A V2X intersection collision warning method
By constructing driving paths with various intersection shapes and vehicle crossing patterns, and combining them with a multi-level early warning mechanism, the problem of low activation rate and accuracy of existing V2X intersection collision warning functions has been solved. This improves the applicability and accuracy of the warnings, ensures that drivers have sufficient reaction time, and enhances driving safety.
Patent Information
- Application Number
- CN202511226755.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-08-29
- Publication Date
- 2026-03-06
- Estimated Expiration
- 2045-08-29
AI Technical Summary
The existing V2X intersection collision warning function only covers perpendicular intersections, has a low activation rate, and has a lot of false alarms and missed alarms. The warning timing that relies on the time difference of arrival at the intersection is inaccurate, and the activation rate is low when relying on MAP messages.
Based on roadside MAP messages and remote vehicle perception messages, the system constructs the driving paths of the local vehicle and remote vehicles as they cross intersections, performs four types of driving trajectory prediction and three types of collision risk assessment, and provides a multi-level early warning mechanism, including L1, L2 and L3 level early warnings. It supports map-free mode and situations where remote vehicles cannot send BSM messages.
The functionality has been expanded to cover various intersection shapes and vehicle crossing patterns, improving the accuracy and activation rate of warnings, ensuring drivers have sufficient reaction time, providing multi-level warnings to address different collision risks, and enhancing driving safety.
Smart Images

Figure CN121034126B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of vehicle networking technology, specifically to a V2X intersection collision warning method. Background Technology
[0002] V2X (vehicle to everything) As an emerging communication technology, warning functions have evolved into a series of warning functions. Compared to ADAS (Advanced Driver Assistance Systems) functions, although warning functions do not have driver assistance capabilities, they play a good supplementary role to ADAS warning functions, especially in intersection scenarios, due to their over-distance warning capability.
[0003] Intersection Collision Warning (ICW) is a typical V2V (vehicle-to-vehicle) collision warning function at urban intersections. This function is of great significance in reducing side-collision injuries and fatalities at urban traffic intersections, especially those without traffic lights.
[0004] The CSAE53 standard defines typical use cases for crossing intersections. Current development of intersection collision warning functions largely follows the CSAE53 intersection collision warning function scenario definitions, which presents the following problems:
[0005] 1. This feature only covers perpendicular intersections, resulting in a low activation rate. It cannot activate the warning for other types of intersection collisions.
[0006] 2. The system only provides early warnings for vertically intersecting movements, resulting in a significant number of missed and false alarms.
[0007] 3. Using the difference in time to arrival at the intersection (TTI) as the warning activation indicator cannot provide the correct warning timing. It can only prove that there is a risk of collision between the two parties, but it cannot guarantee that sufficient driver reaction time and braking distance have been reserved when the warning occurs.
[0008] 4. Reliance on MAP messages (map messages) – that is, local map messages sent by roadside units, which can only be activated when map messages are available, but the current coverage of intersection map messages is still low. Summary of the Invention
[0009] To address at least one aspect of the aforementioned technical problems, the present invention provides a V2X intersection collision warning method.
[0010] The technical solution adopted in this invention is as follows:
[0011] A V2X intersection collision warning method is provided, including the following steps:
[0012] Step 1: Based on the roadside MAP message and / or remote vehicle perception message, construct the driving path of the vehicle and the remote vehicle through the intersection, or provide intersection information.
[0013] Step 2: Predict the driving trajectory based on the driving path:
[0014] Four types of trajectory predictions were performed for this vehicle:
[0015] Type 1: Predict the trajectory of the vehicle while maintaining its current constant speed;
[0016] Type 2: Predict the trajectory of the vehicle, which will first move at a constant speed and then decelerate.
[0017] Type 3: Predicting the vehicle will immediately travel at 2m / s 2 The trajectory of a vehicle braking by deceleration;
[0018] Type 4: Predicts the trajectory of the vehicle that will brake immediately at the minimum safe deceleration to avoid colliding with a distant vehicle;
[0019] Predicting the trajectory of a distant vehicle: Predicting the trajectory of a distant vehicle while maintaining its current constant speed;
[0020] Step 3: Conduct a three-category collision risk assessment based on the predicted driving trajectory:
[0021] Category A Collision Risk: There is a risk of collision between the travel trajectory of this vehicle (Type 1) and that of a distant vehicle.
