WO2022078770A1 - Procédé de production d'un modèle de comportement pour une flotte de véhicules automobiles au moyen d'un dispositif informatique électronique extérieur au véhicule automobile, et dispositif informatique électronique extérieur au véhicule automobile - Google Patents

Procédé de production d'un modèle de comportement pour une flotte de véhicules automobiles au moyen d'un dispositif informatique électronique extérieur au véhicule automobile, et dispositif informatique électronique extérieur au véhicule automobile Download PDF

Info

Publication number
WO2022078770A1
WO2022078770A1 PCT/EP2021/076975 EP2021076975W WO2022078770A1 WO 2022078770 A1 WO2022078770 A1 WO 2022078770A1 EP 2021076975 W EP2021076975 W EP 2021076975W WO 2022078770 A1 WO2022078770 A1 WO 2022078770A1
Authority
WO
WIPO (PCT)
Prior art keywords
motor vehicle
computing device
electronic computing
transmitted
motor
Prior art date
Application number
PCT/EP2021/076975
Other languages
German (de)
English (en)
Inventor
Stefan Studer
Original Assignee
Mercedes-Benz Group AG
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 Mercedes-Benz Group AG filed Critical Mercedes-Benz Group AG
Publication of WO2022078770A1 publication Critical patent/WO2022078770A1/fr

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/005—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 with correlation of navigation data from several sources, e.g. map or contour matching
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/0088—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots characterized by the autonomous decision making process, e.g. artificial intelligence, predefined behaviours
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
    • G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
    • G05D1/02—Control of position or course in two dimensions
    • G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
    • G05D1/0287—Control of position or course in two dimensions specially adapted to land vehicles involving a plurality of land vehicles, e.g. fleet or convoy travelling
    • G05D1/0291—Fleet control
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G06N20/20—Ensemble learning

