CN114444208A

CN114444208A - Method, device, equipment and medium for determining reliability of automatic driving system

Info

Publication number: CN114444208A
Application number: CN202210107033.0A
Authority: CN
Inventors: 周忠贺; 吴振昕; 刘涛; 赵朋刚; 张正龙; 迟霆; 赵思佳; 赵悦岑; 杨渊泽
Original assignee: FAW Group Corp
Current assignee: FAW Group Corp
Priority date: 2022-01-28
Filing date: 2022-01-28
Publication date: 2022-05-06

Abstract

The embodiment of the invention discloses a method, a device, equipment and a medium for determining the reliability of an automatic driving system. The method comprises the steps of obtaining a scene simulation result of the automatic driving system through a built vehicle model, a built sensor model, a built control model of the automatic driving system and a built scene model, building a response surface of a scene variable and the scene simulation result based on the scene simulation result and each scene variable corresponding to the scene model, and further determining a failure probability corresponding to the automatic driving system based on the response surface, a probability density function of each scene variable, a range of each scene variable and a preset failure condition, so that the quantification of the reliability result of the automatic driving system is realized, and a powerful support is provided for the development and improvement of the automatic driving system; and the technical problem that the prior art cannot exhaust all test scenes is solved.

Description

Method, device, equipment and medium for determining reliability of automatic driving system

Technical Field

The embodiment of the invention relates to the technical field of automatic driving, in particular to a reliability determination method, a reliability determination device, reliability determination equipment and reliability determination media for an automatic driving system.

Background

With the development of vehicle intelligence, automatic driving is a great trend, and before the automatic driving automobile is on the road, strict functional safety tests must be carried out. The automatic driving simulation test is to digitally restore an automatic driving application scene in a mathematical modeling mode, establish a system model as close to the real world as possible, and achieve the purpose of testing and verifying an automatic driving system and an algorithm without directly carrying out simulation test on a real vehicle through software. The simulation test has the advantages of high scene coverage, safe test process, high test efficiency and the like.

However, the simulation test is difficult to exhaust and test all scenes, and the prior art cannot realize quantification and evaluation of the reliability of the automatic driving system and the like.

Disclosure of Invention

The embodiment of the invention provides a reliability determination method, a reliability determination device, equipment and a medium of an automatic driving system, which are used for realizing the quantification of a reliability result of the automatic driving system and the prediction of failure probability of the automatic driving system under all test scenes under a certain kind of scenes.

In a first aspect, an embodiment of the present invention provides a reliability determination method for an automatic driving system, where the method includes:

acquiring a scene simulation result corresponding to an automatic driving system based on a pre-constructed vehicle model, a sensor model, a control model of the automatic driving system and a scene model;

determining response surfaces of the scene variables and the scene simulation result based on the scene simulation result and the scene variables corresponding to the scene model;

and determining the failure probability corresponding to the automatic driving system based on the probability density function corresponding to each scene variable, the range of each scene variable, the response surface and a preset failure condition.

Optionally, the method further includes:

acquiring preset scene variables, wherein the scene variables comprise types of the scene variables and ranges of the scene variables;

acquiring a pre-constructed simulation scene template;

and performing first sampling processing on the scene variables based on the types of the scene variables and the ranges of the scene variables, and determining a scene model based on the result of the first sampling processing and the simulation scene template.

Optionally, the method further includes:

constructing a scene static element, wherein the scene static element comprises at least one of road information, lane information and environment information;

setting up a scene dynamic element, wherein the scene dynamic element comprises at least one of traffic characteristic information, own vehicle information, target vehicle information and other traffic participant information;

acquiring a preset environment condition, a simulation duration, a simulation trigger condition and a simulation termination condition, and establishing a simulation scene template based on the scene static element, the scene dynamic element, the preset environment condition, the simulation duration, the simulation trigger condition and the simulation termination condition.

Optionally, the obtaining of the scene simulation result corresponding to the automatic driving system based on the pre-established vehicle model, the sensor model, the control model of the automatic driving system, and the scene model includes:

acquiring vehicle motion information at the current moment sent by the vehicle model and target information at the current moment sent by the sensor model based on the control model;

determining, by the control model, motion control information at a next moment to be sent to the vehicle model based on the vehicle motion information at the current moment, the target information at the current moment, and the scene model;

and acquiring a scene simulation result corresponding to the automatic driving system based on the vehicle motion information at each moment determined by the vehicle model, the motion control information at each moment determined by the control model and the target information at each moment determined by the sensor model.

Optionally, the determining the failure probability corresponding to the automatic driving system based on the probability density function corresponding to each of the scene variables, the range of each of the scene variables, the response surface, and a preset failure condition includes:

determining a sampling output result based on the probability density function corresponding to each scene variable, the range of each scene variable and the response surface;

and determining the failure probability corresponding to the automatic driving system based on the sampling output result, the probability density function and a preset failure condition.

Optionally, the determining a sampling output result based on the probability density function corresponding to each of the scene variables, the range of each of the scene variables, and the response surface includes:

acquiring a probability density function corresponding to each scene variable;

performing second sampling processing on each scene variable based on the probability density function and the range of the scene variable to obtain a sampling variable result of each scene variable;

a sampling output result is determined based on the sampling variable result and the response surface.

Optionally, the method further includes:

calculating sensitivity information corresponding to each scene variable based on the scene simulation result;

and based on the sensitivity information corresponding to each scene variable, eliminating the scene variables which do not meet the preset sensitivity requirement from each scene variable.

