Method and device for checking highly accurate map data for driver assistance functions of a motor vehicle

Abstract

Method for verifying high-precision HD map data for the implementation of driver assistance functions in motor vehicles (3), comprising the following steps: - Providing (S1) HD map data - Performing (S3) a simulation in which a vehicle is virtually moved over one or more routes defined by the HD map data, - Checking (S4) vehicle parameters during the simulation based on predefined test conditions; - Detecting faulty HD card data depending on whether the test conditions are met.

Description

Technical area

[0001] The invention relates to map data that is provided with high accuracy and used to operate autonomous driving functions. Furthermore, the present invention relates to the generation and provision of the highly accurate map data. Technical background

[0002] For semi-autonomous or fully autonomous driving, motor vehicles must be provided with map data for navigation that has a higher accuracy and resolution than the map data used for route planning and driver navigation. Since such high-precision map data (HD map data) provide route resolutions of less than 1 meter, peripheral buildings, signage, and the like, and road layouts can change due to construction work and the like, frequent update cycles of the HD map data are required. These map data are therefore updated very frequently, for example, daily, and made available to motor vehicles for semi-automated and fully automated driving.

[0003] Because automated driving is safety-critical, the provided HD map data must be absolutely accurate and, in conjunction with the semi-automated / fully automated driving function, must not lead to undesirable vehicle behavior or even accidents. HD map data is typically validated through field trials and in-vehicle tests and released only after successful verification. This process typically takes significantly longer than the required update cycles.

[0004] The document DE 10 2007 009 975 A1 discloses a test method for testing navigation systems in vehicles, wherein all environmental signals necessary for the operation of a navigation system are generated by means of simulations and a predetermined number of navigation tasks are carried out with the simulated signals and the reactions of the navigation system are compared with the corresponding reactions of a reference navigation system operated under the same conditions.

[0005] It is therefore a task to provide a way to check and verify the plausibility of HD map data more quickly. Disclosure of the invention

[0006] This object is achieved by the method for checking HD map data for performing semi-autonomous / fully autonomous driving functions according to claim 1 and by a device according to the independent claim.

[0007] Further embodiments are specified in the dependent claims.

[0008] According to a first aspect, a method for verifying high-precision HD map data for performing driver assistance functions in motor vehicles is provided, comprising the following steps: - Providing HD map data; - Carrying out a simulation in which a vehicle is virtually moved over one or more routes defined by the HD map data, - Checking vehicle parameters during simulation based on specified test conditions; and - Detection of faulty HD map data depending on whether the test conditions are met.

[0009] One idea behind the above procedure is to subject the HD map data to be verified to a test in a virtual simulation environment. This can simulate one or more of the driver assistance functions, particularly semi-autonomous / fully autonomous driving functions that directly depend on the correctness and plausibility of the HD map data. Within the framework of a simulation in a virtual test environment, it is possible to identify errors in the HD map data that could lead to undesirable behavior of the driver assistance function, thus confirming the suitability of the HD map data for use in real motor vehicles.

[0010] For testing in the virtual environment, a selected area, specifically a map tile, of HD map data is selected, and a journey by a virtual vehicle is simulated. The selected area represents real routes within a geographical area. The virtual journey proceeds along virtual routes contained in the HD map data, which are composed of connected route sections. A journey by a virtual vehicle is guided according to the average speeds specified for the virtual route sections, so that all possible virtual routes within the area of ​​the HD map data are simulated.

[0011] In this case, the resulting driving state data parameters for the virtual journey are monitored and, if one or more of the driving state parameters deviate from specified normal ranges, an error in the HD map data is detected at a specific position or for a specific virtual route section.

[0012] Fully automated testing and simulation of HD map data allows map suppliers to receive faster feedback on whether the HD map data is correct and suitable for the desired driver assistance functions. Furthermore, it is possible to release HD map data for use in real vehicles in short update cycles, preventing the delivery of faulty HD map data.

[0013] It may be provided that the simulation is carried out with varying environmental parameters and / or varying operating parameters.

[0014] In particular, the varying environmental parameters may include varying road conditions and varying weather conditions, wherein the varying operating parameters include varying vehicle speeds.

[0015] Furthermore, the test conditions can define validity ranges of driving condition parameters and their progressions, whereby the driving condition parameters can include one or more of the following values: - a vehicle speed; - a steering angle; - longitudinal and lateral acceleration (positive / negative) - an over- / under-steering event; and - a friction coefficient or a tire condition.

[0016] According to one embodiment, the test conditions may not be met if at least one of the following conditions is not met: - an acceleration value and / or a deceleration value are within specified limits that define comfortable driving behavior; - a curve radius falls below a specified maximum curve radius that a motor vehicle can drive through at a simulated speed, - a lateral acceleration falls below a maximum lateral acceleration above which the motor vehicle tends to swerve; and - other vehicle parameters are within a predetermined value range resulting from vehicle-specific or function-specific specifications.