Definitions

  • Method for generating a behavior model for a motor vehicle fleet by means of an electronic computing device external to the vehicle, and an electronic computing device external to the vehicle
  • the invention relates to a method for generating a behavior model for a motor vehicle fleet using an electronic computing device external to the vehicle.
  • the invention also relates to an electronic computing device external to the vehicle.
  • EP 3 117 274 A1 discloses that operating data from a number of source systems are used to control a target system, for example a gas or wind power plant or another technical system.
  • the operating data of the source systems are received and differentiated by source system-specific identifiers.
  • a neural model based on the received operating data of the source systems is taken into account of the source system-specific identifiers, wherein a first neural model component is trained on properties shared by the source systems and a second neural model component is trained on properties that vary between the source systems.
  • the trained neural model is further trained based on the operational data of the target system, with further training of the second neural model component being given preference over further training of the first neural model component.
  • the target system is controlled with the help of the further trained neural network.
  • the object of the present invention is to create a method and an electronic computing device external to the motor vehicle, by means of which a behavior model for a motor vehicle fleet can be generated efficiently.
  • One aspect of the invention relates to a method for generating a behavior model for a motor vehicle fleet using an electronic computing device external to the vehicle, in which at least one parameter characterizing a neural network for a control device of a motor vehicle in the motor vehicle fleet is trained using the electronic computing device external to the vehicle and transmitted to the control device and internally in the vehicle is saved.
  • At least one item of information for the behavior model is recorded by means of a recording device of the motor vehicle and a stored motor vehicle-internal neural network is further trained in the motor vehicle by means of the characterizing parameter and the at least one piece of information and at least one weighting parameter of the motor-vehicle further trained neural network is trained by the control device is transmitted to the vehicle-external electronic computing device and the behavior model is generated on the basis of the characterizing parameter transmitted to the control device and on the basis of the transmitted weighting parameter by means of the vehicle-external electronic computing device for the vehicle fleet.
  • the behavior model can thus be generated efficiently and while saving bandwidth.
  • the behavior model is in particular a so-called “world model”.
  • the costs for the bandwidth and for data handling within the electronic computing device external to the motor vehicle can be reduced during data transmission and subsequent processing.
  • the neural network is an example of a machine learning method. Alternative methods can also be used, which can be described by a defined architecture and parameters and learning with pre-trained parameters is possible.
  • the corresponding data or information which can relate to raw data, for example, is processed directly by decentralized training within the control device in the motor vehicle in the sense of edge computing.
  • the transmission bandwidth to the vehicle-external electronic computing device is reduced because, for example, only one weighting parameter of the neural network is transmitted.
  • the knowledge is, so to speak, memorized and compressed in the neural networks.
  • the behavior model can then be created from the trained neural networks, which are internal to the motor vehicle, without any great effort.
  • the control device which is located inside the motor vehicle, can in particular also be an electronic computing device.
  • the method is therefore a computer-implemented method which, in particular, uses a corresponding computer program product with commands to carry out the method on the electronic computing device external to the motor vehicle.
  • the motor vehicle-external electronic computing device has in particular a computer-readable storage medium, which in turn can then have the computer program product.
  • a weighting matrix for the neural network is transmitted to the control device or the neural network as the parameter characterizing the neural network.
  • pre-trained neural networks from the vehicle-external electronic computing device are transmitted to the motor vehicle.
  • a neural network can already be made available inside the motor vehicle, to which only the corresponding weighting matrix is then transmitted.
  • the network architecture for the neural network can then be derived on the basis of the weighting matrix.
  • the behavior model can be generated efficiently.
  • the behavior model is generated by the vehicle-external electronic computing device on the basis of a large number of transmitted weighting parameters from a large number of motor vehicles in the motor vehicle fleet.
  • the motor vehicle fleet thus has a large number of motor vehicles.
  • the large number of motor vehicles can then in turn collect their own information with their own detection devices and then in turn train their motor vehicle-internal neuronal network.
  • the corresponding weighting parameters can then in turn be transmitted to the electronic computing device external to the motor vehicle, in which case the behavior model can then in turn be generated on the basis of the large number of transmitted weighting parameters.
  • a type of swarm knowledge can thus be collected, and on the basis of this swarm knowledge the behavior model for the motor vehicle fleet can in turn be generated.
  • the weighting parameter is transmitted to the electronic computing device external to the motor vehicle.
  • the weighting parameter is first sent from the motor vehicle to the electronic computing device external to the vehicle .
  • the weighting parameter can thus only be transmitted if there is a deviation from the pre-trained neural network, as a result of which the volume of data transmission can be further reduced.
  • the behavior model for the motor vehicle fleet is generated using an ensemble method.
  • the ensemble methodology is in particular a form of machine learning. It takes a finite set of different learning algorithms to get better results than a single learning algorithm. In particular, a result that is already satisfactory can be achieved with a lower depth of calculation.
  • the soft weight sharing method is also known as soft weight sharing.
  • This is another machine learning method. This is a deep learning algorithm that can be used to efficiently generate the behavioral model.
  • the control device carries out a prognosis for a fault cause and/or a prognosis for the use of a navigation device of the motor vehicle and/or a prognosis for a use of an assistance system of the motor vehicle.
  • Forecasts can be carried out inside the motor vehicle using the neural network. Provision can then be made for a behavioral model for the cause of the error and/or for the use of the navigation device and/or for the use of the assistance system of the motor vehicle to be carried out using the electronic computing device external to the motor vehicle. In this way, different behavior models can be generated, which can then in turn be used for the motor vehicle fleet.
  • the applications mentioned here are only examples; in particular, any prognosis in the form of a classification or regression can be represented by this method.
  • a number of measuring points of the detection device and a resulting amount of data to be transmitted are determined and the resulting amount of data is compared with a data amount which describes the weighting parameter and as a function of the comparison by means of the control device a decision is made as to whether the data volume of the measuring points or the data volume of the weighting parameter is transmitted to the electronic computing device external to the motor vehicle.
