Method for guiding a vehicle system in a fully automated manner, and motor vehicle

Patent metadata

Field	Value
Publication number	US-10474151-B2
Application number	US-201616069465-A
Country	US
Kind code	B2
Filing date	Dec 13, 2016
Priority date	Jan 19, 2016
Publication date	Nov 12, 2019
Grant date	Nov 12, 2019

Abstract

The invention relates to a method for operating a motor vehicle system which is designed to guide the motor vehicle in different driving situation classes in a fully automated manner. The method includes ascertaining a current driving situation class from multiple specified driving situation classes using at least some of driving situation data, each driving situation class is assigned at least one analysis function. The method further includes retrieving configuration parameter sets assigned to the analysis functions of the current driving situation class from a database and producing analysis units which carry out the analysis function and which have not yet been provided by configuring configuration objects using the retrieved configuration parameter sets.

First claim

The invention claimed is:

1. A method for operating a motor vehicle system, which is designed to guide a motor vehicle in different driving situation classes in a fully automated manner, the method comprising:
dynamically producing, from a configuration object and using configuration parameter sets, at least some of analysis units designed as neural nets at least partly using software, during runtime,
wherein a computing structure comprising the analysis units is used to ascertain control data as driving situation data for guiding the motor vehicle in the fully automated manner and is used to use the control data to guide the motor vehicle, the control data is ascertained from surroundings data describing surroundings of the motor vehicle and ego data describes a state of the motor vehicle,
wherein each analysis unit is configured to ascertain output data from output data of at least one other analysis unit or the driving situation data, and
wherein the dynamically producing the at least some of analysis units comprises:
ascertaining a current driving situation class from multiple specified driving situation classes using at least some of the driving situation data, wherein each driving situation class is assigned at least one analysis function,
retrieving, from a database, the configuration parameter sets assigned to analysis functions of the current driving situation class, and
producing analysis units that carry out the analysis functions of the current driving situation class and that have not yet been provided by configuring configuration objects using the retrieved configuration parameter sets.

2. The method according to claim 1 , further comprising:
determining the current driving situation class using a neural net.

3. The method according to claim 1 , further comprising:
deactivating analysis units configured as neural nets that perform analysis functions and are not associated with the current driving situation class.

4. The method according to claim 1 , wherein a maximum number of the analysis units that are based on the neural nets are simultaneously active.

5. The method according to claim 1 , wherein at least one analysis unit, which realizes a basic function through a neural net, is always active.

6. The method according to claim 5 , wherein the basic function is an object recognition or a classification function.

7. The method according to claim 1 , wherein the multiple specified driving situation classes describe a country currently driven in, or at least one configuration parameter set specific for at least one country is used.

8. The method according to claim 1 , wherein the retrieving the configuration parameter sets from the database comprises:
retrieving the configuration parameter sets from the database located on a motor vehicle-external central processing unit.

9. The method according to claim 1 , further comprising:
in response to a driver intervention, which indicates a wrong result of an analysis unit, transmitting an error dataset of a current traffic situation to a processing device or a motor vehicle-external central processing unit after the current traffic situation is concluded.

10. The method according to claim 9 , further comprising:
in response to a training condition being met, using the error dataset for training at least one neural net used in the current traffic situation, while the associated configuration parameter set is updated.

11. A motor vehicle, comprising:
a motor vehicle system configured to guide the motor vehicle in different driving situation classes in a fully automated manner; and
a control device configured to:
dynamically produce, from a configuration object and using configuration parameter sets, at least some of analysis units designed as neural nets at least partly using software during runtime,
wherein a computing structure comprising the analysis units is used to ascertain control data as driving situation data for guiding the motor vehicle in the fully automated manner and is used to use the control data to guide the motor vehicle, the control data is ascertained from surroundings data describing surroundings of the motor vehicle and ego data describes a state of the motor vehicle,
wherein each analysis unit is configured to ascertain output data from output data of at least one other analysis unit or the driving situation data, and
wherein to dynamically produce the at least some of analysis units, the control device is configured to:
ascertain a current driving situation class from multiple specified driving situation classes using at least some of the driving situation data, wherein each driving situation class is assigned at least one analysis function,
retrieve, from a database, the configuration parameter sets assigned to analysis functions of the current driving situation class, and
produce analysis units that carry out the analysis functions of the current driving situation class and that have not yet been provided by configuring configuration objects using the retrieved configuration parameter sets.

12. The motor vehicle according to claim 11 , wherein the control device is further configured to:
determine the current driving situation class using a neural net.

13. The motor vehicle according to claim 11 , wherein the control device is further configured to:
deactivate analysis units configured as neural nets that perform analysis functions and are not associated with the current driving situation class.

14. The motor vehicle according to claim 11 , wherein a maximum number of the analysis units that are based on the neural nets are simultaneously active.

15. The motor vehicle according to claim 11 , wherein at least one analysis unit, which realizes a basic function through a neural net, is always active.

16. The motor vehicle according to claim 15 , wherein the basic function is an object recognition or a classification function.

17. The motor vehicle according to claim 11 , wherein the multiple specified driving situation classes describe a country currently driven in, or at least one configuration parameter set specific for at least one country is used.

18. The motor vehicle according to claim 11 , wherein to retrieve the configuration parameter sets from the database, the control device is configured to:
retrieve the configuration parameter sets from the database located on a motor vehicle-external central processing unit.

19. The motor vehicle according to claim 11 , wherein the control device is further configured to:
in response to a driver intervention, which indicates a wrong result of an analysis unit, transmit an error dataset of a current traffic situation to a processing device or a motor vehicle-external central processing unit after the current traffic situation is concluded.

20. The motor vehicle according to claim 19 , wherein the control device is further configured to:
in response to a training condition being met, use the error dataset for training at least one neural net used in the current traffic situation, while the associated configuration parameter set is updated.

Assignees

• Audi Ag

Inventors

• Netter Florian

Classifications

• B60W2050/0057

  Frequency analysis, spectral techniques or transforms · CPC title

• B60W2050/0014

  Adaptive controllers · CPC title

• B60W2050/0004

  In digital systems, e.g. discrete-time systems involving sampling · CPC title

• B60W50/0097

  Predicting future conditions · CPC title

• G08G1/096741

  where the source of the transmitted information selects which information to transmit to each vehicle · CPC title