Computer-implemented method and system for verifying a software-based behavior planner of an automated driving function

Patent metadata

Field	Value
Publication number	US-12559134-B2
Application number	US-202318527122-A
Country	US
Kind code	B2
Filing date	Dec 1, 2023
Priority date	Dec 7, 2022
Publication date	Feb 24, 2026
Grant date	Feb 24, 2026

Abstract

A computer-implemented method for verifying a software-based behavior planner of an automated driving function. The method includes: providing a verification environment model to limit the state space of the behavior planner according to a specifiable traffic scene; providing a formal requirement as a criterion for the correctness of decisions of the behavior planner; generating a model checker representation of the behavior planner taking into account the provided verification environment model; analyzing the model checker representation using a model checking procedure with respect to the formal requirement. The verification environment model is used to determine a physically meaningful parameter interval for at least one location parameter and/or movement parameter of the participants of the given traffic scene. The model checking procedure systematically samples the parameter interval and thus checks the behavior planner for a representative selection of the possible temporal and spatial developments of the given traffic scene.

First claim

What is claimed is:

1 . A computer-implemented method for verifying a software-based behavior planner of an automated driving function, the method comprising the following steps:
providing a verification environment model to limit a state space of the behavior planner according to a specifiable traffic scene; providing a formal requirement as a criterion for correctness of decisions of the behavior planner; generating a model checker representation of the behavior planner taking into account the provided verification environment model; analyzing the model checker representation using a model checking procedure with respect to the formal requirement, wherein the verification environment model is used to determine at least one physically meaningful parameter interval for at least one location parameter and/or movement parameter of participants of the traffic scene, the at least one parameter interval being taken into account in the analysis of the model checker representation of the behavior planner in that the model checking procedure systematically samples the at least one parameter interval and checks the behavior planner for a representative selection of possible temporal and spatial developments of the traffic scene; determining, based on the analyzing, whether the behavior planner is in compliance with the formal requirement; and outputting, based on the determining, an indicator indicating whether the behavior planner is in compliance with the formal requirement.

2 . The method according to claim 1 , wherein the decisions and planning of the behavior planner are based on a runtime environment model which describes the temporal and spatial development of the traffic scene based on the at least one location parameter and/or movement parameter of participants of the traffic scene that are ascertained using a perception module, the determination of the physically meaningful parameter interval for the at least one location parameter and/or movement parameter is based on the accuracy with which the perception module provides the behavior planner with the at least one location parameter and/or movement parameter.

3 . The method according to claim 2 , wherein the method is used for verifying the behavior planner of a lane change function,
wherein a decision to initiate a lane change of an ego vehicle equipped with the behavior planner from a current travel lane to a target lane is made using the corresponding runtime environment model, wherein at least a distance and a relative speed between the ego vehicle and another vehicle following the ego vehicle in the target lane are taken into account, and wherein the distance and the relative speed are determined based on location parameters and speeds of the ego vehicle and the other vehicle ascertained by the perception module; wherein a respective parameter interval for the distance and/or the relative speed is taken into account in the analysis of the model checker representation of the behavior planner, wherein each respective parameter interval for the distance and/or the relative speed is determined based on an accuracy with which the perception module provides the distance or the relative speed to the behavior planner.

4 . The method according to claim 3 , wherein the analysis of the model checker representation of the behavior planner takes into account a respective parameter interval for the location parameters and/or movement parameters of the ego vehicle and the other vehicle which are determined based on physical assumptions of the verification environment model about the temporal and spatial developments of the specifiable traffic scene.

5 . The method according to claim 1 , wherein the determination of the physically meaningful parameter interval for the at least one location parameter and/or movement parameter is based on physically plausible assumptions of the verification environment model about possible behavior of a respective participant of the traffic scene.

6 . The method according to claim 1 , wherein the method is used for verifying the behavior planner in at least one of the following traffic situations: i) lane merging, or ii) entering or exiting expressways or freeways, or iii) behavior at intersections and in a vicinity of traffic lights.

7 . The method according to claim 1 , wherein the physically meaningful parameter interval is determined based on one or more error tolerances associated with the at least one location parameter and/or movement parameter of the participants of the traffic scene.

8 . The method according to claim 1 , further comprising:
outputting an indicator indicating that the behavior planner is in non-compliance with the formal requirement; determining at least one counterexample in which the behavior planner does not meet the given formal requirement, the at least one counterexample being determined in the form of a traffic situation; and outputting the at least one counterexample.

9 . The method according to claim 8 , further comprising displaying the at least one counterexample.

10 . The method according to claim 1 , wherein the generating the model checker representation of the behavior planner further includes translating a plurality of code segments of the behavior planner, wherein the determining includes determining that the behavior planner is in non-compliance with the formal requirement, the method further comprising:
identifying one or more code segments of the plurality of code segments to which the non-compliance can be attributed; and outputting the identified one or more code segments.

11 . A computer-implemented system for verifying a software-based behavior planner of an automated driving function, comprising:
an environment modeling module configured to provide a verification environment model which describes a temporal and spatial development of a specifiable traffic scene based on location and movement parameters of participants of the traffic scene; a requirements module configured to specify a formal requirement as a criterion for correctness of decisions of the behavior planner in the traffic scene; a translation module configured to generate a model checker representation of the behavior planner taking into account the provided verification environment model; a model checker module configured to analyze the model checker representation using a model checking procedure with respect to the given formal requirement, wherein the environment modeling module is configured to determine at least one physically meaningful parameter interval for at least one location parameter and/or movement parameter of participants of the traffic scene based on the verification environment model, and the model checker module is configured to take the at least one parameter interval into account in the analysis of the model checker representation of the behavior planner in that the model checking procedure systematically samples the at least one parameter interval and checks the behavior planner for a representative selection of possible temporal and spatial developments of the traffic scene; a compliance module configured to determine, based on analyzing the model checker representation, whether the behavior planner is in compliance with the formal requirement; and an output module configured to output an indicator indicating whether the behavior planner is in compliance with the formal requirement.

12 . The system according to claim 11 , wherein the physically meaningful parameter interval is determined based on one or more error tolerances associated with the at least one location parameter and/or movement parameter of the participants of the tra

Assignees

• Bosch Gmbh Robert

Inventors

• Koenig Lukas
• Hanselmann Michael
• Messer Michael

Classifications

• B60W40/04 Primary

  Traffic conditions · CPC title

• B60W2556/00

  Input parameters relating to data · CPC title

• B60W2554/80

  Spatial relation or speed relative to objects · CPC title

• B60W60/001 Primary

  Planning or execution of driving tasks · CPC title

• G06F11/3698

  Environments for analysis, debugging or testing of software · CPC title