[0022] Category B Collision Risk: There is a risk of collision between the driving trajectory of this vehicle (Type 2) and the driving trajectory of a distant vehicle;
[0023] Category C collision risk: There is a risk of collision between the driving trajectory of this vehicle (Type 3) and the driving trajectory of a distant vehicle;
[0024] Step 4, Early Warning:
[0025] The warning function is activated when there is a Class A collision risk.
[0026] If a Class B collision risk exists but a Class C collision risk does not exist, and the warning function is activated, an L1 level warning will be issued.
[0027] With the warning function activated, the minimum safe deceleration is less than 3 m / s² when there is a Class C collision risk. 2 Level 2 warning will be issued; minimum safe deceleration is greater than 3 m / s². 2 A Level 3 warning was issued.
[0028] In some implementations, in step one, the remote vehicle sensing message is: a BSM message sent by the remote vehicle, or a remote vehicle message containing remote vehicle location information shared by roadside sensing data.
[0029] In some implementations, step one, which involves constructing the driving paths of the vehicle and the remote vehicle through the intersection based on the roadside MAP message and / or the remote vehicle perception message, or providing an intersection notification message, specifically includes:
[0030] When the vehicle can receive the roadside MAP message and the remote vehicle perception message, it constructs the driving path of each vehicle in the intersection based on the MAP message and the turning intention of the vehicle and other remote vehicles entering the intersection. Each driving path is a straight line connecting the midpoint of the corresponding entrance lane and the midpoint of the exit lane. The intersection of the driving paths of the vehicle and the remote vehicle is the spatial collision point.
[0031] When the vehicle can only receive perception messages from the remote vehicle but cannot receive roadside MAP messages, the driving path constructed within the intersection is a straight line along the heading, and the intersection of the driving paths of the vehicle and the remote vehicle is the spatial collision point.
[0032] When the vehicle can only receive roadside MAP messages and cannot receive remote vehicle perception messages, it does not construct a driving path or execute subsequent steps, but instead uses the roadside MAP messages to prompt the driver with the location information of the intersection ahead.
[0033] In some implementations, the turning intention is one of the following: U-turn, left turn, straight ahead, right turn.
[0034] In some implementations, in step two, the driving trajectory is a sequence of trajectory points, and the trajectory points contain spatiotemporal information, including: the time t corresponding to the trajectory point, the position px of the trajectory point in the x-axis direction, the velocity vx of the trajectory point in the x-axis direction, the acceleration ax of the trajectory point in the x-axis direction, the position py of the trajectory point in the y-axis direction, the velocity vy of the trajectory point in the y-axis direction, the acceleration ay of the trajectory point in the y-axis direction, and the cumulative distance s from the predicted trajectory starting point to the current point.
[0035] In some implementations, the duration of all predicted driving trajectories is greater than 8 seconds.
[0036] In some implementation methods, the method for determining whether there is a risk of collision between the driving trajectory of this vehicle and the driving trajectory of a distant vehicle is as follows:
[0037] A base time series T is given as a unified time reference for the intersection analysis of driving trajectories;
[0038] Perform trajectory intersection analysis at time point t1: At time point t1, perform linear interpolation on the trajectory of the vehicle to obtain trajectory point p1, and perform linear interpolation on the trajectory of the distant vehicle to obtain trajectory point p2; construct a risk region A1 for trajectory point p1 of the vehicle; determine whether trajectory point p2 of the distant vehicle is within the range of A1. If it is, a collision risk is determined; if not, continue with the trajectory intersection analysis at the next time point t2 until the trajectory exceeds the range.
[0039] In some implementations, constructing a risk region A1 for the vehicle's trajectory point p1 specifically includes:
[0040] Construct a square centered at the trajectory point p1 of the vehicle. The area inside the square is the risk area A1, and the side length of the square is L. When the current speed of the vehicle × 1 second > 8.4m, L = current speed of the vehicle × 1 second × 2; when the current speed of the vehicle × 1 second < 8.4m, L = 16.8m.
[0041] In some implementations, when an L1 warning is issued, the driver is reminded to prepare to brake; when an L2 warning is issued, the driver is reminded to immediately intervene in braking; and when an L3 warning is issued, the driver is reminded to brake urgently.
[0042] In some implementations, step four further includes: when issuing L1-L3 level warnings, if the driver accelerates, the warning function is deactivated and activated; then, if the driver cancels the acceleration and remains unaccelerated for more than 1 second, the warning function is reactivated.
