Vehicle control method and device based on road section risk detection, equipment, medium and product

Abstract

The invention provides a vehicle control method and device based on road section risk detection, electronic equipment, a medium and a product. The method comprises the following steps: determining a risk type of a target road section based on at least one of planned trajectory information of the target road section through which an automatic driving vehicle is to pass, predicted trajectory information of a barrier vehicle on the target road section and historical vehicle trajectory information of the target road section; and determining a control strategy for the autonomous vehicle based on the risk type of the target road section. According to the method and the device, the possible vehicle collision conflict of the automatic driving vehicle on the target road section can be comprehensively and accurately detected, the risk of the vehicle collision conflict of the automatic driving vehicle on the target road section can be effectively reduced, and the safe driving of the automatic driving vehicle is ensured.

Description

Technical Field

[0001] Embodiments of the present disclosure relate to the field of computer technologies, and in particular, to a vehicle control method based on road section risk detection, a vehicle control device based on road section risk detection, an electronic device, a computer-readable storage medium, and a computer program product. Background Art

[0002] The driving environment of autonomous vehicles is complex, with a wide variety of vehicles on the road, and the randomness and uncertainty of vehicle driving behaviors are very large, which pose great challenges to the behavior planning of autonomous vehicles. At present, the behavior planning of autonomous vehicles mainly calculates based on vehicle prediction information. For example, when detecting a risk of collision with the predicted trajectory of a dynamic obstacle, a dynamic response decision is planned. However, the existing planning methods can detect limited road risks and cannot meet the needs of behavior planning for autonomous vehicles in a complex driving environment. Summary of the Invention

[0003] A vehicle control method, device, electronic device, medium, and product based on road section risk detection provided by embodiments of the present disclosure can solve or partially solve the above deficiencies in the prior art or other deficiencies in the prior art.

[0004] A vehicle control method based on road section risk detection according to a first aspect of the present disclosure includes: determining a risk type of a target road section based on at least one of planning trajectory information of a target road section that an autonomous vehicle is to pass through, predicted trajectory information of obstacle vehicles on the target road section, and historical vehicle trajectory information of the target road section; and determining a control strategy for the autonomous vehicle based on the risk type of the target road section.

[0005] A vehicle control device based on road section risk detection according to a second aspect of the present disclosure includes: a risk detection module configured to determine a risk type of a target road section based on at least one of planning trajectory information of a target road section that an autonomous vehicle is to pass through, predicted trajectory information of obstacle vehicles on the target road section, and historical vehicle trajectory information of the target road section; and a risk processing module configured to determine a control strategy for the autonomous vehicle based on the risk type of the target road section.

[0006] An electronic device according to a third aspect of the present disclosure may include: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the vehicle control method based on road section risk detection according to the first aspect of the present disclosure.

[0007] The computer-readable storage medium provided according to the fourth aspect of the present disclosure includes a computer program, and when the computer program is executed by a processor, it implements the vehicle control method based on road segment risk detection described in the first aspect of the present disclosure.

[0008] The computer program product provided according to the fifth aspect of the present disclosure stores a computer program, and when the computer program is executed by a processor, it implements the vehicle control method based on road segment risk detection described in the first aspect of the present disclosure.

[0009] The vehicle control method, device, electronic device, medium and product based on road segment risk detection provided according to the embodiments of the present disclosure determine the risk type of the target road segment based on at least one of the planned trajectory information of the target road segment that the autonomous vehicle is to pass through, the predicted trajectory information of the obstacle vehicle on the target road segment, and the historical vehicle trajectory information of the target road segment. By combining multiple types of information to detect the risk of the target road segment, it is possible to comprehensively and accurately detect vehicle collision conflicts that may occur to the autonomous vehicle on the target road segment. Based on the risk type of the target road segment, a control strategy for the autonomous vehicle is determined to control the autonomous vehicle, which can effectively reduce the risk of vehicle collision conflicts occurring to the autonomous vehicle on the target road segment and ensure the safe driving of the autonomous vehicle.

[0010] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it used to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. Description of the Drawings

[0011] By reading the detailed description of the non-limiting embodiments with reference to the following drawings, other features, purposes and advantages of the present disclosure will become more obvious. The drawings are used to better understand the solution and do not constitute a limitation to the present disclosure. Among them: Figure 1 is a flowchart of the vehicle control method based on road segment risk detection according to the embodiments of the present disclosure; Figure 2 is a flowchart of determining the risk type of the target road segment according to some embodiments of the present disclosure; Figure 3 is a flowchart of determining the risk type of the target road segment according to some other embodiments of the present disclosure; Figure 4 is a flowchart of determining the risk type of the target road segment according to still some other embodiments of the present disclosure; Figure 5 is a schematic diagram of aligning the historical vehicle trajectory information according to some embodiments of the present disclosure; Figure 6It is a schematic diagram of an application scenario of a vehicle control method based on road segment risk detection according to an embodiment of the present disclosure; Figure 7 It is a block diagram of a vehicle control device based on road segment risk detection according to an embodiment of the present disclosure; Figure 8 It is a block diagram of an electronic device that can be used to implement an example of an embodiment of the present disclosure. Detailed implementation manners

[0012] The following describes exemplary embodiments of the present disclosure in conjunction with the accompanying drawings. Various details of the embodiments of the present disclosure are included to facilitate understanding, and they should be considered merely exemplary. Therefore, those of ordinary skill in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, the description of well-known functions and structures is omitted below.

[0013] It should be noted that, without conflict, the embodiments in the present disclosure and the features in the embodiments can be combined with each other. The present disclosure will be described in detail below with reference to the drawings and in conjunction with the embodiments.

[0014] An exemplary system architecture for implementing the speed planning method provided by the present disclosure may include a terminal device, a network, and a server. The network is used to provide a communication link between the terminal device and the server, and may include various connection types, for example, wired communication links, wireless communication links, or fiber optic cables, etc.

[0015] Users can use the terminal device to interact with the server through the network to receive or send information, etc. Various client applications can be installed on the terminal device, for example, map-based, navigation-based, entertainment-based, etc. client applications.

[0016] The terminal device may be, for example, the in-vehicle system of a vehicle such as an autonomous vehicle or a delivery robot. This system can be implemented in a hardware manner, or in a software manner, or in a manner combining hardware and software.

[0017] The server can be hardware or software. When the server is hardware, it can be implemented as a distributed server cluster composed of multiple servers, or as a single server. When the server is software, it can be implemented as multiple software or software modules (for example, used to provide distributed services), or as a single software or software module. No specific limitation is made here.

[0018] It should be noted that the execution subject of the speed planning method provided in this disclosure (hereinafter simply referred to as the "execution subject") can be the server in the above system architecture, or the terminal device in the above system architecture, or the server and the terminal device in the above system architecture.

[0019] When the speed planning method is executed by the server, the server can determine the risk type of the target road section based on at least one of the planned trajectory information of the target road section that the autonomous vehicle is to pass through, the predicted trajectory information of the obstacle vehicle on the target road section, and the historical vehicle trajectory information of the target road section. Then, based on the risk type of the target road section, the server determines the control strategy for the autonomous vehicle and sends the control command to the terminal device, such as the in-vehicle system of the autonomous vehicle, and the autonomous vehicle is controlled according to the control command through the in-vehicle system.

[0020] Another applicable scenario is that the speed planning is directly completed by the terminal device, such as the in-vehicle system of the autonomous vehicle, without passing through the server. At this time, the in-vehicle system can determine the risk type of the target road section based on at least one of the planned trajectory information of the target road section that the autonomous vehicle is to pass through, the predicted trajectory information of the obstacle vehicle on the target road section, and the historical vehicle trajectory information of the target road section. Then, based on the risk type of the target road section, the in-vehicle system determines the control strategy for the autonomous vehicle and controls the autonomous vehicle.

[0021] In addition, the speed planning can also be jointly completed by the server and the terminal device, such as the in-vehicle system of the autonomous vehicle. This disclosure does not limit the operations performed by the server and the terminal device at this time. For example, the server can determine the risk type of the target road section based on at least one of the planned trajectory information of the target road section that the autonomous vehicle is to pass through, the predicted trajectory information of the obstacle vehicle on the target road section, and the historical vehicle trajectory information of the target road section. Then, the server sends the risk type of the target road section to the in-vehicle system, and the in-vehicle system can determine the control strategy for the autonomous vehicle based on the risk type of the target road section and control the autonomous vehicle.

