COMPUTER-IMPLEMENTED METHOD FOR AUTOMATICALLY TESTING AND ENABLED VEHICLE FUNCTIONALITY - Patent application

Abstract

Disclosed is a computer-implemented method for automatically testing functions, particularly safety functions, that are integrated into the end-to-end process from data collection in a vehicle to updating the driving function back to the vehicle and / or integrated into an at least semi-autonomous virtual homologation of the driving function under test.

Description

[Technical field]

[0001] The present invention relates to a computer-implemented method for automatically testing and enabling functions, in particular safety functions, and / or for virtually homologating, at least semi-autonomously, a driving function under test, which is integrated into an end-to-end process from data collection in a vehicle to updating the driving function back to the vehicle.

[0002] The invention further relates to a test unit for automatically testing and enabling functions, in particular safety functions, of a vehicle and / or for virtually homologating, at least semi-autonomously, the driving functions to be tested, a computer program and a computer-readable data carrier. [Background technology]

[0003] Today, vehicles are tested and approved as a unit consisting of hardware and software components. Driver assistance systems, such as adaptive cruise control and / or functions for highly automated or autonomous driving, can be demonstrated or verified using various test methods. In particular, hardware-in-the-loop (HIL), model-in-the-loop (MIL), software-in-the-loop (SIL), simulation and / or test drives can be used here. The X-in-the-loop (XIL) method is preferred here, where X stands for the model, software or hardware of the open-loop and closed-loop control in the test. What XIL aims to achieve is a seamless transition between MIL, SIL, HIL and physical environments by reusing models, tests, data and tools.

[0004] Subsequent changes to the vehicle software, i.e. driving functions, require a homologation relevance assessment and, in the case of such an assessment, a supplementary approval is required. At the same time, subsequent changes raise safety concerns for changes to safety-relevant vehicle components / driving functions. As a result, such changes to safety-critical components are generally only made within the framework of recalls and factory visits. Apart from this, development advances are only incorporated into new vehicle generations, which themselves remain on the market for an average of 10 years. Due to the existing data available from vehicles in real operation, there is great potential to use these data for further optimization and homologation, especially virtual homologation.

[0005] The costs, especially the time and / or financial costs, of testing such vehicle functions using the above-mentioned testing methods are typically quite high due to the large number of potential driving situations that must be tested.

[0006] Only on roads with over several billion kilometers of travel, it is not possible for time and cost reasons to test at least semi-autonomous means of transport / vehicles later for optimization purposes. Moreover, many redundant test kilometers may occur, but critical and abnormal situations related to the capabilities of at least semi-autonomous vehicles do not occur.

[0007] This can result in particularly high costs both for test drives and for simulations. DE 10 2017200180 A1 shows a method for verifying and / or validating a vehicle function that is configured for autonomously guiding the vehicle in the longitudinal and / or lateral direction.

[0008] The method includes determining test control instructions for a vehicle function to an actuator of the vehicle based on ambient data about a surrounding of the vehicle, where the test control instructions are not executed by the actuator.

[0009] The method further includes simulating a hypothetical traffic situation that may occur when the test control instruction is executed based on the surrounding data and using a road user model for at least one road user in the vicinity of the vehicle.

[0010] The method further includes providing test data for a virtual traffic situation, where vehicle functions are passively operated in the vehicle to determine the test control instructions.

[0011] A drawback of this method is that the vehicle must be physically operated to determine the necessary data for the purpose of verifying and / or validating the vehicle functionality, and enabling of driving functions is not part of this method.

[0012] The present invention begins after the vehicle has been initially enabled and initial vehicle data can be collected through normal operation of the vehicle. However, the process of new enabling of the modified driving function should be done efficiently and resource-savingly. Thus, the present invention utilizes data collected from the previous operation of the vehicle and defines a full homologation of the optimized driving function by fully virtual and / or partially virtual and / or real tests.

[0013] Due to the significant changes and high development rates, the sale of an original vehicle now implies more than ever the simultaneous ongoing provision of services for the operation, maintenance and further development of the already supplied vehicle systems. Due to the broadening functional scope of today's vehicles and the associated complexity, the number of subsequent modifications to an operational vehicle continues to grow. This increasing intertwining of the various phases and their further development of the product life cycle represents a significant challenge for vehicle manufacturers. To master this challenge, innovative approaches are required for the continuous development of highly complex digital systems.