[0043] Compared with the prior art, the beneficial effects of the present invention are as follows:
[0044] When vehicles pass through an intersection, they collide, which can be categorized into various forms based on the shape of the intersection. This patent expands upon the intersection shapes in CSAE53, including not only orthogonal crossroads but also oblique crossroads, orthogonal T-junctions, and oblique T-junctions, thus relaxing the activation conditions for the function.
[0045] Cross-traffic collisions when vehicles pass through intersections can be categorized into various forms based on the vehicles' travel trajectories. This patent expands the collision types beyond CSAE53, including not only collisions between vehicles traveling straight and crossing each other, but also collisions between vehicles turning left and oncoming vehicles traveling straight, and collisions between vehicles traveling straight and oncoming vehicles turning left—all scenarios involving cross-traffic collisions, significantly broadening the scope of functional activation.
[0046] Considering the current low coverage of intersection map messages, this patent supports a map-free mode, requiring only BSM messages between the local vehicle and distant vehicles without MAP messages, thus improving the activation rate of the function in map-free situations. In map-enabled scenarios, the map is fully utilized to improve the accuracy of trajectory prediction and warnings, thereby enhancing functional performance.
[0047] Considering that some vehicles are not equipped with OBU and therefore do not have the ability to send BSM, in order to reduce the risk of collisions at potential intersections ahead, this patent adds a sub-function to remind drivers of the presence of an intersection ahead. After receiving the reminder, drivers can reduce their speed and observe the surroundings to improve safety when passing through potential intersections and avoid accidents caused by high-speed passage.
[0048] To address the diverse needs of drivers and improve intersection safety, the system should ensure that the warning level is increased when a driver enters the final safe distance after receiving the initial warning but failing to respond. This patent proposes a graded warning scheme with multiple warning levels to meet this requirement. Attached Figure Description
[0049] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.
[0050] Figure 1 A schematic diagram of a vehicle path model at a crossroads in one embodiment of the present invention is shown;
[0051] Figure 2 A schematic diagram of the collision risk area construction in one embodiment of the present invention is shown;
[0052] Figure 3 A schematic diagram of the state transition logic of the early warning function in one embodiment of the present invention is shown. Detailed Implementation
[0053] The above-mentioned and other technical features and advantages of the present invention will be described in more detail below with reference to the accompanying drawings.
[0054] The term “comprising” and its variations as used herein signify open inclusion, i.e., “including but not limited to”. Unless otherwise stated, the term “or” means “and / or”. The term “based on” means “at least partially based on”. The terms “one example embodiment” and “one embodiment” mean “at least one example embodiment”. The term “another embodiment” means “at least one additional embodiment”. Other explicit and implicit definitions may also be included below.
[0055] This invention designs a collision warning function for intersections based on the following ideas:
[0056] Premise: This invention is designed for urban intersections and does not mandate the use of roadside MAP messages (map messages – local map messages sent by roadside units, RSUs) or remote vehicle perception messages (primarily containing remote vehicle location information). Remote vehicle perception messages can be obtained via BSM messages (Basic Safety Messages – shared between vehicles) or through roadside perception data sharing; this solution does not limit the method of obtaining target vehicle information. At intersections with well-deployed roadside perception equipment, even if the remote vehicle does not have an OBU (i.e., cannot send BSM messages), the function will still function normally. When remote vehicle perception messages are received but roadside MAP messages (primarily containing lane-level structured information within the relevant area) are not, this function can still be activated, but false alarms and missed alarms may occur. If both roadside MAP messages and remote vehicle perception messages are received simultaneously, the function performs optimally. If only roadside MAP messages are provided without remote vehicle perception messages, this function can only provide limited intersection location hints to reduce risk but cannot provide accurate warnings.
[0057] Predicting the trajectories of the local and remote vehicles at the intersection is crucial for the functional warning system. Based on the roadside MAP message and steering intention, the system predicts the vehicle's travel path after entering the intersection, thereby distinguishing certain collision scenarios.
[0058] When there is no map (no roadside MAP messages can be received), the trajectory prediction for both the local and remote vehicles uses the heading as the reference path. To reduce false alarms, the warning function provides an option to activate the warning only when a map is available. When a map is available (roadside MAP messages can be received), the trajectory prediction for both the local and remote vehicles uses the road topology as the reference path: for the road segment before the intersection, the center line of the lane is used as the reference path; for the subsequent road segment after the intersection, the center line of the segment is used as the reference path; within the intersection, the straight line connecting the endpoints of the above two reference paths is used as the reference path (the constructed driving path).
