WO2023025613A1 - Validation d'une fonction de commande de conduite pour le fonctionnement automatique d'un véhicule - Google Patents

Validation d'une fonction de commande de conduite pour le fonctionnement automatique d'un véhicule Download PDF

Info

Publication number
WO2023025613A1
WO2023025613A1 PCT/EP2022/072802 EP2022072802W WO2023025613A1 WO 2023025613 A1 WO2023025613 A1 WO 2023025613A1 EP 2022072802 W EP2022072802 W EP 2022072802W WO 2023025613 A1 WO2023025613 A1 WO 2023025613A1
Authority
WO
WIPO (PCT)
Prior art keywords
vehicle
virtual
control function
driving control
simulation
Prior art date
Application number
PCT/EP2022/072802
Other languages
German (de)
English (en)
Inventor
Viktor LIZENBERG
Ulrich Eberle
Original Assignee
Psa Automobiles Sa
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Psa Automobiles Sa filed Critical Psa Automobiles Sa
Publication of WO2023025613A1 publication Critical patent/WO2023025613A1/fr

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/36—Preventing errors by testing or debugging software
    • G06F11/3668—Software testing
    • G06F11/3672—Test management
    • G06F11/3688—Test management for test execution, e.g. scheduling of test suites

Definitions

  • the invention relates to a system for a vehicle for verifying and validating a driving control function for the automatic operation of the vehicle, and a method for verifying and validating the driving control function for the automatic operation of the vehicle.
  • US 2020/0353943 A1 relates to a system with which video data is determined which shows a bird's eye view of moving vehicles. A traffic scenario is generated on the basis of this data, with the scenario having information about at least one dynamic object. A machine learning network is trained based on the traffic scenario.
  • the description of US 2020/0353943 A1 cites a simulation of a three-dimensional driving environment.
  • WO 2020/264276 A1 also relates to a method in which logged data of a vehicle are recorded while driving through an environment, with a scenario being created in a simulated environment based on at least some of this data. Based on at least part of the scenario data, an instantiation region is determined, which is linked to a simulated object in the simulated environment.
  • DE 10 2017 007 136 A1 relates to a method for training self-learning algorithms for an automated vehicle with a specified automation module by generating learning situations, the learning situations being generated as follows: Carrying out a traffic simulation in which a virtual ego vehicle is placed in a virtual scenario with the automation module of the real vehicle, the scenario comprising a route structure with a predetermined route, further comprehensively automatically generated further virtual moving objects with individually specifiable object properties and behavior models, the objects interacting independently and adaptively with one another as the simulation progresses interact on the basis of the respective object properties and behavior models, carry out a vehicle dynamics simulation based on the Automation module as well as virtual sensor signals of the moving objects of a virtual sensor system assigned to the ego vehicle, which correspond to a sensor system of the actually existing vehicle, in which reactions of the ego vehicle are generated, identification of a relevant learning situation using selection criteria that are based on specifiable metrics are determined.
  • a first aspect of the invention relates to a system for a vehicle for verifying and validating a driving control function for the automatic operation of the vehicle, having a simulation unit, a sensor interface, and an analysis unit, the simulation unit being designed during real operation of the vehicle on a Street to simulate the behavior of at least a first virtual road user and from the simulated behavior to generate a virtual vehicle-to-vehicle communication and / or virtual sensor data, wherein the sensor interface is designed to the virtual vehicle-to-vehicle communication and / or to make the virtual sensor data available to the driving control function in such a way that the virtual vehicle-to-vehicle communication for the driving control function cannot be distinguished from real vehicle-to-vehicle communication and the virtual sensor data reflect the original vehicle-specific S Superimpose or replace sensor data depending on the type of sensor, and the analysis unit is designed to monitor the output variables of the driving control function.
  • the vehicle is in particular a passenger car, but can also be a truck, bus or the like.
  • the vehicle is equipped with a drive control function that provides or contributes to the automatic operation of the vehicle.
  • a driving control function is often safety-critical, in particular when it actively intervenes in the driving control of the vehicle, for example Acceleration processes, braking processes or steering processes are carried out.
  • the driving control function reacts depending on the sensor data recorded, which can also affect other road users.
  • an automatic distance control assistant can automatically maintain a speed-dependent, predefined distance from the vehicle in front
  • a lane departure warning assistant can only change lanes if there are no other road users in a certain area, or another system can impose a certain dynamic behavior on the vehicle to minimize the risk of traffic jams, especially when the vehicle is traveling in an autonomous mode without a driver.