[0022] In addition, in the technical solutions involved in this disclosure, the acquisition, storage, use, processing, transportation, provision, and disclosure of vehicle speeds, trajectory information, etc. all comply with the provisions of relevant laws and regulations and do not violate public order and good customs.

[0023] An embodiment of the present disclosure provides a vehicle control method 100 based on road section risk detection.

[0024] Figure 1 The flowchart of the vehicle control method 100 based on road section risk detection according to the embodiment of the present disclosure is shown. As Figure 1As shown, the vehicle control method 100 based on road segment risk detection may include the following steps: S101. Determine the risk type of the target road segment based on at least one of the planned trajectory information of the target road segment that the autonomous vehicle is to pass through, the predicted trajectory information of the obstacle vehicle on the target road segment, and the historical vehicle trajectory information of the target road segment.

[0025] S102. Determine the control strategy for the autonomous vehicle based on the risk type of the target road segment.

[0026] In an embodiment of the present disclosure, the execution subject may determine the risk type of the target road segment based on at least one of the planned trajectory information of the target road segment that the autonomous vehicle is to pass through, the predicted trajectory information of the obstacle vehicle on the target road segment, and the historical vehicle trajectory information of the target road segment.

[0027] In an embodiment of the present disclosure, the execution subject may first obtain the trajectory information of the autonomous vehicle. Here, the trajectory information may refer to the planned trajectory information obtained by performing trajectory planning on the target road segment that the autonomous vehicle is to pass through. Among them, the planned trajectory information may include information such as the pose and speed of the vehicle. For example, the execution subject may obtain the planned trajectory information of the autonomous vehicle from the planning module of the autonomous vehicle. The execution subject may also obtain the trajectory information of the obstacle vehicle. Here, the obstacle vehicle may refer to a vehicle traveling on the target road segment that the autonomous vehicle is to pass through. Here, the trajectory information may refer to the predicted trajectory information obtained by performing trajectory prediction on the obstacle vehicle. Among them, the predicted trajectory information may include information such as the pose and speed of the vehicle. For example, the execution subject may obtain the predicted trajectory information of the obstacle vehicle from the prediction module of the autonomous vehicle. The execution subject may also obtain the historical vehicle driving trajectory information of the target road segment. Here, the historical vehicle driving trajectory information may refer to the historical driving trajectory information of the vehicle that has passed through the target road segment. Among them, the historical vehicle driving trajectory information may include information such as the position of the vehicle. For example, the execution subject may obtain the historical vehicle driving trajectory information of the target road segment from the electronic map server through the map application of the autonomous vehicle.

[0028] Then, the execution entity can perform risk detection on the target road section based on at least one of the planned trajectory information, predicted trajectory information, and historical vehicle trajectory information, and determine whether there is a risk of vehicle collision conflict on the target road section. If there is a risk of vehicle collision conflict on the target road section, the risk type of the vehicle collision conflict occurring on the target road section is determined. Among them, different road risk types can be predefined according to the information about vehicle collision conflicts that may occur on the road, and the embodiments of the present disclosure do not limit the setting principles of the road risk types. In an alternative example, based on at least one of the planned trajectory information, predicted trajectory information, and historical vehicle trajectory information, it can be determined whether there is a type of vehicle collision conflict that may occur on the target road section. If there is a type of vehicle collision conflict that may occur on the target road section, the risk type of the target road section is determined. For example, the types of vehicle collision conflicts that may occur on the road can include explicit conflicts, potential conflicts, etc., and the embodiments of the present disclosure do not limit this. In another alternative example, based on at least one of the planned trajectory information, predicted trajectory information, and historical vehicle trajectory information, it can be determined whether there is a reason for the vehicle collision conflict that may occur on the target road section. If there is a reason for the vehicle collision conflict that may occur on the target road section, the risk type of the target road section is determined. For example, the reasons for the vehicle collision conflicts that may occur on the road can include abnormal vehicle driving behaviors, abnormal non-vehicle driving behaviors, etc., and the embodiments of the present disclosure do not limit this.

[0029] After that, the execution entity can determine the control strategy for the autonomous vehicle according to the risk type of the target road section and control the autonomous vehicle. Among them, different vehicle control strategies can be predefined according to the information about vehicle collision conflicts that may occur on the road, and the embodiments of the present disclosure do not limit the setting principles of the vehicle control strategies.

[0030] The vehicle control method based on road section risk detection provided by the embodiments of the present disclosure determines the risk type of the target road section based on at least one of the planned trajectory information of the target road section that the autonomous vehicle is to pass through, the predicted trajectory information of the obstacle vehicle on the target road section, and the historical vehicle trajectory information of the vehicle passing through the target road section. By combining multiple pieces of information to perform risk detection on the target road section, it can comprehensively and accurately detect the vehicle collision conflicts that may occur to the autonomous vehicle on the target road section. Determining the control strategy for the autonomous vehicle according to the risk type of the target road section and controlling the autonomous vehicle can effectively reduce the risk of vehicle collision conflicts occurring to the autonomous vehicle on the target road section and ensure the safe driving of the autonomous vehicle.

[0031] It should be understood that the steps shown in Method 100 are not exclusive, and other steps may be performed before, after, or between any of the shown steps. In addition, some of the shown steps may be performed simultaneously or may be performed in an order different from Figure 1 that shown.

[0032] Figure 2 A flowchart showing the determination of the risk type of a target road segment according to some embodiments of the present disclosure is shown. As Figure 2 shown, step S101 determines the risk type of the target road segment based on at least one of the planned trajectory information of the target road segment that the autonomous vehicle is to pass through, the predicted trajectory information of the obstacle vehicle on the target road segment, and the historical vehicle trajectory information of the target road segment, and may include the following steps: S201. Based on the planned trajectory information and the predicted trajectory information, determine whether there is a vehicle conflict area on the target road segment.

[0033] In response to the existence of a vehicle conflict area on the target road segment, perform operation S202. Otherwise, end the process.

[0034] S202. Based on the planned trajectory information, determine whether the autonomous vehicle has a specific driving behavior.

[0035] In response to the autonomous vehicle not having a specific driving behavior, perform operation S203. In response to the autonomous vehicle having a specific driving behavior, perform operation S204.

[0036] S203. Determine that the risk type of the target road segment is the first type of risk.

[0037] S204. Determine that the risk type of the target road segment is the second type of risk.

[0038] In the embodiments of the present disclosure, a situation where the autonomous vehicle will collide with other vehicles except for a specific driving behavior may be defined as the first type of risk. A situation where the autonomous vehicle will collide with other vehicles with a specific driving behavior may be defined as the second type of risk. Among them, the specific driving behavior can be set as needed, and the embodiments of the present disclosure do not limit this, for example, the specific driving behavior may include turning, lane changing, diverging, etc.

[0039] In an embodiment of the present disclosure, the execution entity may determine whether there is a collision conflict area between the autonomous vehicle and the obstacle vehicle on the target road section according to the planned trajectory information and the predicted trajectory information. When it is determined that there is no collision conflict area between the autonomous vehicle and the obstacle vehicle on the target road section, the process of the vehicle control method based on road section risk detection may be ended. When it is determined that there is a collision conflict area between the autonomous vehicle and the obstacle vehicle on the target road section, it may be determined that there is an obvious conflict on the target road section. Further, according to the planned trajectory information, it may be determined whether the autonomous vehicle has a specific driving behavior. When it is determined that the autonomous vehicle does not have a specific driving behavior, it may be determined that the risk type of the target road section is the first type of risk. When it is determined that the autonomous vehicle has a specific driving behavior, it may be determined that the risk type of the target road section is the second type of risk.

[0040] By detecting the specific driving behavior of the autonomous vehicle in this embodiment, the cause of the collision conflict can be determined, and the behaviors with risks of the autonomous vehicle can be focused on, which is beneficial to controlling the autonomous vehicle and reducing the risk of vehicle collision conflicts of the autonomous vehicle.

