CN114162133A

CN114162133A - Risk assessment method and device for driving scene and computer readable storage medium

Info

Publication number: CN114162133A
Application number: CN202111324305.4A
Authority: CN
Inventors: 李怡康; 黄思渊; 黄文锋
Original assignee: Shanghai AI Innovation Center
Current assignee: Shanghai AI Innovation Center
Priority date: 2021-11-10
Filing date: 2021-11-10
Publication date: 2022-03-11
Anticipated expiration: 2041-11-10
Also published as: CN114162133B

Abstract

The embodiment of the application discloses a risk assessment method and device for a driving scene and a computer-readable storage medium. The method comprises the following steps: acquiring driving parameters and objects in a sensing range of a vehicle; and determining scene risk characteristics comprising different types of characteristics according to the driving parameters of the object and the driving parameters of the vehicle so as to adapt to high dynamics and diversity of driving scenes. And scene risk characteristics are predicted from the angle of the vehicle, so that the accuracy of the scene risk characteristics is improved. Respectively determining a time sequence risk factor and a space risk factor according to different characteristics and characteristic combinations in the scene risk characteristics; and the scene risk degree in the space dimension and the time dimension is quantified by combining various scene risk characteristics, so that the accuracy of the risk factor is improved. By fusing the time sequence risk factor, the space risk factor and the risk factors of the two dimensions, the scene risk is comprehensively evaluated, and the accuracy of the scene risk index is improved.

Description

Risk assessment method and device for driving scene and computer readable storage medium

Technical Field

The present application relates to the field of computer technologies, and in particular, to a method and an apparatus for risk assessment in a driving scenario, and a computer-readable storage medium.

Background

With the progress of science and technology and the continuous improvement of the living standard of people, the demand of vehicles is more and more increased. The increasing number of traffic participants on roads, including but not limited to vehicles, pedestrians, obstacles, traffic lights, unknown objects, etc., makes road conditions complex and diversified. Therefore, the risk of the driving scene is evaluated, better driving experience can be provided for the driver, and safe and reliable information basis can be provided for the automatic driving system.

In the prior art, the risk assessment result of the driving scene is obtained by collecting historical accident statistical data on a road and combining a high-precision map to analyze and count the driving scene.

However, the risk assessment method in the prior art depends on a high-precision map and long-time historical accident statistical data, does not consider the complexity and diversity of real-time roads and traffic participants, and for example, for the situations of traffic accident burst, rush hour on duty and the like, the real-time scene information of the roads cannot be accurately reflected, and the accuracy of risk assessment of a driving scene is reduced.

Disclosure of Invention

The embodiment of the application expects to provide a risk assessment method and device of a driving scene and a computer-readable storage medium, starting from a vehicle angle, calculating scene risk characteristics by combining driving parameters, quantifying scene risk degree in space dimension and time dimension by combining various scene risk characteristics, fusing risk factors of the two dimensions, and comprehensively assessing scene risks, so that the accuracy of the driving scene risk assessment is improved.

The technical scheme of the embodiment of the application is realized as follows:

in a first aspect, an embodiment of the present application provides a risk assessment method for a driving scenario, where the method includes: acquiring driving parameters of a self vehicle and driving parameters of an object in a self vehicle sensing range; determining scene risk characteristics according to the driving parameters of the object and the driving parameters of the self-vehicle, wherein the scene risk characteristics comprise at least one characteristic of a longitudinal safe distance, a transverse safe distance, a safe region coincidence factor and a future collision probability; respectively determining a time sequence risk factor and a space risk factor according to different characteristics and characteristic combinations in the scene risk characteristics, wherein the time sequence risk factor reflects the scene risk degree in a time dimension, and the space risk factor reflects the scene risk degree in a space dimension; and fusing the time sequence risk factor and the space risk factor to obtain a scene risk index.

In a second aspect, an embodiment of the present application provides a risk assessment apparatus for a driving scenario, where the apparatus includes: the acquisition module is used for acquiring the driving parameters of the self-vehicle and the driving parameters of the object in the sensing range of the self-vehicle; the scene risk characteristic module is used for determining scene risk characteristics according to the driving parameters of the object and the driving parameters of the self-vehicle, wherein the scene risk characteristics comprise at least one characteristic of a longitudinal safe distance, a transverse safe distance, a safe region overlapping factor and a future collision probability; the risk factor module is used for respectively determining a time sequence risk factor and a space risk factor according to different characteristics and characteristic combinations in the scene risk characteristics, wherein the time sequence risk factor reflects the scene risk degree in the time dimension, and the space risk factor reflects the scene risk degree in the space dimension; and the evaluation module is used for fusing the time sequence risk factor and the space risk factor to obtain a scene risk index.

In a third aspect, an embodiment of the present application provides a risk assessment apparatus for a driving scenario, where the apparatus includes a memory for storing executable instructions, and a processor for implementing the risk assessment method for the driving scenario when the executable instructions stored in the memory are executed.

In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, on which executable instructions are stored, and when the computer-readable storage medium is executed by a processor, the method for risk assessment of a driving scenario is implemented.

The embodiment of the application provides a risk assessment method and device for a driving scene and a computer-readable storage medium. According to the scheme provided by the embodiment of the application, the driving parameters and the objects in the sensing range of the vehicle are obtained; determining scene risk characteristics according to the driving parameters of the object and the driving parameters of the vehicle, wherein the scene risk characteristics comprise at least one of longitudinal safe distance, transverse safe distance, safe region coincidence factors and future collision probability; the accuracy of the scene risk characteristics is improved by directly constructing the scene of the self-vehicle, calculating the coincidence factors of the safe distance and the safe region from the angle of the self-vehicle and predicting the future collision probability. Respectively determining a time sequence risk factor and a space risk factor according to different characteristics and characteristic combinations in the scene risk characteristics, wherein the time sequence risk factor reflects the scene risk degree in a time dimension, and the space risk factor reflects the scene risk degree in a space dimension; in consideration of the change of scene risks brought by different driving scenes, the scene risk characteristics including different types of characteristics are determined through the driving parameters so as to adapt to the high dynamic and diversity of the driving scenes, the scene risk degree in space dimension and time dimension is quantized by combining various scene risk characteristics, and the accuracy of risk factors is improved. By fusing the time sequence risk factor, the space risk factor and the risk factors of the two dimensions, the scene risk is comprehensively evaluated, and the accuracy of the scene risk index is improved.

Drawings

Fig. 1 is a flowchart illustrating optional steps of a risk assessment method for a driving scenario according to an embodiment of the present disclosure;

FIG. 2 is an exemplary schematic diagram of travelable areas for different object types provided by an embodiment of the present application;

fig. 3 is an exemplary schematic diagram of a selection rule of a spatial risk factor according to an embodiment of the present application;

FIG. 4 is a flowchart illustrating optional steps of another method for risk assessment in a driving scenario according to an embodiment of the present disclosure;

fig. 5 is a schematic structural diagram of a risk assessment device for a driving scenario according to an embodiment of the present application;

fig. 6 is a schematic structural diagram of a risk assessment device for a driving scenario according to an embodiment of the present application.

Detailed Description

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. It should be understood that some of the embodiments described herein are only for explaining the technical solutions of the present application, and are not intended to limit the technical scope of the present application.

In order to better understand the risk assessment method for the driving scenario provided in the embodiment of the present application, prior to introducing the technical solution of the embodiment of the present application, a description is given to a related technology.

In the related technology, the intersection scene is modeled by a roadside camera and remote sensing data, and long-time historical statistics of data such as accidents and high-precision maps is required. The reaction time of the driver, during which the vehicle is out of control, is not taken into account when an accident is encountered. However, the change of the scene risk often changes in real time along with the change of the scene, a dynamic interaction process is not started from a self-vehicle angle in the related technology, the map and long-time historical statistical data of a specific area are relied on, the generalization capability is not strong, and the accuracy of the risk assessment of the driving scene is reduced.

The present application actually provides a risk assessment method for a driving scenario, as shown in fig. 1, fig. 1 is a flowchart illustrating steps of the risk assessment method for a driving scenario provided in an embodiment of the present application, where the risk assessment method for a driving scenario includes the following steps:

s101, obtaining the driving parameters of the vehicle and the driving parameters of the object in the sensing range of the vehicle.

In the present embodiment, the driving parameters characterize factors that may affect objects that are driving or stationary on the road. The driving parameters of the self vehicle can include but are not limited to speed, acceleration, heading angle, type of the self vehicle and shape of the self vehicle, and the driving parameters of the self vehicle can be acquired by a power system of the self vehicle.

In the present embodiment, the driving parameters of the object include, but are not limited to, speed, acceleration, heading angle, object type, object shape, and object position relative to the vehicle (in front of or behind the vehicle). The object types include, among others, dynamic (bicycles, cars, pedestrians) and static (obstacles, traffic lights, unknown objects), and the cars include, but are not limited to, cars, off-road vehicles, trucks, buses, taxis, ambulances, motorcycles, and electric vehicles. When the driving parameters of the object are obtained, the driving parameters can be obtained, for example, by a sensor network and a perception algorithm of the vehicle, wherein vehicle positioning, static obstacle mapping, moving obstacle detection and tracking, road mapping, traffic signal detection and identification and the like can be realized by the perception algorithm. The radar speed measuring device and the camera of the traffic monitoring system are used for monitoring objects on a road to obtain running parameters of the objects, the traffic monitoring system adopts the radar speed measuring device and the camera to determine the sensing range of a vehicle, and the running parameters of the objects in the sensing range of the vehicle are sent to the vehicle through the wireless module. Or the self-vehicle determines the sensing range of the self-vehicle through a sensing algorithm, and acquires the driving parameters of the corresponding object from the traffic monitoring system through the wireless module. The embodiment of the present application is not limited as long as the driving parameters of the object can be acquired.

It should be noted that the own vehicle may also be referred to as the own vehicle, the vehicle to be evaluated, the vehicle to be analyzed, the driving vehicle, and the like, and the object may also be referred to as a traffic participant, which is not limited in this embodiment of the present application. The sensing range of the vehicle can be set by a person skilled in the art according to actual needs, or can be calculated by a sensing system of the vehicle according to real-time road parameters (weather, road conditions, information of communication parameters on the road, and driving time) by adopting a sensing algorithm. The embodiment of the application is not limited, as long as the effectiveness of the collected driving parameters of the object can be ensured.

In the embodiment of the application, the objects on the road are very many, some objects are far away from the self-vehicle, or the future movement of the self-vehicle is not influenced, and the driving parameters of all the objects on the road do not need to be acquired. Therefore, by obtaining the driving parameters of the object in the sensing range of the vehicle, the data processing amount is reduced, and the calculation efficiency is improved.

S102, determining scene risk characteristics according to the driving parameters of the object and the driving parameters of the vehicle, wherein the scene risk characteristics comprise at least one characteristic of a longitudinal safe distance, a transverse safe distance, a safe region coincidence factor and a plurality of future collision probabilities.

In the embodiment of the application, in the face of high dynamic and diversity of driving scenes, the evaluation device of the embodiment of the application does not depend on the main parameters, high-precision maps and long-time accident statistical data of a driver, and directly constructs the scenes of the self-vehicle. From the angle of the vehicle, the motion states (namely driving parameters) of different types of traffic participants in the vehicle and the perception range of the vehicle are utilized to calculate scene risk characteristics. The scene risk characteristics are used for characterizing characteristics which can influence the risk between the own vehicle and the object, and include at least one of characteristics of longitudinal safe distance, transverse safe distance, safe region coincidence factor and future collision probability.