[0059] The vehicle trajectory prediction is divided into four types: prediction based on uniform speed motion (TTC model), prediction based on uniform speed followed by braking (CSAE braking model), prediction based on immediate standard braking (standard braking model), and prediction based on immediate safe braking (safe braking model). The distant vehicle trajectory is predicted based on the current vehicle speed uniform speed motion (TTC model).
[0060] Based on the collision risks of the predicted trajectories of the vehicle and the distant vehicle, the risk level is assessed and the function output is performed. First, it is determined whether the TTC model trajectories of the vehicle and the distant vehicle intersect, which is a prerequisite for intersection collision warning. Under this premise: when there is only a collision risk in the CSAE braking model, a Level 1 warning is issued (reminding the driver to prepare to brake); when there is a collision risk in the standard braking model and the required deceleration does not exceed 3 m / s², a Level 1 warning is issued. 2 When the required deceleration is greater than 3 m / s², a Level 2 warning is issued (requesting the driver to apply the brakes immediately); 2 At that time, a Level 3 warning was issued (requesting the driver to apply emergency braking).
[0061] When the collision warning function is activated, if the driver accelerates, the function will enter a state of suppression until the driver stops accelerating, at which point it will exit the state. The function is designed with three sub-functions: advance warning, request for braking, and advance intersection warning. Options are provided: the driver can set the warning sensitivity. When set to high sensitivity, the warning time is earlier, allowing more braking distance, but false alarms are more likely. Conversely, when set to low sensitivity, the warning time is delayed, requiring more urgent braking, and the interference caused by false alarms is reduced.
[0062] In one specific embodiment, the V2X intersection collision warning method of the present invention includes:
[0063] 1. Based on roadside MAP messages and / or remote vehicle perception messages, construct the driving path of the vehicle and remote vehicles through the intersection, or provide intersection information.
[0064] See Figure 1 , Figure 1 The present invention uses a typical crossroads scenario to illustrate the vehicle path model within the intersection, but the solution of the present invention is not limited to crossroads. It can be an orthogonal crossroads, an oblique crossroads, an orthogonal T-junction, or an oblique T-junction.
[0065] like Figure 1 As shown. When this vehicle enters an intersection from a certain direction, other vehicles may enter the intersection from three directions. For a T-junction, vehicles may enter from a maximum of two directions.
[0066] After entering the intersection, the possible turning intentions of this vehicle are: U-turn, left turn, straight ahead, and right turn.
[0067] After the vehicle on the other side (to the left of this vehicle) enters the intersection, there are a total of 7 intersection points where it may collide with this vehicle. The corresponding turning intentions of the other vehicle are: U-turn (1 intersection point), left turn (2 intersection points), straight (3 intersection points), and right turn (1 intersection point).
[0068] After the vehicle to the right of this vehicle enters the intersection, there are a total of 7 intersection points where it may collide with this vehicle. The corresponding turning intentions of the vehicle to the other vehicle are: U-turn (1 intersection point), left turn (3 intersection points), straight (2 intersection points), and right turn (1 intersection point).
[0069] When a vehicle on the right (going in the opposite direction) enters the intersection, there are a total of 6 intersection points where it may collide with your vehicle. The corresponding turning intentions of the vehicle on the right are: U-turn (1 intersection point), left turn (2 intersection points), straight (2 intersection points), and right turn (1 intersection point).
[0070] In summary, there are a total of 20 possible collision points (intersections) between the vehicle and the distant vehicle at an intersection, covering all collision scenarios between the vehicle and the distant vehicle in intersection scenarios. This includes all scenarios of intersection collisions and left-turn assist collisions in CSAE53, as well as all collision scenarios in orthogonal intersections, oblique intersections, orthogonal T-junctions, and oblique T-junctions.
[0071] In reality, vehicles travel in smooth arcs within intersections. This scheme assumes that vehicles travel along point-to-point straight lines within intersections, which does not affect the determination of the number of intersections and can correctly distinguish whether driving paths intersect under different turning intentions. This ensures the accuracy of subsequent collision judgments based on driving trajectory. Although this assumption also reduces the accuracy of driving trajectory (within an acceptable range), it also brings the benefit of reducing the difficulty of subsequent driving trajectory algorithms.
[0072] When the vehicle can receive the roadside MAP message and the remote vehicle perception message, it constructs the driving path of each vehicle in the intersection based on the MAP message and the turning intention of the vehicle and other remote vehicles entering the intersection. Each driving path is a straight line connecting the midpoint of the corresponding entrance lane and the midpoint of the exit lane. The intersection of the driving paths of the vehicle and the remote vehicle is the spatial collision point.