[0041] In some alternative embodiments of the present disclosure, the distance between two vehicles approaching a certain distance at the same time may be defined as a vehicle collision conflict. A safety distance may be preset in advance. When the distance between two vehicles at the same time is less than the preset safety distance, it may be considered that these two vehicles have a collision conflict. The vehicle conflict area may refer to the area where the distance between the trajectories of two vehicles is less than the safety distance at the same time. Optionally, step S201 of determining whether there is a vehicle conflict area on the target road section based on the planned trajectory information and the predicted trajectory information may include the following steps: determining the distance between the planned trajectory and the predicted trajectory at the same time based on the coordinates of each trajectory point in the planned trajectory information and the coordinates of each trajectory point in the predicted trajectory information; determining whether the distance between the planned trajectory and the predicted trajectory at the same time is less than the preset safety distance; and in response to the distance between the planned trajectory and the predicted trajectory at the same time being less than the preset safety distance, determining that there is a vehicle conflict area on the target road section.

[0042] In an alternative example, the process of determining whether there is a vehicle conflict area on the target road section is as follows: a. Trajectory equations of two vehicles: Trajectory equation of vehicle 1: ; Trajectory equation of vehicle 2: 。

[0043] Among them, vehicle 1 may be an autonomous vehicle, and vehicle 2 may be an obstacle vehicle.

[0044] b. Distance between vehicle trajectories: The Euclidean distance between the trajectories of two vehicles: .

[0045] Among them, ( t ) is the abscissa of the trajectory point of vehicle 1, is the ordinate of the trajectory point of vehicle 1, is the abscissa of the trajectory point of vehicle 2, is the ordinate of the trajectory point of vehicle 2.

[0046] c. Vehicle conflict area: The vehicle conflict area can be determined by the following formula 1: (Formula 1) Among them, the condition for a collision conflict is , is an empty set, is the safety distance.

[0047] In some alternative embodiments of the present disclosure, the specific driving behavior of an autonomous vehicle can be narrowly defined as a driving state in which the planned trajectory of the autonomous vehicle fluctuates and the autonomous vehicle is prone to collision conflicts with other vehicles. For example, trajectory fluctuations caused by driving states such as lane change, diversion, lane changing, borrowing a lane to bypass, and turning at an intersection. It is possible to determine whether an autonomous vehicle has a specific driving behavior by performing change point detection on the planned trajectory information in the world coordinate system. Optionally, step S202 for determining whether an autonomous vehicle has a specific driving behavior based on the planned trajectory information may include the following steps: determining the curvature of each trajectory point based on the time information and position information of each trajectory point in the predicted trajectory information; respectively determining the standard score of each trajectory point based on the position information and curvature of each trajectory point; determining whether the standard score of each trajectory point is greater than a preset first threshold; in response to the standard score of a trajectory point among each trajectory point being greater than the preset first threshold, determining that the autonomous vehicle has a specific driving behavior.

[0048] In this embodiment, by using the method of change point detection in the world coordinate system, it is possible to accurately identify the specific driving behavior of the autonomous vehicle, providing support for controlling the autonomous vehicle and reducing the risk of vehicle collision conflicts of the autonomous vehicle.

[0049] In an alternative example, a smoothing algorithm using the standard score, also known as the Z-score, can be used to analyze the planned trajectory of the autonomous vehicle to determine whether the autonomous vehicle has a specific driving behavior. The process is as follows: a. Data preparation: Obtain the trajectory data of the vehicle, including the timestamp of the trajectory point, the abscissa x and the ordinate y in the world coordinate system; b. Calculate the curvature: Curvature is an important index used to describe the degree of bending of a trajectory. For the trajectory equation of a two-dimensional curve, its curvature κ can be calculated by the following formula 2: (Formula 2) Wherein, is the first derivative of x with respect to time, is the derivative of y with respect to time, is the second derivative of x with respect to time, is the second derivative of y with respect to time.

[0050] The first derivative and the second derivative can be calculated using the central difference method. The specific formulas are as follows: (Formula 3) (Formula 4) (Formula 5) (Formula 6) Where i is the trajectory point for calculating the derivative, and Δt is the time interval between two adjacent trajectory points.

[0051] c. Calculate the standard score: To identify outliers, the Z-score of each trajectory point can be calculated. The Z-score can represent the degree of deviation of each trajectory point from the mean value, and its calculation formula is as follows: (Formula 7) Where, X = { x , y , κ} is the coordinate and curvature of the trajectory point ,μ is the mean value of each trajectory point in the trajectory, σ is the standard deviation of each trajectory point in the trajectory. The Z-score is calculated for the x, y, and κ values of each trajectory point respectively.

[0052] d. Outlier judgment: Thresholds for x, y, and κ of the trajectory point can be preset to determine whether there are abnormal trajectory points in the trajectory. Its calculation formula is as follows: (Formula 8) Where is the threshold of the Z-score. The appropriate threshold of the Z-score can be selected according to the specific application scenario, and then the risk assessment of the corresponding scenario can be carried out.

[0053] Figure 3 shows a flowchart for determining the risk type of a target road section according to some other embodiments of the present disclosure. AsFigure 3 As shown, step S101 determines the risk type of the target road section based on at least one of the planned trajectory information of the target road section that the autonomous vehicle is to pass through, the predicted trajectory information of the obstacle vehicle on the target road section, and the historical vehicle trajectory information of the target road section, and may include the following steps: S301. Based on the predicted trajectory information and the historical vehicle trajectory information, determine whether the obstacle vehicle has abnormal driving behavior.

[0054] In response to the obstacle vehicle having abnormal driving behavior, perform operation S302; in response to the obstacle vehicle not having abnormal driving behavior, perform operation 303.

[0055] S302. Determine that the risk type of the target road section is the third type of risk.

[0056] S303. Based on the planned trajectory information and the predicted trajectory information, determine whether there is a vehicle conflict area on the target road section.

[0057] In response to there being a vehicle conflict area on the target road section, perform operation S304. Otherwise, end the process.

[0058] S304. Based on the planned trajectory information, determine whether the autonomous vehicle has a specific driving behavior.

[0059] In response to the autonomous vehicle not having a specific driving behavior, perform operation S305. In response to the autonomous vehicle having a specific driving behavior, perform operation S306.

[0060] S305. Determine that the risk type of the target road section is the first type of risk.

[0061] S306. Determine that the risk type of the target road section is the second type of risk.

[0062] In the embodiments of the present disclosure, a situation where the autonomous vehicle may collide with other vehicles except for specific driving behaviors can be defined as the first type of risk. A situation where the autonomous vehicle may collide with other vehicles with specific driving behaviors can be defined as the second type of risk. A situation where abnormal driving of other vehicles may collide with the autonomous vehicle can be defined as the third type of risk. For example, the behavior of other vehicles overtaking in reverse by borrowing the oncoming lane may cause a collision conflict with the autonomous vehicle.

[0063] In an embodiment of the present disclosure, the execution entity may determine whether an obstacle vehicle passes through a target section according to the predicted trajectory information and the historical vehicle trajectory information. When it is determined that the obstacle vehicle does not pass through the target section according to the historical vehicle trajectory of the target section, it may be determined that the obstacle vehicle has an abnormal driving behavior, and there is a potential conflict in the target section. It may be further determined that the risk type of the target section is the third type of risk. When it is determined that the obstacle vehicle passes through the target section according to the historical vehicle trajectory of the target section, it may be determined that the obstacle vehicle does not have an abnormal driving behavior, and it may be further determined whether there is a collision conflict area between an autonomous vehicle and the obstacle vehicle in the target section according to the planned trajectory information and the predicted trajectory information. Among them, for the description of analyzing whether there is a vehicle collision conflict area in the target section according to the planned trajectory information and the predicted trajectory information, reference may be made to Figure 2 the relevant content recorded therein, which will not be elaborated here.

[0064] By detecting the abnormal driving behavior of the obstacle vehicle in this embodiment, the potential risks existing in the target section can be identified, so that the autonomous vehicle can be effectively controlled according to the cause of the risk, which is beneficial to reducing the risk of vehicle collision conflict of the autonomous vehicle and ensuring the safe driving of the autonomous vehicle.

