicted future trajectories of other traffic participants obtained in Step 3). i =(x i ,y i ), where i = 1, 2, ..., N, and N is the number of intersection driving conflict points generated by the trajectories of the main vehicle and other traffic participants. It should be noted that the trajectory includes information such as position, speed, and acceleration.
[0059] Step 42) The intersection conflict point P in step 41) is used for driving. i With R as the center, iGenerate the intersection driving conflict zone δ based on the radius. i , Where K is the gain coefficient, v ego The speed of the main vehicle, v i The speeds of traffic participants whose future trajectories clash with the main vehicle's trajectory at the intersection are represented by A and B, respectively, which serve as weighting coefficients for the speeds of the main vehicle and other traffic participants, satisfying A + B = 1. In this embodiment, considering that the main vehicle is controllable while other traffic participants are not, A < B in general. In other words, the speed of other traffic participants crossing the intersection has a greater impact on the size of the intersection conflict area.
[0060] Step 43) Define the distance L = |P| between the driving conflict areas of the i-th and (i+1)-th intersections. i P i+1 | If there exists a relation L≤R i +R i+1 +s0 means that two adjacent conflicting regions can be considered as the same conflicting region. Further judgment is made if L≤R i +R i+1 , merge two adjacent but intersecting conflict regions into a single conflict region δ i , has δ i =δ i ∪δ i+1 If R i +R i+1 ≤L≤R i +R i+1 +s0, then the connected component Δδ is generated. i This connected component enables two adjacent but non-intersecting conflict regions to be connected into a single conflict region δ. i , has δ i =δ i ∪δ i+1 ∪Δδ i If L > R i +R i+1 If +s0, no action is taken. Where R... i R i+1 Let be the radii of the driving conflict zones at the i-th and (i+1)-th intersections, respectively, and s0 be the safety distance, primarily determined by the vehicle's dimensions. By iterating through all intersection driving conflict zones, the driving conflict zone δ for the vehicle is obtained. j Where j = 1, 2, ..., M, M ≤ N, and M is the final number of intersection traffic conflict zones. Step 5) Calculate the occupancy time of the conflict zone, including the time when traffic participants at the intersection who are conflicting with the main vehicle are about to enter and leave the conflict zone.
[0061] The occupancy time of a conflict zone can be represented by a time interval, T. i=[ta i ,tl i ], ta i =S i / v i +t0, tl i =ta i +S′ i / v i , where, i=1,2,...,N, ta i tl i S represents the time when the traffic participant at the intersection where the conflict with the main vehicle is about to enter and leave the conflict area, respectively. i S′ i These represent the distance from the current location of the traffic participant in the direction of travel or movement to the point where they are about to enter the conflict zone, and the distance from the point where they are about to enter the conflict zone to the point where they are about to leave the conflict zone, respectively. i t0 represents the speed at which traffic participants travel or move at the intersection, and t0 represents the time corresponding to the current moment.
[0062] For the main vehicle, T_ego j =[ta_ego j ,tl_ego j ], ta_ego j =S_ego j / v ego +t0,ta_ego j =ta_ego j +S′_ego j / v ego Where j = 1, 2, ..., M, ta_ego j tl_ego j Let S and S′ represent the times when the main vehicle is about to enter and leave the j-th conflict zone, respectively. Let S and S′ represent the distances from the main vehicle's current position to when it is about to enter the j-th conflict zone and the distances from when it is about to enter the j-th conflict zone to when it is about to leave the j-th conflict zone, respectively, along the driving direction. ego The speed of the main vehicle.
[0063] Since the traffic participant detection in step 2) usually uses a box to represent the general outline of the detection target, the definition of a traffic participant about to enter the conflict area is that the foremost point of the box corresponding to the object enters the conflict area for the first time in the direction of movement and intersects the boundary representing the conflict area; the definition of a traffic participant about to leave the conflict area is that the last point of the box corresponding to the object leaves the conflict area in the direction of movement and intersects the boundary representing the conflict area.
[0064] Step 6) Adjust the speed of the controlled master vehicle.
[0065] Based on the conflict zone occupancy time calculated in step 5), determine whether a collision event has occurred. This indicates the conflict zone δ between the main vehicle and the intersection. j A collision event will occur if the occupancy time of the first traffic participant overlaps with the occupancy time of n other traffic participants passing through the conflict zone. This indicates that there is no overlap in the time occupied, and a collision event will not occur. Step S6) Adjusting the speed to allow the vehicle to pass through the intersection is mainly divided into two stages: Stage 1) before entering the intersection, and Stage 2) after entering the intersection.
[0066] Phase 1) The main vehicle travels at a speed v ego If driving in the straight-ahead buffer zone, At this point, adjust the speed v of the main vehicle. ego ′=v ego -Δv d ,until Among them, v ego ' represents the speed of the master vehicle in the next decision cycle, Δv d The deceleration step size is adjusted accordingly; otherwise, the main vehicle continues to travel at the originally planned speed.