[0073] When the vehicle can only receive perception messages from the remote vehicle but not MAP messages from the roadside, it cannot determine the driving path by using steering intentions and map information. It can only assume that when the vehicle and the remote vehicle enter the intersection, they are going straight along the heading direction. At this time, only cross-traffic collisions can be detected, with a total of 2 collision points.
[0074] When the vehicle can only receive roadside MAP messages and cannot receive remote vehicle perception messages, it cannot construct the driving path of the remote vehicle and can only provide limited intersection location prompts to reduce risks but cannot provide collision warnings.
[0075] 2. Predict the driving trajectory based on the driving path.
[0076] Based on the driving paths constructed within the above intersections, a predicted driving trajectory is further generated, with spatiotemporal information added to the trajectory. How specifically the predicted driving trajectory is generated based on the driving paths is not the core focus of this solution; the specific calculation method is not provided in this embodiment. Various driving trajectory prediction schemes already exist in the prior art, and those skilled in the art can design their own based on considerations such as warning accuracy. However, to meet the requirements of subsequent solutions in this patent, the following general requirements apply to driving trajectory prediction:
[0077] 1. All trajectory prediction times should be greater than 8 seconds. When approaching an intersection, the next road segment after exiting the intersection should be determined based on the turning intention and map information (roadside MAP message). The reference path should be able to cover the exit point of the intersection.
[0078] 2. The trajectory should be a sequence of trajectory points, with the trajectory point format as follows: (t, px, py, vx, vy, ax, ay, s), where the x-axis and y-axis directions are set according to map information. t is the time corresponding to the trajectory point, px is the position of the trajectory point along the x-axis, vx is the velocity of the trajectory point along the x-axis, ax is the acceleration of the trajectory point along the x-axis, py is the position of the trajectory point along the y-axis, vy is the velocity of the trajectory point along the y-axis, ay is the acceleration of the trajectory point along the y-axis, and s is the cumulative distance from the predicted trajectory starting point to the current point.
[0079] The vehicle's trajectory prediction is divided into four types:
[0080] Type 1: Predicted trajectory of the vehicle maintaining its current speed: This trajectory predicts the driver's movement while maintaining the current speed, used to detect the risk of a cross-traffic collision.
[0081] Type 2: Predicts the vehicle's trajectory of first moving at a constant speed and then decelerating: This trajectory predicts the driver's braking process after a delay in reaction time upon receiving the warning signal, and is used to determine the timing of the warning. Moving at a constant speed first means maintaining the current vehicle speed, and decelerating means braking at a certain deceleration. As for the duration of the constant speed movement and the value of the deceleration, those skilled in the art can set them according to the consideration of warning sensitivity, such as whether to reserve more braking distance.
[0082] Type 3: Predicting the vehicle will immediately travel at 2m / s 2 The braking trajectory of the vehicle at a deceleration of 2 m / s²: This trajectory predicts the vehicle's immediate braking speed. 2 The braking process during braking is used to distinguish different levels of warning.
[0083] Type 4: Predicts the driving trajectory of the vehicle when it immediately brakes at the minimum safe deceleration to ensure no collision with the distant vehicle; this trajectory predicts the braking process when the vehicle immediately brakes at the deceleration to ensure no collision. The magnitude of the deceleration is determined by solving the collision risk of the vehicle's trajectory and the intersection point. The predicted trajectory corresponding to the deceleration when no collision occurs is used as the safe braking trajectory, and the corresponding deceleration is used to distinguish different levels of warning.
[0084] Trajectory prediction for distant vehicles: Assuming the distant vehicle maintains its current speed, trajectory prediction is performed.
[0085] Third, conduct a three-category collision risk assessment based on the predicted driving trajectory.
[0086] For intersection collision warning functionality, a collision risk assessment between the vehicle and a distant vehicle is required to determine if there is a spatial or temporal intersection point between them. Based on this, a collision risk assessment between the vehicle and the intersection point is then necessary to determine the timing and severity of the warning.
[0087] Risk assessment algorithm steps:
[0088] 1. Inputs required for the algorithm: trajectory of this vehicle, trajectory of distant vehicles or intersections, and parameters of the vehicle's risk area.
[0089] 2. A base time series T is given as the unified time reference for trajectory analysis.