[0065] In some alternative embodiments of the present disclosure, the historical trajectory dataset of vehicle driving in the target section may be clustered to obtain each trajectory cluster of vehicle driving, and a vehicle through corridor is formed. The abnormal driving behavior of the obstacle vehicle may be defined as that the predicted trajectory of the obstacle vehicle does not travel according to any type of trajectory cluster of the vehicles passing through this section in the past. The Kernel Density Estimation (KDE) method may be used to evaluate whether the predicted trajectory information belongs to the known trajectory clusters to determine whether the obstacle vehicle has an abnormal driving behavior. Optionally, step S301 of determining whether the obstacle vehicle has an abnormal driving behavior based on the predicted trajectory information and the historical vehicle trajectory information may include the following steps: clustering the historical vehicle trajectory information to obtain each trajectory cluster of the historical vehicle trajectory information; performing kernel density estimation on each trajectory cluster respectively to obtain the probability density function of each trajectory cluster; determining the probability density value of the predicted trajectory information belonging to each trajectory cluster based on the probability density function of each trajectory cluster and the coordinates of each trajectory point in the predicted trajectory information; determining whether the probability density value of the predicted trajectory information belonging to each trajectory cluster is greater than a preset second threshold; in response to the probability density value of the predicted trajectory information belonging to each trajectory cluster being less than or equal to the preset second threshold, determining that the obstacle vehicle has an abnormal driving behavior.

[0066] In an optional example, the set of trajectory clusters obtained by clustering is ( ), where each clustering Contains multiple trajectories, and each trajectory can be represented by a series of point trajectories The process of determining whether an obstacle vehicle has abnormal driving behavior is as follows: a. Data preparation: Obtain the set of trajectory points of a trajectory and a type of trajectory cluster in the clustered trajectory clusters ; b. Kernel density estimation: Merge the trajectory points of all trajectories in the trajectory cluster into a set , and its calculation formula is as follows: (Formula 9) where is the th trajectory in the trajectory cluster j . Then, perform kernel density estimation on the set of each trajectory cluster to obtain the corresponding probability density function, and its calculation formula is as follows: (Formula 10) where is the number of trajectory points of the trajectory cluster , h is the bandwidth parameter, K ( u, v ) is the kernel function, and a two-dimensional Gaussian kernel function can be selected, and its calculation formula is as follows: (Formula 11) where , .

[0067] c. Evaluate the trajectory: For each trajectory point of trajectory A, calculate its probability density value in the trajectory cluster , and its calculation formula is as follows: (Formula 12) where m is the number of trajectory points in the trajectory A .

[0068] d. Trajectory discrimination: It can be expected to set a threshold τ. If ( A ) is greater than this threshold, it is considered that the trajectory A belongs to the trajectory cluster , and its calculation formula is as follows: if (Formula 13) e. Abnormality determination: Traverse the trajectory clusters of all clusters, and repeat steps a to d to determine whether the trajectory A belongs to each trajectory cluster.

[0069] If the trajectory A does not belong to any trajectory cluster, it is considered that the driving behavior of the trajectory A is abnormal.

[0070] In some alternative embodiments of the present disclosure, clustering the historical vehicle trajectory information to obtain each trajectory cluster of the historical vehicle trajectory information may include the following steps: aligning, segmenting, and sampling the historical vehicle trajectory information based on the planned trajectory information; clustering based on the coordinates of the sampling points of each segmented trajectory segment to obtain each trajectory cluster of the historical vehicle trajectory information. Optionally, K-means clustering may be used to cluster the historical vehicle trajectory information. Optionally, clustering based on the coordinates of the sampling points of each segmented trajectory segment to obtain each trajectory cluster of the historical vehicle trajectory information may include: traversing each clustering number in the preset set, performing K-means clustering with the corresponding clustering number based on the coordinates of the sampling points of each segmented trajectory segment, and determining the DB index of each trajectory cluster obtained by clustering; comparing the DB indexes of the K-means clustering of each clustering number, and taking the trajectory clusters obtained by the K-means clustering with the smallest DB index as each trajectory cluster of the historical vehicle trajectory information.

[0071] In an alternative example, to facilitate the comparison of historical vehicle trajectory information, the historical vehicle trajectory information may be aligned at a suitable position with the normal direction of the road vertical section as the positive direction according to the environmental road information. The environmental road information may be obtained from the server of the electronic map. The starting point of the shortest trajectory in the historical vehicle trajectory information in the same direction may be selected for alignment. As Figure 5 shown, taking an intersection as an example, among trajectory 501 and trajectory 502, trajectory 502 is relatively shorter, and the starting point of trajectory 502 is selected to align trajectory 501 and trajectory 502; among trajectory 502 and trajectory 503, trajectory 503 is relatively shorter, and the starting point of trajectory 503 is selected to align trajectory 502 and trajectory 503; among trajectory 504 and trajectory 505, trajectory 504 is relatively shorter, and the starting point of trajectory 504 is selected to align trajectory 504 and trajectory 505.

[0072] After aligning the historical vehicle trajectory information, the historical vehicle trajectory information can be segmented into N segments starting from the aligned position according to the reference line direction of the planned trajectory information, and each segmented trajectory segment is sampled respectively. Using the position coordinates of the sampling points as the features of the trajectory, the similarity between two trajectories can be measured by calculating the distances between the trajectory segments. Among them, the number and length of the trajectory segments obtained by segmenting the historical vehicle trajectory information can be set as needed, and the embodiments of the present disclosure do not limit this. The number of sampling points for sampling each trajectory segment can be set as needed, and the embodiments of the present disclosure do not limit this. For example, all historical trajectories of the target road section can be segmented into trajectory segments with the same number and length, and the same number of sampling points can be sampled for the trajectory segments obtained by segmenting all historical trajectories of the target road section.

[0073] In an optional example, the process of implementing K-means clustering on the historical trajectories of the target road section is as follows: a. Select the number of clusters K; b. Select the initial cluster centers: Randomly select K trajectories from the dataset as the initial cluster centers; c. Assign data points: For each trajectory, calculate its distances from all cluster centers and assign it to the cluster center with the closest distance. Repeat this process until all trajectories are assigned to the corresponding cluster centers; d. Update the cluster centers: After all trajectories are assigned to the corresponding cluster centers, update the cluster center of each class. The new cluster center is the average value of all trajectories within the class. For example, for a class containing trajectories A and B , the new cluster center is the average trajectory of trajectories A and B; e. Iteration: Repeat steps c and d until the cluster centers no longer change or change very little, indicating that the trajectory clusters obtained by clustering have converged, and calculate its DB index; f. Sampling: Traverse the number of clusters in the cluster number set, execute steps a to e, and select the clustering result obtained with the cluster number k with the smallest DB index as the final clustering result of the trajectories.

[0074] Optionally, for the distance metric to determine the similarity degree, the Euclidean distance can be used to calculate the similarity between multiple trajectory segments, and its calculation formula is as follows: (Formula 14) Among them, and are the coordinates of specific sampling points between two trajectories, N is the number of sampling points. For the distance between two trajectory clusters, the centroid distance of the two trajectory clusters can be used for measurement.

[0075] Optionally, the process of calculating the DB index for the trajectory clusters obtained by clustering is as follows: a. Calculate the scatter of each trajectory cluster: For the trajectory cluster , calculate the scatter of the internal trajectories ; b. Calculate the distance between trajectory clusters: For any two types of trajectory clusters and , calculate the distance between them , that is, the centroid distance between the two types of trajectory clusters; c. Calculate the similarity measure: For the trajectory clusters and , calculate their similarity , and its calculation formula is as follows: (Formula 15) d. Select the best similarity measure: For the trajectory cluster , find the type of trajectory cluster that is most similar to it, and record the similarity measure between the two as the DB index, and its calculation formula is as follows: (Formula 16) e. Calculate the final DB index: Calculate the average value of the DB indices of all trajectory clusters as the final DB index, and its calculation formula is as follows: (Formula 17) Among them, the DB index is used to judge the local optimal number of clusters. For a given trajectory data set X, it can be divided into k classes , for the trajectory cluster , is the number of samples in the trajectory cluster , is the centroid of the trajectory cluster , is the distance measure between the trajectory cluster and , ， is the average scatter of the trajectory cluster , usually defined as , is any trajectory in the trajectory cluster , is the number of trajectories in the trajectory cluster .

[0076] Figure 4 shows a flowchart for determining the risk type of the target road section according to some further embodiments of the present disclosure. As Figure 4As shown, step S101 determines the risk type of the target road section based on at least one of the planned trajectory information of the target road section that the autonomous vehicle is to pass through, the predicted trajectory information of the obstacle vehicle on the target road section, and the historical vehicle trajectory information of the target road section, and may include the following steps: S401. Based on the planned trajectory information and the predicted trajectory information, determine whether there is a vehicle conflict area on the target road section.

[0077] In response to the existence of a vehicle conflict area on the target road section, perform operation S402; in response to the non-existence of a vehicle conflict area on the target road section, perform operation S403.