[0014] It is foreseeable that the diverse and practical approaches for data collection, development of driving functions and delivery of driving functions are a major barrier to innovation. Dedicated test drives, due to their limited capabilities and high costs, can only represent a relatively small percentage of relevant traffic situations. As a result, they are not a sufficient basis for the development of data-driven driving functions. Data sets generated by simulation and synthesis can compensate for these shortfalls, but with a certain degree of uncertainty about representability and fidelity to the prototype. For effective and efficient solutions, it is essential to use a productive and growing vehicle fleet. These are still underutilized resources for the majority of established car manufacturers, although a well-structured and medium- to long-term exploitation of this potential should lead to clear competitive advantages.

[0015] This requires intelligent test creation and test re-tuning, where driving data from real driving operation is made available and fed into the test process, e.g. evaluating the simulation results and, if necessary, changing the parameter settings of the test / simulation. Summary of the Invention [Problem to be solved by the invention]

[0016] The object of the present invention is therefore to provide a method, a test unit, a computer program and a computer-readable data carrier for creating a suitable test environment and test / simulation for automated testing of functions based on real vehicle data. Due to the high degree of automation of this test execution in parallel with the operation of vehicle fleets with (semi-)automated driving functions and the continuous checking of safety criteria of these fleets in a simulated environment (digital twin / digital behavior twin), the enabling and virtual homologation of driving functions is made possible for the first time in the present invention based on simulation methods such as XIL and reprocessing, i.e. reprocessing of existing data sets. The digital behavior twin, as well as the simulation environment, is automatically created and parameterized corresponding to the test tasks of the driving functions (software functions or electronic equipment functions) and the necessary tests related to the enabling (e.g. test scenarios for approval according to official and / or legal approval criteria). At the same time, test execution is started in the various XIL and reprocessing environments and the necessary information is supplied to the test data management, which enables the enabling based on this information. [Means for solving the problem]

[0017] The object of the present invention is achieved by a computer-implemented method for automatically testing and enabling functions, in particular safety functions, and / or for virtually homologating, at least semi-autonomously, a driving function to be tested, which is integrated into an end-to-end process from data collection in a vehicle to updating the driving function back to said vehicle, according to a computer-implemented method according to claim 1, a test unit according to claim 11, a computer program according to claim 12 and a computer-readable data carrier according to claim 13.

[0018] Autonomous and / or semi-autonomous vehicles contain several control units. Each individual control unit and their complexes must be thoroughly tested during development and for homologation for fault-free functioning in all traffic situations, especially in critical traffic situations.

[0019] However, since today's traffic environment is not only highly complex and dynamic, but also subject to change (e.g. due to new traffic modes, changes in development or changes in traffic regulations), there is no fixed point in time at which all relevant data can be collected, also with a view to finding persistently valid input parameters for the development of driving functions. Rather, it should be assumed that existing driving functions have to be continually readjusted based on new and changing functional requirements. Likewise, the qualitative development of existing driving functions must also be taken into account.

[0020] Therefore, according to the present invention, a synthetic data set is generated based on a combination of newly collected real data sets and past real data sets, which serves to expand the training database. This is done against the background of a driving function test that is required later. In order to have a sufficient real data set available for the corresponding validation, these real data can only be used to a limited extent for model training. Therefore, for further development process, the real data is enriched by a synthetic data set (for example, by superimposing a scene with rain or light and shadow). In this case, the synthesis method by which these data are created plays an important role.

[0021] For driving situation testing, various test methods can be used, such as step-based testing, demand-based testing and / or scenario-based testing, among others.

[0022] In scenario-based testing, the driving modes of the vehicle are analyzed in as realistic a traffic situation as possible. The analyzed aspects of the traffic situation and their evaluation depend on the system under test. For this purpose, in scenario-based testing of systems and system components for autonomously guiding a vehicle, scenarios, which can also be called abstractions of the traffic situation, are defined. Then, for each scenario, test cases can also be executed. Here, a logical scenario is an abstraction of the traffic situation by road, driving behavior and surrounding traffic without determining concrete parameter values. A concrete scenario is selected from this logical scenario by selecting concrete parameter values. Such a concrete scenario corresponds to each individual traffic situation.