[0017] Furthermore, the test conditions can take into account one or more of the following specified aspects: - faulty geometries, whereby permissible values ​​result from the specific functional characteristics of the driver assistance function in question, semi-automated driving, pure warning function without active interventions and the like; - Occlusion of landmarks for locating the motor vehicle, whereby it is checked whether a minimum number of landmark objects can always be identified from each vehicle position or whether occlusions are potentially occurring; - sufficient information density in the HD map data, whereby for each route section it is checked whether the information density is sufficient to be able to recognise a certain number of landmark objects for each vehicle position.

[0018] It may be provided that the simulation is carried out with regard to one or more driver assistance functions.

[0019] According to a further aspect, a device for checking high-precision HD map data for carrying out driver assistance functions in motor vehicles is provided, wherein the device is designed to carry out a simulation in which a vehicle is moved virtually over one or more routes defined by HD map data, to check vehicle parameters during the simulation based on predetermined test conditions and to detect faulty HD map data depending on whether the test conditions are met. Brief description of the drawings

[0020] Embodiments are explained in more detail below with reference to the attached drawings. They show: Fig. 1 a schematic representation of a system with a central location that provides HD map data and motor vehicles that have driver assistance functions and a function for semi-autonomous / fully autonomous driving; and Fig. 2 is a flowchart illustrating a method for verifying HD map data. Description of embodiments

[0021] Fig. 1 schematically shows a system 1 with a central station 2, which provides HD map data for retrieval and transmission to motor vehicles 3. In the system 1 of the Fig. 1, the HD map data are provided by a central location 2. The HD map data are transmitted via a wireless communication connection 4, in particular a radio connection, to the motor vehicles 3 that are to use the HD map data.

[0022] The HD map data corresponds to highly accurate map data that is particularly suitable for or required for certain driver assistance functions, in particular for semi-autonomous / fully autonomous driving (HAF / VAF). The driver assistance functions are provided in motor vehicles 3 and use the most up-to-date HD map data. The timeliness of the HD map data used is particularly important for safety-relevant functions such as HAF / VAF. Therefore, map data suppliers provide HD map data in rapid update cycles. However, before the HD map data is released, it must be verified. This usually involves driving the routes specified in the HD map data with a real test vehicle and checking the HD map data against environmental data recorded by the test vehicle's environmental detection sensors.This process is lengthy and usually takes up to several weeks.

[0023] For the verification and release of the HD cards from Central Office 2, a procedure is provided as shown in the flow chart of the Fig. 2.

[0024] In step S1, HD map data is provided by the map supplier.

[0025] In step S2, an area of ​​the HD map data is selected for review. The area can represent a geographical area and correspond to a tile containing virtual routes defined by the HD map data. The virtual routes comprise interconnected virtual route sections in a route network, each of which is provided with parameters such as topological information, for example, gradient, road surface condition, road width, peripheral development, signage, and the like. The HD map data thus describes the routes in great detail and has a high spatial resolution, in particular a resolution of less than 1 m.

[0026] In a simulation environment, a virtual vehicle is then moved along the virtual route in step S3. This is done using typical operating parameters for the vehicle, such as a speed specified by the average speed for each route section. Driving state parameters for the virtual vehicle are determined for each route section.

[0027] In a simulation environment, the interaction of all vehicle-internal sensors and functions is simulated with a simulated reality and predefined boundary conditions (weather, route sections, road conditions) and corresponding results are generated.

[0028] The virtual route sections described by the HD map data are assigned to route parameters that are generally checked for plausibility by the map supplier before being provided, particularly before the actual driving. For this purpose, individual route parameters can be checked to determine whether their values ​​lie within predetermined validity ranges. The simulation also provides information on how the interaction of the multitude of route parameters affects the behavior of a virtual vehicle when it is driven along the virtual routes of the HD map data area. For example, during the simulation, a vehicle can be driven along the routes of the map data at a predetermined speed for the HAF / VAF. The driving state parameters are monitored.

[0029] In step S4, a check is carried out to determine whether an error can be detected based on the driving state parameters. If this is the case (alternative: yes), faulty HD map data is signaled in step S5. Otherwise (alternative: no), the HD map data is marked as error-free in step S6 and released for transmission to the motor vehicles 3. Test conditions are defined for the driving state parameters; these must be met for the HD map data to be released. The test conditions can, for example, specify validity ranges for the driving state parameters.

[0030] Such an error is detected if at least one of the specified test conditions is not met. These test conditions can include: - Acceleration and deceleration values ​​are within specified limits that define comfortable driving behavior. - A curve radius falls below a specified maximum curve radius that a motor vehicle can negotiate at a simulated speed, - a lateral acceleration falls below a maximum lateral acceleration above which the vehicle tends to skid. - Other vehicle parameters lie within a predetermined value range resulting from vehicle-specific or function-specific specifications.