[0078] S402. Based on the planned trajectory information, determine whether the autonomous vehicle has a specific driving behavior.

[0079] In response to the non-existence of a specific driving behavior of the autonomous vehicle, perform operation S404. In response to the existence of a specific driving behavior of the autonomous vehicle, perform operation S405.

[0080] S404. Determine that the risk type of the target road section is the first type of risk.

[0081] S405. Determine that the risk type of the target road section is the second type of risk.

[0082] S403. Based on the historical vehicle trajectory information, determine whether there is a conflict-prone area on the target road section.

[0083] In response to the existence of a conflict-prone area on the target road section, perform S406. Otherwise, end the process.

[0084] S406. Determine that the risk type of the target road section is the fourth type of risk.

[0085] In an embodiment of the present disclosure, the situation where the autonomous vehicle will have a collision conflict with other vehicles except for specific driving behaviors can be defined as the first type of risk. The situation where the autonomous vehicle will have a collision conflict with other vehicles with specific driving behaviors can be defined as the second type of risk. The situation where a road section is marked as a collision conflict-prone section according to the empirical data of the road but no possible collision conflict is found according to the current data can be defined as the fourth type of risk.

[0086] In an embodiment of the present disclosure, the executing entity may determine whether there is a collision conflict area between the autonomous vehicle and the obstacle vehicle on the target road section according to the planned trajectory information and the predicted trajectory information. When it is determined that there is a collision conflict area between the autonomous vehicle and the obstacle vehicle on the target road section, it may be determined that there is an obvious conflict on the target road section. Further, according to the planned trajectory information, it may be analyzed whether the autonomous vehicle has a specific driving behavior. When it is determined that the autonomous vehicle does not have a specific driving behavior, it may be determined that the risk type of the target road section is the first type of risk. When it is determined that the autonomous vehicle has a specific driving behavior, it may be determined that the risk type of the target road section is the second type of risk. When it is determined that there is no collision conflict area between the autonomous vehicle and the obstacle vehicle on the target road section, further, according to the historical vehicle trajectory information, it may be analyzed whether there is a conflict-prone area on the target road section. When it is determined that there is no conflict-prone area on the target road section, the process of the vehicle control method based on road section risk detection may be ended. When it is determined that there is a conflict-prone area on the target road section, it may be determined that there is a potential conflict on the target road section, and it may be determined that the risk type of the target road section is the fourth type of risk.

[0087] In this embodiment, the conflict-prone area is determined through the historical vehicle trajectory information, and the potential risks existing on the target road section can be identified, so that the autonomous vehicle can be effectively controlled according to the causes of the risks, which is beneficial to reducing the risk of vehicle collision conflicts of the autonomous vehicle and ensuring the safe driving of the autonomous vehicle.

[0088] In some alternative embodiments of the present disclosure, for the first type of risk or the second type of risk, the driving trajectory of the autonomous vehicle may be adjusted according to the fuzzy control strategy, and at the same time, a general risk prompt information of "pay attention to risks" may be generated for risk indication. Optionally, step 102 of determining the control strategy for the autonomous vehicle based on the risk type of the target road section may include the following steps: in response to the risk type of the target road section being the first type of risk or the second type of risk, based on the preset fuzzy rules, the distance and speed difference between the autonomous vehicle and the obstacle vehicle, determine the membership degrees of the output states of the autonomous vehicle as acceleration, deceleration, and constant speed; determine the centroid value of the membership degrees, and adjust the driving speed of the autonomous vehicle based on the centroid value.

[0089] In this embodiment, the fuzzy control strategy is adopted to adjust the driving speed of the autonomous vehicle, which can effectively reduce the risk of obvious vehicle collision conflicts on the target road section and ensure the safe driving of the autonomous vehicle.

[0090] Optionally, the preset fuzzy rules may include: If the distance is close and the speed difference is large, the output is to moderately accelerate or maintain the speed; If the distance is close and the speed difference is moderate, the output is to maintain the speed or moderately decelerate; If the distance is moderate and the speed difference is moderate, the output is to maintain speed; If the distance is far and the speed difference is moderate, the output is to moderately accelerate or maintain speed; If the distance is near and the speed difference is small, the output is to maintain speed or moderately decelerate; If the distance is moderate and the speed difference is small, the output is to moderately accelerate or maintain speed; If the distance is far and the speed difference is small, the output is to moderately accelerate or maintain speed.

[0091] In an optional example, since the characteristics of the first type of risk and the second type of risk are that the conflict process is clear, active safety processing can be carried out. The driving speed of the autonomous vehicle can be adjusted according to the fuzzy control strategy to avoid collision conflicts. The implementation process is as follows: a. Input variables: distance between vehicles and speed difference; b. Output variables: Acceleration: refers to the degree of vehicle acceleration; Deceleration: refers to the degree of vehicle deceleration; Maintain Speed: maintain the current speed under specific circumstances; Among them, the goal of control is to maintain a safe distance by adjusting the driving trajectory of the vehicle, and the output variables should be directly related to the goal of control.

[0092] c. Fuzzy membership function: For the distance between vehicles

[0093] Near: indicates that the distance between the autonomous vehicle and the collision vehicle is small, (Formula 18) Among them, is the upper limit of "Near" for the distance, is the distance between vehicles is the membership degree of the fuzzy set (Near) representing that the distance between the autonomous vehicle and the collision vehicle is small.

[0094] Medium: indicates that the distance between the autonomous vehicle and the collision vehicle is moderate, (Formula 19) Among them, is the central value of "Medium" for the distance, is the distance between vehicles is the membership degree of the fuzzy set (Medium) representing that the distance between the autonomous vehicle and the collision vehicle is moderate.

[0095] Far: It indicates a large distance between the autonomous vehicle and the collision vehicle. (Formula 20) Where is the lower limit of "Far" for the distance, is the distance between vehicles and is the membership degree of the fuzzy set (Far) representing a large distance between the autonomous vehicle and the collision vehicle.

[0096] d. Fuzzy membership function: For the speed difference, the speed difference is a relative value, which is the difference between the speeds of the autonomous vehicle and the collision vehicle, and its value is V .

[0097] Small: It indicates a small speed difference between the autonomous vehicle and the collision vehicle. (Formula 21) Where is the upper limit of "small" for the speed difference, is the speed difference and is the membership degree of the fuzzy set (small) representing a small speed difference between the autonomous vehicle and the collision vehicle.

[0098] Moderate: It indicates a moderate speed difference between the autonomous vehicle and the collision vehicle. (Formula 22) Where is the central value of "moderate" for the speed difference, is the speed difference and is the membership degree of the fuzzy set (moderate) representing a moderate speed difference between the autonomous vehicle and the collision vehicle.

[0099] Big: It indicates a large speed difference between the autonomous vehicle and the collision vehicle. (Formula 23) Where is the upper limit of "big" for the speed difference, is the speed difference and is the membership degree of the fuzzy set (big) representing a large speed difference between the autonomous vehicle and the collision vehicle.

[0100] e. Fuzzy rules: According to the membership functions of the output variable and the input variables, corresponding fuzzy rules are written.

[0101] Rule 1: If the distance is "close" and the speed difference is "big", then the output is "Accelerate moderately or maintain speed" (A-K).

[0102] Rule 2: If the distance is "close" and the speed difference is "moderate", then output "Maintain speed or decelerate moderately" (K-D).

[0103] Rule 3: If the distance is "moderate" and the speed difference is "moderate", then output "Maintain speed" (K).

[0104] Rule 4: If the distance is "far" and the speed difference is "moderate", then output "Accelerate moderately or maintain speed" (A-K).

[0105] Rule 5: If the distance is "close" and the speed difference is "small", then output "Maintain speed or decelerate moderately" (K-D).

[0106] Rule 6: If the distance is "moderate" and the speed difference is "small", then output "Accelerate moderately or maintain speed" (A-K).

[0107] Rule 7: If the distance is "far" and the speed difference is "small", then output "Accelerate moderately or maintain speed" (A-K).

[0108] f. Fuzzy sets: Normalize the fuzzy membership values of the speed difference and the distance.

[0109] The distance normalization results are as follows: (Equation 24) (Equation 25) (Equation 26) The speed difference normalization results are as follows: (Equation 27) (Equation 28) (Equation 29) Define the corresponding fuzzy sets for each output variable. The fuzzy sets can use membership functions to represent different states of these output variables. We assume that the above fuzzy rules and their corresponding outputs are uniform, then the calculation of the fuzzy membership degree is as follows.