[0090] 3. At time point t under time T, perform linear interpolation on the trajectory of this vehicle to obtain p1;
[0091] 4. At time point t of time T, perform linear interpolation on the trajectory of the distant vehicle or the intersection to obtain p2;
[0092] 5. Construct a risk area A1 for the trajectory point p1 of this vehicle;
[0093] like Figure 2 As shown, the risk zone setting for this vehicle is as follows: the risk zone is rectangular, preferably a square with all four sides of equal width, and the center of the square is the geometric center point of the vehicle. By setting the risk zone, collisions from various angles at intersections can be identified. Considering sufficient safety for the driver, the radius of the rectangle (i.e., half the side length) should reflect the time distance. The rectangle radius is set to: vehicle speed * 1 second. At low speeds, the rectangle radius may be too small, so a minimum safe distance R_min needs to be determined as the minimum rectangle radius: based on a safe distance of 5m between the two vehicles, plus half the vehicle length and half the vehicle width, R_min is determined to be 8.4m, meaning the rectangle radius should not be less than 8.4m.
[0094] 6. Determine if the trajectory point p2 of the distant vehicle is within the range of A1. If so, determine a trajectory collision and obtain the collision point time, distance, and remaining distance. For the risk area setting of distant vehicles or intersections: set it as a point.
[0095] 7. If not, continue to check the intersection of the trajectories at the next time point t until t exceeds the range of a certain trajectory and no collision is found, then return to no collision risk.
[0096] Risk assessment results:
[0097] This vehicle has multiple predicted driving trajectories, therefore there are multiple corresponding risk assessments:
[0098] Class A collision risk assessment: There is a risk of collision between the driving trajectory of this vehicle (Type 1) and the driving trajectory of the distant vehicle.
[0099] Input information: The trajectory of this vehicle moving at a constant speed, and the trajectory of the distant vehicle moving at a constant speed.
[0100] This assessment indicates whether there is a risk of a collision with a vehicle traveling at the current speed. This collision risk is a necessary condition for a collision warning. The point of collision between this vehicle and the vehicle traveling at the distance is considered the intersection point.
[0101] Category B Collision Risk Assessment: There is a risk of collision between the driving trajectory of this vehicle (Type 2) and the driving trajectory of the distant vehicle.
[0102] Input information: The trajectory of this vehicle's motion, from constant speed to deceleration, and the trajectory at the intersection point.
[0103] The assessment results indicate whether the driver has sufficient distance to stop the vehicle when applying standard braking after receiving a warning.
[0104] Class C Collision Risk Assessment: There is a risk of collision between the driving trajectory of this vehicle (Type 3) and the driving trajectory of the distant vehicle.
[0105] Input information: The vehicle's immediate standard deceleration trajectory, and the intersection trajectory.
[0106] The assessment results indicate whether the vehicle has sufficient distance to stop immediately when braking at standard deceleration.
[0107] Category D assessment:
[0108] Input information: The trajectory of this vehicle's braking motion at a certain safe deceleration, and the trajectory at the intersection point.
[0109] The assessment results indicate the minimum safe deceleration and trajectory required for the vehicle to avoid a cross-traffic collision.
[0110] IV. Early Warning: Risk Classification and Functional Output.
[0111] Risk classification and functional output of this vehicle:
[0112] 1. When the above-mentioned Category A risks exist, this vehicle is at risk of cross-collision, which is a prerequisite for function activation.
[0113] 2. If there is a Class A risk, and there is a Class B risk as described above, but no Class C risk, the classification is L1. In this case, the driver should be reminded to prepare to brake.
[0114] 3. Given the presence of Category A risks, when the aforementioned Category C risks also exist, and the minimum safe deceleration required by the safe deceleration model is less than 3 m / s², 2 The level is L2, at which point the driver should be asked to immediately apply the brakes.
[0115] 4. Given the presence of Category A risks, when the aforementioned Category C risks also exist, and the minimum safe deceleration required by the safe deceleration model is greater than 3 m / s², 2 The classification is L3, at which point the driver should be asked to apply emergency braking.
[0116] refer to Figure 3 The main functional states of the V2X intersection collision warning function state machine of the present invention are described as follows:
[0117] ON: The warning function is enabled but not activated; it is in a waiting state.
[0118] ACTIVE1: Sub-function 1 is activated, corresponding to the early warning at the L1 level mentioned above;
[0119] ACTIVE2: Sub-function 2 is activated, corresponding to the above-mentioned L2~L3 level request braking warning;
[0120] PASSIVE: The alert function does not need to be activated at certain times;
[0121] ACTIVE3: Sub-function 3 activated, providing an alert for upcoming intersections.