It should be noted that the longitudinal safe distance represents the longitudinal required safe distance, and the larger the required distance is, the more likely the collision between two objects in the transverse direction is, i.e. the more dangerous. The lateral safe distance represents the lateral required safe distance, and the larger the required distance is, the more probable the collision between two objects in the lateral direction is, namely, the more dangerous. The safety region coincidence factor represents the size of the coincidence region between two objects, and a larger coincidence value indicates a higher risk. The future collision probability represents the probability of collision of two objects under the premise that the two objects keep the current motion state within a plurality of seconds in the future, and the larger the probability value is, the more dangerous the object is.

In the embodiment of the application, the scene risk characteristics including different types of characteristics are determined through the driving parameters in consideration of the change of the scene risk caused by different driving scenes, so that the method is suitable for the high dynamic and diversity of the driving scenes, the calculation of the safe distance and the safe region overlapping factor is carried out from the angle of the vehicle, the prediction of the future collision probability is carried out, and the accuracy of the scene risk characteristics is improved.

S103, respectively determining a time sequence risk factor and a space risk factor according to different characteristics and characteristic combinations in the scene risk characteristics, wherein the time sequence risk factor reflects the scene risk degree in the time dimension, and the space risk factor reflects the scene risk degree in the space dimension.

In the embodiment of the application, the scene risk characteristics include a plurality of different characteristics, and a part of the characteristics can be selected as the spatial risk factors according to a preset rule. Illustratively, taking an expression form that a longitudinal safe distance, a transverse safe distance, a safe region coincidence factor and a future collision probability are numerical values as an example, when the space risk factor is determined, for each feature, the feature larger than a preset value is selected; or selecting a preset number of features; or, removing the maximum value, removing the minimum value, and selecting the characteristic at the intermediate value; or selecting some representative characteristics; or, a plurality of numerical values are directly used as the space risk factors without selection. As for the selection manner of the spatial risk factor, the embodiment of the present application is not limited as long as the spatial risk factor can reflect the scene risk degree in the spatial dimension.

In the embodiment of the application, if a part of the features are selected as the spatial risk factors according to the preset rule, the number of the features corresponding to the selected part of the features can be reduced, and then the data calculation amount is reduced and the calculation efficiency is improved when the scene risk index is calculated according to the spatial risk factors. If each feature in the scene risk features is directly used as a space risk factor, the comprehensiveness and accuracy of the space risk factor are improved.

In the embodiment of the application, the scene risk characteristics corresponding to each moment are stored, and the scene risk characteristics correspond to a plurality of different moments. In determining the time-series risk factor, for example, a plurality of preset number (e.g., 6) of scene risk features at different times are selected, or a plurality of interval times (e.g., 1 st time, 3 rd time, 5 th time) of scene risk features are selected, or a plurality of different times within a preset time period (e.g., within 5 s) are selected. Then, for each feature, features at a plurality of times are combined, for example, an average value, a maximum value, a minimum value, a change rate, and the like are calculated, and a time-series risk factor is generated.

In the embodiment of the application, the determined spatial risk factors can quantify the scene risk degree in the spatial dimension by combining the features of the scene risk features at different moments, so that the scene risk factors are fused in the follow-up process, and the accuracy of the scene risk index is improved.

And S104, fusing the time sequence risk factor and the space risk factor to obtain a scene risk index.

For example, taking an expression form in which the longitudinal safe distance, the transverse safe distance, the safe region overlapping factor, and the future collision probability are numerical values, the obtained time sequence risk factor and the space risk factor are also numerical expression forms, and therefore, the time sequence risk factor and the space risk factor need to be weighted and fused to obtain a scene risk index, and the risk index can also be understood as a risk score. If the time sequence risk factor and the space risk factor respectively comprise the characteristics of the longitudinal safe distance, the transverse safe distance, the safe region overlapping factor, the future collision probability and the like, the characteristic weight is respectively set for the longitudinal safe distance, the transverse safe distance, the safe region overlapping factor and the future collision probability. And multiplying the characteristic weights corresponding to the risk factors and summing to obtain the scene risk index. Or setting a maximum threshold value for each risk factor, counting the probability of the risk factors larger than the maximum threshold value, weighting and fusing the risk factors in a probability mode, and obtaining a scene risk index. The embodiment of the present application is not limited as long as fusion of risk factors in multiple aspects can be achieved.

In the embodiment of the application, the risk factors of the two dimensions are fused, the scene risk is comprehensively evaluated, and the accuracy of the scene risk index is improved.

In some embodiments, the driving parameters of the object include an object type and a speed, and the driving parameters of the host vehicle include a speed, and when the longitudinal safe distance and the lateral safe distance in S102 are calculated, the following may be performed. Determining a preset driving limit parameter of the object according to the type of the object; and calculating the longitudinal safe distance and the transverse safe distance according to the speed of the object, the speed of the self-vehicle and respective preset driving limit parameters of the object and the self-vehicle, wherein the preset driving limit parameters represent the driving limit parameters in the process of uniform acceleration or uniform deceleration motion, and the preset driving limit parameters of the self-vehicle are determined according to the type of the self-vehicle.

In the embodiment of the application, the physical motion of the vehicle is modeled into uniform acceleration or uniform deceleration motion, so that preset driving limit parameters of different object types are generated. The object types correspond to preset driving limit parameters one by one, and the preset driving limit parameters can represent the driving limit parameters in the process of uniform acceleration or uniform deceleration movement. The object type represents the category of the object, including dynamic (bicycles, automobiles, pedestrians) and static (obstacles, traffic lights and unknown objects), the preset driving limit parameters of the object are obtained according to the object type, and the preset driving limit parameters of the self-vehicle are also obtained according to the type of the self-vehicle.

In the embodiment of the application, when the preset driving limit parameters are set, the preset driving limit parameters of the unknown dynamic objects are set to be the same as those of the automobile, and the preset driving limit parameters of the static objects (such as traffic lights, green belts, road shoulders and signs) are all set to be 0. Illustratively, the preset travel limit parameters include a minimum runaway time, a longitudinal maximum acceleration, a longitudinal maximum deceleration, a longitudinal minimum deceleration, a lateral maximum acceleration, and a lateral minimum deceleration. As shown in tables 1 to 3, table 1 provides the minimum runaway time for different object types for the embodiments of the present application, table 2 provides the longitudinal driving limit parameter for different object types for the embodiments of the present application, and table 3 provides the lateral driving limit parameter for different object types for the embodiments of the present application.

TABLE 1

Type of object	Minimum out of control time
		Automobile	1 second
Bicycle with a wheel	2 seconds
		Pedestrian	3 seconds
Stationary object	0 second
		Unknown dynamic object	1 second

TABLE 2

TABLE 3

Type of object	Maximum acceleration in the transverse direction (m/s)	Transverse minimum deceleration (m/s)
			Automobile	7	3
Bicycle with a wheel	1.5	0.5
			Pedestrian	1.5	0.5
Stationary object	0	0
			Unknown dynamic object	7	3

It should be noted that the object types may include dynamic objects and static objects in other expressions, and are not limited to the object types listed in tables 1 to 3. The above tables 1 to 3 provide preset driving limit parameters for different object types according to the embodiments of the present application. The preset driving limit parameters may be set appropriately by those skilled in the art according to actual requirements, or may be determined by analyzing a large amount of experimental data, and are not limited to the parameters in tables 1 to 3, and the embodiment of the present application is not limited thereto.

In the embodiment of the application, before the longitudinal safety distance and the transverse safety distance are calculated, the out-of-control acceleration speed, the maximum deceleration distance, the minimum deceleration distance and the out-of-control acceleration distance need to be calculated according to the speed of the object, the speed of the vehicle and the respective preset driving limit parameters of the object and the vehicle, so as to calculate the safety distance. Namely, the embodiment of the application models the physical motion of the vehicle into uniform acceleration and uniform deceleration motion, and the uniform acceleration and uniform deceleration motion is used for subsequently calculating the safe distance and is expressed by the following 4 formulas. The vehicle speed in the 4 equations represents the vehicle speed.

Speed of runaway acceleration ═ vehicle speed + maximum acceleration x minimum time of runaway (1)

Maximum deceleration distance is vehicle speed × vehicle speed/(2 × maximum deceleration) (2)

Minimum deceleration distance is vehicle speed × vehicle speed/(2 × minimum deceleration) (3)

Runaway acceleration distance ═ minimum runaway time × (vehicle speed + runaway acceleration vehicle speed)/2 (4)

In the embodiment of the application, the preset driving limit parameters of the object are determined according to the type of the object; and calculating the longitudinal safe distance and the transverse safe distance according to the speed of the object, the speed of the self vehicle and the respective preset driving limit parameters of the object and the self vehicle. The safety distance between the object and the self-vehicle is calculated by the physical motion parameters of different object types and the current speed of the object and the self-vehicle, so that the accuracy of the safety distance is improved.

In some embodiments, the driving parameters of the object and the host vehicle comprise a heading angle, the longitudinal safe distance comprises a longitudinal co-directional driving safe distance and/or a longitudinal counter-directional driving safe distance, and the calculation of the longitudinal safe distance can be realized in the following manner. Determining an object which runs in the same direction and/or opposite direction to the vehicle in the object according to the course angle of the object and the course angle of the vehicle; calculating a longitudinal equidirectional running safety distance corresponding to the equidirectional running object according to the speed of the equidirectional running object and a preset running limit parameter; and/or calculating the longitudinal opposite-direction running safe distance corresponding to the opposite-direction running object according to the speed of the opposite-direction running object, the speed of the vehicle and the preset running limit parameters of the opposite-direction running object and the vehicle.

For example, according to the heading angle of the object, whether the object and the host vehicle travel in the same direction or in opposite directions can be identified, and the safe distance is calculated by considering the longitudinal same-direction travel and the longitudinal opposite-direction travel respectively.

In the embodiment of the application, when the safe distance for longitudinal equidirectional running is calculated, the two safe distances are run in the same direction, so that only longitudinal equidirectional running objects need to be considered, and the own vehicle does not need to be considered. Namely, according to the speed of the equidirectional running object and the preset running limit parameter, the longitudinal equidirectional running safety distance corresponding to the equidirectional running object is calculated. Among them, in the longitudinally equidirectional traveling objects, it is also necessary to consider a vehicle located in front of the own vehicle and a vehicle located behind the own vehicle, respectively. When the longitudinal direction opposite direction travel safe distance is calculated, since both the longitudinal direction opposite direction travel safe distance and the longitudinal direction opposite direction travel safe distance are opposite direction travel, not only the longitudinal direction same direction travel object but also the own vehicle need to be considered. That is, the longitudinal-direction opposing-travel safe distance is calculated based on the speed of the opposing-direction traveling object, the speed of the own vehicle, and the respective preset travel limit parameters of the opposing-direction traveling object and the own vehicle.

In the embodiment of the application, whether the object runs in the same direction or in opposite direction is distinguished in the object; and respectively aiming at different driving directions, the longitudinal equidirectional driving safe distance and the longitudinal opposite driving safe distance are calculated according to different parameters, so that the accuracy of the longitudinal driving safe distance is improved. It can be understood that only one case of longitudinal co-directional driving or longitudinal counter-directional driving may be considered, which is not limited in the embodiment of the present application, and thus when determining the spatial risk factor and the temporal risk factor, the data processing amount is reduced, and the calculation efficiency is improved.