In a second aspect, an embodiment of the present invention further provides a reliability determination apparatus for an automatic driving system, where the apparatus includes:

the simulation module is used for acquiring a scene simulation result corresponding to the automatic driving system based on a vehicle model, a sensor model, a control model of the automatic driving system and a scene model which are constructed in advance;

the response surface construction module is used for determining the response surfaces of the scene variables and the scene simulation result based on the scene simulation result and the scene variables corresponding to the scene model;

and the failure probability determining module is used for determining the failure probability corresponding to the automatic driving system based on the probability density function corresponding to each scene variable, the range of each scene variable, the response surface and a preset failure condition.

In a third aspect, an embodiment of the present invention further provides an electronic device, where the electronic device includes:

one or more processors;

a storage device for storing one or more programs,

when executed by the one or more processors, cause the one or more processors to implement a method of determining reliability of an autopilot system as provided in any embodiment of the invention.

In a fourth aspect, embodiments of the present invention further provide a computer-readable storage medium, on which a computer program is stored, which when executed by a processor, implements a reliability determination method of an automatic driving system as provided in any of the embodiments of the present invention.

The embodiment of the invention has the following advantages or beneficial effects:

the method comprises the steps of obtaining a scene simulation result of the automatic driving system through a pre-constructed vehicle model, a pre-constructed sensor model, a pre-constructed control model of the automatic driving system and a pre-constructed scene model, constructing a response surface of a scene variable and the scene simulation result based on the scene simulation result and each scene variable corresponding to the scene model, and further determining a failure probability corresponding to the automatic driving system based on the response surface, a probability density function of each scene variable, a range of each scene variable and a preset failure condition; in addition, the method realizes the prediction of the failure probability of the automatic driving system under all test scenes under a certain kind of scenes through the probability density function of the scene variable, and solves the technical problem that the prior art can not exhaust all test scenes.

Drawings

In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, a brief description is given below of the drawings used in describing the embodiments. It should be clear that the described figures are only views of some of the embodiments of the invention to be described, not all, and that for a person skilled in the art, other figures can be derived from these figures without inventive effort.

Fig. 1A is a schematic flowchart of a reliability determination method of an automatic driving system according to an embodiment of the present invention;

fig. 1B is a schematic diagram of a response surface according to an embodiment of the present invention;

fig. 2 is a schematic flowchart of a reliability determination method of an automatic driving system according to a second embodiment of the present invention;

fig. 3 is a schematic flowchart of a reliability determination method for an automatic driving system according to a third embodiment of the present invention;

fig. 4 is a schematic flowchart of a reliability determination method of an automatic driving system according to a fourth embodiment of the present invention;

fig. 5 is a schematic structural diagram of a reliability determination apparatus of an automatic driving system according to a fifth embodiment of the present invention;

fig. 6 is a schematic structural diagram of an electronic device according to a sixth embodiment of the present invention.

Detailed Description

The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not to be construed as limiting the invention. It should be further noted that, for the convenience of description, only some of the structures related to the present invention are shown in the drawings, not all of the structures.

Example one

Fig. 1A is a schematic flowchart of a method for determining reliability of an automatic driving system according to an embodiment of the present invention, where this embodiment is applicable to a case where reliability of the automatic driving system is analyzed, and is particularly applicable to a case where a failure probability of the automatic driving system is calculated according to a vehicle model, a sensor model, a control model of the automatic driving system, and a scenario model that are constructed in advance, the method may be executed by a reliability determining apparatus of the automatic driving system, the apparatus may be implemented by hardware and/or software, and the method specifically includes the following steps:

s110, acquiring a scene simulation result corresponding to the automatic driving system based on a pre-constructed vehicle model, a pre-constructed sensor model, a pre-constructed control model of the automatic driving system and a pre-constructed scene model.

The vehicle model may be a dynamic model that simulates the motion characteristics of the vehicle in actual operation. The vehicle model may be constructed from preset body parameters, aerodynamic parameters, and transmission parameters. By way of example, parameters such as full car height, full car width, frontal area, air density, distance of the center of mass of the sprung mass from the front axle, height of the center of mass of the sprung mass from the inside, vehicle wheelbase, etc. may be used.

The sensor model may be built based on sensor information of actual test vehicle installations. For example, the sensor model may include a laser radar model, a camera model, a millimeter wave radar model, and the like. In the construction process of the sensor model, the types of the sensor model can be selected, such as a physical level, a signal level, a true value level, the installation positions of the sensors, the configuration of basic parameters and the like. Alternatively, the vehicle model and the sensor model may be pre-constructed in the autopilot simulation software.

In the present embodiment, the Automatic driving system may be a control system such as an Automatic Emergency Braking system (AEB), an uphill assist system, an antilock Braking system, a congestion following system, a high-speed ride-on system, an Automatic parking system, and the like. Specifically, the control model of the autonomous driving system may be a model for vehicle control according to an autonomous driving algorithm in the autonomous driving system. The control model of the autopilot system may be integrated with other models in the Simulink environment.

The scene model may be a model constructed from each simulation scene to be tested by the autopilot system. The scene model may be pre-constructed in the autopilot simulation software. For example, taking an automatic driving system as an automatic emergency Braking system as an example, the scene model may be a Car-to-Car reader Moving (CCRm) model, a Car-to-Car reader Braking (CCRb) model, or the like, for a straight-ahead deceleration running vehicle. The scene model can be constructed based on different values of each scene variable, and comprises a plurality of simulation test scenes. For example, for the CCRm model, the scene variables may be the speed of the vehicle, the speed of the target vehicle, the bias rate, and the like, and the scene model including a plurality of simulation test scenes may be obtained by assigning the speed of the vehicle, the speed of the target vehicle, the bias rate, and the like for a plurality of times.