  • the output variables of the driving control function are in particular actuator commands, for example to operate a steering system, a brake, an accelerator pedal or other functions of the vehicle.
  • the simulation unit ensures that another road user can be simulated during regular operation of the vehicle on a physical road in such a way that from the point of view of the driving control function it is not recognizable that it is a simulated road user instead of a real road user.
  • a corresponding virtual vehicle-to-vehicle communication and/or virtual sensor data are generated by the simulation unit, which are fed in via the sensor interface in order to be supplied to the driving control function.
  • a camera image of the surroundings of the vehicle can be overlaid with the image of the simulated other road user, or vehicle-to-vehicle communication that is otherwise not predominant can be generated. It is characteristic of the first virtual road user that he is an interactive road user and can therefore react to the reactions and actions of the vehicle himself, in particular because the driving control function of the vehicle actively makes entries on the vehicle itself.
  • the original vehicle-specific sensor data are those sensor data that are actually determined by the vehicle's sensors.
  • these original on-board sensor data are replaced or overlaid as if the first virtual road user were physically present in reality.
  • the simulation unit is designed to adapt the behavior of the first virtual road user to the behavior of the vehicle, so that the driving control function creates a mutual interaction between the vehicle and the first virtual road user.
  • the simulation is therefore carried out iteratively in particular, since reactions of the driving control function in turn lead to reactions of the first virtual road user and vice versa.
  • this advantageously corresponds very realistically to a real scenario in which road users react to one another on both sides, regardless of whether they are guided by human or machine control.
  • the first virtual road user is a virtual other vehicle
  • the simulation of the behavior of the first virtual road user includes a driving dynamics simulation, which includes equations of motion of the virtual other vehicle, and in the case of a simulated manual operation of the virtual other vehicle, an additional action -/ Includes reaction model of the driver of the virtual other vehicle.
  • the equations of motion represent basic equations of physics that connect acceleration, mass and acting forces and moments. This includes, in particular, frictional forces in relation to the air and in relation to the ground, drive torques and others. If it is also assumed that the first virtual road user is a human driver, the dead time for a required reaction (typically 0.1 seconds) and the sluggish and possibly incorrect behavior of this driver are modeled accordingly.
  • the simulation unit is designed to simulate the behavior of a number of first virtual road users during actual operation of the vehicle on the road.
  • all first virtual road users are virtual other vehicles.
  • the simulation unit is designed to simulate the behavior of the first virtual road users with one another with interactions between the first virtual road users acting on both sides.
  • the simulation unit is designed to simulate the behavior of at least one second virtual road user during actual operation of the vehicle on the road, with the behavior of the second virtual road user being determined solely on the basis of a traffic flow simulation, with the traffic flow simulation being determined independently of the Behavior of the vehicle is such that there is no mutual interaction between the second virtual road user and the vehicle.
  • the behavior of the second virtual road user does not lead to a two-way interaction between the vehicle and the second virtual road user, since the second virtual road user obeys a predetermined pattern in its behavior.
  • the simulation unit is arranged in the vehicle and connected via the sensor interface via a wired signal line or wireless signal line to a computing unit for executing the driving control function.
  • the simulation unit is arranged outside of the vehicle in an external server and is connected via a wireless signal line to a computing unit for executing the driving control function via the sensor interface.
  • the system also has a display unit, with the actual journey of the vehicle on the road taking place at least partially through manual operation by a driver or personal monitoring by a passenger in the vehicle, with the driver or passenger in the vehicle being the first virtual road users is shown on the display unit.
  • the display unit is one of the following: screen, head-up display, virtual or augmented reality glasses.
  • a further aspect of the invention relates to a method for verification and validation a driving control function for the automatic operation of the vehicle, with the behavior of at least a first virtual road user being simulated by a simulation unit during real operation of the vehicle on a road, and virtual vehicle-to-vehicle communication and/or virtual sensor data being generated from the simulated behavior are generated, wherein the virtual vehicle-to-vehicle communication and/or the virtual sensor data of the driving control function are made available through a sensor interface in such a way that the virtual vehicle-to-vehicle communication is indistinguishable for the driving control function a real vehicle-to-vehicle communication and the virtual sensor data superimpose or replace the original vehicle-specific sensor data depending on the type of sensor, and the output variables of the driving control function are monitored by an analysis unit.