In some embodiments, the driving parameters of the object include position information of the object relative to the vehicle, and the preset driving limit parameters include a minimum time to runaway, a maximum acceleration, a maximum deceleration, and a minimum deceleration; when calculating the longitudinal equidirectional travel safe distance, the following method can be implemented. Determining objects located in front of the vehicle and objects located behind the vehicle from among co-directional traveling objects along a traveling direction of the vehicle according to position information of the objects relative to the vehicle; calculating a longitudinal maximum deceleration distance of the object located in front of the own vehicle based on the speed and the maximum deceleration of the object located in front of the own vehicle; calculating a longitudinal out-of-control acceleration distance and a longitudinal minimum deceleration distance of the object behind the self-vehicle according to the speed, the minimum out-of-control time, the maximum acceleration and the minimum deceleration of the object behind the self-vehicle; and calculating the longitudinal equidirectional running safety distance according to the longitudinal maximum deceleration distance of the object positioned in front of the self-vehicle, and the longitudinal uncontrolled acceleration distance and the longitudinal minimum deceleration distance of the object positioned behind the self-vehicle.

In the embodiment of the present application, it is identified whether the object is located in front of or behind the own vehicle among the equidirectional running objects along the running direction of the own vehicle, based on the position information of the object with respect to the own vehicle. When calculating the safe distance, the safe distance is considered from the aspects of being positioned in front of the bicycle and being positioned behind the bicycle respectively. For convenience of description, the former car means an object located in front of the own car, and the latter car means an object located behind the own car. The longitudinal equidirectional travel safety distance comprises: the maximum deceleration distance of the front vehicle, the out-of-control acceleration distance of the rear vehicle and the minimum deceleration distance of the rear vehicle.

For example, the front vehicle longitudinal maximum deceleration distance is calculated according to the above formula (2), that is, the front vehicle longitudinal maximum deceleration distance is the front vehicle speed × the front vehicle speed/(2 × the front vehicle longitudinal maximum deceleration), and the front vehicle longitudinal maximum deceleration can be obtained from table 2, for example, the front vehicle is an automobile, the front vehicle longitudinal maximum deceleration is 8 m/s, the front vehicle is a pedestrian, and the front vehicle longitudinal maximum deceleration is 2 m/s.

For example, the rear vehicle runaway acceleration distance is calculated according to the above equation (1) and equation (3), the rear vehicle runaway acceleration speed is equal to the rear vehicle speed + the rear vehicle longitudinal maximum acceleration × the rear vehicle minimum runaway time, and the rear vehicle runaway acceleration distance is equal to the rear vehicle minimum runaway time × (rear vehicle speed + rear vehicle runaway acceleration speed)/2. The minimum runaway time of the rear vehicle can be obtained from table 1, and the longitudinal maximum acceleration of the rear vehicle can be obtained from table 2.

For example, the rear vehicle minimum deceleration distance is calculated according to the above formula (4), and the rear vehicle longitudinal minimum deceleration distance is the rear vehicle speed × rear vehicle speed/(2 × rear vehicle longitudinal minimum deceleration), and the rear vehicle longitudinal minimum deceleration is obtained from table 2.

Illustratively, the longitudinal equidirectional running safety distance is equal to the longitudinal maximum deceleration distance of the front vehicle, the longitudinal uncontrolled acceleration distance of the rear vehicle and the longitudinal minimum deceleration distance of the rear vehicle.

The above method for calculating the safe distance for longitudinal equidirectional travel is only an exemplary description, and may also be obtained by other calculation methods, and the embodiment of the present application is not limited thereto.

In the embodiment of the application, along the driving direction of the self-vehicle, whether the object is positioned in front of the self-vehicle or behind the self-vehicle is distinguished in the equidirectional driving objects; and respectively calculating the longitudinal maximum deceleration distance, the longitudinal out-of-control acceleration distance and the longitudinal minimum deceleration distance according to different parameters and calculation modes aiming at different physical positions, and then comprehensively considering to obtain the longitudinal same-direction running safety distance, so that the accuracy of the longitudinal same-direction running safety distance is improved.

In some embodiments, the preset driving limit parameters include a minimum time to runaway, a maximum acceleration, and a minimum deceleration, and when calculating the longitudinal counter-driving safe distance, may be implemented in the following manner. And calculating the longitudinal opposite-direction running safe distance according to the speed, the minimum out-of-control time, the maximum acceleration and the minimum deceleration of the self vehicle and the opposite-direction running object.

In the embodiment of the present application, the longitudinal counter-travel safe distance includes: the distance of the car out of control from the car and the distance of the car out of control from the opposite car. And (4) calculating the out-of-control distance of the self vehicle and the out-of-control distance of the opposite vehicle according to the formulas (1), (3) and (4).

Exemplarily, the distance of out-of-control of the vehicle is the longitudinal acceleration distance of out-of-control of the vehicle + the longitudinal minimum deceleration distance of the vehicle; calculating the longitudinal out-of-control acceleration distance of the self-vehicle according to the formulas (1) and (4), wherein the out-of-control acceleration speed of the self-vehicle is the speed of the self-vehicle + the longitudinal maximum acceleration of the self-vehicle multiplied by the minimum out-of-control time of the self-vehicle, and the longitudinal out-of-control acceleration distance of the self-vehicle is the minimum out-of-control time of the self-vehicle multiplied by (the speed of the self-vehicle + the out-of-control acceleration speed of the self-vehicle)/2; calculating a vehicle longitudinal minimum deceleration distance, which is a vehicle speed × a vehicle speed/(2 × a vehicle longitudinal minimum deceleration), according to the above formula (3); the minimum runaway time of the vehicle can be obtained from table 1, and the maximum acceleration in the longitudinal direction of the vehicle and the minimum deceleration in the longitudinal direction of the vehicle can be obtained from table 2.

For example, when the distance away from the vehicle is calculated, the calculation mode is consistent with that of the vehicle, and the detailed description is omitted here.

Illustratively, the longitudinal opposite-direction driving safety distance is the self-vehicle out-of-control distance + the opposite-direction vehicle out-of-control distance.

In the embodiment of the application, the out-of-control distance of the self vehicle and the out-of-control distance of the opposite vehicle are respectively calculated according to the respective speed, the minimum out-of-control time, the maximum acceleration and the minimum deceleration of the self vehicle and the opposite traveling object, and then the longitudinal opposite traveling safety distance is comprehensively considered, so that the accuracy of the longitudinal opposite traveling safety distance is improved.

In some embodiments, the driving parameters of the object include position information of the object relative to the vehicle, and the preset driving limit parameters, namely, the minimum runaway time, the maximum acceleration and the minimum deceleration, can be realized in the following manner when the lateral safety distance is calculated. Determining objects positioned at the left side and the right side of the self-vehicle from the objects along the driving direction of the self-vehicle according to the position information of the objects relative to the self-vehicle; according to the respective speeds, the minimum out-of-control time, the maximum acceleration and the minimum deceleration of the self-vehicle and the objects positioned at the left side and the right side of the self-vehicle, the respective transverse out-of-control acceleration distance and the transverse minimum deceleration distance of the self-vehicle and the objects positioned at the left side and the right side of the self-vehicle are calculated; and calculating the transverse safe distance according to the transverse out-of-control acceleration distance and the transverse minimum deceleration distance of the self-vehicle and the objects positioned at the left side and the right side of the self-vehicle.

In the embodiment of the present application, the left and right relative orientations of two objects in the horizontal direction may be regarded as a case, and the objects on the left and right sides of the own vehicle may be distinguished according to the position information of the objects relative to the own vehicle, with the longitudinal axis located at the center point of the own vehicle as a boundary. For convenience of description, the other vehicle is used to represent objects located on both left and right sides of the own vehicle. The objects on the left side and the right side of the bicycle are not limited to the automobile, and can be pedestrians, bicycles, static objects or unknown dynamic objects, and the transverse safe distance comprises: lateral runaway acceleration distance and lateral minimum deceleration distance.

For example, the vehicle lateral runaway acceleration distance is calculated according to the above equation (1) and equation (4), where the vehicle lateral runaway acceleration speed is the vehicle speed + the vehicle lateral maximum acceleration × the vehicle minimum runaway time, and the vehicle lateral runaway acceleration distance is the vehicle minimum runaway time × (vehicle speed + vehicle runaway acceleration speed)/2. The vehicle lateral direction minimum deceleration distance, which is the vehicle speed × the vehicle speed/(2 × the vehicle lateral direction minimum deceleration), is calculated according to the above equation (3). Wherein the minimum runaway time of the vehicle is obtained from table 1, and the lateral maximum acceleration and lateral minimum deceleration of the vehicle are obtained from table 3.

For example, when the lateral runaway acceleration distance of the other vehicle is calculated, the calculation mode is consistent with that of the own vehicle, and details are not repeated.

Illustratively, the lateral safety distance is the lateral uncontrolled acceleration distance of the self vehicle + the lateral minimum deceleration distance of the self vehicle + the lateral uncontrolled acceleration distance of the other vehicle + the lateral minimum deceleration distance of the other vehicle.

In the embodiment of the application, whether the objects are located on the left side and the right side of the self-vehicle or not is further distinguished in the objects along the driving direction of the self-vehicle, the transverse out-of-control acceleration distance and the transverse minimum deceleration distance of the self-vehicle and the objects located on the left side and the right side of the self-vehicle are calculated according to the speed, the minimum out-of-control time, the maximum acceleration and the minimum deceleration of the self-vehicle and the objects located on the left side and the right side of the self-vehicle, and then the transverse safety distance is comprehensively considered, so that the accuracy of the transverse safety distance is improved.

In some embodiments, the driving parameters of the object include a heading angle and an object type, and the driving parameters of the host vehicle include a heading angle, and when calculating the safety region coincidence factor in S102, the following can be performed. Determining an initialized travelable area of the object according to the type of the object; adjusting the direction of the initialized travelable area of the object according to the course angle of the object, and determining the travelable area of the object; calculating a safety region overlapping factor according to the respective drivable regions of the object and the vehicle; the travelable area of the own vehicle is determined according to the type of the own vehicle and the heading angle of the own vehicle.

In the embodiment of the present application, the types of objects are different, the required sizes of the safe driving areas are also different, the initialized drivable area of an object is determined according to the type of the object, the drivable area can also be understood as a drivable safe area, and the larger the area of the drivable safe area is, the more likely to collide with other objects, that is, the more dangerous, as shown in fig. 2, fig. 2 is an exemplary schematic diagram of a safe area of a different type of object provided in the embodiment of the present application. Fig. 2 (1) shows an initialization travelable region of a bicycle, fig. 2 (2) shows an initialization travelable region of an automobile, and fig. 2 (3) shows an initialization travelable region of a pedestrian. Wherein the parameters in fig. 2 are shown in table 4, and table 4 provides a travelable region parameter for the embodiment of the present application.

TABLE 4

Travelable region parameter	Value of
		Width 1	10 m
Length 1	20 m
		Length 2	10 m
Corner 1	60 degree
		Radius 1	60 m
Radius 2	10 m
		Radius 3	30 m
Diameter 1	5 m

It should be noted that the shape and size of the initialization travelable region in fig. 2 may be set by those skilled in the art according to actual situations, and may also be other forms of travelable regions, and the embodiment of the present application is not limited thereto. The area parameters in table 4 can be set appropriately by those skilled in the art according to actual needs, or can be determined by analyzing a large amount of experimental data, and are not limited to the parameters listed in table 4, and the embodiments of the present application are not limited thereto.