[0023] In order to distinguish between multiple traffic scenarios or scenarios in scenario-based testing, not only static parameters such as environment, buildings or road width are used, but also the driving behavior of individual road users in particular. The movement of road users and thus the driving behavior are represented by trajectories. The trajectories represent paths in both the spatial and temporal directions. The movement of road users can be distinguished by parameters such as speed.

[0024] Irrespective of the test method chosen, driving data and also synthetic data are crucial to be able to sufficiently well test and adapt driving functions by simulation. Here, the various driving situations / scenarios / applications must be covered by the corresponding safety standards. These are, for example, indicated by SOTIF (ISO / PAS 21448).

[0025] One application example is object identification, and thus object identification components in vehicles. Such components and / or driving functions can be tested for whether pedestrians are correctly identified in various scenarios. If this is done, the current model configuration can be maintained. If deviations occur, the model is further adapted until a formulated minimum performance is reached.

[0026] Another example is the so-called cut-in scenario. In such a scenario, driving functions and driver assistance systems that maintain a required safety distance from other road users can be tested. A cut-in scenario can be described as a traffic situation in which a highly automated or autonomous vehicle drives in a predefined lane and another vehicle enters the ego lane from another lane at a lower speed compared to the ego vehicle and at a predefined distance. In this specification, the ego vehicle represents the vehicle under test (SUT).

[0027] The speed of the ego vehicle and the speed of the other vehicle, also called the fellow vehicle, are constant. The speed of the ego vehicle is higher than the speed of the fellow vehicle, so the ego vehicle must be braked to avoid a collision between the two vehicles.

[0028] In order to be able to meaningfully measure these above-mentioned embodiments, the test object must be integrated into a known simulation environment and so-called XIL testing, i.e. mixed testing consisting of model testing, software testing (SIL) and hardware-in-the-loop (HIL) testing, must be performed. Due to the relatively high effort of HIL testing, it is only available at selected points in the process. However, taking into account the cyber-physical nature of vehicle systems, HIL testing is essential to obtain meaningful results for later use in the production vehicle system. Here, co-simulation approaches are particularly useful in hybrid simulation scenarios to ensure the intended and consistent use of the significantly limited test hardware. The more efficiently this division into multiple test scenarios can be structured, the more strongly the overall product lifecycle can be accelerated.

[0029] That is, the virtual validation process consists of an iterative sequence of simulation, model training, and generation of key performance indicators (KPIs) to evaluate model performance, which is divided into individual driving function components and different XIL test categories.

[0030] What is key here is a continuous understanding of the existing base of real data, which can be guaranteed via providing a behavioral / digital twin for the relevant fleet systems and other necessary data sources (e.g. infrastructure elements). Only in this way can the adaptation / configuration and / or further development of driving functions and their behavioral criteria be possible and tested logically and consistently, thus enabling the functions via virtual homologation.

[0031] In the method according to the invention, the term Key Performance Indicator (KPI) or performance indicator refers to values ​​(KPI values) based on which the progress or achievement of important goals or critical success factors can be measured and / or determined after or during a simulation of the at least semi-autonomous vehicle. The KPIs and / or KPI values ​​allow the evaluation of the simulation and / or simulation steps for test readjustment, thus making the tests more targeted, more resource-saving and more time-efficient.

[0032] Homologation solutions are needed in order to cover the entire end-to-end process, consisting of all sub-processes which are necessary to meet specific customer needs or specific requirements and which follow one another in a chronological and logical manner, and which can be correspondingly automated. In particular, the introduction of virtual homologation, i.e. enabling and its evaluation based on virtual simulation results and / or partly on virtual simulation results, is important here. In principle, a distinction can be made here between homologation-neutral and homologation-relevant functions.

[0033] In particular, vehicle manufacturers must guarantee such homologation relevance of driving function changes. When determining and classifying functions as homologation-relevant, it is currently necessary to introduce supplementary methods. It is clear that the required process speed can only be achieved to a certain extent by integrating the test instances in the framework of the overall process. For this, corresponding interfaces between the development and test systems and the homologation process must be created, so that independent testers from technical services have direct access to the subsystems relevant for the test.

