Traffic state judgment method and system based on intersection traffic flow and vehicle

Abstract

The invention provides a traffic state judgment method and system based on intersection traffic flow and a vehicle. The method comprises the following steps: acquiring own vehicle information and establishing an intersection coordinate system; projecting the traffic flow to a projection axis of the coordinate system, and calculating a projection sum in each direction; and calculating the confidence coefficient of the current passing state according to the projection sum of the passing direction of the own vehicle and the total vehicle projection sum. And performing mutual exclusion judgment according to the traffic light identification result and the current traffic state confidence so as to perform decision execution. Correspondingly, the invention further provides a system and a vehicle carrying the system. A more accurate result can be provided by predicting the state of the traffic light through dynamic analysis of the traffic flow, and the prediction method based on traffic flow perception not only improves the perception accuracy, but also provides an effective supplementary means for the traffic light identification module, does not depend on a visual identification result any more, and can adapt to a networking area which is not upgraded.

Description

Technical Field

[0001] The present invention relates to the field of driverless technology, and particularly relates to a method, a system and a vehicle for judging the passing state based on intersection traffic flow. Background Art

[0002] During the research and development of driverless vehicles, the traffic light perception module is an important component in operation, which can identify the traffic lights at intersections and thereby achieve autonomous judgment of passing through intersections. At the same time, this module also needs to take into account the passing efficiency.

[0003] Currently, there is a large amount of research content in the field of traffic light control on the road traffic side. For example, Chinese Patent Application CN114419906A discloses an intelligent traffic implementation method, device, and storage medium based on big data. It collects vehicle images on the road, identifies whether the image is the front part or the rear part of the vehicle according to the vehicle image, and judges the traffic flow direction according to the recognition result; detects the traffic flow volume on the horizontal lane close to the traffic light intersection and the vertical lane intersecting with the horizontal lane; calculates the time required for vehicles on the horizontal lane and the vertical lane to pass through the traffic light intersection respectively; and distributes the traffic light cycle of the horizontal lane and the traffic light cycle of the vertical lane according to the time required for vehicles on the horizontal lane and the vertical lane to pass through the traffic light intersection. However, on the vehicle side, the recognition of traffic lights still mainly relies on visual recognition, and a small part realizes online information acquisition through an improved road traffic Internet of Things.

[0004] However, despite multiple optimizations and improvements, the existing traffic light perception modules still face numerous technical challenges, and their in-vehicle systems still exhibit unstable performance under different environmental conditions. As previously mentioned, Chinese Patent Application CN112053578A discloses a traffic light warning method and its detection device based on machine vision technology, which uses machine vision technology to determine the traffic light state of the current traffic light image information. Although it can be used for driverless vehicles, in complex lighting environments, the perception module may misjudge the signal light state due to strong sunlight or low light. At the same time, the dirt on the vehicle lens will also affect the accuracy of traffic light recognition. In addition, in rainy, foggy weather or when there are objects blocking, the accuracy of the module drops significantly, and there may even be cases where the signal light recognition fails. In the scenario of obtaining the signal light state online through the Internet of Things, such as Chinese Patent CN114519936B discloses a method and system for intelligent control of engine start and stop at traffic light intersections based on V2X, which uses the V2X on-vehicle unit of 5G to receive real-time road environment information sent from the cloud to the vehicle end and can be used for obtaining traffic light information. Especially in some areas, the online function has been realized through this network. However, its main problem lies in the unstable network connection, and the current application coverage is relatively small, and it is necessary to synchronously upgrade traffic facilities and in-vehicle devices. For the problem that most areas lack the application of V2X technology, it still cannot be solved. Therefore, under the condition of not relying entirely on vision and the Internet of Things, there has also emerged a technology that relies on artificial intelligence for auxiliary judgment. For example, Chinese Patent Application CN119091612A discloses a method for estimating traffic light cycles based on machine learning. The method for estimating traffic light cycles based on machine learning can accurately estimate the traffic light cycles of signal lights and handle cycle changes by comprehensively analyzing vehicle driving trajectory data. This method can detect changes in traffic light cycles and use Fourier transform and CUSUM methods to identify the moment of cycle switching and the parameters of the new and old cycles. Although the above technical solution is used to understand the change cycle of traffic lights not connected to the network, it can still provide a relatively accurate basis for judging the signal light state for vehicles in the form of a wireless network. However, it should be noted that if this solution is used on the vehicle side, it is obviously in conflict with the current mainstream technology of adjusting signal lights based on artificial intelligence (as described above), and thus cannot provide a reliable judgment result.