Specifically, after a vehicle model, a sensor model, a control model of an automatic driving system and a scene model are established, data interaction can be performed among the models, and simulation of a test scene is performed through the interactive data to obtain a scene simulation result. For example, the vehicle model may input information such as a speed of the own vehicle, an acceleration of the own vehicle, a steering wheel angle, an opening degree of an accelerator and a brake pedal to the control model, the sensor model may input information such as a relative distance, a relative speed, and a relative angle of the target vehicle or the target object to the own vehicle to the control model, the scene model may input information related to the target vehicle or the target object to the sensor model, the vehicle model may input pose information of the vehicle to the scene model, and the control model may transmit braking information of the vehicle to the vehicle model. And scene simulation results of the automatic driving system can be extracted through data output by each model.

The scene simulation result may include various data related to the test scene, and the data included in the scene simulation result under different test scenes (scene models) are different, for example, for the CCRm model, the scene simulation result may include data such as time to collision (TTC, which may be calculated by a distance between two vehicles and a relative speed of the two vehicles), whether a collision occurs, and a minimum relative distance between the two vehicles.

For example, the obtaining a scene simulation result corresponding to an automatic driving system based on a pre-constructed vehicle model, a sensor model, a control model of the automatic driving system, and a scene model includes: acquiring vehicle motion information at the current moment sent by the vehicle model and target information at the current moment sent by the sensor model based on the control model; determining, by the control model, motion control information at a next moment to be sent to the vehicle model based on the vehicle motion information at the current moment, the target information at the current moment, and the scene model; and acquiring a scene simulation result corresponding to the automatic driving system based on the vehicle motion information at each moment determined by the vehicle model, the motion control information at each moment determined by the control model and the target information at each moment determined by the sensor model.

The target information may include information such as a position, a speed, an acceleration, a relative distance from the host vehicle, a relative speed, a relative angle, and the like of the target vehicle or the target object; the vehicle motion information may include information on the speed of the vehicle, acceleration, steering wheel angle, opening degree of an accelerator pedal and a brake pedal, and the like. Specifically, the control model can calculate the vehicle motion information at the next moment according to the vehicle motion information, the target information and the scene model at the current moment through an automatic driving algorithm, and sends the vehicle motion information as the braking information to the vehicle model so that the vehicle model moves according to the braking information at the next moment; the control model can be connected with the vehicle model, the sensor model and the scene model to form a closed loop. According to the information output by each model at each moment, a scene simulation result of the automatic driving system can be obtained.

And S120, determining a response surface of the scene variable and the scene simulation result based on the scene simulation result and each scene variable corresponding to the scene model.

Each scene variable corresponding to the scene model may be a variable required for constructing the scene model. For example, for the CCRm model, the scene variables may be host vehicle speed, host vehicle acceleration, host vehicle initial position, host vehicle travel direction, host vehicle end position, target vehicle speed, target vehicle acceleration, target vehicle initial position, target vehicle travel direction, target vehicle end position, and other traffic participant states.

Specifically, in this embodiment, a response surface between the scene variable and the scene simulation result output by the simulation system may be constructed through mathematical optimization analysis software. And constructing a response surface by using any two scene variables and a scene simulation result. The scene variable is equivalent to input X to a certain extent, the scene simulation result output by the simulation system is equivalent to Y, and the response surface can be a function for describing the X-Y relation in a three-dimensional space.

For example, as shown in fig. 1B, a schematic diagram of a response surface is shown, taking a CCRm model as an example, a scene variable uses a target vehicle speed and a host vehicle speed, a scene simulation result uses a minimum relative distance between two vehicles, that is, the target vehicle speed and the host vehicle speed are used as inputs, and the minimum relative distance between two vehicles is used as an output, so as to construct the response surface. In this embodiment, the number of the response surfaces may be one or more.

In an embodiment, it is considered that the number of the scene variables may be multiple, a part of the scene variables has a larger influence on the scene simulation result, and a part of the scene variables has a smaller influence on the scene simulation result, for example, the influence of the bias rate on the scene simulation result is very small. For scene variables without or with small influence, the scene variables can be omitted in subsequent response surface construction and failure probability analysis, so that calculation is simplified, and the reliability analysis efficiency of the automatic driving system is improved. Namely, the method further comprises: calculating sensitivity information corresponding to each scene variable based on the scene simulation result; and based on the sensitivity information corresponding to each scene variable, eliminating the scene variables which do not meet the preset sensitivity requirement from each scene variable.

The sensitivity information may describe, among other things, the effect of uncertainty of scene variables in the autopilot system on the scene simulation results. For example, the sensitivity information corresponding to each scene variable is calculated by a variance-based method. By the method, the sensitivity of each scene variable can be analyzed, so that the importance of each scene variable to the scene simulation result is determined and quantified, the subsequent response surface construction or the failure probability analysis is carried out only for the scene variables meeting the preset sensitivity requirement, the calculation is simplified, and the failure probability analysis efficiency is improved. It should be noted that the step of calculating the sensitivity information corresponding to each of the scene variables may be performed before constructing the response surface, so as to construct the response surface for the important scene variables, reduce the amount of calculation, and improve the efficiency of analyzing the failure probability.

S130, determining the failure probability corresponding to the automatic driving system based on the probability density function corresponding to each scene variable, the range of each scene variable, the response surface and a preset failure condition.

The probability density function of the scene variable can be obtained based on big data analysis or vehicle networking data extraction and analysis. For example, according to big data analysis, the vehicle speed substantially follows a normal distribution on a road, and therefore, the probability density function of the vehicle speed may be a normal distribution function. The range of the scene variable may be a value range of the scene variable, for example, 0 to 100; it should be noted that the scene variables are of various types, such as continuous, discrete, constant, or function, and the value ranges of the scene variables of different types may be different.

