Method for controlling vehicle to avoid crossing vulnerable traffic participants at intersection

Abstract

The invention relates to a method for controlling a vehicle to avoid crossing vulnerable traffic participants at an intersection, comprising the following steps: acquiring traffic direction information and traffic light information of each traffic lane at the intersection under the condition that the intersection within a preset distance range in front of the vehicle (1) is monitored; assessing the existence risk of vulnerable traffic participants crossing the intersection based on the traffic signal lamp information of each traffic lane; in the case of assessing the presence of a risk of a vulnerable traffic participant traversing the intersection, motion state information of a surrounding vehicle on an adjacent lane of the vehicle is detected, and the vehicle is controlled at least based on the motion state information of the surrounding vehicle to avoid the vulnerable traffic participant traversing the intersection, the traffic directions of the current lane and the adjacent lane are not opposite. According to the invention, the vehicle can actively avoid the vulnerable traffic participants which cannot be detected due to the shielding of surrounding vehicles, and the safety of the vehicle passing through the intersection is effectively improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of autonomous driving, and in particular to a method for controlling a vehicle to yield to a vulnerable road user crossing at an intersection, an assisted driving system, a vehicle comprising the assisted driving system according to the present application, and a computer program product. BACKGROUND

[0002] When a vehicle is driving through an intersection with traffic lights, it often encounters vulnerable road users crossing the road, including pedestrians, cyclists and / or e-bike riders, etc., and the phenomenon of not obeying traffic regulations is common, especially when the traffic light is just switched, the vulnerable road users often choose to run a red light or a yellow light to continue their journey across the road. Since the surrounding vehicles on each lane of the intersection block the view of the vehicle, the vehicle is difficult to detect the vulnerable road user crossing in front in time and yield, which causes the frequent occurrence of "ghost probe" traffic accidents.

[0003] Against this background, it is desirable to provide an assisted driving control strategy for a vehicle, aiming to control the vehicle to actively yield to a vulnerable road user crossing at an intersection. SUMMARY

[0004] The purpose of the present application is to provide a method for controlling a vehicle to yield to a vulnerable road user crossing at an intersection, an assisted driving system, a vehicle comprising the assisted driving system according to the present application, and a computer program product, to at least partially solve the problems in the prior art.

[0005] According to a first aspect of the present application, a method for controlling a vehicle to yield to a vulnerable road user crossing at an intersection is provided, which can comprise: - In the case of monitoring an intersection within a predetermined distance range in front of the vehicle, the traffic direction information and traffic light information of each lane of the intersection can be acquired; - Based on the traffic light information of each lane, the existence risk of a vulnerable road user crossing in the intersection is evaluated; - In the case of evaluating the existence risk of a vulnerable road user crossing in the intersection, the motion state information of the surrounding vehicles on the adjacent lane of the vehicle can be detected, and the vehicle is controlled based at least on the motion state information of the surrounding vehicles to yield to the vulnerable road user crossing in the intersection, wherein the traffic direction of the current lane of the vehicle is not opposite to that of the adjacent lane.

[0006] The core idea of the present application is that when a vehicle is driving through a front intersection, the driving scenarios in which a weak traffic participant crossing in the intersection is possible are predicted based on traffic signal information of each lane, and the vehicle is controlled in these driving scenarios by using motion state information of surrounding vehicles in adjacent lanes, especially the vehicle is controlled to brake in time in the case that it is determined that the surrounding vehicles perform avoidance behaviors for the weak traffic participant crossing, so that the vehicle can actively avoid the weak traffic participant that cannot be detected due to the shielding of the surrounding vehicles, and the safety of the vehicle driving through the intersection is effectively improved.

[0007] According to an optional embodiment of the present application, whether the surrounding vehicle performs an avoidance behavior for the weak traffic participant crossing can be determined based on the motion state information of the surrounding vehicle on the adjacent lane, and the vehicle is controlled to brake in the case that it is determined that the surrounding vehicle performs the avoidance behavior. Optionally, the tail indicator of the vehicle can also be turned on and / or the steering wheel of the vehicle can be locked.