[0067] Phase 2) The main vehicle travels at a speed v ego When going straight or turning at an intersection, the resulting intersection driving conflict area δ j ,like Adjust the speed v of the main vehicle ego ′=v ego +Δv a ,until Where, Δv a To increase the step length; otherwise, the main vehicle continues to travel at the originally planned speed.
[0068] Speed adjustments should be made based on the road conditions of the main vehicle. The lower the coefficient of friction, the lower the planned speed. This must meet the vehicle's power and handling stability requirements while also adhering to the speed limits at the intersection, ensuring the vehicle safely passes through the intersection without traffic lights.
[0069] This invention also provides an intelligent connected vehicle decision-making system for intersections without traffic lights, such as... Figures 3 to 5 It consists of four parts: a sensing unit, a server unit, a computing unit, and a control unit.
[0070] The perception unit acquires road conditions based on computer vision, detects traffic participants at intersections, and collects information on the type, location coordinates, speed, acceleration, and direction of travel of traffic participants, transmitting the above information to the computing unit in real time.
[0071] The server unit includes a roadside server and a vehicle-side server. It realizes bidirectional transmission of vehicle-road information through wireless communication technology. The roadside server can send a takeover request to the vehicle-side server. After obtaining permission from the vehicle-side server, it can send control commands. The vehicle-side server can send the planned trajectory and vehicle information to the roadside server and can receive control commands sent by the roadside server.
[0072] The computing unit includes a prediction module and a decision module. The prediction module predicts the future trajectories of traffic participants at the intersection based on the information collected by the perception unit. The decision module generates conflict zones and calculates the occupancy time based on the predicted future trajectories and the vehicle planning trajectories received by the roadside server, optimizes the planned speed of vehicles, and generates control commands for vehicles.
[0073] The control unit receives and executes control commands to ensure the vehicle operates normally.
[0074] The roadside server in the sensing unit and server unit belongs to the roadside device, while the vehicle-side server and control unit belong to the vehicle-side device. The computing unit can be either a component of the roadside device or a component of the vehicle-side device. It should be noted that the aforementioned roadside device and vehicle-side device together constitute the intelligent connected vehicle decision-making device for traffic light-free intersections provided by this invention.
[0075] For the main vehicle, if it supports intelligent connectivity and can be controlled, the computing unit is part of the roadside device; if it supports intelligent connectivity and autonomous driving, the computing unit is part of the vehicle-side device.
[0076] When the computing unit is part of the vehicle-mounted device, it can simultaneously receive information from the sensing unit located on the roadside device and the vehicle-mounted sensing unit, that is, the environmental sensing sensors carried by the vehicle itself, such as lidar, millimeter-wave radar, and cameras, and perform sensing fusion to improve the accuracy and precision of sensing, thereby making more scientific and reasonable decisions.
[0077] It should be noted that the computing unit is part of the vehicle-side device and only refers to the hardware architecture of the decision-making device of the intelligent connected vehicle at the intersection without traffic lights provided by the present invention when the method is used for intelligent connected and autonomous driving vehicles. It does not mean that there is no computing unit in the hardware facilities of the intersection. The computing unit in the hardware facilities of the intersection always exists to help intelligent connected and controllable vehicles make decisions.
[0078] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A decision-making method for intelligent connected vehicles at intersections without traffic lights, characterized in that, Includes the following steps: Step 1: Obtain information on the road environment and traffic participants at the intersection; Step 2: Predict the future trajectory of the traffic participants based on the information obtained in Step 1; Step 3: Based on the future trajectory predicted in Step 2, generate the intersection driving conflict point between the main vehicle and other traffic participants, and generate the intersection driving conflict area with the conflict point as the center; calculate the distance between the conflict areas of adjacent intersections and choose to merge them or not. Traverse all intersection conflict zones to obtain the intersection conflict zones from the perspective of the primary vehicle; specifically including: Step 31: Generate intersection conflict points based on the intersection of the vehicle's planned trajectory and the predicted future trajectories of other traffic participants. ,in, , The number of intersection driving conflict points caused by the trajectories of the main vehicle and other traffic participants; Step 32, with With the center as the center, Generate intersection driving conflict zone with radius , ,in, This is the gain coefficient. The speed of the main vehicle, The speed of traffic participants at the intersection where their future trajectories and the main vehicle's trajectory will conflict. , These are used as weighting coefficients for the speeds of the main vehicle and other traffic participants, respectively, to satisfy... ; Step 33, define the first The and the first Distance of the traffic conflict zone at the intersection ,like If two adjacent conflict areas are considered as the same conflict area, further judgment is made. Then, two adjacent but intersecting conflict regions are merged into a single conflict region. ,like Then a connected component is generated. , ;like If no action is taken, then no action is taken. , , The first The center, radius, and driving conflict area of each intersection. To maintain a safe distance; Step 34: Traverse all intersection driving conflict areas and process them to obtain the intersection driving conflict area from the perspective of the main vehicle. ,in, , This represents the final number of intersection traffic conflict zones. Step 4: Calculate the time the conflict zone is occupied, including the time when traffic participants at the intersection who are about to enter and leave the conflict zone; determine whether a collision event has occurred based on the time the conflict zone is occupied, and adjust the speed of the controlled vehicle.