  • the simulation is repeated with different numbers of first virtual road users.
  • a strict strategy for validating and verifying the driving control function is followed, since a wide parameter space can be tested over different scenarios.
  • the simulation is carried out repeatedly with automatically generated test scenarios in that the road users potentially relevant for the interaction with the driving control function are detected in the traffic flow simulation and transferred to the driving dynamics simulation.
  • the level of detail of the simulation of these road users is preferably increased in an adaptive manner according to the requirements of the scenario, which means limited in a suitable manner in terms of time and space.
  • the detection of the relevant road users can be implemented, for example, by artificial intelligence.
  • the generation of new test scenarios can be implemented in a user-friendly manner immediately during the test drives of the automatically operated vehicle for the purpose of validation and verification of the driving control function.
  • Fig. 2 Another situation created by a simulation of the system according to another embodiment of the invention.
  • Fig. 3 A method of using the system for validation and verification of the driving control function.
  • the system 1 shows a system 1 for a vehicle 3 for verifying and validating a driving control function for the automatic operation of the vehicle 3.
  • the driving control function reacts to other road users, for example to give priority to an intersection when another road user drives into this intersection.
  • the system 1 has a simulation unit 5 , a sensor interface 7 and an analysis unit 9 .
  • simulation unit 5 generates the simulated behavior of a first virtual road user 11 and uses the simulated behavior to determine, among other things, visual virtual sensor data as artificially generated camera data.
  • the driving control function makes available to the driving control function via the sensor interface 7 in such a way that the virtual sensor data overlays the original vehicle-specific visual sensor data from a vehicle camera, so that the image of the first virtual road user 11 is overlaid in the image of the recorded environment.
  • the image of the first virtual road user 11 is reduced or enlarged and rotated and distorted in an optical flow according to the simulation by corresponding matrix transformations, as if the first virtual road user 11 (a vehicle such as the vehicle 3) were actually in the vicinity of the vehicle 3 available.
  • the driving control function reacts to the presence of this first virtual road user 11 and implements corresponding commands for the autonomous vehicle 3, for example stopping in front of an intersection where the first virtual road user 11 has priority.
  • the analysis unit 9 monitors the output variables of the drive control function by storing the Output variables and validate and verify with a model behavior so that it can be checked whether the driving control function makes the right decision and executes the correct behavior.
  • a “driver” is present in the vehicle 3, which is operated autonomously for the purpose of validating and verifying the driving control function, in order to be able to intervene if the vehicle 3 behaves incorrectly.
  • This driver of the vehicle 3 is shown an image of the first virtual road user 11 from the simulation data in real time on a head-up display 15, so that he can immediately carry out a plausibility check with regard to the behavior of the vehicle 3 based on the operation of the driving control function.
  • FIG. 2 shows another situation that can be created for the vehicle 3 .
  • a first virtual road user 11 and a second virtual road user 13 are on the road that the vehicle 3 is actually currently driving on. Both the first virtual road user 11 and the second virtual road user 13 originate from the simulation and are generated as in the method described in FIG. 1 .
  • the virtual first road user 11 and the second virtual road user 13 differ in that the first virtual road user 11 shows an interactive behavior towards the vehicle 3 and can behave cooperatively. There is vehicle-to-vehicle communication between the first virtual road user 11 and the vehicle 3, so that there is an interaction between these two vehicles that acts on both sides.
  • the second virtual road user 13 is simulated under different paradigms, so that there is no mutual interaction between the second virtual road user 13 and the vehicle 3, but the behavior of the second virtual road user 13 is only generated by a traffic flow simulation and it does not react to the behavior of the vehicle Vehicle 3 runs.
  • Fig. 3 shows a method for using the system 1 for the validation and verification of the driving control function of the vehicle 3.
  • a virtual road and a virtual traffic with their conditions are selected S1, as well as the rules for the detection of the scenarios according to set to test requirements S2.
  • the system 1 is then activated S3.
  • the test drives with the test scenarios generated by the simulation unit 5 are repeatedly carried out S4 and the analysis unit 9 records data that is then ultimately processed and evaluated S5.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Hardware Design (AREA)
  • Quality & Reliability (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Traffic Control Systems (AREA)