[0031] Furthermore, the test conditions can take into account one or more of the following specified aspects: - faulty geometries, whereby permissible values ​​result from the specific functional characteristics of the driver assistance function in question, semi-automated driving, pure warning function without active interventions and the like; - Occlusion of landmarks for locating the motor vehicle, whereby it is checked whether a minimum number of landmark objects can always be identified from each vehicle position or whether occlusions are potentially occurring; - Comprehensive system testing within the ADAS network. The card is evaluated in conjunction with all ADAS systems and components within a test bench; - sufficient information density in the HD map data, whereby for each route section it is checked whether the information density is sufficient to be able to recognise a certain number of landmark objects for each vehicle position.

[0032] It is possible to simulate the impact of map parameters on driver assistance functions, including their behavior and interpretation of the map parameters. This is not possible with a purely static evaluation of the map parameters, as these are evaluated independently of function-specific conditions.

[0033] Additionally, a simulation can simulate errors or failures of individual vehicle sensors, and variable route parameters, such as environmental parameters such as temperature, road friction coefficients, and the like, can be simulated by performing various virtual driving tests on the virtual routes. This allows the effects of variable environmental conditions to be taken into account, something that would not be possible by driving the routes with a real test vehicle.

[0034] If the simulation, after driving all routes within the HD map data area one or more times with a virtual vehicle, shows that all driving state parameters for every position on the route sections meet the test conditions, the HD map data can be released. Release occurs by replacing a previously valid area of ​​the HD map data with the verified area of ​​the HD map data in the HD map data transmitted to motor vehicle 3 for implementation of the driver assistance functions.

[0035] Overall, by simulating driver assistance functions that require a direct dependency on HD map data, it is possible to detect errors that would lead to undesirable behavior of the driver assistance functions. This allows the suitability of the HD map data for provision to motor vehicles to be determined. If errors occur, the HD map data can be returned to the map supplier at an early stage, allowing for a faster cycle for correcting errors in HD map data. Overall, the use of automated tests and simulations can ensure the release of up-to-date maps or prevent the delivery of faulty HD map data. List of reference symbols 1 system 2 central unit 3 Motor vehicle 4 Communication connection

Claims

[1] Method for verifying high-precision HD map data for performing driver assistance functions in motor vehicles (3), comprising the following steps: - Providing (S1) HD map data - Carrying out (S3) a simulation in which a vehicle is virtually moved over one or more routes defined by the HD map data, - Checking (S4) vehicle parameters during the simulation based on specified test conditions; - Detection of faulty HD map data depending on whether the test conditions are met. [2] Method according to claim 1, wherein the simulation is carried out with varying environmental parameters and / or varying operating parameters. [3] The method of claim 2, wherein the varying environmental parameters include varying road conditions and varying weather conditions, and wherein the varying operating parameters include varying vehicle speeds. [4] Method according to one of claims 1 to 3, wherein the test conditions define validity ranges of driving condition parameters and their progressions, wherein driving condition parameters comprise one or more of the following driving condition parameters: - a vehicle speed; - a steering angle; - longitudinal and lateral acceleration (positive / negative); - an over- / under-steering event; and - a friction coefficient or a tire quality specification. [5] Method according to one of claims 1 to 4, wherein the test conditions are not met if at least one of the following conditions is not met: - an acceleration value and / or a deceleration value are within specified limits that define comfortable driving behavior; - a curve radius falls below a specified maximum curve radius that a motor vehicle can drive through at a simulated speed, - a lateral acceleration falls below a maximum lateral acceleration above which the motor vehicle tends to swerve; and - other vehicle parameters are within a predetermined value range resulting from vehicle-specific or function-specific specifications. [6] Method according to one of claims 1 to 5, wherein the test conditions take into account one or more of the following predetermined aspects: - faulty geometries, whereby permissible values ​​result from the specific functional characteristics of the driver assistance function in question, semi-automated driving, pure warning function without active interventions; - Occlusion of landmarks for locating the motor vehicle, whereby it is checked whether a minimum number of landmark objects can always be identified from each vehicle position or whether occlusions are potentially occurring; - sufficient information density in the HD map data, whereby for each route section it is checked whether the information density is sufficient to be able to recognise a certain number of landmark objects for each vehicle position. [7] Method according to one of claims 1 to 6, wherein the simulation is carried out with respect to one or more driver assistance functions. [8] Device for checking high-precision HD map data for carrying out driver assistance functions in motor vehicles (3), wherein the device is designed to carry out a simulation in which a vehicle is moved virtually over one or more routes defined by HD map data, to check vehicle parameters during the simulation on the basis of predetermined test conditions and to detect faulty HD map data depending on whether the test conditions are met. [9] A computer program product which, when executed on a data processing device, carries out a method according to any one of claims 1 to 7.

Images