[0008] According to another optional embodiment of the present application, if it is detected that the acceleration of the surrounding vehicle changes from a positive acceleration value to a negative acceleration value and returns to zero, and it is detected that the speed of the surrounding vehicle becomes zero, it is determined that the surrounding vehicle performs an avoidance behavior for the weak traffic participant crossing, wherein the change curve of the acceleration of the surrounding vehicle with respect to time is, for example, in a sine waveform.

[0009] According to another optional embodiment of the present application, in the case that a pilot assisted driving function of the vehicle is activated, the intersection within a predetermined distance range in front of the vehicle can be monitored based on navigation information of the vehicle, and the traffic direction information and traffic signal information of each lane of the intersection can be obtained from the navigation information.

[0010] According to another optional embodiment of the present application, in the case that the pilot assisted driving function of the vehicle is not activated, the intersection within a predetermined distance range in front of the vehicle can be monitored based on perception fusion information collected about the driving environment of the vehicle, and the traffic direction information and traffic signal information of each lane of the intersection can be obtained from the perception fusion information. In the case that it is determined based on the perception fusion information that a zebra crossing section perpendicular to the driving direction of the vehicle is within a predetermined distance range in front of the vehicle, the intersection within the predetermined distance range in front of the vehicle can be monitored.

[0011] According to another optional embodiment of the present application, the traffic signal information includes, for example, the state of the traffic light of each lane and the time information thereof. If the green light opening duration of the traffic light of the current lane of the vehicle from the time when the traffic light turns green is less than a predetermined time period, it is determined that there is a risk of the presence of a weak traffic participant crossing in the intersection.

[0012] According to another optional embodiment of the present application, the motion state information of the surrounding vehicle can be determined based on the perception fusion information about the surrounding vehicle on the adjacent lane collected by the millimeter wave radar and / or the vehicle-mounted camera, wherein the motion state information of the surrounding vehicle comprises, for example, the acceleration change process of the surrounding vehicle with respect to time and the speed change process of the surrounding vehicle with respect to time.

[0013] According to a second aspect of the present application, an assisted driving system is provided, which can comprise the following components: - a vehicle-mounted navigation unit configured to provide navigation information of the vehicle, and / or an environmental perception unit configured to collect perception fusion information about the driving environment of the vehicle; - a vehicle-mounted control unit configured to execute the method according to the present application.

[0014] According to another optional embodiment of the present application, the environmental perception unit can comprise a vehicle-mounted camera, and / or a millimeter wave radar, and / or a laser radar, etc.

[0015] According to another optional embodiment of the present application, the vehicle-mounted control unit can be configured as a domain controller.

[0016] According to a third aspect of the present application, a vehicle is provided, which can comprise the assisted driving system according to the present application.

[0017] According to a fourth aspect of the present application, a computer program product, for example a computer-readable program carrier, containing or storing computer program instructions which, when executed by a processor, at least assist in implementing the steps of the method according to the present application. BRIEF DESCRIPTION OF DRAWINGS

[0018] The principles, features and advantages of the present application can be better understood by the following detailed description of the application, with reference to the accompanying drawings. The drawings show: Figure 1 a workflow diagram of a method for controlling a vehicle to yield to a weak traffic participant crossing at an intersection according to one exemplary embodiment of the present application; Figure 2 a driving scenario diagram according to one exemplary embodiment of the present application; Figure 3 a speed and acceleration change curve diagram of a surrounding vehicle according to one exemplary embodiment of the present application; Figure 4 a driving scenario diagram according to one exemplary embodiment of the present application; Figure 5A schematic view of a vehicle according to one example embodiment of the present application is shown. DETAILED DESCRIPTION

[0019] In order to make the technical problems to be solved by the present application, technical solutions and beneficial technical effects clearer, the present application will be further described in detail below in conjunction with the drawings and multiple example embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, but not to limit the protection scope of the present application.

[0020] Figure 1 A workflow diagram of a method for controlling a vehicle to yield to a vulnerable road user crossing at an intersection according to one example embodiment of the present application is shown. The following example embodiments describe the method according to the present application in more detail.