Abstract

L'invention concerne un système (1) pour un véhicule (3) pour vérifier et valider une fonction de commande de conduite pour le fonctionnement automatique du véhicule (3), comprenant une unité de simulation (5), une interface de capteur (7) et une unité d'analyse (9), l'unité de simulation (5) est configurée pour simuler le comportement d'au moins un premier utilisateur de route virtuelle (11) pendant le fonctionnement réel du véhicule (3) sur une route et pour générer une communication de véhicule à véhicule virtuelle et/ou des données de capteur virtuel à partir du comportement simulé, l'interface de capteur (7) est configurée pour rendre la communication de véhicule à véhicule virtuelle et/ou les données de capteur virtuel disponibles pour la fonction de commande de conduite de telle sorte que la communication de véhicule à véhicule virtuelle pour la fonction de commande de conduite ne peut pas être distinguée d'une communication réelle de véhicule à véhicule et que les données de capteur virtuel recouvrent ou remplacent les données de capteur interne de véhicule d'origine, en fonction du type de capteur, et l'unité d'analyse (9) est configurée pour surveiller les valeurs de sortie de la fonction de commande de conduite.
PCT/EP2022/072802 2021-08-26 2022-08-16 Validation d'une fonction de commande de conduite pour le fonctionnement automatique d'un véhicule WO2023025613A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
DE102021209394.9A DE102021209394A1 (de) 2021-08-26 2021-08-26 Validieren einer Fahrsteuerungsfunktion für den automatischen Betrieb eines Fahrzeugs
DE102021209394.9 2021-08-26

Publications (1)

Publication Number Publication Date
WO2023025613A1 true WO2023025613A1 (fr) 2023-03-02

Family

ID=83232680

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/EP2022/072802 WO2023025613A1 (fr) 2021-08-26 2022-08-16 Validation d'une fonction de commande de conduite pour le fonctionnement automatique d'un véhicule

Country Status (2)

Country Link
DE (1) DE102021209394A1 (fr)
WO (1) WO2023025613A1 (fr)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE102017007136A1 (de) 2017-07-27 2019-01-31 Opel Automobile Gmbh Verfahren und Vorrichtung zum Trainieren selbstlernender Algorithmen für ein automatisiert fahrbares Fahrzeug
US20200353943A1 (en) 2019-05-07 2020-11-12 Foresight Ai Inc. Driving scenario machine learning network and driving environment simulation
WO2020264276A1 (fr) 2019-06-28 2020-12-30 Zoox, Inc Générateur de scénarios synthétiques basé sur des attributs
EP3958129A1 (fr) * 2020-08-17 2022-02-23 Volvo Car Corporation Procédé et système de validation de logiciel de commande autonome pour un véhicule autonome

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE102014221011A1 (de) 2014-04-07 2015-10-08 Volkswagen Aktiengesellschaft Modul zur Simulation eines Sondereinsatzfahrzeuges in einer Verkehrsflusssimulation und Verfahren zur Durchführung einer Verkehrsflusssimulation
DE102015001972B4 (de) 2014-04-28 2023-09-14 Iav Gmbh Ingenieurgesellschaft Auto Und Verkehr Verfahren und Vorrichtung zum Testen von C2X-Kommunikation nutzenden Anwendungen

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE102017007136A1 (de) 2017-07-27 2019-01-31 Opel Automobile Gmbh Verfahren und Vorrichtung zum Trainieren selbstlernender Algorithmen für ein automatisiert fahrbares Fahrzeug
US20200353943A1 (en) 2019-05-07 2020-11-12 Foresight Ai Inc. Driving scenario machine learning network and driving environment simulation
WO2020264276A1 (fr) 2019-06-28 2020-12-30 Zoox, Inc Générateur de scénarios synthétiques basé sur des attributs
EP3958129A1 (fr) * 2020-08-17 2022-02-23 Volvo Car Corporation Procédé et système de validation de logiciel de commande autonome pour un véhicule autonome