In the embodiment of the present application, the travelable region of the own vehicle is determined according to the type of the own vehicle, that is, (2) the car shown in fig. 2. Fig. 2 shows the initialized travelable area of the object facing the front, and during actual traveling, the traveling direction of the object may face any direction, so the direction of the initialized travelable area of the object needs to be adjusted according to the heading angle of the object to determine the travelable area of the object. For example, taking fig. 2 (1) as an example for explanation, if the heading angle of the bicycle indicates that the forward direction of the bicycle is 30 degrees on the front right, it is necessary to rotate the entire initialization travelable region of fig. 2 (1) clockwise by 30 degrees around the central point to obtain the travelable region.

In the embodiment of the present application, after obtaining the travelable areas of the object and the host vehicle, the intersection area (i.e., the overlap area) between the object and the travelable area of the object in the sensing range is obtained according to the position and the shape of the travelable area of the object and the host vehicle, and the safety area overlap factor is calculated according to the intersection area.

In the embodiment of the application, compared with the method of considering only collision avoidance at linear distance, the evaluation device of the driving scene also considers the travelable area of the traffic participant, and the accuracy of the scene risk characteristics is improved. Determining an initialization drivable area of a vehicle and an object; and then adjusting the heading angles according to the respective heading angles to obtain travelable areas of the vehicle and the object. And then, calculating the safety region overlapping factor according to the intersection region of each travelable region, thereby improving the accuracy of the safety region overlapping factor.

In some embodiments, when calculating the safety region coincidence factor based on the travelable regions of the object and the host vehicle, respectively, it may be implemented in the following manner. Determining the intersection area of the object and the travelable area of the vehicle according to the position information of the travelable areas of the object and the vehicle; and determining the ratio of the area of the intersection area to the minimum area in the travelable areas of the object and the vehicle as a safety area coincidence factor.

In the embodiment of the present application, there may be an intersection between the object and the travelable region of the host vehicle during traveling, and the maximum overlapping area between the object and the travelable region of the host vehicle is the minimum area in the travelable regions of the object and the host vehicle, for example, if the area of the travelable region of the object is 300 square meters and the area of the travelable region of the host vehicle is 500 square meters, then the maximum overlapping area between the object and the travelable region of the host vehicle is 300 square meters, that is, the minimum area in the travelable regions of the object and the host vehicle.

Illustratively, the minimum area in the travelable areas of the object and the vehicle is selected, the area of the intersection area is determined according to the position and the area of the safety area of the vehicle and the object in the sensing range, and the safety area coincidence factor is the area of the intersection area/the minimum area, namely the safety area coincidence factor is the area of the intersection area/the maximum coincidence area. The range of the safe coincidence factor value range is [0,1], and the larger the value is, the more dangerous the value is. If the object is completely overlapped with the drivable area of the self-vehicle, the overlapping factor of the safety area is 1; if the object and the travelable area of the own vehicle are not overlapped completely, the safety region overlapping factor is 0, and if the object and the travelable area of the own vehicle are partially overlapped, the safety region overlapping factor is a value between 0 and 1.

In the embodiment of the application, the safety region overlapping factor is calculated according to the ratio of the area of the intersection region to the minimum area, so that the accuracy of the safety region overlapping factor is improved.

In some embodiments, the driving parameters of the object and the host vehicle include a heading angle, a speed and an acceleration, and the calculation of the future collision probability in S102 may be implemented as follows. Predicting a plurality of respective driving track segments of the object and the vehicle in a future preset time period according to the respective course angle, speed and acceleration of the object and the vehicle, wherein the future preset time period comprises a plurality of future moments, and one future moment corresponds to one driving track segment; predicting the future collision times according to the respective travel track segments of the object and the own vehicle at the same time in the future; and determining the ratio of the future collision frequency to the number of the plurality of driving track segments in the future preset time period as the future collision probability.

In the embodiment of the application, the object and the self-vehicle keep the current motion state within the next several seconds, and each second motion of the object and the self-vehicle corresponds to one travel track segment. Travel path segments can also be understood as travel path segments. And predicting a plurality of travel track segments of the object and the vehicle in a future preset time period according to the speed, the acceleration and the course angle of the vehicle and the object. And if the own vehicle running track segment and the object running track segment intersect at the same time, judging that the collision is generated, adding 1 to the future collision frequency, and setting the future collision probability as the future collision frequency/the maximum collision frequency. The future collision probability value range is [0,1], and the larger the value is, the more dangerous the value is.

Illustratively, taking the case that the preset time period in the future is 5 seconds (including 5 future times) and 1 second corresponds to one track segment, the number of collisions is 5 times at most in the preset time period in the future, that is, the maximum number of collisions coincides with the number of travel track segments, and if the travel track segment of the own vehicle intersects with the travel track segment of the object in the 1 st s and the 4 th s in the future, the number of collisions in the future is predicted to be 2, and the probability of collision in the future is 2/5.

It should be noted that the specific values of the future preset time period and the future time can be set by those skilled in the art according to actual needs. For example, the future preset time period is 8 seconds (including 5 future times), 2 seconds corresponds to one travel track segment, and the travel track segments of the object and the own vehicle are predicted at 2s, 4s, 6s and 8s, respectively, which is not limited in this embodiment of the application.

In the embodiment of the application, a plurality of travel track segments of an object and a vehicle are predicted in a future preset time period, whether the travel track segments intersect at the same time in the future is judged, so that the future collision frequency is predicted, the future collision probability is determined by combining the number of the plurality of travel track segments, and the accuracy of the future collision probability is improved.

In some embodiments, the driving parameters of the object and the host vehicle each include a shape, and when predicting the number of future collisions, this may be achieved in the following manner. Determining respective driving track areas of the object and the self vehicle according to the respective shapes of the object and the self vehicle and by combining the respective corresponding driving track segments; and for the same time in the future, if the object intersects with the driving track area of the own vehicle, predicting the future collision of the object and the own vehicle until the prediction of the future preset time period is completed, and accumulating the number of the future collision of the object and the own vehicle to obtain the number of the future collision.

In the embodiment of the application, different types of objects are different in shape and different in occupied area, and for automobiles and pedestrians, the shapes of the automobiles are far larger than those of the pedestrians, so that the driving track segments of the automobiles and the pedestrians are not intersected, but the automobiles and the pedestrians collide. Therefore, in order to predict the number of future collisions, it is necessary to determine the respective travel track areas of the object and the host vehicle by considering the shapes of the object and the host vehicle and combining the respective corresponding travel track segments, and determine that the collision occurs if the travel track area of the host vehicle intersects the travel track area of the object at the same time, and add 1 to the number of future collisions.

In the embodiment of the application, the shapes of the object and the vehicle are combined, the travel track segments are converted into travel track areas, and then whether the travel track areas intersect at the same time in the future is judged, so that the future collision frequency is predicted, and the accuracy of the future collision frequency is improved.

In some embodiments, the driving parameters of the object comprise position information of the object relative to the vehicle, and the number of the longitudinal safe distance, the transverse safe distance, the safe region coincidence factor and the future collision probability is multiple; when the spatial risk factor is determined in S103 described above, this can be achieved in the following manner. Determining objects in front of the self-vehicle in the driving direction of the self-vehicle according to the position information of the objects relative to the self-vehicle; selecting a first preset number of longitudinal safety distances from a plurality of longitudinal safety distances corresponding to objects in front of the self-vehicle according to a first preset rule; selecting a second preset number of transverse safe distances from the plurality of transverse safe distances according to a second preset rule; selecting a third preset number of safety region coincidence factors from the plurality of safety region coincidence factors according to a third preset rule; selecting a future collision probability greater than zero from a plurality of future collision probabilities; the space risk factor comprises at least one of a first preset number of longitudinal safe distances, a second preset number of transverse safe distances, a third preset number of safe region overlapping factors and a future collision probability greater than zero.

In the embodiment of the application, the scene risk characteristics include a plurality of different characteristics, such as a longitudinal safe distance, a transverse safe distance, a safe region coincidence factor and a future collision probability. And respectively selecting a plurality of values meeting the rule conditions according to the rules as the space risk factors of the current moment. As shown in fig. 3, fig. 3 is an exemplary schematic diagram of a selection rule of a spatial risk factor according to an embodiment of the present application.

For example, the evaluation device of the driving scene may focus more on the object located in front of the own vehicle at the present time than on the object located behind the own vehicle, avoiding rear-end collision with the object located in front. Therefore, when selecting a plurality of longitudinal safety distances, it is determined whether the object is located in front of the vehicle according to the position information of the object relative to the vehicle, and then a plurality of maximum values (e.g., 3) are selected from the longitudinal safety distances corresponding to the objects located in front of the vehicle.

Illustratively, when the transverse safe distance and the safe region coincidence factor are selected, the selection mode is consistent with the selection mode of the longitudinal safe distance, and the details are not repeated here. For example, from a plurality of lateral safety distances, a number of maximum values (e.g., 3) are selected; from the plurality of safety-region coincidence factors, several maximum values (e.g., 3) are selected.

Illustratively, when selecting the future collision probability, it is determined whether the future collision probability is greater than 0, and all values having a probability greater than 0 are selected.

It should be noted that, in the embodiments of the present application, the first and second are only for distinguishing names, do not represent sequential relationships, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features, for example, the first preset rule, the second preset rule and the third preset rule; a first preset number, a second preset number, and a third preset number. The first preset rule, the first preset number, the second preset rule, the second preset number, the third preset rule and the third preset number can be set appropriately by those skilled in the art according to actual needs. Illustratively, the preset rule is to select features larger than a preset value, and the preset number is related to a specific actual situation; or, the preset rule is to select a preset number of features; alternatively, the preset rule is to select some representative features, and the preset number can be freely set, for example, 3 or 4. The embodiments of the present application are not limited thereto.

In the embodiment of the application, the preset number of features are selected from the scene risk features according to the preset rule and are used as the space risk factors, so that the diversity and the comprehensiveness of the space risk factors are improved.

In some embodiments, when determining the time-series risk factor in S103 described above, this may be achieved in the following manner. Acquiring scene risk characteristics of a plurality of first moments in a first preset time period; sliding in the scene risk characteristics of a plurality of first moments in the first preset time period by taking the second preset time period as a time window and taking the third preset time period as a step length to determine a plurality of time windows; the second preset time period and the third preset time period are both smaller than the first preset time period; counting the scene risk characteristics of a plurality of second moments in each time window to obtain a statistical result corresponding to each time window; the statistical result represents the time sequence risk of the scene risk characteristics in the time window, and the time sequence risk factor comprises the statistical results corresponding to the time windows.

In some embodiments, the statistics include at least one of an average value, a maximum value, and a rate of change, wherein the average value characterizes scene macro risk, the maximum value characterizes scene worst risk, and the rate of change characterizes scene stability.

In the embodiment of the application, the scene risk characteristics at each time are also stored, and the stored scene risk characteristics at a plurality of first times can be used for determining the time sequence risk factor. During storage, a short-time storage mode can be adopted, for example, 45 seconds and 60 seconds, so that the data effectiveness is ensured, the data storage space can be reduced, and the resource utilization rate is improved.