[0005] Therefore, these above problems seriously affect the traffic efficiency and safety of vehicles and become a key obstacle to the reliable operation of driverless vehicles.

[0006] In summary, it is still necessary to develop a fallback algorithm that is not easily interfered by natural factors and can autonomously judge according to traffic scenarios, so as to provide basic-level signal light recognition guarantee when the machine learning model misjudges or fails. Summary of the Invention

[0007] In view of the problems existing in the prior art, the present invention provides an effective and fast method for accurately judging the traffic light state.

[0008] To achieve the above object, the technical solution adopted by the present invention is as follows: On the one hand, the present invention provides a method for judging the passing state based on the traffic flow at an intersection, which includes the following steps: Obtain the vehicle information of the self-vehicle, establish a coordinate system of the current intersection based on the vehicle information of the self-vehicle, and multiple vehicle operation projection axes are provided in the coordinate system; Calculate the sum of vehicle projections on each vehicle operation projection axis by the projection method respectively; Calculate the sum of vehicle projections of vehicles running in the same direction, and the sum of vehicle projections of vehicles running in the same direction is the sum of the vehicle projections on the vehicle operation projection axes in the same direction as the passable direction of the self-vehicle; Calculate the confidence level of the current passing state, and the confidence level of the current passing state is the ratio of the sum of vehicle projections of vehicles running in the same direction to the total sum of vehicle projections in the coordinate system; the total sum of vehicle projections is the sum of the vehicle projections on each vehicle operation projection axis; Based on the confidence level of the current passing state and the recognition result of the traffic lights at the current intersection, perform passing decision execution through mutual exclusion judgment.

[0009] Optionally, the passing decision execution through mutual exclusion judgment includes: Read the recognition result of the traffic lights at the current intersection. When the recognition result is a green light, verify the confidence level of the current passing state: If the confidence level of the current passing state is greater than or equal to the passing threshold, allow passing; If the confidence level of the current passing state is less than the threshold, trigger a stop command and report data to request manual takeover; Read the recognition result of the traffic lights at the current intersection. When the recognition result is recognition failure or the recognition confidence level is lower than the preset threshold, verify the confidence level of the current passing state: If the confidence level of the current passing state is greater than or equal to the passing threshold, pass according to the green light state; If the confidence level of the current passing state is less than the threshold, stop according to the red light state.

[0010] Optionally, the vehicle information of the self-vehicle includes the position of the self-vehicle, the orientation of the self-vehicle, and the driving state of the self-vehicle; the driving state of the self-vehicle at least includes the speed of the self-vehicle; The establishment of the coordinate system of the current intersection includes the following steps: Take the lane direction closest to the orientation of the self-vehicle as the positive direction of the Y-axis, and select the lane direction with the largest included angle with the Y-axis as the positive direction of the X-axis; Take the center of the intersection point of the straight-through lane dividing lines at the current intersection or the center of the set of lane intersection points as the center point of the current intersection; Set a plurality of the vehicle running projection axes to adapt to the positive and negative directions of the X-axis and Y-axis, as well as the directions of the angular bisectors of each quadrant.

[0011] Optionally, before calculating the vehicle projections on each vehicle running projection axis, perform data cleaning on the observable vehicles at the intersection, including the following steps: Eliminate vehicles with low reference value based on the running parameters of the observable vehicles at the intersection in the coordinate system of the current intersection; The running parameters of the observable vehicles at the intersection include the vehicle speed, position quadrant, speed orientation, and distance from the intersection center of the observable vehicles at the intersection; The vehicles with low reference value include right-turn vehicles, U-turn vehicles, and edge slow vehicles.

[0012] Optionally, eliminating vehicles with low reference value includes the following steps: For a vehicle whose distance from the intersection center exceeds 1 / 3 of the intersection radius and the quadrant where the speed orientation is located is the vehicle position quadrant plus 1, it is determined as the right-turn vehicle; For a vehicle whose distance from the intersection center exceeds 1 / 2 of the intersection radius and the speed orientation is opposite to the straight-ahead direction of the lane where it is located, it is determined as the U-turn vehicle; For a vehicle whose vehicle speed is lower than the preset speed threshold and is located at the intersection boundary, it is the edge slow vehicle.