The preset failure condition may be a preset failure condition of the automatic driving system. For example, for a CCRm scenario of an AEB system, the predetermined failure condition may be a collision between two vehicles, i.e., a minimum relative distance between two vehicles is less than 0.

Specifically, the failure probability of the automatic driving system under the scene model can be calculated by using a probability method in the mathematical optimization analysis software through the probability density function of each scene variable, the range of each scene variable, each constructed response surface and a preset failure condition. For example, values of the scene variables can be sampled within the range of the scene variables, a simulation result corresponding to the sampled value of each scene variable is determined according to the response surface, and further, the failure probability of the automatic driving system is calculated according to the simulation result corresponding to each sampled value, the preset failure condition and the probability density function.

In this embodiment, joint simulation may be performed by using the autopilot simulation software and the mathematical optimization analysis software, a control model, a vehicle model, a sensor model, and a scene model of the autopilot system are integrated based on the autopilot simulation software, a simulation scene for testing an autopilot control algorithm in the autopilot system is established, a response surface is constructed based on the mathematical optimization analysis software, and a probability density function of a scene variable is defined and then a failure probability is calculated. The system can perform combined simulation based on the automatic driving simulation software and the mathematical optimization analysis software, can realize automatic test, can also utilize cloud accelerated test, and is beneficial to improving efficiency and reducing cost; the method also has the characteristics of high coverage and high safety, particularly, the test scene can be flexibly configured, the simulation test and the subsequent calculation analysis process can be carried out on extreme and dangerous working conditions, and the automatic driving function test blind area can be effectively covered; in addition, the method can also realize the quantification of the analysis result, the reliability analysis of the automatic driving system is carried out by adopting a probability-based method, the reliability analysis result can be explained and quantified, and powerful support can be provided for the development and improvement of an automatic driving algorithm.

According to the technical scheme of the embodiment, a scene simulation result of the automatic driving system is obtained through a pre-constructed vehicle model, a sensor model, a control model of the automatic driving system and a scene model, a response surface of a scene variable and the scene simulation result is constructed based on the scene simulation result and each scene variable corresponding to the scene model, and further a failure probability corresponding to the automatic driving system is determined based on the response surface, a probability density function of each scene variable, a range of each scene variable and a preset failure condition; in addition, the method realizes the prediction of the failure probability of the automatic driving system under all test scenes under a certain kind of scenes through the probability density function of the scene variable, and solves the technical problem that the prior art can not exhaust all test scenes.

Example two

Fig. 2 is a schematic flow chart of a reliability determination method for an automatic driving system according to a second embodiment of the present invention, where on the basis of the foregoing embodiment, optionally, the method further includes: acquiring preset scene variables, wherein the scene variables comprise the types of the scene variables and the ranges of the scene variables; acquiring a pre-constructed simulation scene template; and performing first sampling processing on the scene variables based on the types of the scene variables and the ranges of the scene variables, and determining a scene model based on the result of the first sampling processing and the simulation scene template. Wherein explanations of the same or corresponding terms as those of the above embodiments are omitted. Referring to fig. 2, the method for determining the reliability of the automatic driving system according to the present embodiment includes the following steps:

s210, obtaining preset scene variables, wherein the scene variables comprise types of the scene variables and ranges of the scene variables.

The preset scene variables can be defined according to the type of the scene to be tested. For example, taking the test AEB system as an example, if the scene to be tested is CCRm, the scene variables may be defined to include information such as the speed of the host vehicle, the speed of the target vehicle, and the bias rate, where the bias rate is a proportion of the host vehicle to the overlapping portion of the host vehicle and the target vehicle, the reference line defined by overlapping is the center line of the host vehicle, and in the case of 100% overlap, the center lines of the host vehicle and the target vehicle are aligned.

Of course, while defining each scene variable, the type and range of the scene variable may also be defined, where the type of the scene variable may be a type such as continuous, discrete, constant, or function, and the range of the scene variable may be a value range. For example, the type of host vehicle speed or target vehicle speed may be a continuous variable; the type of road property (e.g., the attachment coefficient of the road) may be a discrete variable.

S220, acquiring a pre-constructed simulation scene template, performing first sampling processing on each scene variable based on the type of the scene variable and the range of the scene variable, and determining a scene model based on the result of the first sampling processing and the simulation scene template.

Specifically, in this embodiment, the first sampling processing may be performed on the scene variable according to the type of the scene variable and the range of the scene variable, so as to obtain different values of the scene variable. In the embodiment, random sampling can be performed according to the type and range of the scene variable, the strategy and the number of sampling can be determined according to the scene model, a simple random sampling strategy can be generally adopted, and the number of sampling is 10²～10³The magnitude can meet the requirements.

In this way, a plurality of values of each scene variable are obtained as a result of the first sampling process. Further, substituting the result of the first sampling treatment, namely a plurality of values of each scene variable into a pre-constructed simulation scene template to obtain each simulation test scene in the scene model. Specifically, after each scene variable takes one value, the scene variables which take the value are combined to form a sample, and the sample is substituted into a simulation scene template, which is equivalent to establishing a simulation test scene; each simulation test scenario constitutes a scenario model. Illustratively, a1, b1, c1 constitute one sample; a2, b2, c2 constitute one sample, and so on.

Illustratively, the simulation scenario template may be constructed in the following manner, that is, optionally, the method further includes: constructing a scene static element, wherein the scene static element comprises at least one of road information, lane information and environment information; setting up a scene dynamic element, wherein the scene dynamic element comprises at least one of traffic characteristic information, own vehicle information, target vehicle information and other traffic participant information; acquiring a preset environment condition, a simulation time length, a simulation trigger condition and a simulation termination condition, and establishing a simulation scene template based on the scene static element, the scene dynamic element, the preset environment condition, the simulation time length, the simulation trigger condition and the simulation termination condition.