For example, the first preset time period is 45 seconds, 1 second corresponds to a scene risk feature at a first time, the second preset time period is 5 seconds, and the third preset time period is 1 second. Acquiring scene risk characteristics of 45 first moments within 45 seconds; and sliding by taking 5 seconds as a time window and 1 second as a step size, and determining a plurality of time windows, for example, the 1 st time window comprises the 1 st to 5 th scene risk features, the 2 nd time window comprises the 2 nd to 6 th scene risk features, and … … to the 41 th time window comprise the 41 st to 45 th scene risk features. And (4) counting the scene risk characteristics at 5 moments in 41 time windows to obtain the statistical results corresponding to the 41 time windows. The time sequence risk factor comprises statistical results corresponding to 41 time windows, and the statistical results represent time sequence risks of scene risk features in the time windows.

In the embodiment of the application, when the scene risk characteristics of the plurality of second moments in each time window are counted and the statistical result corresponding to each time window is obtained, at least one of the average value, the maximum value and the change rate is counted.

Illustratively, each time window includes 5 time scene risk features, and 1 time scene risk feature includes features such as a longitudinal safe distance, a transverse safe distance, a safe region coincidence factor, and a future collision probability, and the 1 st time window includes 1 st to 5 th second scene risk features, and the longitudinal safe distance, the transverse safe distance, the safe region coincidence factor, and the future collision probability are respectively averaged, maximized, and a change rate are obtained to obtain a statistical result of the 1 st time window. The statistical result of the 1 st time window comprises an average value, a maximum value and a change rate obtained for 5 longitudinal safe distances, an average value, a maximum value and a change rate obtained for 5 transverse safe distances, an average value, a maximum value and a change rate obtained for 5 safe region overlapping factors, and an average value, a maximum value and a change rate obtained for 5 future collision probabilities. In the above manner, the statistical result of each window can be obtained.

It should be noted that the first and second embodiments in the present application are only for distinguishing names, do not represent sequential relationships, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features, for example, the first time and the second time. The scene risk characteristics at the multiple first moments represent the scene risk characteristics included in the first preset time period, and the scene risk characteristics at the multiple second moments represent the scene risk characteristics included in one time window.

In the embodiment of the application, a time window with a fixed size is selected, the time window is continuously slid, the average value, the maximum value and the change rate of four types of characteristic values are counted for a plurality of longitudinal safe distances, a plurality of transverse safe distances, a plurality of safe region overlapping factors and a plurality of future collision probabilities included in scene risk characteristics in each time window, and the four types of characteristic values are used as descriptions of the time sequence risk factors. The average value describes scene macro risk, the maximum value describes scene worst risk, and the change rate describes scene stability. The scene risk of the self-vehicle in the driving process is described in multiple angles, and the diversity and the accuracy of time sequence factors are improved.

In some embodiments, the scene risk characteristics of each second time in each time window include a plurality of longitudinal safety distances, a plurality of transverse safety distances, a plurality of safety region coincidence factors and a plurality of future collision probabilities, and the statistics of the scene risk characteristics of the plurality of second times in each time window to obtain the statistical results corresponding to each time window can be implemented in the following manner, wherein the maximum value of the plurality of safety region coincidence factors is selected according to the scene risk characteristics of each second time to obtain the maximum value of the safety region coincidence factor; summing the multiple future collision probabilities to obtain a target future collision probability; and respectively carrying out statistics on the plurality of longitudinal safe distances, the plurality of transverse safe distances, the maximum value of the safe region coincidence factor and the future collision probability of the target to obtain a statistical result corresponding to each time window.

In the embodiment of the application, the average value, the maximum value and the change rate of the four types of feature values included in the scene risk feature are counted. And for the scene risk characteristics of each second moment, obtaining the maximum value of the safety region coincidence factor, and summing the future collision probabilities of the second moments to obtain the target future collision probability. As such, the scene risk characteristics for each second time instance may include a plurality of longitudinal safe distances, a plurality of lateral safe distances, a safe region coincidence factor maximum, and a target future collision probability. Then, the average value, the maximum value and the change rate of the four types of characteristic values included in the scene risk characteristics of the plurality of second moments in each time window are counted to obtain the statistical result corresponding to each time window.

In the embodiment of the application, the maximum value of the safe region overlapping factor and the future collision probability of the target accurately reflect the danger degree in the time dimension, and the data volume is reduced and the calculation efficiency is improved by solving the combination mode of summing the maximum value of the safe region overlapping factor and the future collision probability.

In some embodiments, the temporal risk factor comprises a plurality of factors characterizing different features in the temporal dimension, and the spatial risk factor comprises a plurality of factors characterizing different features in the spatial dimension; when the risk factors are fused to obtain the scene risk index in S104, the method can be implemented in the following manner. Respectively selecting factors of each characteristic greater than a respective corresponding preset threshold value from the time sequence risk factor and the space risk factor; calculating the ratio of the factor of each characteristic to the total number of the corresponding factors to obtain the probability of each characteristic; calculating products between preset feature weights corresponding to factors of the features and the feature probabilities of the factors to obtain a plurality of feature products; and determining a scene risk index according to the characteristic products.

In the embodiment of the present application, based on the spatial risk factor and the time sequence risk factor, the uncertainty of the scene risk is considered, and preset thresholds are also set for different features in the scene risk features, as shown in tables 5 and 6, table 5 provides a threshold corresponding to the spatial risk factor for the embodiment of the present application. Table 6 provides a threshold corresponding to a time-series risk factor for the embodiment of the present application.

Illustratively, the preset thresholds of the spatial risk factors in table 5 include a zone coincidence factor (corresponding to a safety zone coincidence factor), a collision warning factor (corresponding to a future collision probability), a lateral/longitudinal required safety distance (corresponding to a lateral/longitudinal safety distance), and a lateral/longitudinal maximum risk distance (corresponding to a lateral/longitudinal safety distance), which can be understood as being considered twice for the lateral/longitudinal safety distance. Illustratively, if a certain lateral safety distance is 30 meters, the statistical occurrence number of the lateral safety distance is 2. The preset thresholds of the time-series risk factors in table 6 include a zone coincidence change rate (corresponding to the maximum value of the safety zone coincidence factor), a maximum value of the lateral/longitudinal required safety distance, a mean value, a change rate (corresponding to the lateral/longitudinal safety distance), and a front lateral/longitudinal required safety distance change rate (corresponding to the lateral/longitudinal safety distance), and may be understood as considering the lateral/longitudinal safety distance change rate of the object located in front of the host vehicle twice. For example, if the lateral safe distance change rate of an object located in front of the own vehicle is 0.3, the statistical occurrence number of the lateral safe distance change rate is 2, that is, the estimation device of the driving scene may pay more attention to the preceding vehicle, that is, pay more attention to whether the own vehicle may collide with the preceding vehicle.

In the embodiment of the application, for each feature, the numerical value occurrence probability that the feature is larger than the corresponding preset threshold value is counted to obtain each feature probability. Preset feature weights are also set for different features in the scene risk features, as shown in table 7 and table 8, table 7 provides a feature weight corresponding to a spatial risk factor for the embodiment of the present application, and table 8 provides a feature weight corresponding to a timing risk factor for the embodiment of the present application. The feature weights in tables 7 and 8 correspond to the features in tables 5 and 6, respectively.

In the embodiment of the application, calculating the product between the preset feature weight corresponding to the factor of each feature and the feature probability of each feature to obtain a plurality of feature products; and determining the sum of the feature products as the scene risk index.

It should be noted that, the above tables 5 to 8 provide a preset threshold and a characteristic weight corresponding to a time sequence risk factor for the embodiment of the present application. The preset threshold and the weight of the feature weight can be set appropriately by those skilled in the art according to actual needs, and can also be determined by analyzing a large amount of experimental data, and are not limited to the parameters in tables 5 to 8, and the embodiment of the present application is not limited thereto.

In the embodiment of the application, the preset threshold corresponding to each feature is used as a lower bound, the numerical value occurrence probability greater than the preset threshold in each feature is counted, the time sequence risk factor and the space risk factor are weighted and fused in a probability form, and the scene risk is comprehensively evaluated, that is, the scene risk index is the probability that the feature weight is multiplied by the feature greater than the corresponding threshold, so that the accuracy of the scene risk index is improved.

TABLE 5

Spatial risk factor	Threshold value
		Longitudinal required safety distance	25 m
Lateral required safety distance	20 m
		Longitudinal maximum hazard distance	50 m
Lateral maximum hazard distance	25 m
		Area coincidence factor	0.5
Collision warning factor	0.3

TABLE 6

Time sequence risk factor	Threshold value
		Maximum value of safety distance required in longitudinal direction	50 m
Mean value of longitudinal required safety distance	25 m
		Longitudinal required safe distance rate of change	0.2
Maximum value of safety distance required in transverse direction	25 m
		Mean value of lateral required safety distance	20 m
Lateral required safety distance rate of change	0.2
		Rate of change of area overlap	0.1
Forward longitudinal required safe distance rate of change	0.2
		Rate of change of safety distance required in the front lateral direction	0.2

TABLE 7

Spatial risk factor	Feature weights
		Longitudinal required safety distance, maximum distance factor	0.3
Lateral required safety distance, maximum distance factor	0.3
		Area coincidence factor	0.2
Collision warning factor	0.2

TABLE 8

Time sequence risk factor	Feature weights
		Maximum value, mean value and change rate of longitudinal required safety distance factor	0.25
Maximum value, mean value and change rate of transverse required safety distance factor	0.25
		Rate of change of area overlap factor	0.1
Forward longitudinal required safe distance factor rate of change	0.2
		Rate of change of forward lateral required safety distance factor	0.2

In some embodiments, the spatial risk factor includes a plurality of horizontal/vertical safety distances, and when a factor of each characteristic greater than a respective preset threshold is selected from the spatial risk factors, the following steps may be further included. If the object is subjected to first attention, selecting a transverse/longitudinal safety distance larger than a first preset threshold value from the plurality of transverse/longitudinal safety distances; if the object is subjected to second attention, selecting a transverse/longitudinal safety distance larger than a second preset threshold value from the plurality of transverse/longitudinal safety distances; the attention degree corresponding to the first attention is larger than that corresponding to the second attention, and the first preset threshold is smaller than the second preset threshold.

In the embodiment of the application, for a novice driver, a higher attention degree needs to be set for traffic participants on a road, or for a high-risk traffic participant, a higher attention degree also needs to be set. The evaluation equipment of the driving scene realizes the attention of high-risk traffic participants by considering the extreme transverse/longitudinal safety distance. When the probability of occurrence of the numerical value of each feature larger than the corresponding preset threshold is counted to obtain the probability of each feature, the preset threshold corresponding to the horizontal/longitudinal safety distance (e.g., the horizontal/longitudinal required safety distance and the horizontal/longitudinal maximum dangerous distance in table 5) needs to be set lower, for example, the longitudinal required safety distance is changed from the original 25 meters to 20 meters; alternatively, a new preset threshold value is set for the lateral/longitudinal safety distance, for example, a parameter is newly added in table 5 above: the maximum safe distance threshold for the transverse/longitudinal direction may be understood as considering the transverse/longitudinal safe distance twice.

Illustratively, taking the parameters in table 5 as an example, the parameters are newly added: and if a certain transverse/longitudinal maximum safety distance is 30 meters, the statistical occurrence frequency of the transverse safety distance is 3 times, so that the characteristic probability corresponding to the transverse safety distance in the space risk factor is increased, and the attention to high-risk traffic participants is realized.

It should be noted that the first and second embodiments in the present application are only for name differentiation and do not represent sequential relationships, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features, for example, the first attention and the second attention, the first preset threshold and the second preset threshold.