[0013] Optionally, calculating the vehicle projection sum includes the following steps: Project the speed vector of the vehicle onto the nearest vehicle running projection axis, and calculate the projection value by taking the absolute value; add the projection values on the vehicle running projection axis to obtain the vehicle projection sum; For a left-turn critical state vehicle, divide the projection value of the left-turn critical state vehicle equally between the straight-ahead axis and the left-turn axis of the left-turn critical state vehicle; For the lane direction that does not exist at a T-shaped intersection, reclassify the vehicles belonging to the corresponding vehicle running projection axis to the adjacent vehicle running projection axis.

[0014] Optionally, calculating the confidence level of the current traffic state further includes the following steps: Calculate the vehicle projection sum of vehicles running in the same direction; the addends of the vehicle projection sum of vehicles running in the same direction include the projection sum in the direction of the self-vehicle and the projection sum in the co-existing direction; The co-existing direction is the straight-ahead vehicle direction and the left-turn vehicle direction allowed to pass in the same traffic light phase.

[0015] Optionally, the method further includes the following steps: periodically calculating a real-time traffic state confidence level when the host vehicle enters the intersection range, continuing until leaving the intersection range, and storing the calculation result in an information queue; When performing the traffic decision execution, extracting the calculation result of the information queue and obtaining the current traffic state confidence level through weighted average.

[0016] On the other hand, the present invention also provides a traffic state determination system based on intersection traffic flow based on the above method; The system includes: a traffic light sensing module, a host vehicle information reading module, an intersection coordinate system establishment module, an obstacle reading module, and a data processing module; The traffic light sensing module is used to identify the traffic light state of the current intersection; The host vehicle information reading module is used to obtain host vehicle information; The intersection coordinate system establishment module is used to establish the coordinate system of the current intersection and calculate the vehicle running projection axis of the current intersection; The obstacle reading module is used to screen dynamic obstacles; The data processing module is used to calculate the current traffic state confidence level and perform traffic decision execution.

[0017] On yet another aspect, the present invention also provides a vehicle, which is deployed with the above traffic state determination system based on intersection traffic flow.

[0018] Compared with the prior art, the present invention has the following beneficial effects: The present invention can provide more accurate results by dynamically analyzing the traffic flow to predict the traffic light state. Moreover, this prediction method based on traffic flow perception not only improves the accuracy of perception, but also provides an effective supplementary means for the traffic light recognition module, no longer relying on visual recognition results, and can adapt to un-upgraded networked areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0020] Figure 1 is the method flowchart in a specific embodiment of the present invention; Figure 2 is the traffic flow screening flowchart in a specific embodiment of the present invention; Figure 3 is the vehicle projection schematic diagram in a specific embodiment of the present invention.

[0021] In the figure: 1. Own vehicle, 2. Leading vehicle, 3. Vehicles traveling in the same direction, 4. Oncoming vehicles, 5. Left-turning vehicles, 6. Right-turning vehicles, 7. Slow-moving vehicles at the edge. Detailed implementation manners

[0022] To make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0023] It should be noted that similar reference numerals and letters denote similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0024] It is worth noting that the methods used in the present invention are all conventional methods unless otherwise specified; the raw materials and devices used are all conventional commercially available products unless otherwise specified, and their sources are not specifically limited.

[0025] It also needs to be noted that in the specific embodiments of the present invention, for the convenience of understanding, the method steps are described in a certain order. However, those skilled in the art can adjust the order of the steps according to actual needs. Therefore, this cannot be used as a limiting condition.

[0026] Please refer to Figure 1 and Figure 2 As shown, this embodiment provides a method for judging the traffic flow state at an intersection, which mainly includes the following steps: The in-vehicle system continuously obtains the information of the own vehicle according to the in-vehicle sensors. Among them, the information of the own vehicle includes the position, orientation and driving state of the own vehicle; further, the driving state of the own vehicle at least includes the speed of the own vehicle.

[0027] Optionally, in order to reduce resource requirements and improve response speed, it is first necessary to judge whether the intersection area is reached according to the information of the own vehicle, and determine whether the own vehicle is in the intersection and has a speed. If the intersection is not reached or the own vehicle has a driving speed, it means that there is no need to judge the traffic flow state, so this task is skipped; otherwise, the subsequent steps are continued.

