Computer-implemented method and system for classifying traffic situation

Abstract

To provide a method for classifying a preset traffic situation included by a data set of environment data of a motor vehicle.SOLUTION: The method comprises: the step (S2) of applying a directed graph (G) to a first data set (DS1), where nodes of the directed graph segment the first data set into at least one segment of movement behavior of an ego vehicle and / or a fellow vehicle relative to a vehicle environment each according to a first condition (14) met in a time interval, and where edges of the directed graph represent links between the nodes; the step (S3) classifying the preset traffic situation, if all of the determined segments meet a second condition (16) of the preset traffic situation; and the step (S4) outputting a class (K) representing the preset traffic situation and / or corresponding start and end times of a second data set (DS2) comprising the segment representing the preset traffic situation.SELECTED DRAWING: Figure 1a

Description

[Technical field]

[0001] The present invention relates to a computer-implemented method for classifying pre-defined traffic situations encompassed by a dataset of vehicle environmental data.

[0002] The invention further relates to a system for classifying pre-defined traffic situations encompassed by a dataset of vehicle environmental data. [Background technology]

[0003] To create test scenarios for the simulation, test drives have to be carried out, after which the resulting sensor data is abstracted into logical scenarios.

[0004] The input data here are raw data, i.e. sensor data from measured drives in the sense of radar echo recordings, 3D point clouds from lidar measurements and image data. The result data are driving scenarios that can be simulated, including the environment on the one hand and the trajectory on the other hand.

[0005] The document "Szenario-Optimierung fur die Absicherung von automatisierten und autonomen Fahrsystemen (Florian Hauer, Bernd Holzmueller, arXiv:1901.05680)" discloses a method for verifying and validating automated and autonomous driving systems, in particular finding suitable test scenarios for virtual protection.

[0006] The testing methodology envisages the adaptation of a metaheuristic search to optimise scenarios. For this purpose, a reasonable search space and an appropriate quality function need to be established. Starting from an abstract description of the system functionality and application cases, parameterised scenarios are derived.

[0007] This starts from the fact that certain parameters have a large influence on the situation. For example, the situation in which emergency braking is applied is mainly determined by the speed of both road users. It is therefore crucial for the user to understand the distribution of values ​​of these parameters and to know which parts are not covered by the data or the simulation. The missing data points have to be collected either in reality or by simulation.

[0008] It is therefore desirable for users to effectively evaluate the generated data sets to identify critical situations and test slightly altered versions of the same simulation scenario.

[0009] Furthermore, many of the scenarios created for testing autonomous driving functions are determined a priori and are based on expert knowledge of which scenarios are required for which test purposes.

[0010] On the other hand, real-data based approaches are based on machine learning algorithms, which are difficult to generalize to other cases where similar traffic situations are desired.

[0011] As a result, there is a need to improve existing methods for analyzing driving scenario-datasets, thereby enabling effective identification and classification of critical situations. Summary of the Invention [Problem to be solved by the invention]

[0012] It is therefore an object of the present invention to provide a method for classifying predefined traffic situations encompassed by a data set of vehicle environmental data, which allows for an effective identification and classification of critical traffic situations of interest. [Means for solving the problem]

[0013] This problem is solved according to the invention by a computer-implemented method for classifying pre-defined traffic situations encompassed by a dataset of vehicle environmental data having the features of claim 1.

[0014] This problem is further solved according to the invention by a system for classifying predefined traffic situations which are encompassed by a dataset of vehicle environmental data having the features of claim 14.

[0015] The present invention relates to a computer-implemented method for classifying pre-defined traffic situations encompassed by a dataset of vehicle environmental data.

[0016] The method includes providing a first data set, in particular a data stream, of sensor data including a plurality of traffic situations of a journey of an ego vehicle and / or fellow vehicles captured by at least one vehicle-side environment identification sensor.

[0017] The method further includes applying a directed graph to the first data set, wherein the nodes of the directed graph differentiate the first data set into at least one partition of motion behavior of the ego-vehicle and / or companion vehicles relative to the vehicle environment depending on a first condition satisfied in each time interval, and wherein the edges of the directed graph represent connections between each node.