In the embodiment of the application, the flexibility of driving scene estimation is improved by setting different attention degrees.

In some embodiments, after determining the scenario risk index, the method for risk assessment of a driving scenario further comprises the following steps. And converting the scene risk index into a risk grade according to a preset mapping relation.

In the embodiment of the application, after the scene risk index is determined, the driving scene is classified according to the preset mapping relation, the scene risk index is converted into the corresponding risk grade, the risk grade can be used for reminding a driver, and in the field of automatic driving, the risk grade can also be used for feeding back to the driving system, so that the driving system plans the driving route at the next moment according to the risk grade.

It should be noted that the preset mapping relationship may be set by a person skilled in the art according to actual situations, as long as the risk degree of the driving scene can be represented. As shown in table 9, table 9 provides a corresponding relationship between a scene risk index and a risk level for the embodiments of the present application.

TABLE 9

Risk index	Risk rating
		(0,0.1]	Level 1
(0.1,0.15]	Stage 2
		(0.15,0.2]	Grade 3
(0.2,0.25]	4 stage
		(0.25,0.3]	Grade 5
(0.3,0.35]	Grade 6
		(0.35,0.4]	Stage 7
(0.4,0.5]	Stage 8
		(0.6,0.7]	Grade 9
(0.7,+∞)	Grade 10

Next, an exemplary application of the embodiment of the present application in a practical application scenario will be described.

In the embodiment of the present application, the object is represented by a traffic participant, as shown in fig. 4, and fig. 4 is a flowchart illustrating optional steps of another driving scenario risk assessment method provided in the embodiment of the present application.

S401, acquiring object types, speeds, accelerations, relative own vehicle positions, shape sizes and heading angles of other traffic participants in a perception range, and acquiring the speed, the acceleration, the heading angles and the shape sizes of the own vehicle.

The traffic participants comprise all dynamic objects on the road, and static objects such as obstacles and traffic lights. Meanwhile, the shape and size of the own vehicle are known, and the speed, the acceleration and the heading angle of the own vehicle are obtained from the power system of the own vehicle during running.

And S402, generating a travelable area according to the object type and the course angle.

And knowing the object type and the course angle of the own vehicle and other traffic participants, and generating an initialized travelable area according to the object type and the type of the own vehicle. The object types can be classified into bicycles, automobiles, pedestrians, stationary objects, and unknown dynamic objects, wherein, in order to ensure driving safety, in the embodiment of the present application, the travelable regions of the obstacles and the unknown dynamic objects are set to be consistent with the travelable region of the automobile. The initialized travelable area for each type of object is shown in fig. 2 above, and the parameters of side length, angle, etc. are shown in table 4 above. And adjusting the directions of the respective initialized travelable areas according to the course angle of the object and the course angle of the own vehicle, ensuring the movement directions of the object and the own vehicle to be consistent, and obtaining the respective travelable areas. The driving area is used for calculating a safety area coincidence factor between the vehicle and other objects.

And S403, generating a minimum runaway time, a horizontal/longitudinal maximum acceleration, a horizontal/longitudinal minimum deceleration and a longitudinal maximum deceleration according to the type of the object.

And (3) the object types of the own vehicle and other traffic participants are known from S401, and preset driving limit parameters of different types of objects are generated initially, wherein the preset driving limit parameters comprise minimum out-of-control time, longitudinal maximum acceleration, longitudinal maximum deceleration, longitudinal minimum deceleration, transverse maximum acceleration and transverse minimum deceleration. The default parameter values of the preset driving limit parameters may refer to tables 1 to 3. The above parameters are used to calculate the longitudinal required safety distance and the lateral required safety distance.

And S404, acquiring and generating a plurality of data of the own vehicle and other traffic participants from S401-S403, and calculating scene risk characteristics of the own vehicle and other traffic participants. The scene risk characteristics include a longitudinal safe distance, a lateral safe distance, a safe zone coincidence factor, and a future collision probability.

And S405, storing the scene risk characteristics obtained by calculation in a short time.

The driving scene snapshot data (i.e., the scene risk characteristics) at the current moment are stored in the snapshot database for time sequence modeling of the subsequent scene risk characteristics, and it should be noted that the scene risk characteristics within a short time (e.g., 45 seconds) are stored in the snapshot database, that is, the scene risk characteristics at multiple moments can be stored.

S406, selecting a plurality of values which accord with the rule conditions as the space risk factors of the current moment.

According to the longitudinal required safety distance, the transverse required safety distance, the safety region overlapping factor and the future collision probability of the scene risk feature in S404, a plurality of values meeting the rule condition are respectively selected according to a preset rule as the space risk factor at the current moment, and the selection mode can refer to fig. 3.

S407, selecting a time window with a fixed size, continuously sliding the time window, and counting the statistical mean, the maximum and the change rate of the four types of risk characteristic values in the time window to be used as time sequence risk factors.

Traversing the scene risk characteristics stored in the snapshot database at each moment in S405, wherein for the scene risk characteristics at each moment, a maximum value of the safety region coincidence factor is obtained, and the future collision probability of the own vehicle and all other objects at each moment is summed. In order to realize multi-angle description of the scene risk of the self-vehicle in the driving process, a time window with a fixed size is selected, the time window is continuously slid, and the statistical mean value, the maximum value and the change rate of the four types of risk characteristic values in the time window are counted to be used as the description of the time sequence risk factor. In a time sequence scene, the average value describes scene macro risks, the maximum value describes scene worst risks, and the change rate describes scene stability.

And S408, performing weighted fusion by combining corresponding characteristic weights according to the relationship between the numerical values of the space risk factor and the time sequence risk factor and the corresponding preset threshold value, and calculating a risk index.

Based on the space risk factor and the time sequence risk factor calculated in S406 and S407, considering uncertainty of the scene risk, setting thresholds by the evaluation equipment of the driving scene aiming at risk features of different scenes, wherein the thresholds corresponding to the features are lower bounds, counting the numerical value occurrence probability of the corresponding features larger than the thresholds, weighting and fusing the space risk factor and the time sequence risk factor in a probability mode, and comprehensively evaluating the scene risk. The scene risk index is the feature weight × the probability that the feature is larger than the corresponding threshold, and the corresponding threshold may refer to table 5 and table 6 described above, and the feature weight may refer to table 7 and table 8 described above.

In the embodiment of the present application, the situation of the required safety distance in the extreme horizontal/longitudinal direction is also considered, and the maximum horizontal/longitudinal safety distance threshold is further added in table 5, so as to realize the attention of the high-risk traffic participants.

And S409, obtaining a risk grade according to the risk index.

And grading the driving scene according to the risk index obtained in the step S408, and subdividing the driving scene into ten grades, wherein the reference table 9 refers to the corresponding relation between the risk index and the risk grade.

In the embodiment of the application, in the face of high dynamic and diversity of scenes in the driving process, the evaluation method of the driving scene provided by the embodiment of the application does not depend on drivers, maps and long-time accident statistical data, calculates scene risk characteristics by utilizing the motion states, the out-of-control time, the travelable areas and the future collision probability of different types of traffic participants in the own vehicle and the sensing range thereof, stores and records driving scene snapshot data for a period of time, designs a multi-index weighted fusion model in space-time, quantifies the risk index of the driving scene by combining various scene risk factors, realizes the evaluation of the risk index of the driving scene, and improves the accuracy of evaluation results. And further classifying and dividing the scene risk level according to the risk index, and feeding back the scene risk level to the user.

In order to implement the risk assessment method for the driving scenario of the embodiment of the present application, an embodiment of the present application further provides a risk assessment device for the driving scenario, as shown in fig. 5, fig. 5 is a schematic structural diagram of the risk assessment device for the driving scenario provided by the embodiment of the present application, and the risk assessment device 50 for the driving scenario includes: the acquiring module 501 is used for acquiring the driving parameters of the self-vehicle and the driving parameters of the object in the sensing range of the self-vehicle;

a scene risk characteristic module 502, configured to determine a scene risk characteristic according to the driving parameters of the object and the driving parameters of the host vehicle, where the scene risk characteristic includes at least one characteristic of a longitudinal safe distance, a transverse safe distance, a safe region overlapping factor, and a future collision probability;

a risk factor module 503, configured to determine a time sequence risk factor and a space risk factor respectively according to different features and feature combinations in the scene risk features, where the time sequence risk factor reflects a scene risk degree in a time dimension, and the space risk factor reflects a scene risk degree in a space dimension;

and the evaluation module 504 is configured to fuse the time-series risk factor and the spatial risk factor to obtain a scene risk index.

In some embodiments, the driving parameters of the object include an object type and a speed, and the driving parameters of the host vehicle include a speed;

the scene risk characteristic module 502 is further configured to determine a preset driving limit parameter of the object according to the object type; and calculating the longitudinal safe distance and the transverse safe distance according to the speed of the object, the speed of the self-vehicle and respective preset driving limit parameters of the object and the self-vehicle, wherein the preset driving limit parameters represent the driving limit parameters in the process of uniform acceleration or uniform deceleration motion, and the preset driving limit parameters of the self-vehicle are determined according to the type of the self-vehicle.

In some embodiments, the driving parameters of the object and the host vehicle each include a heading angle, and the longitudinal safe distance includes a longitudinal co-directional driving safe distance and/or a longitudinal counter-directional driving safe distance;

the scene risk characteristic module 502 is further used for determining an object which runs in the same direction and/or opposite direction with the own vehicle in the object according to the course angle of the object and the course angle of the own vehicle; calculating a longitudinal equidirectional running safety distance corresponding to the equidirectional running object according to the speed of the equidirectional running object and a preset running limit parameter; and/or calculating the longitudinal opposite-direction running safe distance corresponding to the opposite-direction running object according to the speed of the opposite-direction running object, the speed of the vehicle and the preset running limit parameters of the opposite-direction running object and the vehicle.

In some embodiments, the driving parameters of the object include position information of the object relative to the vehicle, and the preset driving limit parameters include a minimum time to runaway, a maximum acceleration, a maximum deceleration, and a minimum deceleration;

the scene risk characteristic module 502 is further configured to determine, from the position information of the object relative to the host vehicle, an object located in front of the host vehicle and an object located behind the host vehicle among co-directional traveling objects along the traveling direction of the host vehicle; calculating a longitudinal maximum deceleration distance of the object located in front of the own vehicle based on the speed and the maximum deceleration of the object located in front of the own vehicle; calculating a longitudinal out-of-control acceleration distance and a longitudinal minimum deceleration distance of the object behind the self-vehicle according to the speed, the minimum out-of-control time, the maximum acceleration and the minimum deceleration of the object behind the self-vehicle; and calculating the longitudinal equidirectional running safety distance according to the longitudinal maximum deceleration distance of the object positioned in front of the self-vehicle, and the longitudinal uncontrolled acceleration distance and the longitudinal minimum deceleration distance of the object positioned behind the self-vehicle.

In some embodiments, the preset travel limit parameters include a minimum time to runaway, a maximum acceleration, and a minimum deceleration;

the scene risk characteristic module 502 is further configured to calculate a longitudinal opposite-direction travel safety distance according to the speed, the minimum out-of-control time, the maximum acceleration, and the minimum deceleration of the host vehicle and the opposite-direction travel object, respectively.