[0021] As Figure 1 shown, the method can include steps S1 to S3. In step S1, in the case that an intersection within a predetermined distance range in front of the vehicle is monitored, the traffic direction information and traffic signal information of each lane of the intersection can be obtained. In the current embodiment of the present application, a vehicle 1 equipped with an assisted driving system of an assisted driving level of L2 and above mostly has a Navigate on Autopilot (NOA) function, which is an assisted driving function that combines navigation data and driving control technology, and by which the vehicle 1 can be controlled to autonomously change lanes, overtake, brake, etc. on urban roads according to a planned navigation route. In the case that the NOA function of the vehicle is activated, the navigation system 11 of the vehicle 1 can plan a navigation route for the vehicle 1 according to the current position of the vehicle and the destination input by the user, and in the process of controlling the vehicle 1 to travel according to the planned navigation route, the intersection within a predetermined distance range in front of the vehicle 1 can be monitored based on the navigation information of the vehicle 1, and the traffic direction information and traffic signal information of each lane of the intersection can be obtained from the navigation information. As Figure 2 A driving scene schematic diagram according to one example embodiment of the present application is shown, in which the vehicle 1 monitors a crossroad in front of the vehicle during driving on a straight lane, and obtains the three lanes of the crossroad as a straight lane, a left turn lane and a right turn lane from the navigation information including high-precision map information, a planned navigation route and current position information of the vehicle. A traffic signal 4 is installed at the crossroad to indicate the right of way of each lane in real time. The traffic signal 4 is usually connected to the background server of the vehicle navigation unit 11, so that the traffic signal information of the straight lane, the left turn lane and the right turn lane can be obtained from the navigation information.

[0022] Without activating the navigation assistance function of vehicle 1, vehicle 1 can collect perception fusion information about the vehicle's driving environment through the environmental perception unit 12—including onboard cameras, and / or millimeter-wave radar, and / or lidar, etc.—while driving along a straight lane. Based on this perception fusion information, vehicle 1 can monitor intersections within a pre-defined distance range ahead of the vehicle. Specifically, if a zebra crossing segment perpendicular to the vehicle's direction of travel is identified within the pre-defined distance range ahead of the vehicle based on the perception fusion information, it can be determined that an intersection within the pre-defined distance range ahead of the vehicle has been detected. Furthermore, the perception fusion information can be used to obtain the traffic direction information and traffic light information for each lane at the intersection—including straight lanes, left-turn lanes, and right-turn lanes.

[0023] In step S2, the risk of vulnerable road users crossing the intersection can be assessed based on traffic light information for each lane. This traffic light information may include the red and green light status and timing information for each lane—including straight-ahead lanes, left-turn lanes, and right-turn lanes. If the green light duration for vehicle 1's current lane is less than a pre-defined time period (e.g., 5 seconds), it means the traffic light for the straight-ahead lane has just changed from red to green. There may be vulnerable road users (including pedestrians, cyclists, and / or electric scooter riders) on the crosswalk who have not yet completed their crossing. The traffic light information has a weak binding force on these vulnerable road users, who often choose to ignore the prohibition signal and continue crossing the road. Therefore, the risk of vulnerable road users crossing the intersection is determined, meaning there is a high probability that vulnerable road users are crossing the intersection.

[0024] In step S3, after assessing the risk of a vulnerable road user crossing the intersection, the motion status information of surrounding vehicles in the adjacent lane of vehicle 1's current lane can be detected, and vehicle 1 can be controlled, at least based on the motion status information of the surrounding vehicles, to avoid the vulnerable road user crossing the intersection, wherein the current lane of vehicle 1 and the adjacent lane do not have opposite travel directions. Figure 2As shown, a first surrounding vehicle 21 is parked in the adjacent left-turn lane, which obstructs the view of the first vulnerable traffic participant 31 to the left front of vehicle 1. This makes it difficult for vehicle 1's onboard camera and lidar to detect the first vulnerable traffic participant 31. Since vulnerable traffic participants usually stop and start when crossing the road while running a red light or yellow light, the point cloud data about the first vulnerable traffic participant 31 collected by vehicle 1's millimeter-wave radar will repeatedly switch between static and dynamic point clouds. This results in low confidence of the millimeter-wave radar's detection results of vulnerable traffic participants crossing the road. Therefore, the various environmental perception units 12 of vehicle 1 cannot reliably detect the first vulnerable traffic participant 31, which is obscured by the first surrounding vehicle 21.