[0122] Main state machine transition conditions:
[0123] Jump 12: ON jump to ACTIVE1. Sub-function 1 is activated when all of the following conditions are met: Vehicle speed 5-70 km / h; Class A collision risk exists; Class B collision risk exists but there is no Class C collision risk.
[0124] Jump 13: ON jump to ACTIVE2. Sub-function 2 is activated when all of the following conditions are met simultaneously: Vehicle speed 5~70kmph; There is a Class A collision risk; There is a Class C collision risk.
[0125] Jump 15: ACTIVE to PASSIVE. The driver's acceleration behavior warning function is not activated, and the entry function is suppressed.
[0126] Jump 16: PASSIVE to ON. If the driver does not accelerate (acceleration is canceled) and the suppression state lasts for more than 1 second, the warning function will be reactivated.
[0127] Jump 17: ON jump to ACTIVE3. Sub-function 3 is activated when all of the following conditions are met: vehicle speed 5~70kmph; intersection within 100m ahead (determined by MAP message).
[0128] The intersection collision warning function proposed in this invention has the following advantages:
[0129] 1. This solution can adapt to intersections of various shapes, thus expanding its functional applicability.
[0130] 2. This solution can accurately analyze collisions of various trajectory intersections based on steering intentions, thus improving the accuracy of early warnings.
[0131] 3. This plan adopts a conservative strategy to ensure that, in the worst-case scenario (when the vehicle and the distant vehicle cannot stagger their passage times at the intersection), the vehicle can eventually stop before the intersection, thereby avoiding spatial overlap.
[0132] 4. This solution predicts multiple trajectories for the vehicle, including a risk analysis of the trajectory (first constant speed then constant deceleration) and the intersection trajectory to determine the timing of the initial warning. This is a safety distance-based warning model, which is safer than the TTC warning model because for functions requiring absolute stopping power, such as cross-traffic collisions, simply allowing a certain amount of TTC time is insufficient to guarantee safety. Furthermore, the TTC warning model requires a method to determine the TTC limit. The safety distance-based warning model also allows for setting the duration of the constant speed movement, thus supporting personalized settings for different drivers.
[0133] 5. The function supports different levels of output (divided into sub-function 1 and sub-function 2) according to the degree of collision risk, giving the driver a certain reaction time and different levels of urgency indication, thus improving the user experience.
[0134] 6. The function can be activated even when a MAP is not available, or when a remote vehicle cannot send BSM, thus improving its practicality and usability.
[0135] The above are merely preferred embodiments of the present invention and are illustrative rather than restrictive. The structure and connection methods of the components in the present invention can be varied, and any equivalent transformations and improvements made based on the technical solution of the present invention should not be excluded from the protection scope of the present invention.
Claims
1. A V2X intersection collision warning method, characterized in that, The method comprises the following steps: Step one, based on the road end MAP message and / or the remote vehicle perception message, the driving path of the vehicle and the remote vehicle crossing the intersection is constructed, or the intersection information is prompted; Step two, based on the driving path, the driving trajectory is predicted: Four types of driving trajectory prediction are performed on the vehicle: Type 1: Predict the driving trajectory of the vehicle maintaining the current vehicle speed at a constant speed; Type 2: Predict the driving trajectory of the vehicle first moving at a constant speed and then decelerating; Type 3: Predicts a travel trajectory in which the host vehicle immediately brakes at a deceleration of 2 m / s 2 . Type 4: Predict the driving trajectory of the vehicle immediately braking at the minimum safe deceleration to ensure no collision with the remote vehicle; The driving trajectory of the remote vehicle is predicted: the driving trajectory of the remote vehicle maintaining the current vehicle speed at a constant speed is predicted; Step three, three types of collision risk assessment are performed according to the predicted driving trajectory: A type of collision risk: the driving trajectory of the vehicle type 1 has a collision risk with the driving trajectory of the remote vehicle; B type of collision risk: the driving trajectory of the vehicle type 2 has a collision risk with the driving trajectory of the remote vehicle; C type of collision risk: the driving trajectory of the vehicle type 3 has a collision risk with the driving trajectory of the remote vehicle; Step four, warning: When there is A type of collision risk, the warning function is activated; Under the premise that the warning function is activated, when there is B type of collision risk but no C type of collision risk, L1 level warning is performed. When the pre-warning function is activated, the minimum safety deceleration is less than 3 m / s when there is a C-class collision risk 2 , and an L2-level pre-warning is performed Minimum safe deceleration greater than 3 m / s 2 L3 level warning is given. 2.The V2X intersection collision warning method of claim 1, wherein, In step one, the remote vehicle perception message is: the BSM message sent by the remote vehicle, or the remote vehicle message containing the remote vehicle position information shared by the roadside perception data.
