Autonomous Driving Test Scenario Generation System and Method for Data Infrastructure for Testing and Evaluation of Autonomous Driving Systems

Abstract

To provide an autonomous travel test scenario generation system for data infrastructure for testing and evaluating an autonomous travel system, and a method.SOLUTION: An autonomous travel test scenario generation system includes: a tracking object processing part that generates a tracking objet by using tracking object data acquired by an autonomous vehicle while traveling, and identifies some of untracked objects among the tracked objects; an untracked factor reflection part that reflects untracked factors to untracked object data; and a scenario generation part that generates autonomous travel test scenarios by using elements within an operational design scope of the autonomous travel test scenarios and the untracked object data obtained by reflecting the untracked factors. Accordingly, scenarios that can test not only decision-making and control functions but also cognitive functions based on data obtained from autonomous travel are generated, enabling more comprehensive and delicate testing and evaluation of autonomous travel systems that perform dynamic driving tasks.SELECTED DRAWING: Figure 1

Description

Technical Field

[0004] , ,

[0006] , , , ,

[0005] , , , , , ,

[0003]

[0001] The present invention relates to the generation of autonomous driving test scenarios, and more particularly, to a system and method for processing data collected from autonomous driving vehicles and generating scenarios for autonomous driving tests and evaluations.

Background Art

[0002] In order to introduce a vehicle equipped with an autonomous driving system into the market, reasonable verification and evaluation of the autonomous driving system are necessary. For testing the autonomous driving system, a scenario-based testing method is used.

[0003] In scenario-based testing, in order to confirm whether there is any abnormality in operating within the operating design range (range for roads, weather, and traffic), which is the range in which an autonomous driving system that performs dynamic driving tasks including cognitive, judgment, and control functions that can replace the driver's driving can operate, scenarios of various situations are generated and tested.

[0004] Conventional scenario-based testing focuses only on testing the judgment and control of the autonomous driving system, and does not generate situations in which cognitive errors occur from scenarios.

[0005] However, for the correct judgment and control of the autonomous driving system, accurate recognition must be a prerequisite, so a measure for supplementing it is necessary. In particular, it is required that a cognitive test based on data corresponding to the autonomous driving situation, rather than a cognitive test using general data, be performed.

Summary of the Invention

Problems to be Solved by the Invention

[0006] Therefore, the present invention has been made in view of the above problems, and the object of the present invention is to provide a system and method for generating a scenario that can evaluate an autonomous driving system that performs dynamic driving tasks including perception, judgment, and control functions, based on data that can be obtained from autonomous driving. [Means for solving the problem]

[0007] A method for generating an autonomous driving test scenario according to one embodiment of the present invention for achieving the above objectives includes the steps of: generating tracking objects using tracking object data acquired while an autonomous vehicle is driving; identifying untracked objects and supplementing untracked sections; reflecting untracked factors in the untracked object data; identifying elements within the operational design range of the autonomous driving test scenario; and generating an autonomous driving test scenario using the identified elements and the untracked object data reflecting the untracked factors.

[0008] The identified elements may include dynamic elements, static elements, and environmental elements.

[0009] The generation step may use the identified elements, and for untracked objects among the dynamic elements identified during autonomous driving test scenario generation, use untracked object data that reflects the untracked factors.

[0010] The completion step may involve identifying untracked objects as those where the interval between the end of tracking and the start of tracking is less than or equal to a certain time, and the interval between the end of tracking and the subsequent start of tracking is less than or equal to a certain distance.

[0011] The reflection step may include the step of analyzing the factors causing untracked objects to remain untracked, and the step of reflecting the analyzed factors in the untracked object data.

[0012] Untracked factors may include the color and reflectance of the untracked object.

[0013] The autonomous driving test scenario generation method according to the present invention may further include the step of identifying and filtering out erroneously tracked objects.

[0014] The filtering step may involve identifying false tracking objects based on the behavior of the tracking object.

[0015] The autonomous driving test scenario generation method according to the present invention may further include the step of organizing tracking object data for filtered and completed tracking objects.

[0016] According to another aspect of the invention, an autonomous driving test scenario generation system is provided, which includes an object generation unit that generates tracked objects using tracked object data acquired by an autonomous vehicle while it is driving; a completion unit that identifies untracked objects and completes untracked sections; a reflection unit that reflects untracked factors in the untracked object data; an identification unit that identifies elements within the operational design range of the autonomous driving test scenario; and a scenario generation unit that generates an autonomous driving test scenario using the identified elements and the untracked object data with the untracked factors reflected. [Effects of the Invention]

[0017] As described above, according to the present invention, by generating scenarios that can test not only judgment and control functions but also cognitive functions based on data acquired during autonomous driving, it becomes possible to perform more comprehensive and detailed testing and evaluation of autonomous driving systems that perform dynamic driving tasks. [Brief explanation of the drawing]

[0018] [Figure 1] This figure shows an autonomous driving test scenario generation system according to one embodiment of the present invention. [Figure 2] This diagram shows the detailed configuration of the tracking object processing unit. [Figure 3]It is a diagram showing the detailed configuration of the untracked factor reflection unit. [Figure 4] It is a diagram showing the detailed configuration of the scenario area identification unit. [Figure 5] It is a diagram showing an autonomous driving taste scenario generation method according to an embodiment of the present invention. [Figure 6] It is a diagram showing an autonomous driving taste scenario generation method according to an embodiment of the present invention.

Mode for Carrying Out the Invention

[0019] Hereinafter, the present invention will be described in more detail with reference to the drawings.

[0020] In an embodiment of the present invention, an autonomous driving test scenario generation system and method for a data base for testing and evaluating an autonomous driving system are presented.

[0021] Based on the actual data that can be obtained from autonomous driving, by generating a scenario that can test not only judgment and control functions but also cognitive functions, it is a technology that enables more comprehensive and delicate testing and evaluation of an autonomous driving system that performs dynamic driving tasks.

[0022] FIG. 1 is a diagram showing an autonomous driving test scenario generation system according to an embodiment of the present invention. The autonomous driving test scenario generation system according to an embodiment of the present invention includes, as shown in the figure, an autonomous driving DB 110, a tracking object processing unit 120, an untracked factor reflection unit 130, a precise map DB 140, a scenario area identification unit 150, a scenario generation unit 160, and a scenario DB 170.

[0023] The autonomous driving DB 110 is a database in which sensor data, tracking object data, internal / external status data, and weather data collected while an autonomous driving vehicle actually runs are stored.

[0024] The tracking object processing unit 120 generates tracking objects using the tracking object data stored in the autonomous driving DB 110 and performs the necessary subsequent processing on them. The detailed configuration of the tracking object processing unit 120 is shown in Figure 2.

[0025] As shown in the figure, the tracking object processing unit 120 is composed of a tracking object generation unit 121, a mistracking object filtering unit 122, an untracking object completion unit 123, and a tracking object organization unit 124.

[0026] The tracking object generation unit 121 generates tracking objects using tracking object data stored in the autonomous driving DB 110. The attributes of the tracking object data include object classification data, size data, and position data, but do not include data regarding specific shape, color, or material.

[0027] The erroneous tracking object filtering unit 122 identifies and filters (removes) erroneous tracking objects from among the multiple objects generated by the tracking object generation unit 121. The identification of erroneous tracking objects is performed based on the operational status of the tracking objects.

[0028] Specifically, the system identifies the state of irrational individual tracking objects and the state between irrational tracking objects, and removes tracking objects that are considered to be tracking incorrectly. Examples of irrational individual tracking object states include short maximum tracking times and unrealistic speed changes. Examples of irrational inter-tracking object states include collisions continuing for a certain period of time or collisions with multiple tracking objects.

[0029] The untracked object completion unit 123 filters out some untracked objects from the remaining objects using the mistracked object filtering unit 122 and completes the untracked intervals. Some untracked objects are objects that have been tracked → untracked → tracked, such as tracked objects where 1) the interval between the end of tracking and the start of tracking is less than or equal to a certain time, and 2) the interval between the end of tracking and the subsequent start of tracking is less than or equal to a certain distance.

[0030] The untracked object completion unit 123 completes / interpolates the untracked intervals for such objects to generate tracked object data for consecutive intervals.

[0031] The tracking object sorting unit 124 consolidates tracking object data for tracking objects that have been filtered by the mistracking object filtering unit 122 and interpolated by the untracking object completion unit 123.

[0032] Specifically, the tracking object sorting unit 125 selects a single value from the attributes of the tracking object data, such as size, using a statistical method such as the median or average, and then organizes the tracking object data.

[0033] Returning to Figure 1, the explanation continues. The untracked factor reflection unit 130 reflects the untracked factors in the untracked object data and transmits this information to the scenario generation unit 160. The detailed configuration of the untracked factor reflection unit 130 is shown in Figure 3.

[0034] As shown in the figure, the untracked factor reflection unit 130 is composed of an untracked factor analysis unit 131 and an untracked object data completion unit 132.

[0035] The untracking factor analysis unit 131 analyzes the factors causing untracking of untracked objects identified by the untracking object completion unit 123. To this end, the untracking factor analysis unit 131 uses the autonomous driving DB 110 to acquire and analyze sensor data of the untracked object at the time of untracking and internal status data of the autonomous vehicle.

[0036] Specifically, the untracked factor analysis unit 131 determines whether the untracked factor is due to one of the following: the performance limits of the sensor (maximum recognition distance, blind spot, etc.), the characteristics of the object (color, reflectivity, etc.), or factors caused by another object (blockage, etc.).

[0037] The untracked object data completion unit 132 reflects the analyzed untracked factors into the untracked object data. Specifically, if the untracked factors analyzed by the untracked factor analysis unit 131 are object characteristics (such as the object's color or reflectance), the untracked object data completion unit 132 adds the object's color and reflectance to the tracked object data of the untracked object, thereby reflecting the untracked factors in the untracked object.

[0038] Let's return to Figure 1 and continue the explanation. The precision map DB140 is a database that stores precision map data that represents roads on a lane-by-lane basis, as used by autonomous vehicles while they are driving.

[0039] The scenario domain identification unit 150 identifies elements within the operational design domain of the autonomous driving test scenario. The operational design domain refers to the ODD (Operational Design Domain) according to ISO 34503:2023.

[0040] Figure 4 shows the detailed configuration of the scenario region identification unit 150. As shown in the figure, the scenario region identification unit 150 is composed of a dynamic element identification unit 151, an environmental element identification unit 152, and a static element identification unit 153.

[0041] The dynamic element identification unit 151 identifies detailed items of dynamic elements within the operational design scope from the tracking object data compiled by the tracking object organization unit 124.

[0042] The environmental element identification unit 152 identifies detailed items of environmental conditions within the operational design range from the weather data of the autonomous driving DB 110.

[0043] The static element identification unit 153 identifies detailed items of the static area (Scenery) within the operational design range from the external status data of the autonomous driving DB 110 and the precision map data of the precision map DB 140.

[0044] Returning to Figure 1, the explanation continues. The scenario generation unit 160 generates an autonomous driving test scenario using the dynamic elements, environmental elements, and static elements identified by the scenario area identification unit 150. However, when generating the scenario, for untracked objects among the dynamic elements, the untracked object data completion unit 132 uses untracked object data that reflects the untracked factors.

[0045] The autonomous driving test scenarios generated by the scenario generation unit 160 are stored in the scenario DB 170 and used for testing and evaluating the overall dynamic driving tasks, including the perception, judgment, and control functions of the autonomous driving system.

[0046] Figures 5 and 6 are flowcharts showing a method for generating autonomous driving taste scenarios according to another embodiment of the present invention.

[0047] To generate autonomous driving test scenarios, as shown in Figure 5, sensor data, tracked object data, internal / external status data, weather data, etc., collected while the autonomous vehicle is actually driving are first saved to the autonomous driving DB110 (S210).

[0048] Then, the tracking object generation unit 121 of the tracking object processing unit 120 generates tracking objects using the tracking object data saved in step S210 (S220), and the erroneous tracking object filtering unit 122 identifies and filters out erroneous tracking objects from the objects generated in step S220 (S230).

[0049] Next, the untracked object completion unit 123 identifies some untracked objects from the remaining objects filtered in step S230 and completes the untracked intervals (S240). Then, the tracked object organization unit 124 consolidates the tracked object data for the tracked objects for which filtering in step S230 and completion in step S240 have been completed (S250).

[0050] On the other hand, as shown in Figure 6, the untracked factor analysis unit 131 of the untracked factor reflection unit 130 analyzes the untracked factors of the untracked objects identified in step S240 (S260), and the untracked object data supplementation unit 132 supplements the untracked factors of the tracked object data of the untracked objects (S270).

[0051] The scenario domain identification unit 150 then identifies the dynamic elements, environmental elements, and static elements within the operational design range of the autonomous driving test scenario (S280).

[0052] Then, the scenario generation unit 160 generates an autonomous driving test scenario using the dynamic elements, environmental elements, and static elements identified in step S280. Among the dynamic elements, for untracked objects, it uses untracked object data that reflects the untracked factors in step S270 (S290).

[0053] Next, the autonomous driving test scenario generated in step S290 is saved to scenario DB170 (S295).

[0054] To date, a system and method for generating autonomous driving test scenarios for a data infrastructure used in testing and evaluating autonomous driving systems have been provided.

[0055] In the above embodiment, by generating scenarios that can test not only decision-making and control functions but also cognitive functions based on data obtained from autonomous driving, it becomes possible to obtain more comprehensive and detailed data and evaluation for autonomous driving systems performing dynamic driving tasks.

[0056] On the other hand, the technical ideas of the present invention can also be applied to computer-readable recording media incorporating a computer program that performs the functions of the apparatus and method according to this embodiment. Furthermore, the technical ideas of various embodiments of the present invention may be realized in computer-readable code format recorded on a computer-readable recording medium. A computer-readable recording medium can be any data storage device that can be read by a computer and store data. For example, a computer-readable recording medium may be ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical disk, hard disk drive, etc. Furthermore, computer-readable code or program stored on a computer-readable recording medium may be transmitted via a network connected between computers.

[0057] Although preferred embodiments of the present invention have been described in detail above with reference to the attached drawings, the present invention is not limited to these embodiments. It is clear to any person with ordinary skill in the art to which the present invention belongs that various modifications or alterations can be conceived within the scope of the technical intent described in the claims, and these are also understood to fall within the technical scope of the present invention.

Claims

1. A computer, The autonomous vehicle generates tracking objects using tracking object data acquired while driving, The steps include identifying untracked objects and filling in the untracked intervals, Steps to reflect the untracking factors in untracked object data, Steps include identifying elements within the operational design scope of the autonomous driving test scenario, The steps include generating an autonomous driving test scenario using identified elements and untracked object data that reflects untracked factors, and A method for generating autonomous driving test scenarios, characterized by causing the following to be executed.

2. The identified elements are, The method for generating an autonomous driving test scenario according to claim 1, characterized in that it includes dynamic elements, static elements, and environmental elements.

3. The steps to generate are: The method for generating an autonomous driving test scenario according to claim 2, characterized in that, using the identified elements, for untracked objects among the dynamic elements identified during the generation of the autonomous driving test scenario, data of untracked objects that reflect the untracked factors is used.

4. The complementary steps are: The method for generating an autonomous driving test scenario according to claim 1, characterized in that it identifies an untracked object as one in which the interval between the end of tracking and the start of tracking is less than or equal to a certain time, and the interval between the end of tracking and the subsequent start of tracking is less than or equal to a certain distance.

5. The steps to implement this are: Steps to analyze the reasons for untracked objects, Steps to reflect the analyzed untracked factors in the untracked object data and The method for generating an autonomous driving test scenario according to claim 4, characterized by including the following:

6. Factors that were not tracked: The method for generating an autonomous driving test scenario according to claim 5, characterized in that it includes the color and reflectance of an untracked object.

7. The method for generating an autonomous driving test scenario according to claim 1, further characterized in that the computer is made to perform the step of identifying and filtering out erroneous tracking objects.

8. The filtering step is, The method for generating an autonomous driving test scenario according to claim 7, characterized in that it identifies a mis-tracked object based on the operating status of the tracked object.

9. The method for generating an autonomous driving test scenario according to claim 7, further characterized in that the computer is made to perform the step of organizing the tracking object data for the filtered and completed tracking objects.

10. An object generation unit that generates tracked objects using tracked object data acquired by an autonomous vehicle while it is in motion, A interpolation unit that identifies untracked objects and fills in the untracked sections, A reflection section that reflects the untracking factors in untracked object data, An identification unit that identifies elements within the operational design scope of the autonomous driving test scenario, A scenario generation unit generates autonomous driving test scenarios using identified elements and untracked object data that reflects untracked factors. An autonomous driving test scenario generation system characterized by including the following.

Images