[0028] Establish a coordinate system for the current intersection based on the information of the own vehicle. Specifically, considering that some intersections are not standard cross-shaped intersections, the coordinate system adopts a non-rectangular coordinate system to adapt to different intersection requirements. Thus, a plane non-rectangular coordinate system is established according to the actual situation of the intersection; calculate the center point of the current intersection and the vehicle running projection axis.

[0029] The establishment of the non-rectangular coordinate system of the current intersection includes the following steps: (1) Read all lanes within the intersection, and screen out the straight-through motor vehicle lanes from them, and record their orientations. Specifically, for orientations that are not much different, they will not be stored repeatedly, that is, after screening, it is judged whether the orientation is within the tolerance interval of each recorded orientation. If none of them are within the interval, then record it. If the number of obtained orientations is less than 4, then the starting and ending orientations of the left-turn lanes are screened in the same way. Taking the orientation of the vehicle itself as the positive direction of the Y-axis, select the lane direction with the largest angle with the Y-axis as the positive direction of the X-axis; that is, find the orientation with the smallest angle (without direction) with the orientation of the vehicle itself among all orientations as the positive direction of the Y-axis of the intersection. Then take the orientation with the largest angle (clockwise) with the positive direction of the Y-axis as the positive direction of the X-axis of the intersection.

[0030] (2) Take the center of the intersection of the dividing lines of the straight-through lanes of the current intersection or the center of the set of lane intersections as the center point of the current intersection; specifically, take the center of the four intersections of the leftmost dividing lines of the straight-through lanes in the four directions as the center point of the intersection, and take half of the difference between the extreme points of the intersection polygon as the radius of the intersection. When the straight-through lanes in each direction intersect, take the center of the set of intersections of the innermost dividing lines of each lane as the center point of the intersection. Take the length from the center point of the intersection to the side of the intersection polygon and subtract the boundary value as the traffic flow screening radius; if the vehicle itself is in the non-motor vehicle lane and in the first quadrant of the established coordinate system, it is considered to be in the situation of waiting to turn left in the non-motor vehicle lane, and the vehicle head may be misaligned due to obstruction. The module actively resets the negative direction of the x-axis as the new positive direction of the y-axis.

[0031] (3) Set multiple vehicle operation projection axes, and adapt to the positive and negative directions of each axis, as well as the direction of the angle bisector of each quadrant; specifically, taking a crossroads as an example: as Figure 3 shown by the center dotted line, use the positive and negative directions of the X-axis and Y-axis and the angle bisectors of the four quadrants as the four straight-through and four left-turn projection axes, thus forming a projection axis in the shape of a cross.

[0032] Before the projection calculation, first clean the data, and thus eliminate vehicles with low reference value based on the operation parameters of the observable vehicles at the current intersection in the non-rectangular coordinate system of the intersection; among them, the vehicle itself reads all dynamic obstacles from the perception channel according to the data collected by the sensor, such as vehicles within and near the intersection, and screens out other vehicles within the intersection area as the observable vehicles at the intersection. The screening content is whether it is a motor vehicle and whether the speed is greater than a certain value. The operation parameters of the observable vehicles at the intersection include the vehicle speed, position quadrant, speed orientation, and distance from the intersection center of the observable vehicles at the intersection.

[0033] Optionally, if it is determined that there is no traffic flow in the above operation of reading and screening the traffic flow within the intersection, it means that there is no need to perform the traffic state judgment, and it can pass normally or rely on the in-vehicle vision system to identify the traffic lights, that is, in an unobstructed state, so this task is also skipped; on the contrary, the subsequent steps are continued. For example, vehicle 1 is blocked by vehicle 2 within the intersection range, resulting in the image sensor being unable to collect the traffic light signal image.

[0034] Low-reference-value vehicles include right-turn vehicles 6, U-turn vehicles, and edge slow vehicles 7. These vehicles cannot provide effective information for the judgment of the traffic light state. Therefore, the steps for excluding low-reference-value vehicles are as follows: For vehicles whose distance from the intersection center exceeds 1 / 3 of the intersection radius and whose speed direction quadrant is the vehicle position quadrant plus 1 (the fourth quadrant is the first quadrant), they are determined as right-turn vehicles 6; For vehicles whose distance from the intersection center exceeds 1 / 2 of the intersection radius and whose speed direction is opposite to the straight-ahead direction of the lane where they are located, they are determined as U-turn vehicles; Regarding edge slow vehicles 7, vehicles that are at the intersection boundary and have a slow speed are excluded because their states are difficult to determine; thus, vehicles with a speed lower than the preset speed threshold and located at the intersection boundary are edge slow vehicles.