[0034] At least semi-automated driving functions must minimize risks to the safety of vehicle occupants in the vehicle and other road users according to various regulations, e.g. those for automated lane keeping systems (ALKS). This must be guaranteed at least to the level at which a competent, prudent and attentive human driver would minimize the risks. Such a risk balance must be evaluated and verified. In order to be able to estimate such risks consistently, manufacturers in particular must determine risk limit values.

[0035] In general, there are two possibilities for risk monitoring: risk monitoring in absolute values ​​and risk monitoring relative to a particular reference system or behavior. In the first case of absolute form of risk assessment, the probability of the envisaged scenarios, the probability of the subsequent damage-causing causes and the possible dependency between these two factors must be evaluated. In the latter case, representative scenarios are compared with the reference system in terms of damage probability and severity, including their consequences for the system under test. In both approaches (absolute risk assessment and relative risk assessment), it must be taken into account whether the targeted risk threshold is a global threshold or whether there are individual sub-thresholds for the individual scenarios or scenario groups.

[0036] Currently, accident occurrences on public roads are dominated by human-controlled vehicles. That is, today, a positive risk balance can only be demonstrated by demonstrating that the system under test outperforms a human driver, for example in a vehicle equipped with an Advanced Emergency Braking System (AEB), under road conditions that allow for strong decelerations of -0.85G. The main task of AEBS is to prevent rear-end collisions in longitudinally oriented traffic or to reduce the severity of accidents by autonomously reducing the speed (without driver involvement). This requires the calculation of a risk balance from the fulfilment of the requirements for the homologation of driving functions with only a human driver compared to the fulfilment of the requirements for the homologation of driving functions activated without a human driver.

[0037] In this case, the risk balance will ultimately be positive if the driving function provides greater safety, driving comfort and / or energy efficiency than a human driver, and will ultimately be negative if greater safety, driving comfort and / or energy efficiency is achieved by the human driver without the intervention of the driving function.

[0038] This highlights the need and challenge of the present invention, which thus allows for subsequent optimization and new enablement, i.e. virtual homologation, of driving functions by having real driving data available, transitioning to appropriate test systems and / or simulations, and incorporating test instances into the test process.

[0039] The invention further comprises a computer-implemented method for automatically testing and enabling vehicle functions, in particular safety functions, and / or virtually homologating at least semi-autonomous driving functions to be tested.

[0040] Further embodiments of the invention are the subject of the further dependent claims and the following description with reference to the figures.

[0041] The test unit includes means for automatically testing the vehicle's functions, in particular the safety functions, and / or for homologating, at least semi-autonomously, the driving functions to be tested.

[0042] According to another aspect of the invention, a computer program is further configured with a program code for performing the method according to the invention when the computer program is run on a computer. According to yet another aspect of the invention, a data carrier is further configured with a program code of a computer program for performing the method according to the invention when the computer program is run on a computer.

[0043] The features of the computer-implemented method described herein can be used to automatically test and enable vehicle functions, in particular safety functions, and / or for virtually homologating, at least semi-autonomously, the driving functions to be tested.Similarly, the test unit according to the invention is suitable for testing and reconditioning many different devices or control devices, for example, of motor vehicles, transport and / or commercial vehicles, ships or aircraft.

[0044] For a better understanding of the invention and its advantages, reference is made to the following description in conjunction with the accompanying drawings, in which: In the following, the invention is explained in more detail on the basis of exemplary embodiments that are illustrated in the schematic diagrams of the drawings. [Brief description of the drawings]