[0025] If it is determined based on the travel direction information of each lane at the intersection that the current lane of vehicle 1 is not opposite to the travel direction of the adjacent lane, for example, if the current lane is a straight lane, the adjacent lane can be a left-turn lane, a right-turn lane, or a straight lane, as long as the adjacent lane is not the opposite lane of the current lane, then the motion state information of the surrounding vehicles can be obtained based on the perception fusion information about the surrounding vehicles in the adjacent lane collected by millimeter-wave radar and / or vehicle-mounted cameras. The motion state information of the surrounding vehicles can include the change process of the acceleration of the surrounding vehicles with respect to time and the change process of the speed of the surrounding vehicles with respect to time.

[0026] In the post-fusion scheme of vehicle-mounted sensors, the millimeter-wave radar of vehicle 1 can detect the Doppler velocities of surrounding vehicles. Transforming these Doppler velocities from polar coordinates to Cartesian coordinates yields the relative velocities of the surrounding vehicles with respect to vehicle 1. Adding these relative velocities to the vehicle 1's speed gives the absolute speed information of the surrounding vehicles. This absolute speed information can then be optimized using target tracking algorithms—such as Hungarian matching and Kalman filtering. Next, an image processing model—such as convolutional neural networks or large language models based on attention mechanisms—can calculate the moving speed information of the surrounding vehicles based on image information of vehicles in adjacent lanes acquired by the vehicle-mounted camera. Fusing the absolute speed information obtained from the millimeter-wave radar's point cloud data with the moving speed information obtained from the vehicle-mounted camera's image information, for example using Mahalanobis distance, yields the time-varying velocity of the surrounding vehicles. This time-varying velocity information can then be used to calculate the time-varying acceleration of the surrounding vehicles.

[0027] In the front fusion scheme of vehicle sensors, the velocity information of each point cloud data can be extracted based on the point cloud data of surrounding vehicles collected by millimeter-wave radar, and these point cloud data are projected onto the pixels of the image collected by the vehicle camera, thereby assigning velocity information to each pixel. The image processing model—such as convolutional neural networks, large language models based on attention mechanisms, etc.—can obtain the velocity information of surrounding vehicles based on these pixels and the velocity information they are assigned, especially the velocity change process of surrounding vehicles with respect to time, and then obtain the acceleration change process of surrounding vehicles with respect to time based on the velocity change process of surrounding vehicles with respect to time.

[0028] Here, based on the motion state information of surrounding vehicles in adjacent lanes, it can be determined whether the surrounding vehicles have performed yielding behavior towards a vulnerable road user crossing the road. If the acceleration of a surrounding vehicle changes from a positive acceleration value to a negative acceleration value and then returns to zero, and the speed of the surrounding vehicle becomes zero, this means that the surrounding vehicle has undergone a process of acceleration and deceleration to a stop, and thus, it can be determined that the surrounding vehicle has performed yielding behavior towards the vulnerable road user crossing the road. Figure 2 In the driving scenario, the first surrounding vehicle 21 starts moving after the traffic light in the left-turn lane turns green, thus detecting a positive acceleration value for the first surrounding vehicle 21. However, after starting, the first surrounding vehicle 21 recognizes the first vulnerable road user 31 who is still crossing the zebra crossing to its left and performs a braking operation until the first surrounding vehicle 21 comes to a complete stop. Therefore, the acceleration of the first surrounding vehicle 31 changes from a positive acceleration value to a negative acceleration value and returns to zero, and the speed of the first surrounding vehicle 31 eventually becomes zero. Figure 3 The diagram illustrates the variation curves of the speed V and acceleration a of a surrounding vehicle with respect to time t according to an exemplary embodiment of this application. The acceleration of the first surrounding vehicle 31 varies with time, for example, in a shape similar to a sine wave. The speed of the first surrounding vehicle 31 increases from zero to a certain speed value, then decreases and eventually returns to zero.

[0029] For example, in Figure 4In the illustrated driving scenario diagram according to an exemplary embodiment of this application, the current lane where vehicle 1 is located and the adjacent lane to the right where the second surrounding vehicle 22 is located both have a straight-ahead direction. The second surrounding vehicle 22 starts moving after the traffic light in the straight-ahead lane turns green, thereby detecting a positive acceleration value for the second surrounding vehicle 22. However, after starting, the second surrounding vehicle 22 identifies a second vulnerable road user 32 who is still crossing the zebra crossing to its right and performs a braking operation until the second surrounding vehicle 22 comes to a complete stop. As a result, the acceleration of the second surrounding vehicle 32 changes from a positive acceleration value to a negative acceleration value and returns to zero, and the speed of the second surrounding vehicle 32 eventually becomes zero. Figure 3 As shown, the acceleration of the second surrounding vehicle 32 can also be in the shape of a sine wave, where the speed of the second surrounding vehicle 32 increases from zero to a certain value and then decreases, eventually returning to zero.

[0030] Upon determining the yielding behavior of surrounding vehicles, vehicle 1 can be controlled to brake (i.e., pre-braking can be activated), causing vehicle 1 to stop before the zebra crossing. This allows for timely yielding to vulnerable road users crossing the road in front of vehicle 1—including a first vulnerable road user 31 crossing from the left front of vehicle 1 and / or a second vulnerable road user 32 crossing from the right front of vehicle 1. Optionally, the taillights of vehicle 1 can be activated—including, for example, brake lights and / or hazard warning lights located at the rear of the vehicle—to alert vehicles behind vehicle 1 to brake in time. Optionally, the steering wheel of vehicle 1 can be locked to prevent vehicle 1 from changing lanes into adjacent lanes.

[0031] According to the above embodiments of this application, when a vehicle is driving through an intersection ahead, the system predicts driving scenarios where vulnerable road users may be crossing the intersection based on the traffic light information of each lane. In these driving scenarios, the system controls the vehicle using the motion status information of surrounding vehicles in adjacent lanes. In particular, when it is determined that surrounding vehicles are taking evasive action against vulnerable road users crossing the intersection, the system controls the vehicle to brake in a timely manner. As a result, the vehicle can actively avoid vulnerable road users who cannot be detected due to obstruction by surrounding vehicles, effectively improving the safety of the vehicle when driving through the intersection.

[0032] In addition, it should be noted that the step numbers described herein do not necessarily represent the order of steps, but are merely a reference numeral. The order may be changed depending on the specific circumstances, as long as the technical objective of this application can be achieved.

[0033] Figure 5 A schematic diagram of a vehicle according to an exemplary embodiment of this application is shown. Figure 5As shown, vehicle 1 may be equipped with a driver assistance system 10, which may include the following components: - In-vehicle navigation unit 11 and / or environmental perception unit 12, wherein the in-vehicle navigation unit 11 is configured to provide navigation information for vehicle 1, and the environmental perception unit 12 is configured to collect perception fusion information about the driving environment of the vehicle, wherein the environmental perception unit 12 includes, for example, an in-vehicle camera, and / or millimeter-wave radar, and / or lidar, etc. - An onboard control unit 13 for performing the method according to any one of the preceding claims, wherein the onboard control unit 13 is particularly configured as a domain controller.

[0034] It should be understood that the terms “first,” “second,” “third,” etc., used in this document are for descriptive purposes only and should not be construed as indicating or implying relative importance, nor should they be construed as implicitly specifying the number of technical features indicated.

[0035] If an embodiment includes an "and / or" association between a first feature and a second feature, it should be interpreted as follows: according to one implementation, the embodiment has not only the first feature but also the second feature; according to another implementation, the embodiment has either only the first feature or only the second feature.

[0036] Although specific embodiments have been described above, these embodiments are not intended to limit the scope of this application, even when only a single embodiment is described with respect to a particular feature. The feature examples provided in this application are intended for illustrative purposes and not for limitation, unless otherwise stated. In practice, multiple features may be combined with each other as needed and where technically feasible. Various substitutions, modifications, and alterations are also conceived without departing from the spirit and scope of this application.