  • the amount of data from the measuring points is in particular so-called raw data. If the data volume of the raw data is less than the data volume of the weighting parameter, it can be decided that the data volume of the measuring points is to be transmitted.
  • the weighting parameter is transmitted. Efficient generation of the behavior model can thus be realized.
  • 20 variables can be stored per recording, which corresponds in particular to 20 float values. If, for example, the neural network can then be described by no more than 1893 float values, with a total of more than 94.7 measurement points being available through quotient formation, it is more efficient from a bandwidth perspective to transmit the weights of the neural network instead of the measurement points themselves.
  • the neural network is pre-trained, and the neural network pre-trained with the reduced data set or a weighting matrix is transmitted to the control device.
  • the neural network is generated with reduced data and is only fully trained inside the motor vehicle.
  • This neural network pre-trained with the reduced data set or the weighting matrix is then transmitted to the control device, with the pre-trained neural network then being further trained inside the motor vehicle.
  • the behavior model can thus be generated by using a reduced bandwidth.
  • a further aspect of the invention relates to a computer program product with instructions which, when instructions are executed on an electronic computing device, cause the steps according to the preceding aspect to be carried out. Furthermore, the invention also relates to a computer-readable storage medium with the aforementioned computer program product.
  • Yet another aspect of the invention relates to an electronic computing device external to the motor vehicle for generating a behavior model for a motor vehicle fleet, the electronic computing device external to the motor vehicle being designed to carry out a method according to the preceding aspect.
  • the method is carried out using the electronic computing device external to the motor vehicle.
  • Advantageous embodiments of the method are as advantageous embodiments of the computer program product, the computer-readable View storage medium and the vehicle-external electronic computing device.
  • the motor vehicle-external electronic computing device has physical features that are required to carry out the method or an advantageous embodiment thereof.
  • the single figure shows a schematic side view of an embodiment of an electronic computing device external to the motor vehicle for a motor vehicle fleet.
  • the figure shows a schematic side view of an embodiment of a motor vehicle-external electronic computing device 10.
  • the motor vehicle-external electronic computing device 10 is designed to generate a behavior model 12 for a motor vehicle fleet 14 shown purely schematically.
  • a motor vehicle 16 is also shown in particular, the motor vehicle 16 being to be regarded as part of the motor vehicle fleet 14 .
  • Motor vehicle 16 has a control device 18 , a detection device 20 and a functional unit 22 .
  • Functional unit 22 can be a head unit of motor vehicle 16 or an assistance system of motor vehicle 16, for example. This list is purely exemplary and only serves to understand the invention. Further functional units 22 in motor vehicle 16 can thus also be used.
  • a neural network 24 characterizing parameters 28 for the control device 18 of the motor vehicle 16 of the motor vehicle fleet 14 trained by means of the vehicle-external electronic computing device 10, in particular pre-trained, and transmitted to the control device 18 and stored internally in the vehicle.
  • At least one item of information for the behavior model 12 is recorded by means of the recording device 20 of the motor vehicle 16 and the stored motor vehicle-internal neural network 26 is further trained in the motor vehicle by means of the characterizing parameter 28 and the at least one piece of information and at least one weighting parameter 30 of the motor-vehicle further trained neural network Network 26 is transmitted from the control device 18 to the electronic computing device 10 external to the motor vehicle, and the behavior model 12 is generated for the motor vehicle fleet 14 on the basis of the characterizing parameter 28 transmitted to the control device 18 and on the basis of the transmitted weighting parameter 30 by means of the electronic computing device 10 external to the motor vehicle.
  • error messages can be stored within the control device 18, environmental data within the control device 18, such as temperature, voltage, input variables for the control device 18, and error signals from networked control devices.
  • the aim is to predict the causes of errors.
  • data on the day, time, weather and location can be stored.
  • a prognosis for the use of the navigation device is generated.
  • the assistance system which can in particular also be referred to as ADAS, it is stored which activation of the assistance functions, which driving situation or which location is present.
  • the class of hot spots for example for pedestrian braking, can be generated.
  • a number of measuring points of the detection device 20 and a data volume resulting therefrom and to be transmitted are determined and the resulting data volume is compared with a data volume which describes the weighting parameter 30 and depending on the comparison by means of the control device 18 a decision is made as to whether the data volume of the measuring points or the data volume of the weighting parameter 30 is to be transmitted to the electronic computing device 10 external to the motor vehicle.
  • the neural network 24 in the motor vehicle-external electronic computing device 10 which can also be referred to as the backend, is pre-trained with a reduced data set.
  • the pre-trained neural networks 24 or their weighting matrices, from which the network architecture can then in turn be derived, are transmitted to the control device 18 in the motor vehicle 16 .
  • the control device 18 collects further data and uses it to train the motor vehicle's internal neural network 26 for the behavior observed precisely in this motor vehicle 16, which can also be referred to as edge computing.
  • the weights of the motor vehicle's internal neural network 26 are transferred from this motor vehicle 16 to the vehicle's external electronic Computing device 10 transferred.
  • the behavior model 12 is created from the individual models in the vehicle-external electronic computing device 10 with the aid of a method such as soft weight sharing or the ensemble method. With this behavior model 12, which can also be referred to as a world model, the desired forecasts, such as forecasts for the cause of the error, for the use of the navigation device or other assistance system can then be generated.
  • the data are thus processed directly by the decentralized training in the control device 18 in the sense of edge computing.
  • the transmission bandwidth in the backend is reduced since only the weighting parameter 30 of the motor vehicle's internal neural network 26 is transmitted.
  • the knowledge is so to speak memorized and compressed in the motor vehicle's internal neural networks 26 .
  • the behavior model 12 can then in turn be constructed from the trained neural networks 26 without great effort.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Software Systems (AREA)
  • Automation & Control Theory (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Computing Systems (AREA)
  • Aviation & Aerospace Engineering (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Medical Informatics (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Business, Economics & Management (AREA)
  • Game Theory and Decision Science (AREA)
  • Electric Propulsion And Braking For Vehicles (AREA)