[0018] The method still further includes a step of classifying the predetermined traffic situation if all of the determined segments satisfy a second condition of the predetermined traffic situation, and a step of outputting a second data set having respective start and end times of the classes representing the predetermined traffic situation and / or the segments representing the predetermined traffic situation.

[0019] The invention further relates to a system for classifying pre-defined traffic situations encompassed by a dataset of vehicle environmental data.

[0020] The system includes a data memory configured to provide a first data set, in particular a data stream, of sensor data including a plurality of traffic conditions of travel of the host vehicle and / or fellow vehicles captured by at least one vehicle-side environment identification sensor.

[0021] The system further includes a first computing unit configured to apply a directed graph to the first data set, where the nodes of the directed graph differentiate the first data set into at least one partition of motion behavior of the ego-vehicle and / or companion vehicles relative to the vehicle environment depending on a first condition satisfied in each time interval, where the edges of the directed graph represent connections between each node.

[0022] The system still further includes a second calculation unit configured to classify the predetermined traffic situation if all of the determined segments satisfy a second condition of the predetermined traffic situation, and a data output unit configured to output a second data set having the classes representative of the predetermined traffic situation and / or the respective start and end time points of the segments representative of the predetermined traffic situation.

[0023] The invention further relates to a computer program comprising a program code for implementing the method according to the invention for classifying a pre-defined traffic situation encompassed by a dataset of vehicle environmental data, when the computer program is run on a computer.

[0024] One consideration of the present invention is to perform improved data selection and, possibly based thereon in subsequent steps, improved parameter extraction of traffic conditions contained in a dataset of sensor data of the movement of the host vehicle and / or companion vehicles captured by a first plurality of vehicle-side environment identification sensors.

[0025] This allows the data to be automatically searched for according to the relevant situations and subsequently only the correspondingly extracted traffic situations to be brought into the simulation.

[0026] Therefore, the present invention provides a method for selecting scenarios that can be used to identify traffic situations from real-world data. The proposed method is graph-based and allows to meaningfully define the traffic situations to be identified in the provided raw measurement data.

[0027] The graph-based structure is inherently tied to situation similarity, and similar situations are mapped to corresponding similar graphs.

[0028] The graph-based templates can be easily modified to identify different scenarios, and the graphs can be reused and combined to describe more complex traffic behavior. Thus, multiple graphs used in parallel allow the description of an entire data stream with blocks of identified traffic situations, e.g. cut-ins.

[0029] The coverage of the data stream by the identified traffic situations is directly visible. Besides, the user can specify the templates to be identified without writing any code. The user can decide for himself how simple and general or how complex and detailed the required template should be.

[0030] Traffic situations may be understood here to mean small-scale scenarios, basic scenarios and / or large-scale scenarios or traffic scenarios involving multiple road users.

[0031] Further embodiments of the invention are the subject of further dependent claims and the following description with reference to the drawings.

[0032] According to a preferred embodiment of the invention, it is provided that the predefined traffic situations to be classified are determined in advance, the directed graph determining the classification of the respective predefined traffic situations, and thus it is advantageously possible to determine or binary classify whether the classification encompassed by the respective traffic situation is present in the first data set according to the setting specifications.

[0033] According to a further preferred development of the invention, it is envisaged that each node has at least one input and exactly one output, and that at least one input of each node is provided by an edge of a further node or by the first data set, such that these nodes are able to process one or more inputs of the preceding layer, so that by means of each succeeding layer the vehicle behaviour can be determined at a higher logical or abstraction level.

[0034] According to a further preferred development of the invention, each node has an algorithm applied to the node's input data, where the algorithm is envisaged to classify whether a first condition of the movement behaviour of the ego-vehicle and / or companion vehicles relative to the vehicle environment is fulfilled or not.

[0035] The classification here is a binary classification, ie a determination whether a pre-defined condition that falls into the category is true or false.

[0036] According to a further preferred development of the invention, it is provided that the algorithm of each node outputs the respective start and end times of the segments representing the movement behavior if the first condition is fulfilled, whereby it is advantageously possible to determine the relevant segments in which a particular vehicle behavior occurs.

[0037] According to a further preferred development of the invention, it is envisaged that the section of the movement behavior of the host vehicle and / or companion vehicles relative to the vehicle environment is a time interval including the start and end times of the movement behavior covered by the first data set, in particular the data stream, of the host vehicle and / or companion vehicles relative to the vehicle environment.

[0038] Thereby, starting from a first data set comprising a data stream of sensor data comprising a plurality of traffic situations, a corresponding time period comprising the movement behaviour of the ego-vehicle and / or fellow vehicles of interest can be determined.

[0039] According to a further preferred development of the invention, it is envisaged that the directed graph has a first layer including at least one node and whose input is a first data set, wherein a second layer of the directed graph including at least the first node and a second node differentiates the first data set into at least one category of the motion behavior of the ego-vehicle and / or companion vehicles relative to the vehicle environment, respectively depending on a first condition, and wherein a third layer of the directed graph including at least one node classifies a pre-set traffic situation using the time intervals output by the nodes of the second layer if the combination of determined time intervals satisfies a second condition of the pre-set traffic situation.

[0040] Thus, each layer of the graph fulfills the task of successively differentiating the first data set according to each configuration specification to check whether the higher-level traffic situation is encompassed by the data based on a particular partition.

[0041] According to a further preferred development of the invention, it is provided that the second condition of the preset traffic situation indicates that at least one section of the movement behavior of the host vehicle and / or of the companion vehicle determined according to the first condition occurs in a preset sequence and / or within a preset time interval, thereby making it possible to ensure that the determined sections occur in a corresponding sequence characterizing the preset traffic situation.

[0042] According to a further preferred development of the invention, it is envisaged that the movement behaviour of the ego-vehicle and / or the fellow vehicles with respect to the vehicle environment comprises all vehicle actions that are detectable in the first data set, so that all data captured by the different sensor types as well as events detectable in the data can be used to identify a predefined traffic situation.

[0043] According to a further preferred development of the invention, The motion behavior of the host vehicle and / or companion vehicles with respect to the vehicle environment is the lateral and / or longitudinal behavior of the ego vehicle relative to the traffic infrastructure and / or at least one fellow vehicle, and / or lateral and / or longitudinal behavior of fellow vehicles relative to the traffic infrastructure and / or the at least one ego vehicle, Here, the motion behavior of the own vehicle and / or fellow vehicles relative to the vehicle environment is assumed to include lane keeping, lane changing, turning operations, constant or varying acceleration and resulting speed, activation of brake lights, passing of objects in the vehicle environment and / or identification of traffic signs.

[0044] Thus, the movement behavior of the ego-vehicle detectable in the first data set relative to the traffic infrastructure and / or the at least one fellow vehicle and vice versa comprises a number of different behaviors of the vehicles involved.

[0045] According to a further preferred development of the invention, it is envisaged that the sensor data of the travel of the ego-vehicle and / or the companion vehicle captured by at least one vehicle-side environment identification sensor are position data of GNSS sensors, IMU data, camera data, LiDAR data, radar data and / or ultrasonic data, which can thus be obtained from a number of different data sources.

[0046] According to a further preferred development of the invention, it is provided that a virtual test for verifying the automated driving function of the motor vehicle is carried out based on the output second data set. Thus, the determined traffic situation can form the basis for a virtual test for verifying the automated driving function of the motor vehicle.

[0047] According to a further preferred development of the invention, it is envisaged that the output of the directed graph and the output of the further directed graph are used to classify a further preset traffic situation, whereby the further preset traffic situation is classified if a combination of the outputs of the directed graph and the further directed graph satisfies a third condition of the further preset traffic situation. Thus, a combination of a number of determined traffic situations can advantageously be used to determine a further traffic situation.

[0048] Features of the computer-implemented method for classifying pre-defined traffic situations encompassed by a dataset of vehicle environmental data described herein are equally applicable to a system according to the invention for classifying pre-defined traffic situations encompassed by a dataset of vehicle environmental data, and vice versa.

[0049] For a better understanding of the present invention, and its advantages, reference is now made to the following descriptions taken in conjunction with the accompanying drawings, in which:

[0050] The invention will be explained in more detail below on the basis of exemplary embodiments which are illustrated in the schematic diagrams of the drawings. [Brief description of the drawings]

[0051] [Figure 1a] 1 is a flow chart illustrating a computer-implemented method for classifying pre-defined traffic situations encompassed by a dataset of vehicle environmental data according to a preferred embodiment of the present invention. [Figure 1b] FIG. 1 is a schematic diagram illustrating a cut-in scenario according to a preferred embodiment of the present invention. [Diagram 2] FIG. 13 illustrates a graph for determining a cut-in scenario according to a preferred embodiment of the present invention. [Diagram 3] FIG. 2 is a schematic diagram illustrating a trajectory of an ego-vehicle according to a preferred embodiment of the present invention. [Figure 4] FIG. 2 is a schematic diagram illustrating the trajectory of a companion vehicle according to a preferred embodiment of the present invention. [Diagram 5] FIG. 2 is a schematic diagram illustrating vehicle operation during various time intervals according to a preferred embodiment of the present invention. [Figure 6] 1A-1D are schematic diagrams illustrating different scenarios according to a preferred embodiment of the present invention; [Figure 7] FIG. 13 illustrates a graph for determining a cut-in scenario according to a preferred embodiment of the present invention. [Figure 8] FIG. 1 illustrates a graph for determining a drift-out scenario according to a preferred embodiment of the present invention. [Figure 9] FIG. 2 illustrates a graph for determining an overtaking maneuver in accordance with a preferred embodiment of the present invention. [Figure 10] FIG. 1 is a schematic diagram illustrating a system for classifying pre-defined traffic situations encompassed by a dataset of vehicle environmental data according to a preferred embodiment of the present invention.

[0052] Unless otherwise specified, like reference numbers refer to like elements in the drawings. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0053] FIG. 1a shows a flowchart of a computer-implemented method for classifying pre-defined traffic situations encompassed by a dataset of vehicle environmental data according to a preferred embodiment of the present invention.

[0054] The method comprises a step S1 of providing a first data set DS1, in particular a data stream, of sensor data comprising a number of traffic situations of the travel of the host vehicle 10 and / or companion vehicles 12 captured by at least one vehicle-side environment identification sensor.

[0055] Further, the method includes a step S2 of applying the directed graph G to the first data set DS1, where the nodes 11 of the directed graph G differentiate the first data set DS1 into at least one partition of the motion behavior of the ego-vehicle 10 and / or the companion vehicle 12 relative to the vehicle environment depending on a first condition 14 satisfied in each time interval, where the edges 15 of the directed graph G represent the connections between each node 11.

[0056] Furthermore, the method comprises a step S3 of classifying the pre-defined traffic situation if all of the determined segments satisfy a second condition 16 of the pre-defined traffic situation, and a step S4 of outputting a second data set DS2 having a class K representative of the pre-defined traffic situation and / or respective start and end times of the segments representative of the pre-defined traffic situation.

[0057] The concept according to the invention is explained based on the example of a cut-in scenario shown in Fig. 1b, in which the ego vehicle 10 travels in the left lane until it is in front of another vehicle or fellow vehicle 12 traveling in the center lane, and then it changes into the other vehicle's lane.

[0058] To simplify the description of scenarios like cut-ins, in the framework of the following description Allen's interval algebra is used, in which relations between time intervals can be defined.

[0059] In the context of these scenarios, basic vehicle motion behavior types are used as time intervals to describe the vehicle motion. For example, the motion of the ego-vehicle 10 can be divided into three intervals: an interval for keeping the lane, followed by an interval for changing lanes, followed by another interval for keeping the lane.

[0060] By applying Allen's interval algebra to the time intervals used to describe the vehicle's motion, the cut-in scenario can be described using three simple propositions.

[0061] Vehicle Order: Staying in lane, encountering lane changes, staying in lane.

[0062] Companion Vehicle Order: Maintain lane. Furthermore, the own vehicle sequence also occurs between the peer vehicle sequences.

[0063] Based on this description, a computation graph G is created to determine whether a scenario given as input can be categorized or classified as a cut-in scenario. The resulting graph G for identifying cut-in scenarios is shown in Figure 2.

[0064] The graph G consists of a number of nodes 11 connected by edges 15. Each node 11 has one or more inputs and always produces one output. The inputs of these nodes 11 are defined by the edges 15 connecting them to other nodes 11. Each node 11 performs a test that returns the intervals of the partitions for which the test of the input interval is true.

[0065] The graph G includes different types of nodes 11. The first layer 17a of nodes 11 of the graph G are input data based on the vehicle's motion data, e.g., the vehicle's trajectory. The second layer 17b of the graph G are nodes 11 that analyze the trajectory and divide it into time intervals, e.g., "lane keeping".

[0066] These intervals are then combined by the subsequent node 11 to check whether the time dependencies between the divided intervals correspond to the scenario description, e.g. "meets". After passing the last node 11 of the graph G, the intervals of the identified cut-in scenarios are output.

[0067] In the following, the creation of the graph G is explained. The first step in determining a scenario or traffic situation consists in the differentiation of the road user trajectories into time segments in order to semantically describe the movement. For this purpose, a node 11 is used which implements a differentiation algorithm that allows a differentiation of a trajectory into lateral time segments or segments, for example in a cut-in scenario into a "lane keeping" segment and a "lane changing" segment.

[0068] For the ego vehicle 10 from the cut-in example of Fig. 1, the trajectory can be differentiated into time segments as shown in Fig. 3. As already mentioned above, the differentiation of the trajectory results in three time segments: "Keep in lane", "Change lane" and a further segment "Keep in lane".

[0069] If the same differentiation as in the example is performed on the trajectory of companion vehicle 12, only one "lane keeping" segment will result (Figure 5).

[0070] After differentiation, the time intervals are recombined using a node 11 that performs a check based on Allen's interval algebra. The first node 11 based on Allen's interval algebra used for the graph G in the cut-in scenario is "Meet". This node 11 "Meet" has two inputs: the time intervals "Lane Keeping" and "Lane Changing" from the differentiation node 11 of the previous layer of the graph G.

[0071] The check of node 11 "Meet" provides an interval where a "Lane Keep" interval is directly followed by a "Lane Change" interval. In the example of Figure 5, this is the case for the first interval of the "Lane Keep" differentiation outcome and of the "Lane Change" differentiation outcome. The output of this node 11 is the combined interval of the "Lane Keep" and "Lane Change" intervals.

[0072] The same method is applied to the next "Meets" node 11, which checks whether a further "Lane Keeping" time interval directly follows the output of the first "Meets" node 11. The last node 11 "During" of the graph G checks whether the time interval of the output of the "Meets" node 11 describing the ego vehicle 10 sequence is in the same time interval as the "Lane Keeping" time interval of the identified peer vehicle 12.

[0073] By observing only the time description of the vehicle motion, we can see that it is possible to determine several other scenarios using the graph G designed only for identifying the cut-in scenario (Figure 6). In all four scenarios, the ego vehicle and the peer vehicle perform the same motion, ignoring additional scenario parameters.

[0074] Therefore, additional nodes 11 need to be added to the graph G in order to verify additional scenario parameters. These additional scenario parameters are spatial parameters to describe, in the case of cut-in scenarios, the relative position of the vehicle and the relative lane in which the vehicle is traveling. To distinguish cut-in scenarios, two additional conditions can be added: 1) The vehicle and companion vehicles 10 and 12 reach the goal in the same lane. 2) The host vehicle 10 reaches the finish line ahead of the companion vehicle 12.

[0075] If these two conditions are transformed and added to the drawing to detect the cut-in scenario, the final diagram will look like that shown in Figure 7.

[0076] In the second graph G, there is again a first layer 17a with nodes 11 “Same lane” and “Ego vehicle ahead of fellow vehicle”, which differentiates based on the vehicle trajectories and outputs the time intervals when both vehicles are in the same lane, as well as the time intervals when the ego vehicle 10 is ahead of the fellow vehicle 12.

[0077] In the second layer 17b of the second graph G, it is calculated whether the detected time interval of the first graph G is ended by the time interval of the node output of "Same lane" and the node output of "Own vehicle ahead of fellow vehicle". The last node 11 "During" checks whether the finally generated conclusion time interval takes place with the same duration. If the time interval remains at the end, the cut-in scenario is identified.

[0078] In Figure 8, to identify the drift-out scenario, only three nodes 11 of the graph G for the identification of the cut-in scenario need to be swapped, i.e., from "forward" to "backward" and from "end" to "start".

[0079] In FIG. 9 it is shown how the graph G, once defined for identifying drift-out and cut-in scenarios, can be defined and then combined as nodes 11 for identifying new complex traffic scenarios such as overtaking scenarios.

[0080] FIG. 10 is a schematic diagram of a system for classifying pre-defined traffic situations encompassed by a dataset of vehicle environmental data according to a preferred embodiment of the present invention.

[0081] The system 1 includes a data memory 20 configured to provide a first data set DS1, in particular a data stream, of sensor data comprising a plurality of traffic situations of the travel of the host vehicle 10 and / or companion vehicles 12 captured by at least one vehicle-side environment identification sensor.

[0082] The system 1 further includes a first computation unit 22 configured to apply a directed graph G to the first data set DS1, where the nodes 11 of the directed graph G differentiate the first data set DS1 into at least one partition of the motion behavior of the ego-vehicle 10 and / or the companion vehicle 12 relative to the vehicle environment depending on a first condition 14 satisfied in each time interval, where the edges 15 of the directed graph G represent the connections between each node 11.

[0083] The system 1 further comprises a second calculation unit 24 configured to classify the predetermined traffic situation if all of the determined segments satisfy a second condition of the predetermined traffic situation, and a data output unit 26 configured to output a second data set DS2 having a class K representative of the predetermined traffic situation and / or respective start and end time points of the segments representative of the predetermined traffic situation. [Explanation of symbols]

[0084] 1 System 10 Vehicle 11 Nodes 12 Fellow Vehicles 14 First condition 15 sides 16 Second condition 18 Third condition 20 Data Memory 22 First Calculation Unit 24 Second Computation Unit 26 Data Output Unit A Algorithm G Graph K Class DS1 First Data Set DS2 Second Data Set S1~S4 Processing steps

Claims

1. A computer implementation method for classifying predefined traffic conditions encompassed by a dataset of automotive environmental data, wherein the computer implementation method is Step (S1) provides a first dataset (DS1), in particular a data stream, of sensor data including multiple traffic conditions of the vehicle (10) and / or other vehicles (12) being driven by at least one vehicle-side environmental identification sensor, Step (S2) of applying a directed graph (G) to the first dataset (DS1), wherein the nodes (11) of the directed graph (G) divide the first dataset (DS1) into at least one segment of the motion behavior of the own vehicle (10) and / or fellow vehicles (12) in relation to the vehicle environment, according to a first condition (14) that is satisfied in each time interval, and the edges (15) of the directed graph (G) represent the connections between each of the nodes (11), If all of the determined categories satisfy the second condition (16) of the predetermined traffic conditions, the step (S3) is to classify the predetermined traffic conditions, Step (S4) of outputting a second dataset (DS2) having a start time and an end time for each class (K) and / or category representing a pre-defined traffic condition, Computer implementation methods including

2. The pre-defined traffic conditions to be classified are determined in advance, and the directed graph (G) determines the classification of each pre-defined traffic condition. The computer implementation method according to claim 1.

3. Each node (11) has at least one input and exactly one output, and the at least one input of each node (11) is given by the edges of further nodes (11) or by the first dataset (DS1). The computer implementation method according to claim 1.

4. Each node (11) has an algorithm (A) applied to the input data of the node (11), and the algorithm (A) classifies whether or not a first condition (14) of the motion behavior of the vehicle (10) and / or the other vehicle (12) in relation to the vehicle environment is met. The computer implementation method according to claim 1.

5. The algorithm (A) of each node (11) outputs the start and end times of each section representing the motion behavior when the first condition (14) is met. The computer implementation method according to claim 4.