3. The V2X intersection collision warning method of claim 2, wherein, In step one, the road end MAP message and / or the remote vehicle perception message is used to construct the driving path of the vehicle and the remote vehicle crossing the intersection, or to prompt the intersection information, which specifically comprises: When the vehicle can receive the road end MAP message and the remote vehicle perception message, the driving path of each vehicle in the intersection is constructed based on the MAP message and the turning intention of the remote vehicle entering the intersection from the other entrance, each driving path is a straight line connecting the midpoint of the corresponding entrance lane and the midpoint of the exit lane, and the intersection point of the driving paths of the vehicle and the remote vehicle is a spatial collision point; When the vehicle can only receive the remote vehicle perception message and cannot receive the road end MAP message, the driving path constructed in the intersection is a straight line along the heading, and the intersection point of the driving paths of the vehicle and the remote vehicle is a spatial collision point; When the vehicle can only receive the road end MAP message and cannot receive the remote vehicle perception message, the driving path is not constructed and the following steps are not performed, and the position information of the intersection in front of the vehicle is prompted based on the road end MAP message.
4. The V2X intersection collision warning method of claim 3, wherein, The turning intention is one of the following: U-turn, left turn, straight driving, and right turn.
5. The V2X intersection collision warning method of claim 1, wherein, In step two, the driving trajectory is a sequence of trajectory points, and the trajectory points contain space-time information, including: the time t corresponding to the trajectory point, the position px of the trajectory point in the x-axis direction, the velocity vx of the trajectory point in the x-axis direction, the acceleration ax of the trajectory point in the x-axis direction, the position py of the trajectory point in the y-axis direction, the velocity vy of the trajectory point in the y-axis direction, the acceleration ay of the trajectory point in the y-axis direction, and the cumulative distance s from the starting point to the current point of the predicted trajectory.
6. The V2X intersection collision warning method of claim 5, wherein, In step two, the length of all predicted driving trajectories is greater than 8 seconds.
7. The V2X intersection collision warning method of claim 1, wherein, In step three, the judgment method for whether the driving trajectory of the vehicle has a collision risk with the driving trajectory of the remote vehicle is: Given a reference time sequence T as a unified time reference for the driving trajectory intersection analysis; Performing driving trajectory intersection analysis at time point t1: at time point t1 of T, linear interpolation is performed on the driving trajectory of the host vehicle to obtain host vehicle trajectory point p1, and linear interpolation is performed on the driving trajectory of the remote vehicle to obtain remote vehicle trajectory point p2; a risk area A1 is constructed for the host vehicle trajectory point p1; it is judged whether the remote vehicle trajectory point p2 is within the range of A1, if yes, it is determined that there is a collision risk; if not, the driving trajectory intersection analysis at the next time point t2 is continued until the driving trajectory range is exceeded.
8. The V2X intersection collision warning method of claim 7, wherein, The construction of the risk area A1 for the host vehicle trajectory point p1 specifically includes: A square centered on the host vehicle trajectory point p1 is constructed, the area within the square is the risk area A1, and the side length of the square is L; when the current speed of the host vehicle × 1 second > 8.4 m, L = the current speed of the host vehicle × 1 second × 2; when the current speed of the host vehicle × 1 second < 8.4 m, L = 16.8 m.
9. The V2X intersection collision warning method of claim 1, wherein, When performing L1 level warning, the driver is reminded to prepare for braking; when performing L2 level warning, the driver is reminded to immediately intervene in braking; When performing L3 level warning, the driver is reminded to brake urgently.
10. The V2X intersection collision warning method of claim 1, wherein, In step four, it further includes: when performing L1-L3 level warning, if the driver has an acceleration behavior, the warning function is closed and activated, and then if the driver cancels the acceleration behavior and remains without acceleration for more than 1 second, the warning function is reactivated.
Citation Information
Patent Citations
Braking process-based control method and device considering adhesion coefficient and storage medium
CN116620282A
Collision early warning method, vehicle-mounted terminal and storage medium
CN116935693A