[0110] Acceleration: Represents the state where the autonomous vehicle needs to accelerate moderately, and its membership μA (1) is: (Equation 30) Deceleration: Represents the state where the autonomous vehicle needs to decelerate moderately, and its membership μD( -1) is: (Formula 31) Maintain Speed: It indicates that the autonomous vehicle needs to continue driving at the current speed, and its membership degree μK (0) is as follows:

[0111] (Formula 32)

[0112] g. Defuzzification: The Center of Gravity (CoG) method is used for defuzzification to calculate the center of gravity position of the fuzzy output. Assume that the range of the membership function is [-1, 0, 1], and the corresponding membership values are as follows: Accelerate: (1)= ; Decelerate: (-1)= ; Maintain Speed: (0)= ; Center of gravity position can be calculated by the following formula: (Formula 33) The above formula can be simplified to the form of weighted average: (Formula 34) According to the calculation result and the approximate relationship with [-1, 0, 1], determine the driving strategy that the autonomous vehicle should choose. For example, ≈1, the autonomous vehicle should choose to accelerate.

[0113] h. Output control signal: Design the output variable of the fuzzy controller: , where is the output control signal. Set the expected speeds for acceleration, deceleration, and maintaining speed to be , the stop speed to be , and the current speed to be . Among them, is the set minimum creep speed, and the empirical value can be taken as 0.1. The degree of acceleration or deceleration is determined by the fuzzy center of gravity value , which indirectly affects the time to reach the target speed and thus affects the effect of acceleration or deceleration. ​

[0114] For acceleration, assume that the maximum acceleration of the autonomous vehicle is , then the autonomous vehicle accelerates from the current speed v to the desired speed The required time is , and this time is the fastest acceleration time. In addition, set the slowest acceleration time to , then the acceleration time interval is , so there is the following formula: (Formula 35) Similarly, for deceleration, assume that the maximum deceleration of the autonomous vehicle is , then the autonomous vehicle decelerates from the current speed v to the stop speed The required time is , and this time is the fastest braking time. In addition, set the slowest braking time to , then the deceleration time interval is , so there is the following formula: (Formula 36) For constant-speed driving, to avoid the problem of the vehicle stopping due to the current speed =0, the following formula is set: (Formula 37) Among them, is the output control signal.

[0115] In some alternative embodiments of the present disclosure, for the third type of risk, the autonomous vehicle can be controlled to decelerate and stop according to the strategy of stopping when the distance between the autonomous vehicle and the obstacle vehicle is less than or equal to the preset safety distance. At the same time, a risk prompt message of "Special Attention" can be generated for risk indication. Optionally, step 102 determines the control strategy for the autonomous vehicle based on the risk type of the target road section, which may include: in response to the risk type of the target road section being the third type of risk, controlling the autonomous vehicle to decelerate and stop based on the current speed of the autonomous vehicle, the current speed of the obstacle vehicle, the preset reduction coefficient, the current distance between the autonomous vehicle and the obstacle vehicle, and the preset safety distance.

[0116] In this embodiment, for the potential risks caused by the abnormal driving of the obstacle vehicle, by controlling the autonomous vehicle to decelerate and stop within the safe distance for observation, the risk of vehicle collision and conflict can be minimized to ensure the safe driving of the autonomous vehicle.

[0117] In an optional example, since the third type of risk is characterized by the uncertainty of the possible future conflict process, passive safety processing can be performed. Such risk data, namely the position and speed information of the vehicle, should be reported. Since the driving strategy of the obstacle vehicle deviates greatly from the driving data of the past vehicles, such as the possibility of illegal driving behaviors such as driving in the opposite direction, its future behavior cannot be rationally inferred. Therefore, the worst case scenario should be considered, that is, when the obstacle vehicle is at a safe distance from the autonomous driving vehicle. When the vehicle is within the range, the autonomous vehicle should stop and observe. The implementation process is as follows: For this type of deterministic problem, the current speed of the autonomous vehicle is v , the speed of the abnormally moving obstacle vehicle is The distance between vehicles is d (> ), then according to the deceleration of the abnormally moving obstacle vehicle, the travel distance reduction coefficient λ from the autonomous driving vehicle to the abnormally moving obstacle vehicle is reasonably set, , the calculation formula of the output control signal is as follows: (Formula 38) For example, if an abnormally moving obstacle vehicle slows down, you can set =1; if the abnormally moving obstacle vehicle does not slow down, then it can be set <1.

[0118] In some optional embodiments of the present disclosure, for the fourth type of risk, the autonomous driving vehicle can be controlled to travel at a limited speed according to the two-point boundary value optimal control (Optimal Bundary Value Problem, referred to as OBVP) strategy, and a risk warning message of "pay attention" can be generated to indicate the risk. Optionally, step 102 determines the control strategy for the autonomous driving vehicle based on the risk type of the target road section, and can include the following steps: in response to the risk type of the target road section being the fourth type of risk, the current position of the autonomous driving vehicle is taken as the first point, and the first state is determined based on the current speed and acceleration of the autonomous driving vehicle; the position closest to the conflict-prone area and the first point is taken as the second point, and the second state is determined based on the distance between the conflict-prone area and the first point and the preset speed limit; based on the first state and the second state, the state transition trajectory and state transition time of the quintic spline curve from the first point to the second point are determined to control the autonomous driving vehicle to travel at a limited speed.

[0119] In this implementation, for potential risks arising from conflict-prone areas, a two-point boundary value optimal control strategy is used to control the speed limit of the autonomous driving vehicle, which can effectively reduce the risk of vehicle collision conflicts and ensure the safe driving of the autonomous driving vehicle.

[0120] In an optional example, due to the characteristics of the fourth type of risk, it is known that there is a historical conflict-prone area ahead on the road section, but currently no vehicles with collision conflicts are observed. Therefore, preventive safety measures can be taken. The speed of the autonomous vehicle can be controlled according to the two-point boundary value optimal control strategy, and observations can be made to prevent danger. The implementation process is as follows: Construct a general two-point boundary value optimal control problem and determine the starting state as the current state , assuming that the conflict-prone area is m meters away from the autonomous vehicle, and the speed limit is , then the target state is . The five-degree spline curve is selected for the state transition process, and the state transition time is the output control signal. The calculation formula is as follows: (Formula 39) where is the fixed transfer time.

[0121] Figure 6 shows a schematic diagram of an application scenario of a vehicle control method based on road section risk detection according to an embodiment of the present disclosure. As Figure 6 shown, the specific steps for the autonomous vehicle to perform road section risk detection control are as follows: Step 1: Obtain the historical traffic data 611 of the target road section, including historical vehicle trajectory information and environmental road information; Step 2: Cluster the historical vehicle trajectory information according to the historical traffic data 611, and perform risk detection on the target road section according to the planned trajectory information 612 of the autonomous vehicle and the predicted trajectory information 613 of the obstacle vehicle, specifically including: Step 621: Identify the abnormal driving behaviors of the obstacle vehicles according to the predicted trajectory information and the historical vehicle trajectory information; Step 622: Determine whether there is a vehicle conflict area on the target road section according to the planned trajectory information and the predicted trajectory information; Step 623: Identify the specific driving behaviors of the autonomous vehicle according to the planned trajectory information; Step 624: Determine whether there is a conflict-prone area on the target road section according to the historical vehicle trajectory information; Step 3: Give a risk indication and determine the safety control strategy for the autonomous vehicle according to the risk type, specifically including: Step 631: For the abnormal driving behaviors of the obstacle vehicles, prompt "Pay special attention", decelerate and stop for observation, monitor the risk, and make preparations for emergency handling; Step 632: For the vehicle conflict area on the target road section, prompt "Pay attention to the risk", and take active response measures: adjust the driving trajectory of the autonomous vehicle; Step 633: For specific driving behaviors of the autonomous vehicle, prompt "Pay attention to risks" and take proactive response measures: adjust the driving trajectory of the autonomous vehicle; Step 634: For areas with frequent conflicts in the target section, prompt "Pay attention and observe" and take preventive speed limit measures: control the autonomous vehicle to drive at a speed limit.