Namely, the construction of the simulation scene template comprises the construction of a scene static element, the construction of a scene dynamic element, the definition of a preset environment condition, and the setting of a simulation duration, a simulation trigger condition and a simulation termination condition. Wherein the scene static element may include at least one of road information, lane information, and environment information. The road information may include road geometry (e.g., road length, road width), road surface material, and number of road lanes. The lane information may include lane length, lane width, lane line information. The environmental information may include traffic lights, pavement markers, traffic signs, barricades, fences, road structures (e.g., bridges, tunnels), and the like. The traffic characteristic information can be information such as vehicle density, vehicle speed, pedestrian and vehicle distribution and the like; the vehicle information may include vehicle geometric model information, vehicle motion parameters, vehicle motion route and other information; the target vehicle information can comprise information such as geometric model information of the target vehicle, motion parameters of the target vehicle, motion routes of the target vehicle and the like; other traffic participant information may include motion information, geometric model information, etc. of the participating objects, such as pedestrians, animals, etc. The preset environmental condition may be a setting condition for the simulation environmental element, which includes, but is not limited to, setting of data of weather type, light visibility, and the like. The simulation trigger condition may be a condition that triggers the start of a simulation test, such as the start of braking of the host vehicle. The simulation termination condition may be a condition for terminating the simulation test, such as a time at which the vehicle speed of the host vehicle decreases to 0; of course, in order to ensure that a sufficient scene simulation result is acquired, the simulation end condition may set the vehicle speed of the host vehicle to decrease to 0 and elapse the set delay time. By the method, the simulation scene template is constructed, and each simulation test scene in the scene model can be obtained based on the simulation scene template.

And S230, acquiring a scene simulation result corresponding to the automatic driving system based on a pre-constructed vehicle model, a pre-constructed sensor model, a pre-constructed control model of the automatic driving system and a pre-constructed scene model.

S240, determining a response surface of the scene variable and the scene simulation result based on the scene simulation result and each scene variable corresponding to the scene model.

And S250, determining the failure probability corresponding to the automatic driving system based on the probability density function corresponding to each scene variable, the range of each scene variable, the response surface and a preset failure condition.

According to the technical scheme of the embodiment, each scene variable is defined, the simulation scene template is constructed, first sampling processing is carried out on each scene variable according to the type of the scene variable and the range of the scene variable, and then the scene model is determined according to the result of the first ozone processing and the simulation scene template, so that the establishment of each simulation test scene is realized. In addition, the simulation test and the subsequent calculation analysis process of extreme and dangerous working conditions can be realized through flexible configuration of the simulation test scene, and the dead zone of the automatic driving function test can be effectively covered.

EXAMPLE III

Fig. 3 is a schematic flow chart of a reliability determination method for an automatic driving system according to a third embodiment of the present invention, where, on the basis of the third embodiment, optionally, the determining a failure probability corresponding to the automatic driving system based on a probability density function corresponding to each of the scene variables, a range of each of the scene variables, the response surface, and a preset failure condition includes: determining a sampling output result of each scene variable based on a probability density function corresponding to each scene variable, a range of each scene variable and the response surface; and determining the failure probability corresponding to the automatic driving system based on the sampling output result, the probability density function and a preset failure condition. Wherein explanations of the same or corresponding terms as those of the above embodiments are omitted. Referring to fig. 3, the method for determining the reliability of the automatic driving system according to the present embodiment includes the following steps:

s310, acquiring a scene simulation result corresponding to the automatic driving system based on a pre-constructed vehicle model, a pre-constructed sensor model, a pre-constructed control model of the automatic driving system and a pre-constructed scene model.

S320, determining a response surface of the scene variable and the scene simulation result based on the scene simulation result and each scene variable corresponding to the scene model.

S330, determining a sampling output result based on the probability density function corresponding to each scene variable, the range of each scene variable and the response surface.

The sampling output result may be a scene simulation result corresponding to the sampling value of the scene variable in the response surface. For example, the determining a sampling output result based on the probability density function corresponding to each of the scene variables, the range of each of the scene variables, and the response surface may be: acquiring a probability density function corresponding to each scene variable; performing second sampling processing on each scene variable based on the probability density function and the range of the scene variable to obtain a sampling variable result of each scene variable; a sampling output result is determined based on the sampling variable result and the response surface.

Specifically, the second sampling processing may be performed on the scene variables to obtain different values of each scene variable, and further, a scene simulation result, i.e., a sampling output result, corresponding to the sampling variable result may be quickly obtained through the response surface. It should be noted that the sampling output result may include a simulated scene simulation result, and may also include a result of fitting when a response surface is constructed.

Optionally, second sampling processing may be performed on the scene variable based on a preset constraint condition, where the preset constraint condition may be a preset condition for constraining a value of the sampled scene variable; for example, the vehicle speed is greater than the front vehicle speed. The reliability analysis of the scene variable values with small influence can be avoided through the constraint conditions, so that the calculated amount of the reliability analysis is reduced, and the analysis efficiency is further improved.

Wherein, the method of the second sampling process can be selected according to the number of the scene variables, the form of the failure condition and the number of the samples, and the method of the second sampling process includes but is not limited to Monte Carlo sampling, Latin hypercube sampling, importance sampling and adaptive sampling.