In some embodiments, the running parameters of the object comprise position information of the object relative to the vehicle, and the preset running limit parameters comprise minimum out-of-control time, maximum acceleration and minimum deceleration;

the scene risk characteristic module 502 is further configured to determine, from the position information of the object relative to the vehicle, objects located on the left and right sides of the vehicle along the driving direction of the vehicle; according to the respective speeds, the minimum out-of-control time, the maximum acceleration and the minimum deceleration of the self-vehicle and the objects positioned at the left side and the right side of the self-vehicle, the respective transverse out-of-control acceleration distance and the transverse minimum deceleration distance of the self-vehicle and the objects positioned at the left side and the right side of the self-vehicle are calculated; and calculating the transverse safe distance according to the transverse out-of-control acceleration distance and the transverse minimum deceleration distance of the self-vehicle and the objects positioned at the left side and the right side of the self-vehicle.

In some embodiments, the driving parameters of the object include a heading angle and an object type, and the driving parameters of the host vehicle include a heading angle;

the scene risk characteristic module 502 is further configured to determine an initialized travelable area of the object according to the object type; adjusting the direction of the initialized travelable area of the object according to the course angle of the object, and determining the travelable area of the object; calculating a safety region overlapping factor according to the respective drivable regions of the object and the vehicle; the travelable area of the own vehicle is determined according to the type of the own vehicle and the heading angle of the own vehicle.

In some embodiments, the scene risk characteristic module 502 is further configured to determine an intersection area of the object and the travelable area of the host vehicle according to the position information of the travelable area of each of the object and the host vehicle; and determining the ratio of the area of the intersection area to the minimum area in the travelable areas of the object and the vehicle as a safety area coincidence factor.

In some embodiments, the driving parameters of the object and the host vehicle each include a heading angle, a speed, and an acceleration;

the scene risk characteristic module 502 is further configured to predict a plurality of respective travel track segments of the object and the vehicle within a future preset time period according to the respective heading angle, speed and acceleration of the object and the vehicle, where the future preset time period includes a plurality of future moments, and one future moment corresponds to one travel track segment; predicting the future collision times according to the respective travel track segments of the object and the own vehicle at the same time in the future; and determining the ratio of the future collision frequency to the number of the plurality of driving track segments in the future preset time period as the future collision probability.

In some embodiments, the driving parameters of the object and the host vehicle each include a shape;

the scene risk characteristic module 502 is further configured to determine respective driving track areas of the object and the vehicle according to respective shapes of the object and the vehicle and by combining respective corresponding driving track segments; and for the same time in the future, if the object intersects with the driving track area of the own vehicle, predicting the future collision of the object and the own vehicle until the prediction of the future preset time period is completed, and accumulating the number of the future collision of the object and the own vehicle to obtain the number of the future collision.

In some embodiments, the driving parameter of the object includes position information of the object relative to the own vehicle; the number of the longitudinal safe distance, the number of the transverse safe distance, the safe region overlapping factor and the number of the future collision probability are all multiple;

the risk factor module 503 is further configured to determine, in the object, an object located in front of the host vehicle along the driving direction of the host vehicle according to the position information of the object relative to the host vehicle; selecting a first preset number of longitudinal safety distances from a plurality of longitudinal safety distances corresponding to objects in front of the self-vehicle according to a first preset rule; selecting a second preset number of transverse safe distances from the plurality of transverse safe distances according to a second preset rule; selecting a third preset number of safety region coincidence factors from the plurality of safety region coincidence factors according to a third preset rule; selecting a future collision probability greater than zero from a plurality of future collision probabilities; the space risk factor comprises at least one of a first preset number of longitudinal safe distances, a second preset number of transverse safe distances, a third preset number of safe region overlapping factors and a future collision probability greater than zero.

In some embodiments, the risk factor module 503 is further configured to obtain scene risk features of a plurality of first moments within a first preset time period; sliding in the scene risk characteristics of a plurality of first moments in the first preset time period by taking the second preset time period as a time window and taking the third preset time period as a step length to determine a plurality of time windows; the second preset time period and the third preset time period are both smaller than the first preset time period; counting the scene risk characteristics of a plurality of second moments in each time window to obtain a statistical result corresponding to each time window; the statistical result represents the time sequence risk of the scene risk characteristics in the time window, and the time sequence risk factor comprises the statistical results corresponding to the time windows.

In some embodiments, the scene risk characteristics for each second time instant in each time window include a plurality of longitudinal safe distances, a plurality of lateral safe distances, a plurality of safe-zone coincidence factors, and a plurality of future collision probabilities;

the risk factor module 503 is further configured to select a maximum value of the multiple safety region coincidence factors for the scene risk feature at each second time to obtain a maximum value of the safety region coincidence factors; summing the multiple future collision probabilities to obtain a target future collision probability; and respectively carrying out statistics on the plurality of longitudinal safe distances, the plurality of transverse safe distances, the maximum value of the safe region coincidence factor and the future collision probability of the target to obtain a statistical result corresponding to each time window.

In some embodiments, the temporal risk factor comprises a plurality of factors characterizing different features in the temporal dimension, and the spatial risk factor comprises a plurality of factors characterizing different features in the spatial dimension;

the evaluation module 504 is further configured to select, from the time sequence risk factor and the space risk factor, factors of each feature larger than a respective preset threshold value; calculating the ratio of the factor of each characteristic to the total number of the corresponding factors to obtain the probability of each characteristic; calculating products between preset feature weights corresponding to factors of the features and the feature probabilities of the factors to obtain a plurality of feature products; and determining a scene risk index according to the characteristic products.

In some embodiments, the spatial risk factor comprises a plurality of lateral/longitudinal safety distances;

the evaluation module 504 is further configured to select a lateral/longitudinal safety distance greater than a first preset threshold from the plurality of lateral/longitudinal safety distances if the object is subjected to the first attention; if the object is subjected to second attention, selecting a transverse/longitudinal safety distance larger than a second preset threshold value from the plurality of transverse/longitudinal safety distances; the attention degree corresponding to the first attention is larger than that corresponding to the second attention, and the first preset threshold is smaller than the second preset threshold.

In some embodiments, the risk assessment apparatus 50 for driving scenes further includes a conversion module, which is configured to convert the scene risk index into a risk level according to a preset mapping relationship.

It should be noted that, when the risk assessment device for a driving scenario provided in the above embodiment performs risk assessment for a driving scenario, only the division of the above program modules is taken as an example, and in practical applications, the above processing may be distributed to different program modules according to needs, that is, the internal structure of the device may be divided into different program modules to complete all or part of the above-described processing. In addition, the risk assessment device for the driving scene and the risk assessment method for the driving scene provided by the embodiment belong to the same concept, and specific implementation processes and beneficial effects are detailed in the method embodiment and are not repeated herein. For technical details not disclosed in the embodiments of the apparatus, reference is made to the description of the embodiments of the method of the present application for understanding.

In this embodiment of the present application, fig. 6 is a schematic structural diagram of a risk assessment apparatus for a driving scenario proposed in this embodiment of the present application, and as shown in fig. 6, an apparatus 60 proposed in this embodiment of the present application may further include a processor 601, a memory 602 storing executable instructions of the processor 601, and in some embodiments, the risk assessment apparatus 60 for a driving scenario may further include a communication interface 603, and a bus 604 for connecting the processor 601, the memory 602, and the communication interface 603.

In the embodiment of the present Application, the Processor 601 may be at least one of an Application Specific Integrated Circuit (ASIC), a Digital Signal Processor (DSP), a Digital Signal Processing Device (DSPD), a ProgRAMmable Logic Device (PLD), a Field ProgRAMmable Gate Array (FPGA), a Central Processing Unit (CPU), a controller, a microcontroller, and a microprocessor. It is understood that the electronic devices for implementing the above processor functions may be other devices, and the embodiments of the present application are not limited in particular.

In the embodiment of the present application, the bus 604 is used for connecting the communication interface 603, the processor 601, and the memory 602, and the intercommunication among these devices.

In the embodiment of the present application, the processor 601 is configured to obtain a driving parameter of the host vehicle and a driving parameter of an object in a sensing range of the host vehicle; determining scene risk characteristics according to the driving parameters of the object and the driving parameters of the vehicle, wherein the scene risk characteristics comprise at least one characteristic of a longitudinal safe distance, a transverse safe distance, a safe region coincidence factor and a future collision probability; respectively determining a time sequence risk factor and a space risk factor according to different characteristics and characteristic combinations in the scene risk characteristics, wherein the time sequence risk factor reflects the scene risk degree in the time dimension, and the space risk factor reflects the scene risk degree in the space dimension; and fusing the time sequence risk factor and the space risk factor to obtain a scene risk index.

The memory 602 of the risk assessment device 60 of the driving scenario may be connected to the processor 601, the memory 602 being used for storing executable program code and data, the program code comprising computer operating instructions, the memory 602 may comprise a high speed RAM memory, and may further comprise a non-volatile memory, such as at least two disk memories. In practical applications, the Memory 602 may be a volatile Memory (volatile Memory), such as a Random-Access Memory (RAM); or a non-volatile Memory (non-volatile Memory), such as a Read-Only Memory (ROM), a flash Memory (flash Memory), a Hard Disk (Hard Disk Drive, HDD) or a Solid-State Drive (SSD); or a combination of the above types of memories and provides instructions and data to the processor 601.

In addition, each functional module in this embodiment may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware or a form of a software functional module.

Based on the understanding that the technical solution of the present embodiment essentially or a part contributing to the prior art, or all or part of the technical solution, may be embodied in the form of a software product stored in a storage medium, and include several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) or a processor (processor) to execute all or part of the steps of the method of the present embodiment. And the aforementioned storage medium includes: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.

Embodiments of the present application provide a computer-readable storage medium, on which a program is stored, and when the program is executed by a processor, the method for risk assessment of a driving scenario according to any of the above embodiments is implemented.

For example, the program instructions corresponding to the risk assessment method for a driving scenario in this embodiment may be stored in a storage medium such as an optical disc, a hard disk, or a usb disk, and when the program instructions corresponding to the risk assessment method for a driving scenario in the storage medium are read or executed by an electronic device, the risk assessment method for a driving scenario in any of the above embodiments may be implemented.

As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of a hardware embodiment, a software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, optical storage, and the like) having computer-usable program code embodied therein.

The present application is described with reference to flowchart illustrations and/or block diagrams of implementations of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart block or blocks and/or flowchart block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks in the flowchart and/or block diagram block or blocks.

The above description is only for the preferred embodiment of the present application, and is not intended to limit the scope of the present application.

Claims

1. A method for risk assessment of a driving scenario, the method comprising:

acquiring driving parameters of a self vehicle and driving parameters of an object in a self vehicle sensing range;

determining scene risk characteristics according to the driving parameters of the object and the driving parameters of the self-vehicle, wherein the scene risk characteristics comprise at least one characteristic of a longitudinal safe distance, a transverse safe distance, a safe region coincidence factor and a future collision probability;

respectively determining a time sequence risk factor and a space risk factor according to different characteristics and characteristic combinations in the scene risk characteristics, wherein the time sequence risk factor reflects the scene risk degree in a time dimension, and the space risk factor reflects the scene risk degree in a space dimension;

and fusing the time sequence risk factor and the space risk factor to obtain a scene risk index.