[0122] An embodiment of the present disclosure also provides a vehicle control device based on road section risk detection. Figure 7 The block diagram of a vehicle control device 700 based on road section risk detection according to an embodiment of the present disclosure is shown. The vehicle control device 700 based on road section risk detection according to an embodiment of the present disclosure can execute the above-mentioned vehicle control method 100 based on road section risk detection. As Figure 7 shown, the vehicle control device 700 based on road section risk detection may include: A risk detection module 701, configured to determine the risk type of the target road section based on at least one of the planned trajectory information of the target road section that the autonomous vehicle is to pass through, the predicted trajectory information of the obstacle vehicle on the target road section, and the historical vehicle trajectory information of the target road section; A risk processing module 702, configured to determine the control strategy for the autonomous vehicle based on the risk type of the target road section.

[0123] In some optional embodiments, the risk detection module 701 is further configured to: Based on the planned trajectory information and the predicted trajectory information, determine whether there is a vehicle conflict area on the target road section; In response to the existence of a vehicle conflict area on the target road section, based on the planned trajectory information, determine whether the autonomous vehicle has a specific driving behavior; In response to the autonomous vehicle not having a specific driving behavior, determine that the risk type of the target road section is the first type of risk; In response to the autonomous vehicle having a specific driving behavior, determine that the risk type of the target road section is the second type of risk.

[0124] In some optional embodiments, the risk detection module 701 is further configured to: Based on the predicted trajectory information and the historical vehicle trajectory information, determine whether the obstacle vehicle has abnormal driving behavior; In response to the obstacle vehicle having abnormal driving behavior, determine that the risk type of the target road section is the third type of risk; In response to the obstacle vehicle not having abnormal driving behavior, execute determining whether there is a vehicle conflict area on the target road section based on the planned trajectory information and the predicted trajectory information.

[0125] In some alternative embodiments, the risk detection module 701 is further configured to: In response to the absence of a vehicle conflict area on the target road section, based on the historical vehicle trajectory information, determine whether there is a conflict-prone area on the target road section; In response to the presence of a conflict-prone area on the target road section, determine that the risk type of the target road section is the fourth type of risk.

[0126] In some alternative embodiments, the risk detection module 701 is further configured to: Based on the time information and position information of each trajectory point in the predicted trajectory information, determine the curvature of each trajectory point; Based on the position information and curvature of each trajectory point, determine the standard score of each trajectory point; Determine whether the standard score of each trajectory point is greater than a preset first threshold; In response to the standard score of a trajectory point among each trajectory point being greater than the preset first threshold, determine that the autonomous vehicle has a specific driving behavior.

[0127] In some alternative embodiments, the risk detection module 701 is further configured to: Cluster the historical vehicle trajectory information to obtain each trajectory cluster of the historical vehicle trajectory information; Perform kernel density estimation on each trajectory cluster respectively to obtain the probability density function of each trajectory cluster; Based on the probability density function of each trajectory cluster and the coordinates of each trajectory point in the predicted trajectory information, determine the probability density value of the predicted trajectory information belonging to each trajectory cluster; Determine whether the probability density value of the predicted trajectory information belonging to each trajectory cluster is greater than a preset second threshold; In response to the probability density value of the predicted trajectory information belonging to each trajectory cluster being less than or equal to the preset second threshold, determine that the obstacle vehicle has an abnormal driving behavior.

[0128] In some alternative embodiments, the risk detection module 701 is further configured to: Align, segment, and sample the historical vehicle trajectory information based on the planned trajectory information; Cluster based on the coordinates of the sampling points of each trajectory segment obtained by segmentation to obtain each trajectory cluster of the historical vehicle trajectory information.

[0129] In some alternative embodiments, the risk detection module 701 is further configured to: Traverse each clustering quantity in the preset set, perform K-means clustering with the corresponding clustering quantity based on the coordinates of the sampling points of each trajectory segment obtained by segmentation, and determine the DB index of each trajectory cluster obtained by clustering. Compare the DB indices of the K-means clusterings with different clustering quantities, and use the trajectory clusters obtained by the K-means clustering with the smallest DB index as the trajectory clusters of the historical vehicle trajectory information.

[0130] In some optional implementation manners, the risk detection module 701 is further configured to: Based on the coordinates of each trajectory point in the planned trajectory information and the coordinates of each trajectory point in the predicted trajectory information, determine the distance between the planned trajectory and the predicted trajectory at the same moment. Determine whether the distance between the planned trajectory and the predicted trajectory at the same moment is less than a preset safety distance. In response to the distance between the planned trajectory and the predicted trajectory at the same moment being less than the preset safety distance, determine that there is a vehicle conflict area in the target road section.

[0131] In some optional implementation manners, the risk handling module 702 is further configured to: In response to the risk type of the target road section being the first type of risk or the second type of risk, based on preset fuzzy rules, the distance between the autonomous driving vehicle and the obstacle vehicle, and the speed difference, determine the membership degrees of the output states of the autonomous driving vehicle as accelerating, decelerating, and maintaining a constant speed. Determine the centroid value of the membership degrees, and adjust the driving speed of the autonomous driving vehicle based on the centroid value.

[0132] In some optional implementation manners, the risk handling module 702 is further configured to: In response to the risk type of the target road section being the third type of risk, based on the current speed of the autonomous driving vehicle, the current speed of the obstacle vehicle, a preset reduction coefficient, the current distance between the autonomous driving vehicle and the obstacle vehicle, and the preset safety distance, control the autonomous driving vehicle to decelerate and stop.

[0133] In some optional implementation manners, the risk handling module 702 is further configured to: In response to the risk type of the target road section being the fourth type of risk, use the current position of the autonomous driving vehicle as the first point, and determine the first state based on the current speed and acceleration of the autonomous driving vehicle. Use the position closest to the first point in the conflict-prone area as the second point, and determine the second state based on the distance between the conflict-prone area and the first point and the preset speed limit. Based on the first state and the second state, determine the quintic spline curve state transition trajectory and the state transition time from the first point to the second point, and control the autonomous vehicle to travel at a speed limit.

[0134] In addition, an embodiment of the present disclosure also provides an electronic device, which includes at least one processor and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the above-mentioned vehicle control method 100 based on road section risk detection.

[0135] An embodiment of the present disclosure also provides a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, the above-mentioned vehicle control method 100 based on road section risk detection is implemented.

[0136] An embodiment of the present disclosure also provides a computer program product, including a computer program, and when the computer program is executed by a processor, the above-mentioned vehicle control method 100 based on road section risk detection is implemented.

[0137] Figure 8 The block diagram of an electronic device 800 showing an example that can be used to implement the embodiments of the present disclosure is shown. The electronic device 800 is intended to represent various forms of digital computers. The electronic device 800 can also represent various forms of mobile devices that can run computing programs. The components shown herein, their connections and relationships, and their functions are only examples and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0138] As Figure 8 shown, the electronic device 800 includes a computing unit 810, which can execute various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 820 or the computer program loaded from the storage unit 880 into the random access memory (RAM) 830. In the RAM 830, various programs and data required for the operation of the electronic device 800 can also be stored. The computing unit 810, the ROM 820, and the RAM 830 are connected to each other through a bus 840. The input / output (I / O) interface 850 is also connected to the bus 840.

[0139] Multiple components in the electronic device 800 are connected to the I / O interface 850, including: an input unit 860, such as a touch screen, etc.; an output unit 870, such as various types of displays, speakers, etc.; a storage unit 880, such as a disk, etc.; and a communication unit 890, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 890 allows the electronic device 800 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0140] The computing unit 810 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 810 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 810 executes the various methods and processes described above, such as the vehicle control method 100 based on road segment risk detection. For example, in some embodiments, the vehicle control method 100 based on road segment risk detection can be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as the storage unit 880. In some embodiments, part or all of the computer program can be loaded and / or installed onto the electronic device 800 via the ROM 820 and / or the communication unit 890. When the computer program is loaded into the RAM 830 and executed by the computing unit 810, one or more steps of the vehicle control method 100 based on road segment risk detection described above can be executed. Alternatively, in other embodiments, the computing unit 810 can be configured to execute the vehicle control method 100 based on road segment risk detection in any other suitable manner (e.g., by means of firmware).

[0141] Various embodiments of the systems and techniques described above in this document can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGA), application-specific integrated circuits (ASIC), application-specific standard products (ASSP), systems-on-chip (SOC), complex programmable logic devices (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: implemented in one or more computer programs, the one or more computer programs can be executed and / or interpreted on a programmable system including at least one programmable processor, the programmable processor can be a special or general programmable processor, can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit the data and instructions to the storage system, the at least one input device, and the at least one output device.