[0045] [Figure 1] FIG. 2 is a schematic diagram for distinguishing driving situations according to the present invention; [Diagram 2] FIG. 2 is another schematic diagram for distinguishing driving situations according to the present invention; [Diagram 3]FIG. 2 is a schematic diagram illustrating the boundary between critical and non-critical test results. [Figure 4] FIG. 2 shows an illustration according to the present invention of Key Performance Indicators (KPIs). [Diagram 5] FIG. 2 is a schematic diagram for explaining an optimization process using KPIs according to the present invention. [Figure 6] FIG. 2 is a schematic diagram illustrating the use of behavioral twins in accordance with the present invention. [Figure 7] FIG. 2 is a schematic diagram of a method according to the invention envisaged for automatic enabling; DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0046] In Fig. 1 two different scenarios S1 and S2 are described. Here, one intersection area is shown in each case. In the two scenarios S1 and S2, an ego vehicle (Ego) is shown. In S1, the ego vehicle (Ego) performs a turning maneuver. The ego vehicle is also the system under test (SUT: Subject under Test). Here, four fellow vehicles (F1-F4) use the road. In S2, the ego vehicle follows a straight trajectory without the fellow vehicles using the road. Therefore, there are differences in the surrounding parameters as well as in the driving situation parameters. The goal in these scenarios may be, for example, the testing and simulation of vehicle-to-vehicle control.

[0047] Such vehicle distance control tests and simulations may require many test readjustments in order to obtain valid test results. Using driving data from real driving (fleet operation / normal driving operation), appropriate selections for scenarios can be made and sufficiently good parameter settings can be found, so that meaningful tests can be performed. Over at least one test run and / or over several test runs, many test results are obtained that can be automatically evaluated by automatically and / or manually selected KPIs. Evaluation via KPIs can accelerate or even make the homologation process possible for the first time, thereby realizing a timely enablement of modified driving functions for further use in real driving operation. An example of an analysis via KPIs is the evaluation of the relationship in crash rates, e.g. • If at least ten (10) test drives (N) are carried out, the crash rate is checked for simulated test kilometres. ○In the event of an average of 2 (X) crashes (including near crashes) per 1,000 km, only limited enabling of driving functions is permitted. In the event of three (Y) crashes (including near crashes) over 1,000 km, enabling of the driving functions is permitted only under certain conditions. In the event of four (Z) crashes (including near crashes) over 1,000 km, the current test process would be interrupted and further optimisation of the driving function would be required. ●If all test runs in the first scenario are simulated and no collisions occur (0,A), then enabling can be considered. This allows the evaluation by KPIs to show the fulfillment of the requirements for the function under test and the enabling of the function.

[0048] In Fig. 2, a schematic diagram for distinguishing driving situations / scenarios (S1 to Sn) according to the invention is shown. According to Fig. 2, the scenarios S1 and S2 may be completely different, in particular with respect to vehicle parameters and / or driving situation parameters and / or ambient parameters, may have overlapping vehicle parameters and / or driving situation parameters and / or ambient parameters, or may be the same with respect to the respective parameters. Indeed, in order to represent a sufficient homologation and thus to enable driving functions, the selection of the appropriate driving situations / scenarios, and also the selection of the parameterization of the scenarios, plays an important role. For this purpose, it is important to extract the relevant situations from real data, from which the parameterization to be tested can be derived.

[0049] A function showing the boundary between critical and non-critical test results is shown in Figure 3. The points shown are simulated test results. Alternatively, they may be approximated test results.

[0050] The illustrated function is a safety objective function, which has a numerical value having a minimum value for a safe distance between the ego vehicle (Ego) and another vehicle, i.e., the fellow vehicle, of equal to or greater than VFELLOW*0.55, a maximum value during a collision between the ego vehicle (Ego) and another vehicle, and a numerical value greater than the minimum value for a safe distance between the vehicle and another vehicle of equal to or less than VFELLOW*0.55. Results via the safety objective function can be monitored by KPIs over at least one and / or multiple test runs. Results of the evaluation can be used to make a decision about enabling modified driving functions.

[0051] Alternatively to the safety objective function, for example a comfort objective function or an efficiency objective function can be simulated and / or approximated, which objective function has a numerical value which has a minimum value in the absence of a change in the acceleration of the vehicle, a maximum value in the event of a collision between the ego vehicle (Ego) and another vehicle, and in the event of a change in the acceleration of the ego vehicle (Ego) a numerical value between the minimum and maximum value depending on the absolute value of the change in acceleration. A number of driving situation parameters, in particular the speed VEGO of the ego vehicle (Ego) and the speed VFELLOW of the other vehicle, i.e. the fellow vehicle, are generated within a predefined definition range, for example by simulation.