For example, the second sampling process may be performed on each scene variable by using a latin hypercube sampling method. The Latin hypercube sampling method is used for randomly sampling in a design space by utilizing a layering principle, so that not only can the sampling points be prevented from gathering, but also the good space coverage can be ensured, and the high sampling efficiency can be ensured. The process of performing the second sampling processing on each scene variable by adopting the Latin hypercube sampling method can be as follows: (1) dividing the space of each scene variable into N parts according to a sample point N to be sampled; (2) carrying out equal probability random sampling once in each subspace domain of the scene variable to obtain N data in total; (3) the N data of each scene variable are randomly matched to N sample points (where each data is used only once when matched).

S340, determining the failure probability corresponding to the automatic driving system based on the sampling output result, the probability density function and the preset failure condition.

Specifically, whether a scene simulation result corresponding to each scene variable value in the sampling output result is a failure result or not can be judged through a preset failure condition, the distribution probability of the scene variable corresponding to each failure result in the probability density function is obtained, and the failure probability of the automatic driving system is calculated according to the distribution probability of the scene variable corresponding to each failure result.

According to the technical scheme of the embodiment, the sampling output result is determined through the probability density function corresponding to each scene variable, the range of each scene variable and the response surface, and the failure probability corresponding to the automatic driving system is further determined through the sampling output result, the probability density function and the preset failure condition, so that the reliability analysis result of the automatic driving system is quantized, and the failure probability of the automatic driving system under all test scenes is predicted. For example, a range of a certain scenario variable is [0,100], and the scenario variable is continuous in the range, then, a value of the scenario variable may be infinite, so that the number of simulation test scenarios may also be infinite, however, the prior art cannot test all simulation test scenarios. By adopting the method provided by the embodiment, the sample can be replaced by a whole by a sampling method, and the failure probability is calculated by sampling the output result, the probability density function and the preset failure condition, so that the reliability calculation and analysis of the automatic driving system can be realized by using an efficient sampling method, the technical problem that the prior art cannot exhaust all test scenes is solved, and the accuracy of the reliability analysis result of the automatic driving system is improved.

Example four

Fig. 4 is a schematic flowchart of a method for determining reliability of an automatic driving system according to a fourth embodiment of the present invention, and this embodiment is applicable to a situation where reliability of the automatic driving system is analyzed, and is particularly applicable to a situation where a failure probability of the automatic driving system is calculated according to a vehicle model, a sensor model, a control model of the automatic driving system, and a scene model that are constructed in advance, as shown in fig. 4, the method for determining reliability of an automatic driving system according to this embodiment includes the following steps:

and S410, establishing a data interaction channel of the automatic driving simulation software and the mathematical optimization analysis software.

At present, the automatic driving simulation software and the mathematical optimization analysis software generally provide rich interfaces, and a data interaction channel between the simulation software and the mathematical optimization analysis software can be opened in a direct (the simulation software and the mathematical optimization analysis software interface are communicated) or indirect mode (by means of python programs and other tools).

And S420, building a vehicle model and a sensor model in the automatic driving simulation software, and integrating a control model.

Specifically, the control model may be built in Simulink or implemented in C + + code and integrated with other models (vehicle model, sensor model, etc.). A vehicle model and a sensor model can be built in the automatic driving simulation software. The vehicle model inputs information such as speed, acceleration, steering wheel angle, opening degree of an accelerator and a brake pedal and the like of the vehicle to the control model; the sensor model inputs information such as a relative distance, a relative speed, a relative angle, and the like of the target object and the vehicle to the control model. In this way, the control model (autopilot control algorithm) and the vehicle model are connected to form a closed loop.

And S430, building a scene model in the automatic driving simulation software.

Specifically, the construction of the scene model comprises the following steps:

step 1, determining a test scene type according to a tested automatic driving system, and defining scene variables (including defining the type, the range and the like of the scene variables);

step 2, building a simulation scene template;

the establishing of the simulation scene template comprises the following steps: (1) defining static elements, including establishing a scene road model, setting road length, lane width, road material attribute, lane line, configuration environment information and the like; (2) defining dynamic elements, including setting states of the vehicle, the target vehicle and other traffic participants, such as vehicle speed, initial position, driving direction and the like; (3) setting environmental conditions (light, rain, snow, fog, wind, etc.); (4) and setting simulation duration, simulation trigger conditions and simulation termination conditions.

And 3, sampling the scene variables for the first time, determining each simulation test scene according to the sampling result and the simulation scene template, and forming a scene model according to each simulation test scene.

And S440, performing combined simulation based on the automatic driving simulation software and the mathematical optimization analysis software, and feeding back a scene simulation result to the mathematical optimization analysis software through a data channel through the automatic driving simulation software.

Specifically, a vehicle model, a sensor model, a control model of an automatic driving system and a scene model are simulated through automatic driving simulation software to obtain a scene simulation result, and the automatic driving simulation software transmits the scene simulation result to mathematical optimization analysis software through a data channel.

S450, constructing a response surface between the scene variables and the scene simulation result through mathematical optimization analysis software, and determining the sensitivity information of each scene variable.

Specifically, a response surface between the scene variable and the scene simulation result can be constructed according to the scene variable and the scene simulation result through an internal algorithm in the mathematical optimization analysis software.

And S460, defining the distribution type of the scene variable and the probability density function of the scene variable.

And S470, defining a preset constraint condition and a preset failure condition.

Specifically, a preset constraint condition and a preset failure condition may be defined by using mathematical optimization analysis software. For example, for testing the CCRm scenario of the AEB system, the preset constraint condition may be defined as the speed of the vehicle is greater than the speed of the vehicle ahead, and the preset failure condition may be defined as the two vehicles colliding, i.e. the minimum relative distance between the two vehicles is less than 0.

And S480, sampling the scene variables for the second time, and determining the failure probability according to the probability density function, the preset failure condition and the response surface.