2. The method according to claim 1, wherein the travel parameters of the object include an object type and a speed, and the travel parameters of the own vehicle include a speed;

the determining scene risk characteristics according to the driving parameters of the object and the driving parameters of the self-vehicle comprises the following steps:

determining a preset driving limit parameter of the object according to the object type;

and calculating a longitudinal safe distance and a transverse safe distance according to the speed of the object, the speed of the self-vehicle and respective preset driving limit parameters of the object and the self-vehicle, wherein the preset driving limit parameters represent driving limit parameters in a uniform acceleration or uniform deceleration motion process, and the preset driving limit parameters of the self-vehicle are determined according to the type of the self-vehicle.

3. The method of claim 2, wherein the driving parameters of the object and the host vehicle each comprise a heading angle, and the longitudinal safety distance comprises a longitudinal co-directional driving safety distance and/or a longitudinal counter-directional driving safety distance;

the calculating the longitudinal safe distance according to the speed of the object, the speed of the self-vehicle and the respective preset driving limit parameters of the object and the self-vehicle comprises the following steps:

determining an object which runs in the same direction and/or opposite direction to the vehicle in the object according to the course angle of the object and the course angle of the vehicle;

calculating the longitudinal equidirectional running safety distance corresponding to the equidirectional running object according to the speed of the equidirectional running object and a preset running limit parameter;

and/or the presence of a gas in the gas,

and calculating the longitudinal opposite-direction running safety distance corresponding to the opposite-direction running object according to the speed of the opposite-direction running object, the speed of the self vehicle and the preset running limit parameters of the opposite-direction running object and the self vehicle.

4. The method according to claim 3, wherein the travel parameters of the object include position information of the object relative to the own vehicle, and the preset travel limit parameters include a minimum time to runaway, a maximum acceleration, a maximum deceleration, and a minimum deceleration;

the calculating the longitudinal equidirectional running safety distance corresponding to the equidirectional running object according to the speed of the equidirectional running object and a preset running limit parameter comprises the following steps:

determining objects located in front of the self vehicle and objects located behind the self vehicle in the same-direction driving objects along the driving direction of the self vehicle according to the position information of the objects relative to the self vehicle;

calculating a longitudinal maximum deceleration distance of the object located in front of the own vehicle according to the speed and the maximum deceleration of the object located in front of the own vehicle;

calculating a longitudinal runaway acceleration distance and a longitudinal minimum deceleration distance of the object behind the self-vehicle according to the speed, the minimum runaway time, the maximum acceleration and the minimum deceleration of the object behind the self-vehicle;

and calculating the longitudinal equidirectional running safety distance according to the longitudinal maximum deceleration distance of the object positioned in front of the self-vehicle, and the longitudinal uncontrolled acceleration distance and the longitudinal minimum deceleration distance of the object positioned behind the self-vehicle.

5. A method according to claim 3, characterized in that the preset driving limit parameters comprise a minimum time to runaway, a maximum acceleration and a minimum deceleration;

the calculating the longitudinal opposite-direction running safety distance corresponding to the opposite-direction running object according to the speed of the opposite-direction running object, the speed of the self vehicle and the preset running limit parameters of the object and the self vehicle comprises the following steps:

and calculating the longitudinal counter-running safe distance according to the speed, the minimum runaway time, the maximum acceleration and the minimum deceleration of the self vehicle and the counter-running object.

6. The method according to claim 2, wherein the driving parameters of the object comprise position information of the object relative to the vehicle, and the preset driving limit parameters comprise minimum out-of-control time, maximum acceleration and minimum deceleration;

the calculating the transverse safe distance according to the speed of the object, the speed of the self-vehicle and the respective preset driving limit parameters of the object and the self-vehicle comprises the following steps:

determining objects positioned at the left side and the right side of the self-vehicle in the objects along the driving direction of the self-vehicle according to the position information of the objects relative to the self-vehicle;

according to the respective speeds, the minimum out-of-control time, the maximum acceleration and the minimum deceleration of the self-vehicle and the objects positioned on the left side and the right side of the self-vehicle, the respective transverse out-of-control acceleration distance and the transverse minimum deceleration distance of the self-vehicle and the objects positioned on the left side and the right side of the self-vehicle are calculated;

and calculating the transverse safe distance according to the transverse uncontrolled acceleration distance and the transverse minimum deceleration distance of the self-vehicle and the objects positioned at the left side and the right side of the self-vehicle.

7. The method of claim 1, wherein the driving parameters of the object include a heading angle and an object type, and the driving parameters of the host vehicle include a heading angle;

determining an initialized travelable area of the object according to the object type;

adjusting the direction of the initialized travelable area of the object according to the course angle of the object, and determining the travelable area of the object;

calculating a safety region overlapping factor according to the travelable regions of the object and the self vehicle; and the travelable area of the self-vehicle is determined according to the type of the self-vehicle and the course angle of the self-vehicle.

8. The method according to claim 7, wherein the calculating a safety region coincidence factor from the travelable regions of the object and the host vehicle, respectively, comprises:

determining an intersection area of the object and the travelable area of the self vehicle according to the position information of the travelable areas of the object and the self vehicle;

and determining the ratio of the area of the intersection region to the minimum area in the travelable regions of the object and the vehicle as the safety region coincidence factor.

9. The method according to claim 1, wherein the driving parameters of the object and the host vehicle each include a heading angle, a speed, and an acceleration;

predicting a plurality of running track segments of the object and the self vehicle in a future preset time period according to the course angle, the speed and the acceleration of the object and the self vehicle, wherein the future preset time period comprises a plurality of future moments, and one future moment corresponds to one running track segment;

predicting the future collision times according to the respective travel track segments of the object and the own vehicle at the same time in the future;

and determining the ratio of the future collision frequency to the number of the plurality of driving track segments in the future preset time period as the future collision probability.

10. The method according to claim 9, wherein the driving parameters of the object and the own vehicle each include a shape;

the predicting the future collision times according to the respective travel track segments of the object and the own vehicle at the same time in the future comprises the following steps:

determining respective driving track areas of the object and the self vehicle by combining the respective corresponding driving track segments according to the respective shapes of the object and the self vehicle;

and for the same time in the future, if the object intersects with the driving track area of the self-vehicle, predicting the future collision of the object and the self-vehicle until the prediction of the future preset time period is completed, and accumulating the number of the future collision of the object and the self-vehicle to obtain the number of the future collision.

11. The method according to any one of claims 1 to 10, wherein the travel parameters of the object include position information of the object relative to the own vehicle; the number of the longitudinal safe distance, the number of the transverse safe distance, the safe region coincidence factor and the number of the future collision probability are all multiple;

determining a spatial risk factor according to different features and feature combinations in the scene risk features includes:

determining objects located in front of the self-vehicle in the driving direction of the self-vehicle according to the position information of the objects relative to the self-vehicle;

according to a first preset rule, selecting a first preset number of longitudinal safety distances from the longitudinal safety distances corresponding to the objects in front of the self-vehicle;

selecting a second preset number of transverse safe distances from the plurality of transverse safe distances according to a second preset rule;

according to a third preset rule, selecting a third preset number of safety region coincidence factors from the plurality of safety region coincidence factors;

selecting a future collision probability greater than zero from a plurality of the future collision probabilities;

wherein the spatial risk factor comprises at least one of the first preset number of longitudinal safe distances, the second preset number of lateral safe distances, the third preset number of safe zone overlap factors, and the future collision probability greater than zero.

12. The method according to any one of claims 1-10, wherein determining a time series risk factor according to different features and feature combinations in the scene risk features comprises:

acquiring scene risk characteristics of a plurality of first moments in a first preset time period;

sliding in the scene risk characteristics of a plurality of first moments in the first preset time period by taking a second preset time period as a time window and a third preset time period as a step length to determine a plurality of time windows; the second preset time period and the third preset time period are both smaller than the first preset time period;

counting the scene risk characteristics of a plurality of second moments in each time window to obtain a statistical result corresponding to each time window; the statistical results represent the time sequence risks of the scene risk characteristics in the time windows, and the time sequence risk factors comprise the statistical results corresponding to the time windows.

13. The method of claim 12, wherein the statistics comprise at least one of an average value, a maximum value, and a rate of change, wherein the average value characterizes scene macro risk, the maximum value characterizes scene worst risk, and the rate of change characterizes scene stability.

14. The method of claim 12, wherein the scene risk profile for each second time instant in each time window comprises a plurality of the longitudinal safe distances, a plurality of the lateral safe distances, a plurality of the safe-zone coincidence factors, and a plurality of the future collision probabilities;

counting the scene risk characteristics of a plurality of second moments in each time window to obtain a statistical result corresponding to each time window, wherein the statistical result comprises the following steps:

selecting the maximum value of the safety region coincidence factors according to the scene risk characteristics of each second moment to obtain the maximum value of the safety region coincidence factors;

summing the plurality of future collision probabilities to obtain a target future collision probability;

and respectively carrying out statistics on the plurality of longitudinal safe distances, the plurality of transverse safe distances, the maximum value of the safe region coincidence factor and the future collision probability of the target to obtain a statistical result corresponding to each time window.

15. The method according to any one of claims 1-10, wherein the temporal risk factors include a plurality of factors characterizing different features in a temporal dimension, and the spatial risk factors include a plurality of factors characterizing different features in a spatial dimension;

the fusion of the time sequence risk factor and the space risk factor to obtain a scene risk index includes:

respectively selecting factors of each characteristic greater than a respective corresponding preset threshold value from the time sequence risk factor and the space risk factor;

calculating the ratio of the factors of each characteristic to the total number of the corresponding factors to obtain the probability of each characteristic;

calculating products between preset feature weights corresponding to the factors of the features and the feature probabilities of the factors to obtain a plurality of feature products;

and determining the scene risk index according to the characteristic products.

16. The method of claim 15, wherein the spatial risk factor comprises a plurality of lateral/longitudinal safety distances;

the selecting, from the space risk factors, factors of each feature larger than a respective preset threshold value includes:

if the object is subjected to first attention, selecting a transverse/longitudinal safety distance larger than a first preset threshold value from the plurality of transverse/longitudinal safety distances;

if the object is subjected to second attention, selecting a transverse/longitudinal safety distance larger than a second preset threshold value from the plurality of transverse/longitudinal safety distances;

wherein the attention degree corresponding to the first attention is larger than the attention degree corresponding to the second attention, and the first preset threshold is smaller than the second preset threshold.

17. The method according to any one of claims 1-10, wherein after fusing the temporal risk factor and the spatial risk factor to obtain a scene risk index, the method further comprises:

and converting the scene risk index into a risk grade according to a preset mapping relation.

18. A risk assessment device for a driving scenario, the device comprising:

the acquisition module is used for acquiring the driving parameters of the self-vehicle and the driving parameters of the object in the sensing range of the self-vehicle;

the scene risk characteristic module is used for determining scene risk characteristics according to the driving parameters of the object and the driving parameters of the self-vehicle, wherein the scene risk characteristics comprise at least one characteristic of a longitudinal safe distance, a transverse safe distance, a safe region overlapping factor and a future collision probability;

the risk factor module is used for respectively determining a time sequence risk factor and a space risk factor according to different characteristics and characteristic combinations in the scene risk characteristics, wherein the time sequence risk factor reflects the scene risk degree in a time dimension, and the space risk factor reflects the scene risk degree in a space dimension;

and the evaluation module is used for fusing the time sequence risk factor and the space risk factor to obtain a scene risk index.

19. A risk assessment device for a driving scenario, characterized in that the device comprises a memory storing a computer program executable on a processor, and a processor implementing the method of any one of claims 1-17 when executing the program.

20. A computer-readable storage medium having stored thereon executable instructions for, when executed by a processor, implementing the method of any one of claims 1-17.