[0035] The vehicle projection sum on each vehicle running projection axis is calculated respectively by the projection method. Among them, the way of projection summation can be determined according to actual needs, such as calculating the projection quantity, weighting the projection value, or summing the projection values; so in this embodiment, as Figure 3 shown, it is implemented in the way of projection summation, specifically as follows: First, the speed of the vehicles within the intersection area is collected by the sensor, including the speed magnitude and direction, thereby forming a speed vector; After that, the speed vector is projected onto the nearest vehicle running projection axis, and the projection value is obtained, such as Figure 3 the thick black line segment in; Finally, the projection values on each vehicle running projection axis are summed respectively, so as to obtain the vehicle projection sum of this vehicle running projection axis.

[0036] Furthermore, the nearest vehicle running projection axis is the vehicle running projection axis with the smallest included angle with the corresponding vehicle speed direction.

[0037] Therefore, the confidence level of the current traffic state is calculated. The confidence level of the current traffic state is the ratio of the sum of the projections of vehicles running in the same direction to the sum of the projections of all vehicles. Among them, the sum of the projections of all vehicles is the sum of the projections of vehicles on the 8 vehicle running projection axes in this embodiment; the sum of the projections of vehicles running in the same direction is the sum of the projections of vehicles on the vehicle running projection axes that are in the same direction as the passable direction of the host vehicle. The addends of the sum of the projections of vehicles running in the same direction include the sum of the projections of vehicles in the host vehicle direction and the sum of the projections of vehicles in the co - existent direction; further, the co - existent direction is the straight - ahead vehicle direction and the left - turn vehicle direction allowed to pass in the same traffic light phase; as Figure 3 shown in Figure 3 , there are 3 vehicles running in the same direction in the same lane in the host vehicle direction. The co - existent direction includes vehicles in adjacent non - turning lanes, oncoming vehicles 4, and left - turn vehicles 5. Since these types of vehicles enter the intersection area and have a certain speed, they are regarded as vehicles with reference value, and their projections are projected onto the nearest vehicle running projection axis to obtain traffic flow information.

[0038] Among them, the projection method in this embodiment is to project the velocity vector of each vehicle onto its nearest coordinate axis. Since it is a statistical value, the absolute value is taken without considering the negative sign for calculation, and it is considered to belong to the traffic flow of the traffic state corresponding to this axis.

[0039] Further, the left - turn critical state (i.e., just starting or already turning) is judged, and its projection is evenly divided onto the left - turn and straight - ahead axes for calculation.

[0040] Furthermore, although this embodiment is for cross - shaped intersections, in fact, there are also T - shaped intersections. Therefore, it is necessary to re - classify the traffic flow that actually does not exist in T - shaped intersections.

[0041] Therefore, the ratio of the sum of the projections of vehicles running in the same direction of the host vehicle to the sum of the projections of all vehicles is calculated, and the ratio is used as the confidence level of the current traffic state.

[0042] The above are the calculation steps of the confidence level of the current traffic state. Specifically, in this embodiment, corresponding start conditions need to be designed to save system resources; at the same time, when it is not a non - motor vehicle lane turning left at the same intersection, the coordinate system can be simply verified and reused to reduce calculation consumption. Therefore, when the vehicle enters the intersection area, the above - mentioned calculation process is started, specifically as follows: Based on the confidence level of the current traffic state and the recognition result of the traffic lights at the current intersection, the traffic decision is executed through mutually exclusive judgment.

[0043] Among them, for decision execution: by calculating the traffic flow state in the intersection, the possibility of each traffic state is calculated according to mutual exclusion; A. Read the recognition result of the traffic lights at the current intersection through the traffic light perception module. When the recognition result is green, the confidence level of the current traffic state is verified: If the confidence level of the current traffic state is greater than or equal to the passing threshold, the host vehicle is allowed to pass; If the confidence level of the current driving state is less than the threshold, trigger a parking command, decelerate - brake, and report data to the cloud to request manual takeover; B. Read the recognition result of the traffic lights at the current intersection through the traffic - light perception module. When the recognition result is recognition failure or the recognition confidence level is lower than the preset threshold, verify the confidence level of the current driving state: If the confidence level of the current driving state is greater than or equal to the driving threshold, drive according to the green - light state; If the confidence level of the current driving state is less than the threshold, stop according to the red - light state.