Claims

1. A method for controlling a vehicle to yield to a vulnerable road user crossing an intersection, the method comprising: When an intersection within a predetermined distance range ahead of the vehicle (1) is detected, the travel direction information and traffic light information of each lane at the intersection are obtained. The presence risk of vulnerable road users crossing the intersection is assessed based on traffic signal information for each lane. In assessing the risk of a vulnerable road user crossing the intersection, the motion state information of surrounding vehicles in the adjacent lane of vehicle (1) is detected, and vehicle (1) is controlled at least based on the motion state information of the surrounding vehicles to avoid the vulnerable road user crossing the intersection, wherein the current lane of vehicle (1) is not opposite to the travel direction of the adjacent lane.

2. The method according to claim 1, wherein, Based on the motion state information of surrounding vehicles in adjacent lanes, determine whether the surrounding vehicles perform evasive action against vulnerable traffic participants crossing the road, and if the evasive action of the surrounding vehicles is determined, control the vehicle (1) to brake, optionally turn on the taillight indicator of the vehicle (1) and / or lock the steering wheel of the vehicle (1).

3. The method according to claim 2, wherein, If the acceleration of the surrounding vehicle changes from a positive acceleration value to a negative acceleration value and returns to zero, and the speed of the surrounding vehicle becomes zero, then the avoidance behavior of the surrounding vehicle towards the vulnerable road user crossing the road is determined, wherein the acceleration of the surrounding vehicle changes with time, for example, in the form of a sine wave.

4. The method according to any one of claims 1 to 3, wherein, When the navigation assistance function of the vehicle (1) is activated, the intersections within a pre-given distance range ahead of the vehicle are monitored based on the navigation information of the vehicle (1), and the traffic direction information and traffic light information of each lane of the intersection are obtained from the navigation information.

5. The method according to any one of claims 1 to 3, wherein, Without activating the navigation assistance function of the vehicle (1), the intersections within a predetermined distance range in front of the vehicle are monitored based on the collected perception fusion information about the vehicle's driving environment, and the traffic direction information and traffic light information of each lane of the intersection are obtained from the perception fusion information. In the case that a zebra crossing segment perpendicular to the vehicle's driving direction is determined within a predetermined distance range in front of the vehicle based on the perception fusion information, the intersections within a predetermined distance range in front of the vehicle are monitored.

6. The method according to any one of the preceding claims, wherein, The traffic signal information includes the red and green light status and time information for each lane. If the green light duration of the traffic signal for the current lane of vehicle (1) is less than a pre-given time period from the moment it turns green, then the presence risk of a vulnerable traffic participant crossing the intersection is determined.

7. The method according to any one of the preceding claims, wherein, The motion state information of the surrounding vehicles is obtained based on the perception fusion information about the surrounding vehicles in adjacent lanes collected by millimeter-wave radar and / or vehicle-mounted cameras. The motion state information of the surrounding vehicles includes the change process of the acceleration of the surrounding vehicles with respect to time and the change process of the speed of the surrounding vehicles with respect to time.

8. A driver assistance system (10), the driver assistance system (10) comprising the following components: The vehicle navigation unit (11) and / or the environmental perception unit (12), wherein, The vehicle navigation unit (11) is configured to provide navigation information for the vehicle (1), and the environmental perception unit (12) is configured to collect perception fusion information about the vehicle's driving environment. The vehicle control unit (13) is used to perform the method according to any one of claims 1 to 7.

9. The driver assistance system (10) according to claim 8, wherein, The environmental perception unit (12) includes an onboard camera, and / or millimeter-wave radar, and / or lidar; and / or The vehicle control unit (13) is configured as a domain controller.

10. A vehicle (1) comprising a driver assistance system (10) according to claim 8 or 9.

11. A computer program product, such as a computer-readable program carrier, comprising or storing computer program instructions that, when executed by a processor, at least auxiliaryly implement the steps of the method according to any one of claims 1 to 7.

Images