Abstract

L'invention concerne un procédé de production d'un modèle de comportement (12) pour une flotte de véhicules automobiles (14) au moyen d'un dispositif informatique électronique extérieur au véhicule automobile (10), dans lequel au moins un paramètre (28) d'un véhicule automobile (16), lequel paramètre caractérise un réseau neuronal (24), est entraîné au moyen du dispositif informatique électronique extérieur au véhicule automobile (10) et est transmis au dispositif de commande (18) ; - au moins une information pour le modèle de comportement (12) est saisi au moyen d'un dispositif de capture (20) du véhicule automobile (16) ; - un réseau neuronal à l'intérieur du véhicule automobile stocké (26) est ensuite entraîné, à l'intérieur du véhicule automobile, au moyen du paramètre de caractérisation (28) ; - au moins un paramètre de pondération (30) du réseau neuronal (26) est transféré du dispositif de commande (18) au dispositif informatique électronique extérieur au véhicule automobile (10) ; et - le modèle de comportement (12) est produit pour la flotte de véhicules automobiles (14) au moyen du dispositif informatique électronique extérieur au véhicule automobile (10) sur la base du paramètre de caractérisation (28) transmis et sur la base du paramètre de pondération (30). L'invention concerne également un dispositif informatique extérieur au véhicule automobile (10).
PCT/EP2021/076975 2020-10-12 2021-09-30 Procédé de production d'un modèle de comportement pour une flotte de véhicules automobiles au moyen d'un dispositif informatique électronique extérieur au véhicule automobile, et dispositif informatique électronique extérieur au véhicule automobile WO2022078770A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
DE102020006267.9 2020-10-12
DE102020006267.9A DE102020006267A1 (de) 2020-10-12 2020-10-12 Verfahren zum Erzeugen eines Verhaltensmodells für eine Kraftfahrzeugflotte mittels einer kraftfahrzeugexternen elektronischen Recheneinrichtung, sowie kraftfahrzeugexterne elektronische Recheneinrichtung

Publications (1)

Publication Number Publication Date
WO2022078770A1 true WO2022078770A1 (fr) 2022-04-21

Family

ID=78085884

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/EP2021/076975 WO2022078770A1 (fr) 2020-10-12 2021-09-30 Procédé de production d'un modèle de comportement pour une flotte de véhicules automobiles au moyen d'un dispositif informatique électronique extérieur au véhicule automobile, et dispositif informatique électronique extérieur au véhicule automobile

Country Status (2)

Country Link
DE (1) DE102020006267A1 (fr)
WO (1) WO2022078770A1 (fr)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3117274A1 (fr) 2014-04-22 2017-01-18 Siemens Aktiengesellschaft Procédé, dispositif de commande et produit-programme d'ordinateur permettant de commander un système cible par l'entraînement distinct de premier et second modèles de réseaux neuronaux récurrents qui sont initialement entraînés au moyen de données opérationnelles de systèmes sources
EP3206411A1 (fr) * 2016-02-11 2017-08-16 Volvo Car Corporation Agencement et procédé permettant de prédire le frottement d'une route dans un réseau routier
WO2019001649A1 (fr) * 2017-06-30 2019-01-03 Conti Temic Microelectronic Gmbh Transfert de connaissance entre différentes architectures d'apprentissage profond
DE102019005825A1 (de) * 2019-08-19 2020-03-12 Daimler Ag Verfahren zum Erzeugen von Trainingsdaten für eine kraftfahrzeugexterne elektronische Recheneinrichtung, sowie Fahrerassistenzsystem

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3117274A1 (fr) 2014-04-22 2017-01-18 Siemens Aktiengesellschaft Procédé, dispositif de commande et produit-programme d'ordinateur permettant de commander un système cible par l'entraînement distinct de premier et second modèles de réseaux neuronaux récurrents qui sont initialement entraînés au moyen de données opérationnelles de systèmes sources
EP3206411A1 (fr) * 2016-02-11 2017-08-16 Volvo Car Corporation Agencement et procédé permettant de prédire le frottement d'une route dans un réseau routier
WO2019001649A1 (fr) * 2017-06-30 2019-01-03 Conti Temic Microelectronic Gmbh Transfert de connaissance entre différentes architectures d'apprentissage profond
DE102019005825A1 (de) * 2019-08-19 2020-03-12 Daimler Ag Verfahren zum Erzeugen von Trainingsdaten für eine kraftfahrzeugexterne elektronische Recheneinrichtung, sowie Fahrerassistenzsystem

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
YE DONGDONG ET AL: "Federated Learning in Vehicular Edge Computing: A Selective Model Aggregation Approach", IEEE ACCESS, IEEE, USA, vol. 8, 20 January 2020 (2020-01-20), pages 23920 - 23935, XP011771171, DOI: 10.1109/ACCESS.2020.2968399 *

Also Published As

Publication number Publication date
DE102020006267A1 (de) 2022-04-14

Similar Documents

Publication Publication Date Title
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
DE102017223751A1 (de) Verfahren und Vorrichtung zur Erkennung von Anomalien in einem Datenstrom eines Kommunikationsnetzwerks
DE102014118479A1 (de) Affektive Benutzerschnittstelle in einem autonomen Fahrzeug
DE102017213119A1 (de) Verfahren und Vorrichtung zum Ermitteln von Anomalien in einem Kommunikationsnetzwerk
EP3899682B1 (fr) Surveillance de fonctions de conduite basée sur les réseaux neuronaux
DE102019124018A1 (de) Verfahren zum Optimieren von Tests von Regelsystemen für automatisierte Fahrdynamiksysteme
EP3684015A1 (fr) Dispositif et procédé de classification de données, en particulier pour un réseau de zone de contrôleur ou un réseau ethernet automobile
WO2020187591A1 (fr) Procédé et dispositif de commande d'un robot
DE102018209595A1 (de) Verfahren zum automatischen Bestimmen eines Straßenzustands
DE102018209250A1 (de) Steuereinrichtung, Verfahren zur Steuerung eines Steuergeräts, computerlesbares Speichermedium und Steuersystem
EP2102723B1 (fr) Procédé et dispositif pour diagnostiquer des fonctions et des systèmes de véhicules
EP3748453B1 (fr) Procédé et dispositif de réalisation automatique d'une fonction de commande d'un véhicule
DE102018129563A1 (de) Verfahren zum Bestimmen des Steuermodus eines Lenkrads
EP3460727A1 (fr) Procédé d'analyse d'un comportement fonctionnel d'un système technique et unité d'évaluation
EP3748454B1 (fr) Procédé et dispositif de réalisation automatique d'une fonction de commande d'un véhicule
WO2022078770A1 (fr) Procédé de production d'un modèle de comportement pour une flotte de véhicules automobiles au moyen d'un dispositif informatique électronique extérieur au véhicule automobile, et dispositif informatique électronique extérieur au véhicule automobile
DE102020107001A1 (de) Verfahren und Vorrichtung zur Ermittlung des Zustands eines elektrischen Bordnetzes eines Fahrzeugs
DE102019113958A1 (de) Verfahren zur Leistungssteigerung eines Fahrzeugsystems mit einem neuronalen Netz zum Steuern einer Fahrzeugkomponente
DE102018128892A1 (de) Verfahren, System und Computerprogramm zum Bereitstellen einer Information zum Betreiben eines Fahrzeugs
DE102020204867A1 (de) Verfahren und System zum Betreiben einer Fahrmanöverplanung mindestens eines zumindest teilautomatisiert fahrenden Fahrzeugs
DE102005034047A1 (de) Datenübertragungsverfahren und Datenübertragungssystem
EP4339066A1 (fr) Modèle dynamique pour véhicule ferroviaire
DE102021207493A1 (de) Verfahren zur Unterstützung eines Betriebs eines Fahrzeuges mit einer Sensoreinheit, Computerprogrammprodukt sowie System
WO2023237274A1 (fr) Concept de sélection d'extraits audio de bruit de frein dans un véhicule

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: 21789620

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21789620

Country of ref document: EP

Kind code of ref document: A1