[0044] Optionally, when the vehicle enters the range of the intersection (for example, when the intersection is 30 meters ahead), the above - mentioned process state calculates the real - time driving - state confidence level periodically at 100 ms, continues until leaving the intersection range, and stores the calculation result in the information queue.

[0045] To improve the judgment accuracy, in this embodiment, when making a decision, the calculation result of the information queue is extracted. The queue length is 30, and the data within 3 seconds is recorded. The final result takes the average value of the data frames within 3 seconds as the confidence level of the current driving state.

[0046] On the other hand, based on the above - mentioned method, this embodiment configures a corresponding driving - state judgment system based on the traffic flow at the intersection. Among them, the system includes: a traffic - light perception module, a self - vehicle information reading module, an intersection coordinate - system establishment module, an obstacle reading module, and a data - processing module; The traffic - light perception module is used to identify the traffic - light state at the current intersection; optionally, the traffic - light perception module in this embodiment uses a vision - recognition module, which can identify the traffic - light state at the current intersection under the condition that the visibility meets the shooting conditions and there is no occlusion.

[0047] The self - vehicle information reading module is used to obtain self - vehicle information, which is connected to the in - vehicle sensors of the self - vehicle, including but not limited to speed sensors, positioning modules (such as Beidou, GPS).

[0048] The intersection coordinate - system establishment module is used to establish a non - rectangular coordinate system of the current intersection and calculate the center point of the current intersection and the vehicle - running projection axis.

[0049] The obstacle reading module is used to screen dynamic obstacles and eliminate vehicles with low reference value.

[0050] The data - processing module is used to calculate the confidence level of the current driving state and make decision execution.

[0051] This embodiment also provides a vehicle, which can be an autonomous vehicle or a manned vehicle. The vehicle is equipped with the aforementioned traffic flow-based passing state judgment system at intersections. When it is an autonomous vehicle, it operates according to the processing method of the aforementioned steps; when it is a manned vehicle, it feeds back prompt information to the driver.

[0052] Finally, it should be noted that the above content is only used to illustrate the technical solution of the present invention, rather than to limit the protection scope of the present invention. Any simple modification or equivalent replacement of the technical solution of the present invention by those of ordinary skill in the art shall not depart from the essence and scope of the technical solution of the present invention.

Claims

1. A method for judging the passing state based on intersection traffic flow, characterized in that: It includes the following steps: Obtain the vehicle information of the host vehicle, establish a coordinate system of the current intersection based on the vehicle information of the host vehicle, and the coordinate system is provided with multiple vehicle running projection axes; Calculate the sum of vehicle projections on each vehicle running projection axis respectively by the projection method; Calculate the sum of vehicle projections in the same direction of travel, and the sum of vehicle projections in the same direction of travel is the sum of the vehicle projections on the vehicle running projection axes in the same direction as the passable direction of the host vehicle; Calculate the confidence level of the current passing state, and the confidence level of the current passing state is the ratio of the sum of vehicle projections in the same direction of travel to the total sum of vehicle projections in the coordinate system; the total sum of vehicle projections is the sum of the vehicle projections on each vehicle running projection axis; Based on the confidence level of the current passing state and the recognition result of the traffic lights at the current intersection, perform passing decision execution through mutually exclusive judgment.

2. The traffic flow state determination method based on intersection traffic flow according to claim 1, wherein: The passing decision execution through mutually exclusive judgment includes: Read the recognition result of the traffic lights at the current intersection. When the recognition result is a green light, verify the confidence level of the current passing state: If the confidence level of the current passing state is greater than or equal to the passing threshold, allow passing; If the confidence level of the current passing state is less than the threshold, trigger a stop command and report data to request manual takeover; Read the recognition result of the traffic lights at the current intersection. When the recognition result is a recognition failure or the recognition confidence level is lower than the preset threshold, verify the confidence level of the current passing state: If the confidence level of the current passing state is greater than or equal to the passing threshold, pass according to the green light state; If the confidence level of the current passing state is less than the threshold, stop according to the red light state.