Specifically, each scene variable can be sampled for the second time according to the preset constraint condition, the range of the scene variable and the probability density function, further, a simulation result corresponding to the sampling value of the scene variable is determined according to the response surface, whether the simulation result is a failure result is judged based on the preset failure condition, and then the failure probability is calculated based on the distribution probability of the scene variable value corresponding to the failure result in the probability density function.

According to the technical scheme of the embodiment, a simulation scene template is established according to a tested automatic driving control algorithm, scene variables and the range thereof are defined in the mathematical optimization analysis software according to the established scene, a proper random sampling method is selected according to the types of the scene variables, the scene variables are sampled and combined to form a specific simulation test scene, the generated simulation test scene automatically executes a simulation process in the simulation software, and a scene simulation result is automatically transmitted to the mathematical optimization analysis software. In the mathematical optimization analysis software, a high-quality response surface between the scene variable and the scene simulation result is constructed, and the sensitivity of the scene variable is quantified based on the analysis method of variance, so that the importance of the scene variable is determined. The probability distribution function based on the scene variable and the like applies an efficient sampling method and algorithm, calculates the failure probability of the system by defining the failure condition of the system, and realizes the calculation and quantification of the reliability of the automatic driving system.

EXAMPLE five

Fig. 5 is a schematic structural diagram of a reliability determining apparatus of an automatic driving system according to a fifth embodiment of the present invention, which is applicable to a case of analyzing reliability of the automatic driving system, and is particularly applicable to a case of calculating a failure probability of the automatic driving system according to a vehicle model, a sensor model, a control model of the automatic driving system, and a scene model that are constructed in advance, and the apparatus specifically includes: a simulation module 510, a response surface construction module 520, and a failure probability determination module 530.

The simulation module 510 is configured to obtain a scene simulation result corresponding to an automatic driving system based on a vehicle model, a sensor model, a control model of the automatic driving system, and a scene model that are constructed in advance;

a response surface constructing module 520, configured to determine a response surface between the scene variable and the scene simulation result based on the scene simulation result and each scene variable corresponding to the scene model;

a failure probability determining module 530, configured to determine a failure probability corresponding to the automatic driving system based on a probability density function corresponding to each of the scene variables, a range of each of the scene variables, the response surface, and a preset failure condition.

Optionally, the apparatus further includes a scene model construction module, where the scene model construction module is configured to obtain preset scene variables; acquiring a pre-constructed simulation scene template; performing first sampling processing on each scene variable based on the type of the scene variable and the range of the scene variable, and determining a scene model based on the result of the first sampling processing and the simulation scene template, wherein the scene variable comprises the type of the scene variable and the range of the scene variable.

Optionally, the device further includes a scene template building module, where the scene template building module is configured to build a scene static element, where the scene static element includes at least one of road information, lane information, and environment information; constructing a scene dynamic element, wherein the scene dynamic element comprises at least one of traffic characteristic information, own vehicle information, target vehicle information and other traffic participant information; acquiring a preset environment condition, a simulation duration, a simulation trigger condition and a simulation termination condition, and establishing a simulation scene template based on the scene static element, the scene dynamic element, the preset environment condition, the simulation duration, the simulation trigger condition and the simulation termination condition.

Optionally, the simulation module 510 is specifically configured to:

acquiring vehicle motion information at the current moment sent by the vehicle model and target information at the current moment sent by the sensor model based on the control model; determining, by the control model, motion control information at a next moment to be sent to the vehicle model based on the vehicle motion information at the current moment, the target information at the current moment, and the scene model; and acquiring a scene simulation result corresponding to the automatic driving system based on the vehicle motion information at each moment determined by the vehicle model, the motion control information at each moment determined by the control model and the target information at each moment determined by the sensor model.

Optionally, the failure probability determination module 530 includes a sampling unit and a probability calculation unit; wherein the content of the first and second substances,

the sampling unit is used for determining a sampling output result based on the probability density function corresponding to each scene variable, the range of each scene variable and the response surface;

and the probability calculation unit is used for determining the failure probability corresponding to the automatic driving system based on the sampling output result, the probability density function and a preset failure condition.

Optionally, the sampling unit is specifically configured to:

acquiring a probability density function corresponding to each scene variable; performing second sampling processing on each scene variable based on the probability density function and the range of the scene variable to obtain a sampling variable result of each scene variable; a sample output result is determined based on the sample variable result and the response surface.

Optionally, the apparatus further includes a sensitivity analysis module, where the sensitivity analysis module is configured to calculate, based on the scene simulation result, sensitivity information corresponding to each of the scene variables; and based on the sensitivity information corresponding to each scene variable, eliminating the scene variables which do not meet the preset sensitivity requirement from each scene variable.

In the embodiment, a pre-constructed vehicle model, a sensor model, a control model of an automatic driving system and a scene model are obtained through a simulation module, a scene simulation result of the automatic driving system is determined, a response surface of a scene variable and the scene simulation result is constructed through a response surface construction module based on the scene simulation result and each scene variable corresponding to the scene model, and a failure probability corresponding to the automatic driving system is further determined through a failure probability determination module based on the response surface, a probability density function of each scene variable, a range of each scene variable and a preset failure condition; in addition, the method realizes the prediction of the failure probability of the automatic driving system under all test scenes through the probability density function of the scene variable, and solves the technical problem that the prior art cannot exhaust all test scenes.

The reliability determining device of the automatic driving system provided by the embodiment of the invention can execute the reliability determining method of the automatic driving system provided by any embodiment of the invention, and has corresponding functional modules and beneficial effects of the executing method.