6. The division of the motion behavior of the vehicle (10) and / or the other vehicle (12) in relation to the vehicle environment is a time interval including the start and end times of the motion behavior encompassed by the first data set (DS1), particularly the data stream, of the vehicle (10) and / or the other vehicle (12) in relation to the vehicle environment. The computer implementation method according to claim 1.

7. The directed graph (G) has a first layer (17a) which includes at least one node (11) and whose input is a first dataset, a second layer (17b) of the directed graph (G) which includes at least a first node (11) and a second node (11) which divides the first dataset (DS1) into at least one segment of the motion behavior of the own vehicle (10) and / or fellow vehicles (12) in relation to the vehicle environment, according to the first condition (14), and a third layer (17c) of the directed graph (G) which includes at least one node (11) which classifies the predetermined traffic situation using the time segments output by the node (11) of the second layer (17b) when the determined combination of time segments satisfies the second condition (16) of the predetermined traffic situation. The computer implementation method according to claim 6.

8. The second condition (16) of the pre-set traffic conditions indicates that at least one of the categories of the vehicle (10) and / or the motion behavior of the other vehicle (12) determined according to the first condition (14) will occur in a pre-set order and / or within a pre-set time interval. The computer implementation method according to claim 7.

9. The motion behavior of the vehicle (10) and / or the other vehicle (12) in relation to the vehicle environment includes all vehicle actions that represent motion behavior and are detectable in the first dataset (DS1). The computer implementation method according to claim 1.

10. The motion behavior of the vehicle (10) and / or the other vehicle (12) in relation to the aforementioned vehicle environment is as follows: The lateral and / or longitudinal behavior of the vehicle (10) with respect to traffic infrastructure and / or at least one fellow vehicle (12), and / or the lateral and / or longitudinal behavior of the fellow vehicle (12) with respect to traffic infrastructure and / or at least one fellow vehicle (10), the motion behavior of the vehicle (10) and / or fellow vehicle (12) with respect to the vehicle environment includes lane keeping, lane changing, turning, constant or changing acceleration and the resulting speed, activation of brake lights, passage of objects in the vehicle environment and / or identification of traffic signs, The computer implementation method according to claim 1.

11. Sensor data of the vehicle's (10) and / or a fellow vehicle's (12) movement, captured by at least one vehicle-side environmental identification sensor, includes GNSS sensor position data, IMU data, camera data, LiDAR data, radar data, and / or ultrasonic data. The computer implementation method according to claim 1.

12. Based on the output second dataset (DS2), a virtual test is performed to verify the automated driving function of the vehicle. The computer implementation method according to claim 1.

13. The output of the directed graph (G) and the output of the further directed graph (G) are used to classify further pre-defined traffic conditions, and if the combination of the outputs of the directed graph (G) and the further directed graph (G) satisfies the third condition (18) of the further pre-defined traffic condition, the further pre-defined traffic condition is classified. The computer implementation method according to claim 1.

14. A system (1) for classifying predefined traffic conditions encompassed by a dataset of automotive environmental data, wherein the system (1) A data memory (20) configured to provide a first dataset (DS1) of sensor data, in particular a data stream, which includes multiple traffic conditions of the vehicle (10) and / or other vehicles (12) being driven by at least one vehicle-side environmental identification sensor, A first computing unit (22) configured to apply a directed graph (G) to the first dataset (DS1), wherein the nodes (11) of the directed graph (G) divide the first dataset (DS1) into at least one segment of the motion behavior of the local vehicle (10) and / or fellow vehicles (12) in relation to the vehicle environment, according to a first condition (14) that is satisfied in each time interval, and the edges (15) of the directed graph (G) represent the connections between the respective nodes (11) of the first computing unit (22), A second calculation unit (24) configured to classify a predetermined traffic situation when all of the determined categories satisfy a second predetermined traffic situation, A data output unit (26) configured to output a second dataset (DS2) having start and end points for each class (K) and / or category representing a pre-defined traffic condition, A system (1) equipped with the following features.

15. A computer program comprising, when the computer program is executed on a computer, program code for carrying out a method for classifying a set of predefined traffic conditions that are encompassed by a set of environmental data for automobiles described in any one of claims 1 to 13.

Images