3. The traffic flow-based passing state determination method according to claim 1, wherein: The vehicle information of the host vehicle includes the position of the host vehicle, the orientation of the host vehicle, and the driving state of the host vehicle; the driving state of the host vehicle is at least provided with the speed of the host vehicle; The establishment of the coordinate system of the current intersection includes the following steps: Take the lane direction closest to the orientation of the host vehicle as the positive direction of the Y axis, and select the lane direction with the largest included angle with the Y axis as the positive direction of the X axis; Take the center of the intersection of the straight-through lane dividing lines at the current intersection or the center of the set of lane intersections as the center point of the current intersection; Set multiple vehicle running projection axes to adapt to the positive and negative directions of the X axis and Y axis, and the direction of the angular bisector of each quadrant.

4. The traffic flow state determination method based on intersection traffic flow according to claim 3, wherein: Before calculating the sum of vehicle projections on each vehicle running projection axis, perform data cleaning on the observable vehicles at the intersection, including the following steps: Eliminate vehicles with low reference value based on the running parameters of the observable vehicles at the intersection in the coordinate system of the current intersection; The running parameters of the observable vehicles at the intersection include the vehicle speed, position quadrant, speed orientation, and distance from the intersection center of the observable vehicles at the intersection; The vehicles with low reference value include right-turn vehicles, U-turn vehicles, and edge slow vehicles.

5. The method for judging the traffic state based on the intersection traffic flow according to claim 4, wherein: Eliminating vehicles with low reference value includes the following steps: For vehicles whose distance from the intersection center exceeds 1 / 3 of the intersection radius and the quadrant where the speed orientation is located is the vehicle position quadrant plus 1, they are determined as the right-turn vehicles; For vehicles whose distance from the intersection center exceeds 1 / 2 of the intersection radius and the speed orientation is opposite to the straight-through direction of the lane where they are located, they are determined as the U-turn vehicles; Vehicles with a speed lower than a preset speed threshold and located at the intersection boundary are the edge slow vehicles.

6. The traffic state determination method based on intersection traffic flow according to any one of claims 3-5, characterized in that: Calculating the vehicle projection sum includes the following steps: Project the vehicle's velocity vector onto the nearest vehicle running projection axis, and take the absolute value to calculate the projection value; add up the projection values on the vehicle running projection axis to obtain the vehicle projection sum; For left-turn critical state vehicles, divide the projection value of the left-turn critical state vehicle equally between the straight-ahead axis and the left-turn axis of the left-turn critical state vehicle; For lane directions that do not exist at a T-intersection, reclassify the vehicles belonging to the corresponding vehicle running projection axis to the adjacent vehicle running projection axis.

7. The traffic condition judgment method based on intersection traffic flow according to claim 6, characterized in that: Calculating the confidence level of the current traffic state further includes the following steps: Calculate the projection sum of vehicles running in the same direction; the addends of the projection sum of vehicles running in the same direction include the projection sum in the self-vehicle direction and the projection sum in the coexistent direction; The coexistent direction is the straight-ahead vehicle direction and the left-turn vehicle direction allowed to pass under the same traffic light phase.

8. The traffic state determination method based on intersection traffic flow according to claim 7, characterized in that: It further includes the following steps: when the self-vehicle enters the intersection range, periodically calculate the real-time traffic state confidence level, continue until leaving the intersection range, and store the calculation results in the information queue; When performing the traffic decision execution, extract the calculation results from the information queue and obtain the current traffic state confidence level by weighted average.

9. A traffic flow-based passing state judgment system at intersections, characterized in that: For running the traffic state judgment method based on intersection traffic flow according to any one of claims 1-8; The system includes: a traffic light sensing module, a self-vehicle information reading module, an intersection coordinate system establishment module, an obstacle reading module, and a data processing module; The traffic light sensing module is used to identify the traffic light state of the current intersection; The self-vehicle information reading module is used to obtain self-vehicle information; The intersection coordinate system establishment module is used to establish the coordinate system of the current intersection and calculate the vehicle running projection axis of the current intersection; The obstacle reading module is used to screen dynamic obstacles; The data processing module is used to calculate the current traffic state confidence level and perform traffic decision execution.

10. A vehicle, characterized in that: Deploy the traffic state judgment system based on intersection traffic flow according to claim 9.

Images