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
KHASTGIR SIDDARTHA ET AL: "Identifying a gap in existing validation methodologies for intelligent automotive systems: Introducing the 3xD simulator", 2015 IEEE INTELLIGENT VEHICLES SYMPOSIUM (IV), IEEE, 28 June 2015 (2015-06-28), pages 648 - 653, XP033209796, DOI: 10.1109/IVS.2015.7225758 *
ZOFKA MARC RENE ET AL: "Testing and validating high level components for automated driving: simulation framework for traffic scenarios", 2016 IEEE INTELLIGENT VEHICLES SYMPOSIUM (IV), IEEE, 19 June 2016 (2016-06-19), pages 144 - 150, XP032938956, DOI: 10.1109/IVS.2016.7535378 *

Also Published As

Publication number Publication date
DE102021209394A1 (de) 2023-03-02

Similar Documents

Publication Publication Date Title
DE102006044086B4 (de) System und Verfahren zur Simulation von Verkehrssituationen, insbesondere unfallkritischen Gefahrensituationen, sowie ein Fahrsimulator
EP3970077B1 (fr) Procédé pour l'entraînement d'au moins un algorithme pour un appareil de commande d'un véhicule automobile, produit de programme informatique, véhicule automobile ainsi que système
DE102018128289B4 (de) Verfahren und vorrichtung für eine autonome systemleistung und zur einstufung
DE102019203712B4 (de) Verfahren zum Trainieren wenigstens eines Algorithmus für ein Steuergerät eines Kraftfahrzeugs, Computerprogrammprodukt, Kraftfahrzeug sowie System
DE102017214611B4 (de) Verfahren zum Überprüfen eines Reaktionssignals einer Fahrzeugkomponente sowie Überprüfungsvorrichtung und Kraftfahrzeug
AT523834B1 (de) Verfahren und System zum Testen eines Fahrerassistenzsystems
EP3899682A1 (fr) Surveillance de fonctions de conduite basée sur les réseaux neuronaux
DE4211556A1 (de) Verfahren zur fahreradaptiven, situationsspezifischen Modellierung des Autofahrerverhaltens in realer Fahrumgebung
EP3945338A2 (fr) Chemin de traitement du signal, dispositif de détection de l'environnement et procédé de validation d'un système de conduite à fonctionnement automatisé
EP4226248A1 (fr) Dispositif et système pour tester un système d'aide à la conduite pour un véhicule
DE102019204892A1 (de) Verfahren und Steuergerät zum Erkennen einer Müdigkeit eines Fahrers für ein Fahrassistenzsystem für ein Fahrzeug
EP4022589A1 (fr) Identification automatisée d'un comportement anormal d'un usager de la route
DE102017214542A1 (de) Verfahren, Vorrichtung, Computerprogramm und Computerprogrammprodukt zur Verifikation eines Fahrerassistenzsystems
DE102020210789A1 (de) Analyse eines neuronalen Netzes für ein Fahrzeug
DE102017201796A1 (de) Steuervorrichtung zum Ermitteln einer Eigenbewegung eines Kraftfahrzeugs sowie Kraftfahrzeug und Verfahren zum Bereitstellen der Steuervorrichtung
WO2023099066A1 (fr) Simulation pour valider une fonction d'automatisation de la conduite d'un véhicule
WO2023025613A1 (fr) Validation d'une fonction de commande de conduite pour le fonctionnement automatique d'un véhicule
DE102018125712A1 (de) Fahrunterstützungsverfahrens für ein Fahrzeug
DE102019203205A1 (de) Verfahren zum Auswerten von Fahrzeugdaten sowie Fahrzeugdatenauswertesystem zum Durchführen eines derartigen Verfahrens
DE102020201851B4 (de) Verfahren zum Betreiben eines Assistenzsystems, sowie Assistenzsystem
AT524821A1 (de) Verfahren und System zum Erzeugen von Szenariendaten zum Testen eines Fahrerassistenzsystems eines Fahrzeugs
DE102019218079A1 (de) Bestimmen einer Steuerung eines zumindest teilweise autonom fahrenden Fahrzeugs
DE102019212830A1 (de) Analyse und Validierung eines neuronalen Netzes für ein Fahrzeug
DE102023000357B3 (de) Verfahren zum Erzeugen von Testdaten für eine Simulation eines Assistenzsystems eines zumindest teilweise assistiert betriebenen Kraftfahrzeugs, Computerprogrammprodukt, computerlesbares Speichermedium sowie elektronische Recheneinrichtung
DE10358494A1 (de) Fahrassistenzvorrichtung und Verfahren zur prädiktiven Steuerung eines Fahrzeugs

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22765784

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE