Method for providing a digital road map

Abstract

A computer-implemented method for providing a digital road map is proposed, wherein the digital road map is used for testing at least partially automated vehicle systems, the method comprising the following steps: - accessing a database in which permissible representations of a plurality of road characteristics are stored, - creating at least one road map segment by automatically selecting, for each of the road map segments, one of the possible representations from the database for a first one of the plurality of road characteristics, - providing the digital road map, wherein the digital road map comprises the at least one road map segment.

Description

Technical Field

[0001] The present invention relates to a computer-implemented method for providing a digital road map, wherein the digital road map is used in particular for testing at least semi-automated vehicle systems. Background Art

[0002] Automated vehicle systems must be designed for a wide variety of situations. The safety of such systems poses a significant technical challenge. Due to the infinitely large range of possible driving situations, corresponding testing based on safety requirements is extremely complex and virtually impossible to describe solely through actual testing. Suitable support for testing through simulations is desirable. A particular challenge in such simulations is ensuring the broadest possible and, in the best case, quantifiable coverage of possible situations.

[0003] Road maps for simulation tests of vehicle systems are created by first measuring real roads with high-precision sensor systems and then manually digitally processing the measured data. Another popular option for creating such road maps is an editor, which allows for manual map definition. Tile-based road editors exist, in which prefabricated road elements are provided on square or hexagonal tiles, which can be manually placed next to each other.

[0004] A possible map format for simulation is OpenDRIVE, which is supported by a variety of commonly used simulation environments. Here, the road map can also be directly specified manually using a text editor as XML.

[0005] US 20170316127 A1 describes a method for creating test scenarios for a driverless vehicle. In one embodiment, the OpenDRIVE file format is used for environment simulation, which allows for a logical description of a road network. Summary of the Invention

[0006] A computer-implemented method for providing a digital road map is proposed, wherein the digital road map is used, in particular, for testing at least partially automated vehicle systems. In this method, a database is accessed, in which permissible representations of a plurality of road characteristics are stored. At least one road map segment is created by automatically selecting, for each first road characteristic of the plurality of road characteristics, one of a plurality of possible representations from the database. Finally, at least one digital road map is provided, wherein the digital road map comprises at least one road map segment.

[0007] The described method allows for the automatic calculation, generation, and provision of a digital road map or multiple digital road maps, significantly reducing the effort compared to manual creation in a map editor. Furthermore, it also partially enables the preconfiguration or consideration of boundary conditions or constraints. Automatic parameter variation allows for the automatic generation of a large number of different maps with predefined boundary conditions. Generation based on a database of relevant road characteristics and alternative permissible representations of road characteristics ensures the broadest possible, and possibly quantifiable, coverage of the entire parameter space or guarantees a defined degree of coverage. The approach is arbitrarily parallelizable and allows for a significant reduction in effort resulting from computer-implemented automatic generation.

[0008] The described method allows for the automated generation of 2D or 3D maps using, for example, freely definable and, in particular, automatically variable values from a parameter range for representing road characteristics, such as arbitrary curvature radii, slopes, road types, number of lanes, etc. This significantly exceeds the possibilities of, for example, conventional tile-based methods. For this purpose, the automation can, in particular, employ a random selection of discrete alternatives for the representation or a random selection from a continuous parameter range for the representation.

[0009] Due to the data structure in the database, this method already helps reduce the complexity of large parameter spaces. Optional subconfiguration of the representation or configured constraints allow for the provision of customized digital road maps, thereby further reducing the number of variants. The optional introduction of equivalence classes, based on which the degree of coverage can also be determined, further reduces complexity and enables the systematic provision of maps that cover the available space on the map as well as possible within predefined boundary conditions.

[0010] In some advantageous configurations, multiple road map segments are created and joined together to form a digital map. This allows for a particularly flexible and modular approach. The compatibility of the joined segments and the appropriate transitions between them are automatically ensured.

[0011] Advantageously, impermissible combinations of characteristic representations are already excluded during the calculation and generation of the digital map. Furthermore, the generated map is optimally checked for consistency before being made available. Consequently, despite the high degree of automation, the method achieves a qualitatively high output.

[0012] In some particularly preferred embodiments, the described method is used such that at least one provided map is used to test an at least partially automated vehicle system. In this case, during a simulation test, the vehicle system is subsequently tested while driving along a route on a digital road map provided as described to determine whether at least one predetermined requirement, in particular a safety requirement, for the vehicle system is met during the simulated driving.

[0013] Due to the aforementioned advantages offered by such maps, such as the wide and possibly also guaranteed coverage, a fully automated testing method for a driving system can thus be carried out.

[0014] The described method is implemented on a computer. For this purpose, a processing unit of the computer executes a computer program that is stored on a machine-readable data medium and is provided for carrying out the described method. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Next, embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the drawings:

[0016] Figure 1 An exemplary sequence of a computer-implemented method for providing a digital road map is schematically shown.

[0017] Figure 2 Schematically shows an exemplary provision of a plurality of digital road maps in the case of a subconfiguration of a database with road characteristics,

[0018] Figure 3 A further exemplary provision of a plurality of digital road maps in the context of a subconfiguration of a database with road characteristics is schematically shown. DETAILED DESCRIPTION

[0019] Figure 1 An exemplary sequence of a computer-implemented method for providing a digital road map and its use for simulation testing of a vehicle system is schematically shown.

[0020] In a first step 101 , a randomly selected and possibly predetermined representation 12 for a selected characteristic is extracted from a database 11 , wherein the database stores permissible representations of a specific road characteristic.

[0021] The database includes the assignment of road characteristics and the permissible representations for these road characteristics, wherein the representations are defined in such a way that they are mutually exclusive, i.e., there is no overlap between the representations. The database includes, in particular, Zwicky boxes, morphological boxes, or comparable data structures for combinatorial modeling, which uses dimensions and alternatives within the individual dimensions to combinatorially describe the relevant characteristics of the design space.

[0022] For example, relevant road properties in the database may include road width, road surface or friction parameters, road markings, the presence of specific road infrastructure such as traffic lights, road edge structures, traffic signs, pedestrian crossings, road curvature radius, the presence of intersections and slopes, etc.

[0023] In accordance with some of the characteristics mentioned as examples, the following exemplary representations for road map segments can be stored in the database.

[0024] Road width in meters within the defined parameter range

[0025] Lanes 1, 2, 3, 4, 5, 6

[0026] Traffic light No, Yes

[0027] Pedestrian crossing No, Yes (zebra crossing), Yes (pedestrian signal light)

[0028] No for marginal buildings, Yes - low (up to 1 meter), Yes - high (>1 meter)

[0029] In a particularly preferred representation, the database also stores information about which combinations of representations are meaningless or impermissible for different characteristics. These characteristics can be used when selecting a representation and automatically taken into account during the selection, so that digital maps with such combinations are not provided.

[0030] In particular, at least some of the representations for the relevant characteristics are automatically selected by random selection, wherein selection according to a predetermined algorithm is also possible. The random selection is performed in particular by using a random number generator or a pseudo-random number generator. In addition, in an advantageous embodiment, a sub-configuration of the representations for the characteristics is possible. For this purpose, in the method, Figure 1 In a step not shown in FIG. 1 , a configuration can be received and a portion of the representation can be selected corresponding to the sub-configuration. The remaining representation is automatically determined, thereby determining all relevant properties or all properties necessary for providing a map.

[0031] For example, a subconfiguration can be selected so that a customized map set is provided for the desired purpose. For example, if the highway functions of an at least semi-autonomous vehicle are to be tested in a simulation using a provided digital map, the road characteristics representation "Number of Lanes" can be limited to values of 2, 3, or 4. This also allows the generation of country-specific maps, for example by limiting the representation to those permitted by left-hand traffic regulations. The representation can also be automatically limited to a specific parameter range. This can be done, for example, based on parameters of the vehicle type (e.g., car, truck) to specifically generate difficult maps or maps that are possible for a defined vehicle type, particularly for the purpose of using the provided road map for the corresponding vehicle system.

[0032] In step 102, the representation determined for the relevant road characteristics is transmitted to a computer-implemented map generator 13, which automatically generates and provides at least one digital map 14 based on the transmitted representation in step 103. The calculation, generation, and provision of the at least one digital road map can be configured to determine whether, in optionally given subconfigurations, only one digital characteristic or a specific number of digital characteristics are to be calculated, generated, and provided, or whether a predetermined, for example complete, coverage of the existing parameter space is desired.

[0033] If only one road map is to be generated, a random selection of permissible (and free depending on the possible sub-configurations) representations of the road characteristics can be made, particularly while taking into account impermissible combinations, and the road map can be calculated, generated, and provided on this basis. If a specific number of road maps are to be provided under the same conditions, the steps of calculating, generating, and providing can be repeated correspondingly frequently.

[0034] However, in alternative configurations, complete coverage of the design space or a specific degree of coverage of the design space may also be desired. To this end, equivalence categories can be defined, that is, when two provided maps are considered equivalent, i.e., when they deviate from each other by no more than a predetermined range. Identical maps are always considered equivalent. For inconsistent maps, for example, parameter ranges can be defined within which two maps are considered equivalent with respect to a specific representation of a characteristic. An example of the latter case could be two road maps that differ only slightly in road width, i.e., they are considered equivalent. If the remaining characteristics are also considered equivalent, then the two road maps belong to the same equivalence category. When calculating, generating, and providing digital maps, it can be preconfigured to generate at least one, or precisely one, map for each equivalence category. A degree of coverage can also be defined (e.g., providing road maps in at least 50% of the equivalence categories) that must be met during calculation, generation, and provision.

[0035] When selecting representation 12 in step 101 or calculating a digital road map based thereon in map generator 13, the use of parameter ranges is crucial. While discrete representations (e.g., "Traffic light: yes / no" or "Lane: 1, 2, 3, 4, 5, 6") can be selected directly, representations with continuous parameter ranges (e.g., road width, radius of curvature, location of traffic signs or other infrastructure, intersection location, slope values) require determining which specific values within the parameter range are included in the calculation. A random selection can be made from the corresponding parameter range. Alternatively, a targeted variation of the continuous characteristic can be used to achieve the widest possible coverage.

[0036] In some preferred embodiments, at least one digital road map with altitude information is provided. In one embodiment, this can be achieved by having altitude or slope values available as road characteristics in a database. Alternatively, the already calculated and generated digital map can also be adapted based on a random or predetermined altitude profile before provision.

[0037] When calculating and generating the digital road map, a consistency check can also be performed so that the selected parameters lead to a realistic map.

[0038] In one advantageous embodiment, the calculation and generation of the digital road map includes the creation of one or more road map segments based on a selected or transmitted representation of road characteristics. A plurality of road map segments can be combined into a road map, wherein the transitions between road map sections are automatically refined or adapted so that all elements of the road map segments neatly adjoin one another at their edges (e.g., by inserting clothoids in the case of curves). The selection of the road map segments and the combination of the road map segments can each be performed randomly according to a predefined pattern.

[0039] The provision of digital road maps is carried out in particular by outputting them in a predetermined or configurable machine-readable output format. When the map(s) are further processed in the simulator, they are advantageously output in a data format that can be read by the simulation environment 15, for example in the OpenDRIVE format.

[0040] In an advantageous embodiment, in step 104 the at least one provided digital map 14 can be transmitted to a simulation environment 15 , which tests the vehicle system in a simulation for driving along a route on digital map 14 and outputs a simulation result 16 in step 105 .

[0041] In some advantageous embodiments, if the simulation test identifies errors or weaknesses in the vehicle system by comparison with predetermined requirements for the vehicle system behavior, the vehicle system can be automatically adapted based on the simulation results. In particular, software parameters of the vehicle system software can be adapted.

[0042] The digital road map provided by the described method can be automatically compared with the digital road map of the actual road. Based on this comparison, the road characteristics and the representation can be automatically supplemented with previously unconsidered characteristics, representations, combinations or parameter ranges.

[0043] exist Figure 2 An exemplary database 20 is schematically shown in FIG. For road characteristics 21 or 22 or 23, corresponding representations 211 to 214 or 221 to 224 or 231 to 234 are stored there. While representation 211 is preconfigured for characteristic 21 and representation 223 is preconfigured for characteristic 22, a representation is automatically, in particular randomly, selected only for characteristic 23. When the entire space is covered by maps of each equivalent class, four maps (25) are obtained in this example.

[0044] exist Figure 3An exemplary database 30 is schematically shown in FIG. For road characteristics 31 or 32 or 33, corresponding representations 311 to 314 or 321 to 324 or 331 to 334 are stored there. While representation 323 is preconfigured for characteristic 32, representations are automatically, in particular randomly, selected only for characteristics 31 and 33. If the entire space is covered by maps from each equivalent class, sixteen maps (35) are obtained in this example.

Claims

1. A computer-implemented method for providing a digital road map (14), wherein the digital road map is used for testing at least semi-automated vehicle systems, characterized by the following steps: (a) accessing a database (11) in which are stored, for a plurality of road characteristics, the permissible representations of said road characteristics, (b) creating at least one road map segment by automatically selecting one of the representation forms from the database (11) for each first road feature of the plurality of road features for the road map segment, (c) providing the digital road map (14), wherein the digital road map comprises the at least one road map segment, wherein a plurality of digital road maps (14) are automatically provided successively or in parallel by repeatedly performing steps (a) to (c), wherein the repetition of steps (a) to (c) is interrupted when at least one digital road map (14) is provided in each equivalence class of the road maps to a predetermined coverage range, wherein each equivalence class defines maps that deviate from one another by no more than a predetermined range, wherein consistent maps are always considered equivalent, and for inconsistent maps, the following parameter range is defined: within the parameter range, two maps are considered equivalent with respect to a certain representation of a characteristic with respect to this characteristic, (d) transmitting the provided digital map (14) to a simulation environment (15) for testing vehicle systems in a simulation for driving along a route on the digital map (14), (e) If the simulation test determines errors or weaknesses in the vehicle system by comparison with predetermined requirements for the behavior of the vehicle system, software parameters of the vehicle system software are automatically adapted according to the results of the simulation.

2. The method according to claim 1, characterized in that The representation form is randomly selected for a first one of the plurality of road characteristics.

3. The method according to claim 1, characterized in that In order to create the at least one road map segment, one of the possible representation forms is preconfigured from the database (11) for each second road property of the plurality of road properties.

4. The method according to claim 3, characterized in that In order to create the at least one road map segment, a corresponding number or a corresponding range of possible representations in the database (11) is restricted for a third road characteristic of the plurality of road characteristics.

5. The method according to claim 4, characterized in that The selection of a representation for a third road characteristic from the plurality of road characteristics is performed randomly from a limited number or a limited range of possible representations.

6. The method according to claim 1, characterized in that At least two road map segments are created, and the at least one digital road map (14) provided has the at least two road map segments.

7. The method according to claim 6, characterized in that The at least one digital road map (14) is provided by automatically joining the at least two road map segments.

8. The method according to claim 7, characterized in that During the joining, the road map segments are automatically taken into account: which road map segments are matched and how.

9. The method according to any one of claims 7 or 8, characterized in that During the joining, the transition between the road map segments is automatically adapted or improved.

10. The method according to any one of the preceding claims 1 to 8, characterized in that The representation of the road characteristic exists as discrete alternatives or as a range of alternative values.

11. The method according to claim 10, characterized in that Selection is made from a discrete alternative of representations by randomly selecting from said alternatives.

12. The method according to claim 10, characterized in that The selection from the alternative range of values is made by randomly selecting a value from the range of values.

13. The method according to any one of the preceding claims 1 to 8, characterized in that When creating the at least one road map segment, impermissible combinations of representations of different road characteristics are automatically excluded during selection.

14. The method according to any one of the preceding claims 1 to 8, characterized in that After the at least one digital road map has been provided, an automatic consistency check is performed and inconsistent maps are automatically improved or rejected or replaced.

15. The method according to any one of the preceding claims 1 to 8, characterized in that The road characteristics and their permissible representations are stored in the database as Zwicky boxes.

16. The method according to any one of the preceding claims 1 to 8, characterized in that The at least one road map segment or the at least one digital map (14) is automatically expanded with a randomly selected height profile before providing.

17. The method according to any one of the preceding claims 1 to 8, characterized in that The at least one provided digital road map (14) is compared with a stored digital representation of an actual road and the database (11) is supplemented based on the comparison.

18. The method according to claim 1, wherein The method is interrupted when a predetermined number of digital road maps (14) is reached.

19. A method for testing an at least partially automated vehicle system, characterized in that While driving along a route on a digital road map (14) provided by the method according to any one of claims 1 to 18, the vehicle system is simulated and checked to see whether at least one predetermined safety requirement for the vehicle system is complied with during the simulated driving.

20. The method according to claim 19, characterized in that For one of the road characteristics determined therein, the amount of the corresponding representation or the permissible range for the corresponding representation is automatically adapted or limited depending on the type or the characteristics of the vehicle system to be tested.

21. The method according to any one of claims 19 or 20, characterized in that If the check reveals that at least one predetermined requirement was not complied with during the simulated driving, parameters of the software of the vehicle system are automatically adapted. 22 . A computer program product comprising a computer program, which is configured to perform the method according to claim 1 , when the computer program is run on a processor. 23 . A machine-readable data carrier comprising a computer program which, when executed on a processor, is configured to carry out the method according to claim 1 .

Images