It should be noted that, the units and modules included in the system are merely divided according to functional logic, but are not limited to the above division as long as the corresponding functions can be realized; in addition, specific names of the functional units are only for convenience of distinguishing from each other, and are not used for limiting the protection scope of the embodiment of the present invention.

EXAMPLE six

Fig. 6 is a schematic structural diagram of an electronic device according to a sixth embodiment of the present invention. FIG. 6 illustrates a block diagram of an exemplary electronic device 12 suitable for use in implementing embodiments of the present invention. The electronic device 12 shown in fig. 6 is only an example and should not bring any limitation to the function and the scope of use of the embodiment of the present invention. The device 12 is typically an electronic device that undertakes the functionality of an autopilot system reliability analysis.

As shown in FIG. 6, electronic device 12 is embodied in the form of a general purpose computing device. The components of electronic device 12 may include, but are not limited to: one or more processors or processing units 16, a memory 28, and a bus 18 that couples the various components (including the memory 28 and the processing unit 16).

Bus 18 represents one or more of any of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, and a processor or local bus using any of a variety of bus architectures. By way of example, such architectures include, but are not limited to, an Industry Standard Architecture (ISA) bus, a Micro Channel Architecture (MCA) bus, an enhanced ISA bus, a Video Electronics Standards Association (VESA) local bus, and a Peripheral Component Interconnect (PCI) bus.

Electronic device 12 typically includes a variety of computer-readable media. Such media may be any available media that is accessible by electronic device 12 and includes both volatile and nonvolatile media, removable and non-removable media.

Memory 28 may include computer device readable media in the form of volatile Memory, such as Random Access Memory (RAM) 30 and/or cache Memory 32. The electronic device 12 may further include other removable/non-removable, volatile/nonvolatile computer storage media. By way of example only, the storage device 34 may be used to read from and write to non-removable, nonvolatile magnetic media (which are not shown in FIG. 6 and are commonly referred to as "hard drives"). Although not shown in FIG. 6, a magnetic disk drive for reading from and writing to a removable, nonvolatile magnetic disk (e.g., a "floppy disk") and an optical disk drive for reading from or writing to a removable, nonvolatile optical disk (e.g., a Compact disk-Read Only Memory (CD-ROM), a Digital Video disk (DVD-ROM), or other optical media) may be provided. In these cases, each drive may be connected to bus 18 by one or more data media interfaces. Memory 28 may include at least one program product 40, with program product 40 having a set of program modules 42 configured to carry out the functions of embodiments of the invention. Program product 40 may be stored, for example, in memory 28, and such program modules 42 include, but are not limited to, one or more application programs, other program modules, and program data, each of which examples or some combination may comprise an implementation of a network environment. Program modules 42 generally carry out the functions and/or methodologies of embodiments of the invention as described.

Electronic device 12 may also communicate with one or more external devices 14 (e.g., keyboard, mouse, camera, etc., and display), one or more devices that enable a user to interact with electronic device 12, and/or any devices (e.g., network card, modem, etc.) that enable electronic device 12 to communicate with one or more other computing devices. Such communication may be through an input/output (I/O) interface 22. Also, the electronic device 12 may communicate with one or more networks (e.g., a Local Area Network (LAN), Wide Area Network (WAN), and/or a public Network such as the internet) via the Network adapter 20. As shown, the network adapter 20 communicates with other modules of the electronic device 12 via the bus 18. It should be understood that although not shown in the figures, other hardware and/or software modules may be used in conjunction with electronic device 12, including but not limited to: microcode, device drivers, Redundant processing units, external disk drive Arrays, disk array (RAID) devices, tape drives, and data backup storage devices, to name a few.

The processor 16 executes various functional applications and data processing by executing programs stored in the memory 28, for example, to implement the reliability determination method of the automatic driving system provided by the above embodiment of the present invention, including:

Of course, those skilled in the art will appreciate that the processor may also implement the technical solution of the reliability determination method of the automatic driving system provided in any embodiment of the present invention.

EXAMPLE seven

Seventh, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the steps of the method for determining reliability of an automatic driving system according to any embodiment of the present invention, where the method includes:

Computer storage media for embodiments of the invention may employ any combination of one or more computer-readable media. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the computer readable storage medium would include the following: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.

A computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.

Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

Computer program code for carrying out operations for embodiments of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider).

It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present invention and the technical principles employed. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, although the present invention has been described in greater detail by the above embodiments, the present invention is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A reliability determination method of an autonomous driving system, the method comprising:

2. The method of claim 1, further comprising:

acquiring a pre-constructed simulation scene template;

3. The method of claim 2, further comprising:

4. The method according to claim 1, wherein the obtaining of the scene simulation result corresponding to the automatic driving system based on the pre-constructed vehicle model, the sensor model, the control model of the automatic driving system and the scene model comprises:

5. The method of claim 1, wherein determining the failure probability corresponding to the automatic driving system based on the probability density function corresponding to each of the scene variables, the range of each of the scene variables, the response surface, and a preset failure condition comprises:

6. The method of claim 5, wherein determining a sample output based on the probability density function for each of the scene variables, the range for each of the scene variables, and the response surface comprises:

acquiring a probability density function corresponding to each scene variable;

7. The method of claim 1, further comprising:

8. A reliability determination apparatus of an automatic driving system, characterized in that the apparatus comprises:

9. An electronic device, characterized in that the electronic device comprises:

one or more processors;

a storage device for storing one or more programs,

when executed by the one or more processors, cause the one or more processors to implement a reliability determination method for an autopilot system according to any one of claims 1-7.

10. A computer-readable storage medium, on which a computer program is stored, which program, when being executed by a processor, is adapted to carry out the method for determining the reliability of an autopilot system according to one of the claims 1 to 7.