[0142] The program code for implementing the methods of the present disclosure can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing devices, such that when the program codes are executed by the processor or controller, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The program codes can be executed entirely on the machine, partially on the machine, executed partially on the machine and partially on a remote machine as an independent software package, or executed entirely on a remote machine or server.

[0143] In the context of the present disclosure, a machine-readable medium can be a tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0144] It should be understood that various forms of the flow shown above can be used, steps can be reordered, added, or deleted. For example, the steps recited in the present disclosure can be executed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions disclosed in the present disclosure can be achieved, and no limitations are imposed herein.

[0145] The above specific embodiments do not constitute a limitation on the protection scope of the present disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent replacements, and improvements made within the spirit and principles of the present disclosure shall be included within the protection scope of the present disclosure.

Claims

1. A vehicle control method based on road section risk detection, characterized in that: include: Determine the risk type of the target road section based on at least one of planned trajectory information of the target road section that the autonomous driving vehicle is to pass, predicted trajectory information of obstacle vehicles on the target road section, and historical vehicle trajectory information of the target road section; A control strategy for the autonomous driving vehicle is determined based on the risk type of the target road section.

2. The method according to claim 1, characterized in that The determining the risk type of the target road section based on at least one of planned trajectory information of the target road section to be passed by the autonomous driving vehicle, predicted trajectory information of an obstacle vehicle on the target road section, and historical vehicle trajectory information of the target road section includes: Based on the planned trajectory information and the predicted trajectory information, determining whether there is a vehicle conflict area on the target road section; In response to the presence of a vehicle conflict area on the target road section, determining whether the autonomous driving vehicle has a specific driving behavior based on the planned trajectory information; In response to the autonomous driving vehicle not having a specific driving behavior, determining that the risk type of the target road section is a first type of risk; In response to the autonomous driving vehicle having a specific driving behavior, it is determined that the risk type of the target road section is a second type of risk.

3. The method according to claim 2, characterized in that The step of determining the risk type of the target road section based on at least one of the planned trajectory information of the target road section to be traversed by the autonomous driving vehicle, the predicted trajectory information of the obstacle vehicle on the target road section, and the historical vehicle trajectory information of the target road section further includes: Based on the predicted trajectory information and the historical vehicle trajectory information, determining whether the obstacle vehicle has abnormal driving behavior; In response to the abnormal driving behavior of the obstacle vehicle, the risk type of the target road section is determined to be a third type of risk.

4. The method according to claim 3, characterized in that The step of determining the risk type of the target road section based on at least one of the planned trajectory information of the target road section to be traversed by the autonomous driving vehicle, the predicted trajectory information of the obstacle vehicle on the target road section, and the historical vehicle trajectory information of the target road section further includes: In response to the target road section not having a vehicle conflict area, determining whether the target road section has a conflict-prone area based on the historical vehicle trajectory information; In response to the existence of a conflict-prone area in the target road section, the risk type of the target road section is determined to be a fourth risk type.

5. The method according to any one of claims 2 to 4, characterized in that: The determining, based on the planned trajectory information, whether the autonomous driving vehicle has a specific driving behavior includes: Determining the curvature of each trajectory point based on the time information and the position information of each trajectory point in the predicted trajectory information; Determining a standard score for each trajectory point based on the position information and curvature of each trajectory point; Determining whether the standard score of each trajectory point is greater than a preset first threshold; In response to the existence of a trajectory point among the trajectory points having a standard score greater than a preset first threshold, it is determined that the autonomous driving vehicle has a specific driving behavior.

6. The method according to claim 3 or 4, characterized in that: The determining whether the obstacle vehicle has abnormal driving behavior based on the predicted trajectory information and the historical vehicle trajectory information includes: Clustering the historical vehicle trajectory information to obtain trajectory clusters of the historical vehicle trajectory information; Performing kernel density estimation on each of the trajectory clusters to obtain a probability density function of each of the trajectory clusters; Determining a probability density value that the predicted trajectory information belongs to each trajectory cluster based on the probability density function of each trajectory cluster and the coordinates of each trajectory point in the predicted trajectory information; Determining whether a probability density value of the predicted trajectory information belonging to each trajectory cluster is greater than a preset second threshold; In response to the probability density values ​​of the predicted trajectory information belonging to each trajectory cluster being less than or equal to a preset second threshold, it is determined that the obstacle vehicle has abnormal driving behavior.

7. The method according to claim 6, characterized in that The clustering of the historical vehicle trajectory information to obtain trajectory clusters of the historical vehicle trajectory information includes: Aligning, segmenting and sampling the historical vehicle trajectory information based on the planned trajectory information; Clustering is performed based on the coordinates of the sampling points of each trajectory segment obtained by segmentation to obtain trajectory clusters of the historical vehicle trajectory information.

8. The method according to claim 7, characterized in that The coordinates of the sampling points of each trajectory segment obtained based on segmentation are clustered to obtain each trajectory cluster of the historical vehicle trajectory information, including: Traverse each cluster number in the preset set, perform K-means clustering of the corresponding cluster number based on the coordinates of the sampling points of each trajectory segment obtained by segmentation, and determine the DB index of each trajectory cluster obtained by clustering; The DB indexes of the K-means clusterings of the clustering numbers are compared, and the trajectory clusters obtained by the K-means clustering of the clustering number with the smallest DB index are used as the trajectory clusters of the historical vehicle trajectory information.

9. The method according to any one of claims 2 to 4, characterized in that: The determining whether there is a vehicle conflict area in the target road section based on the planned trajectory information and the predicted trajectory information includes: Determine the distance between the planned trajectory and the predicted trajectory at the same time based on the coordinates of each trajectory point in the planned trajectory information and the coordinates of each trajectory point in the predicted trajectory information; Determine whether the distance between the planned trajectory and the predicted trajectory at the same time is less than the preset safety distance; In response to a distance between the planned trajectory and the predicted trajectory at the same time being less than a preset safety distance, it is determined that a vehicle conflict area exists in the target road section.

10. The method according to any one of claims 2 to 4, characterized in that: The determining of a control strategy for the autonomous driving vehicle based on the risk type of the target road section includes: In response to the risk type of the target road section being the first risk type or the second risk type, determining, based on a preset fuzzy rule, a distance and a speed difference between the autonomous driving vehicle and the obstacle vehicle, a membership degree of acceleration, deceleration and uniform speed as an output state of the autonomous driving vehicle; Determine a center of gravity value of the membership degree, and adjust a driving speed of the autonomous driving vehicle based on the center of gravity value.

11. The method according to claim 3 or 4, characterized in that: The determining of a control strategy for the autonomous driving vehicle based on the risk type of the target road section includes: In response to the risk type of the target road section being the third type of risk, the autonomous driving vehicle is controlled to slow down and stop based on the current speed of the autonomous driving vehicle, the current speed of the obstacle vehicle, a preset reduction factor, the current distance between the autonomous driving vehicle and the obstacle vehicle, and a preset safety distance.

12. The method according to claim 4, characterized in that The determining of a control strategy for the autonomous driving vehicle based on the risk type of the target road section includes: In response to the risk type of the target road segment being the fourth risk type, taking the current position of the autonomous driving vehicle as a first point, and determining a first state based on the current speed and acceleration of the autonomous driving vehicle; The position closest to the first point and the conflict-prone area is taken as the second point, and the second state is determined based on the distance between the conflict-prone area and the first point and the preset speed limit; Based on the first state and the second state, a quintic spline curve state transition trajectory and a state transition time from the first point to the second point are determined to control the automatic driving vehicle to travel at a limited speed.

13. A vehicle control device based on road section risk detection, characterized in that: include: a risk detection module configured to determine a risk type of a target road section to be traversed by the autonomous driving vehicle based on at least one of planned trajectory information of the target road section, predicted trajectory information of an obstacle vehicle on the target road section, and historical vehicle trajectory information of the target road section; The risk processing module is configured to determine a control strategy for the autonomous driving vehicle based on the risk type of the target road section.

14. An electronic device, characterized in that: include: at least one processor; as well as, A memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the vehicle control method based on road section risk detection as described in any one of claims 1 to 12.

15. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the vehicle control method based on road section risk detection as described in any one of claims 1 to 12 is implemented.

16. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the vehicle control method based on road section risk detection as described in any one of claims 1 to 12 is implemented.

Images