[0052] In FIG. 4, an inventive illustration of Key Performance Indicators (KPIs) is shown. If the scenario / driving situation and thus the system under test (SUT) in question is related to the occurrence of a collision (VC), then a KPI (KPI) should be selected for this, such as a safety KPI for evaluating the driving situation. The KPI (KPI) can in particular determine the impact speed (IV) in the event of a collision (VC) or indicate the minimum distance (Min D) between the ego vehicle (Ego) and the fellow vehicle if no collision (VC) occurs. In particular in a cut-in scenario, where for example vehicle distance control is to be checked, these results are relevant for the evaluation of the scenario. For this purpose, a KPI value can be determined from the determined results. If the KPI value is below a defined threshold, the determined KPI evaluation can be used to find another parameter setting for the simulation. However, if the threshold for the KPI value is exceeded, the enablement of the driving function can be established.

[0053] In Fig. 5, a schematic diagram for explaining the optimization process using KPIs according to the present invention is shown. For this purpose, a driving function is changed / optimized (CF). This driving function is then simulated in a simulation environment (Sim). Subsequently, the result of the simulation (TR) is evaluated by at least one KPI (KPI-Eval) and a KPI value is determined. Depending on the result of the evaluation, a further optimization process is started or the test process is considered to be finished. During the simulation phase, different parameters can be used and / or the tests can be based on different driving situations / scenarios.

[0054] FIG. 6 depicts a schematic diagram for explaining the use of a behavioral twin (D_T, Digital Twin) according to the invention. The behavioral twin (D_T) represents a vehicle and / or vehicle components and / or driving functions, which allow for appropriate testing. Vehicle data (1) is acquired from at least one real vehicle (R_V) in real driving operation, the vehicle data including the measured sensors of the vehicle, and is transmitted to the behavioral twin (D_T). At least one driving function is adapted (CF) and tested (2-n) by the behavioral twin (D_T), where XIL, SIL and / or HIL tests can be used. The testing via the digital behavioral twin includes at least one iteration. The respective test / simulation (Sim) generates test results (TR), which are subsequently further analyzed. If in the key performance indicators (KPI-Eval) the thresholds of the KPIs for evaluating the test / simulation and / or at least one simulation step are above and / or below the fulfillment of the requirements for homologation of the function, the enabling process can be started, which ends the iterations of modifying and / or optimizing the function. After enabling (m), the modified driving function is again passed to the real vehicle (R_V) and used in real driving operation.

[0055] In Fig. 7, a schematic diagram of a method according to the invention is shown, which is conceivable for automatically enabling a driving function. For this purpose, in a preferred embodiment, a risk balance is calculated. As a basis for determining the risk balance, existing data (A_D) and results from a KPI evaluation (KPI-Eval) are used, which contain the results (TR) of a simulation (Sim) and the corresponding evaluation of the KPI thresholds. Here, the KPIs can be set manually by the user / automatically by the test system and / or as predefined set values ​​for the danger / risk evaluation.

[0056] Within the framework of virtual homologation (H-DF), the abovementioned data base is compiled and a risk balance is determined, which indicates the degree to which the requirements are fulfilled. The risk balance results in a positive outcome if the driving function provides higher safety, driving comfort and / or energy efficiency than a human driver and a negative outcome if higher safety, driving comfort and / or energy efficiency is achieved by the human driver without the intervention of the driving function. If the risk balance is positive, the driving function under test is transferred back to the real vehicle (R_V).

[0057] In addition to the described embodiment of enabling driving functions, other embodiments in accordance with the present invention are included.

Claims

1. A computer-implemented method for automatically testing and enabling functions, in particular safety functions, and / or for virtually homologating at least semi-autonomous driving functions to be tested, wherein the functions are incorporated into an end-to-end process from data collection in a vehicle to updating the driving function back to the vehicle, and the computer-implemented method comprises the following steps, namely, a) obtaining vehicle data of at least one real vehicle obtained during a real driving operation, the vehicle data including measured sensor values of the vehicle; b) extracting at least one parameter from the vehicle data; c) configuring a simulation with parameter settings based on at least one of the extracted parameters; d) performing at least one test with the configured simulation; e) evaluating the simulation results using at least one manually and / or automatically selected key performance indicator (KPI); f) when the KPI is below and / or above a determined threshold value, optimizing and / or changing the function to be tested and testing the function based on steps d) and e); g) enabling the changed function and / or transmitting the changed function to at least one of the real vehicles; A computer-implemented method having the above steps.

2. The functions of the vehicle include functions for assisting the driver of the vehicle in specific driving situations and / or traffic situations, and the functions can improve safety and / or energy efficiency and / or driving comfort. The computer-implemented method according to Claim 1.

3. The test and / or simulation can be executed virtually and / or partially virtually. The computer-implemented method according to Claim 1.

4. At least one test and / or simulation is determined by at least one parameter extracted from vehicle data, the parameter being retrieved during and / or at the end of the simulation execution, and the parameter particularly has one of the following parameter characteristics, namely, a) a vehicle parameter having sensor values indicating at least one vehicle setting and / or vehicle characteristic; b) ambient parameters having at least one of the characteristics, namely the number and / or width and / or curvature and / or road regulations and / or ambient temperature of the lane markings; c) driving situation parameters describing the number and characteristics of movable objects, having the following characteristics, namely at least one of the number of road users and / or the number of lane changes in the traffic situation and / or the speed of said road users and / or the means of transportation, in particular the characteristics of the vehicle; comprising; The computer-implemented method according to claim 1.

5. The KPI value is determined by the KPI, and based on the KPI value, the current simulation and / or the evaluation of at least one simulation step are carried out, and the threshold value is defined for the KPI. The computer-implemented method according to claim 1.

6. Indicating the sufficiency of the requirements for the virtual homologation and / or the enabling of the function by the defined threshold value of the KPI for evaluating the simulation and / or at least one simulation step. The computer-implemented method according to claim 1.

7. The sufficiency of the requirements for the virtual homologation / enabling of the function is indicated by a risk balance, which results in a plus if higher safety, driving comfort and / or energy efficiency are formed by the driving function than by a human driver, and results in a minus if higher safety, driving comfort and / or energy efficiency are achieved without the intervention of the driving function by a human driver. The computer-implemented method according to claim 1.

8. To configure a virtual and / or partially virtual simulation, a digital behavior twin of the real vehicle and / or the driving function is generated by the simulation environment corresponding to the driving situation and / or the traffic situation and based on the extracted parameters, thereby enabling the actual test and optimization of the function in a time-efficient manner. The computer-implemented method according to claim 1.

9. In response to a test task, i.e., corresponding to the driving function of the object to be tested consisting of at least one software function and / or electronic function, and corresponding to the necessary enable-related test for the driving function of the object to be tested consisting of at least one software function and / or electronic function, and based on the authorization determination criteria for the driving function of the object to be tested, the digital twin is automatically created and parameterized. The computer-implemented method according to claim 8.

10. During the optimization, at least one setting of the function and / or at least one action criterion of the function are repeatedly adapted. The computer-implemented method according to claim 1.

11. A test unit for automatically testing and enabling functions, in particular safety functions, and / or for virtually homologating the driving function of the object to be tested, which is at least semi-autonomous, The function is incorporated into an end-to-end process from data collection in the vehicle to updating the driving function and returning it to the vehicle. The test unit includes the following steps, namely, a) A step of obtaining vehicle data of at least one real vehicle obtained in a real driving operation, wherein the vehicle data includes measured sensor values of the vehicle; b) A step of extracting at least one parameter from the vehicle data; c) A step of configuring a simulation with parameters set by at least one of the extracted parameters; d) A step of performing at least one test by the configured simulation; e) A step of evaluating the simulation result using at least one key performance indicator (KPI) selected manually and / or automatically; f) When the KPI is below and / or above a determined threshold value, optimizing and / or changing the function of the object to be tested, and testing the function based on steps d) and e); g) A step of enabling the changed function and / or transmitting the changed function to at least one of the real vehicles; A test unit having the above steps.

12. The test unit is composed of a control unit in which at least one function of the vehicle is inspected by virtual testing and / or real testing. The test unit according to claim 11.

13. A computer program having program code for implementing the method according to any one of claims 1 to 10 when the computer program is executed on a computer.

Images