2. The intelligent connected vehicle decision-making method for intersections without traffic lights according to claim 1, characterized in that, In step 1, the weather conditions of the intersection at the current moment are identified using computer vision methods, and the road surface conditions corresponding to the current weather are queried, including dry, slippery, and icy / snowy. Cameras are set up at intersections without traffic lights to obtain information on traffic participants at the intersection. The information on traffic participants includes the type of traffic participant, location coordinates, speed, acceleration, and direction of travel. The types of traffic participants include vehicles, pedestrians, and others.
3. The intelligent connected vehicle decision-making method for intersections without traffic lights according to claim 1, characterized in that, In step 2, traffic participants are perceived in advance, their historical location information is obtained, historical driving or movement trajectories are generated, and multimodal trajectory prediction is performed in combination with the previously obtained intersection road environment information.
4. The intelligent connected vehicle decision-making method for intersections without traffic lights according to claim 1, characterized in that, In step 4, the method for calculating the occupancy time of the conflict zone is as follows: using time intervals To indicate the time the conflict zone is occupied. , , ,in, , , The first The time when traffic participants at an intersection who are about to enter or leave the conflict zone when a conflict occurs with the main vehicle. , These refer to the distance a traffic participant travels in the direction of travel or movement, from their current location to where they are about to enter the conflict zone, and from where they are about to enter the conflict zone to where they are about to leave the conflict zone. The speed at which traffic participants travel or move at intersections. This represents the time corresponding to the current moment.
5. The intelligent connected vehicle decision-making method for intersections without traffic lights according to claim 4, characterized in that, In step 4, the method for determining whether a collision event has occurred is as follows: calculate the conflict area between the main vehicle and the intersection. Occupancy time and passage through the conflict zone If the occupancy times of other traffic participants overlap, it indicates that a collision event will occur; otherwise, it indicates that a collision event will not occur.
6. The intelligent connected vehicle decision-making method for intersections without traffic lights according to claim 5, characterized in that, The time the main vehicle occupancy zone is expressed as follows: , , ,in, , , These are the main vehicle about to enter and about to leave. Time in each conflict zone , The positions of the main vehicle in the direction of travel, from its current position to the point where it is about to enter the next... The journey to the conflict zone and the approach to the... The conflict zone was about to be left. The journey through the conflict zone The speed of the main vehicle; when When, it indicates that a collision event will occur; when When this occurs, it indicates that a collision event will not occur.
7. The intelligent connected vehicle decision-making method for intersections without traffic lights according to claim 6, characterized in that, The regulation of the controlled vehicle's speed includes two phases: Phase 1) before entering the intersection, and Phase 2) after entering the intersection. Phase 1) The main vehicle travels at a speed If driving in the straight-ahead buffer zone, At this point, adjust the speed of the main vehicle. ,until ,in, For the speed of the main vehicle in the next decision cycle, To reduce the speed step size; otherwise, the main vehicle will continue to travel at the originally planned speed. Phase 2) The main vehicle travels at a speed When going straight or turning at an intersection, there is a conflict zone between the two traffic lanes. ,like Adjust the speed of the main vehicle ,until ,in, To increase the step length; otherwise, the main vehicle continues to travel at the originally planned speed.
8. A decision system for implementing the decision-making method of claim 1, characterized in that, It includes a sensing unit, a server unit, a computing unit, and a control unit; The perception unit acquires road conditions based on computer vision, detects road environment information and traffic participant information at intersections, and transmits the above information to the computing unit in real time. The server unit includes a roadside server and a vehicle-side server. It realizes bidirectional transmission of vehicle-road information through wireless communication technology. The roadside server sends a takeover request to the vehicle-side server. After obtaining permission from the vehicle-side server, it can send control commands. The vehicle-side server sends the planned trajectory and vehicle information to the roadside server and receives the control commands sent by the roadside server. The computing unit includes a prediction module and a decision module. The prediction module predicts the future trajectories of traffic participants at the intersection based on the information collected by the perception unit. The decision module generates conflict zones and calculates the occupancy time based on the predicted future trajectories and the vehicle planning trajectories received by the roadside server, optimizes the speed of vehicle planning, and generates control commands for the vehicles. The control unit receives and executes control commands to ensure the vehicle operates normally.
9. The decision-making system according to claim 8, characterized in that, The roadside server in the sensing unit and server unit belongs to the roadside device, the vehicle-side server and control unit belong to the vehicle-side device, and the computing unit is a link of the roadside device or a link of the vehicle-side device. If the main vehicle supports intelligent connectivity and can be controlled, the computing unit is part of the roadside device; if the main vehicle supports intelligent connectivity and autonomous driving, the computing unit is part of the vehicle-side device.