Method for Driving Maneuver Assistance of a Vehicle, Device, Computer Program, and Computer Program Product

Abstract

In a method for driving maneuver assistance of a vehicle, a predefined neural network is provided, which is designed to determine whether a predefined driving maneuver is probably possible. A predefined driver model is provided, which is designed to predict a probable future behavior of a vehicle. A current driving situation of the vehicle is determined. Depending on the determined driving situation, the driver model and the neural network, it is determined whether a predefined driving maneuver is possible. Depending on the determination as to whether the driving maneuver is possible, a driver assistance function for the driving maneuver is carried out and/or the driving maneuver is carried out autonomously.

Claims

1-11. (canceled)

12. A method for driving maneuver assistance of a vehicle, comprising: providing a predefined neural network designed to determine whether a predefined driving maneuver is probably possible; providing a predefined driver model designed to predict a probable future behavior of a vehicle; determining a current driving situation of the vehicle; on the basis of the determined driving situation, the driver model and the neural network, determining whether a predefined driving maneuver is possible; and on the basis of the determination of whether the driving maneuver is possible, carrying out a driver assistance function for the driving maneuver and/or autonomously carrying out the driving maneuver.

13. The method according to claim 12, wherein the neural network is bidirectional.

14. The method according to claim 13, wherein the neural network is recurrent.

15. The method according to claim 12, wherein the neural network is recurrent.

16. The method according to claim 12, further comprising: providing a set of driving data, wherein the driving data are representative of past journeys; and training the neural network using the set of driving data.

17. The method according to claim 16, wherein the set of driving data comprises data relating to determined driving situations of the journeys and the neural network is trained on the basis of the data relating to determined driving situations, and the data relating to determined driving situations comprise information relating to whether the driving maneuver can be carried out in the respective driving situation.

18. The method according to claim 17, wherein the data relating to determined driving situations are determined by checking whether the driving maneuver was carried out following the respective driving situation.

19. The method according to claim 18, wherein the data relating to determined driving situations are determined by checking whether the driving maneuver can be theoretically carried out following the respective driving situation.

20. The method according to claim 17, wherein the data relating to determined driving situations are determined by checking whether the driving maneuver can be theoretically carried out following the respective driving situation.

21. The method according to claim 12, wherein the driving maneuver is a lane change or an overtaking maneuver.

22. An apparatus for driving maneuver assistance of a vehicle, comprising: a computer and associated peripherals configured to: provide a predefined neural network designed to determine whether a predefined driving maneuver is probably possible; provide a predefined driver model designed to predict a probable future behavior of a vehicle; determine a current driving situation of the vehicle; on the basis of the determined driving situation, the driver model and the neural network, determine whether a predefined driving maneuver is possible; and on the basis of the determination of whether the driving maneuver is possible, carry out a driver assistance function for the driving maneuver and/or autonomously carry out the driving maneuver.

23. The apparatus according to claim 22, wherein the neural network is bidirectional.

24. The apparatus according to claim 23, wherein the neural network is recurrent.

25. A computer program product comprising a non-transitory computer readable medium having stored thereon executable program code, wherein during execution by a data processing apparatus, the program code is configured to: provide a predefined neural network designed to determine whether a predefined driving maneuver is probably possible; provide a predefined driver model designed to predict a probable future behavior of a vehicle; determine a current driving situation of the vehicle; on the basis of the determined driving situation, the driver model and the neural network, determine whether a predefined driving maneuver is possible; and on the basis of the determination of whether the driving maneuver is possible, carry out a driver assistance function for the driving maneuver and/or autonomously carry out the driving maneuver.

26. The product according to claim 25, wherein the neural network is bidirectional.

27. The product according to claim 26, wherein the neural network is recurrent.

Description

BRIEF DESCRIPTION OF THE DRAWING

[0033] FIG. 1 is a flowchart for driving maneuver assistance of a vehicle.

DETAILED DESCRIPTION OF THE DRAWING

[0034] FIG. 1 shows a flowchart of a program for driving maneuver assistance of a vehicle.

[0035] The program can be executed by an apparatus. The apparatus is implemented, for example, in a control apparatus of a vehicle.

[0036] The apparatus can also be referred to as an apparatus for driving maneuver assistance of a vehicle.

[0037] For this purpose, the apparatus has, in particular, a computing unit, a program and data memory and one or more communication interfaces, for example. The program and data memory and/or the computing unit and/or the communication interfaces can be formed in one structural unit and/or in a manner distributed among a plurality of structural units.

[0038] A program for driving maneuver assistance of a vehicle, in particular, is stored in the program and data memory of the apparatus for this purpose.

[0039] The program is started in a step S1 in which variables can be initialized, if necessary.

[0040] In a step S3, a predefined neural network is provided and is designed to determine whether a predefined driving maneuver is probably possible.

[0041] In order to train the neural network, a set of driving data is provided, for example, wherein the driving data are representative of past journeys. The network is then trained using the set of driving data.

[0042] The set of driving data comprises, for example, data relating to determined driving situations of the journeys. The neural network is trained, for example, on the basis of the data relating to determined driving situations, wherein the data relating to determined driving situations comprise information relating to whether the driving maneuver can be carried out in the respective driving situation.

[0043] The data relating to determined driving situations are determined, for example, by checking whether the driving maneuver was carried out following the respective driving situation.

[0044] Alternatively or additionally, the data relating to determined driving situations are determined by checking whether the driving maneuver can be theoretically carried out following the respective driving situation.

[0045] The neural network is bidirectional and/or recurrent, for example.

[0046] In a step S5, a predefined driver model is provided and is designed to predict a probable future behavior of a vehicle.

[0047] In a step S7, a current driving situation of the vehicle is determined.

[0048] For this purpose, data from a plurality of sensors are evaluated, for example. A driving situation is distinguished, for example, by one or more of the following parameters: [0049] current lane and/or speed of the ego vehicle, [0050] number (and position) of vehicles around the ego vehicle, [0051] behavior (current speeds and/or current accelerations and/or current lane changes) of these other vehicles, [0052] information relating to whether a lane, to which it is intended to change, is currently free, [0053] information relating to whether safety distances are complied with, [0054] information relating to a future course of the route.

[0055] In a step S9, it is determined whether a predefined driving maneuver is possible on the basis of the determined driving situation, the driver model and the neural network.

[0056] In a step S11, a driver assistance function for the driving maneuver is carried out and/or the driving maneuver is autonomously carried out on the basis of the determination of whether the driving maneuver is possible.

[0057] The driving maneuver is, for example, a lane change and/or an overtaking maneuver.

[0058] Finally, the program is ended in a step S13 and can be started again in step S1, if necessary.

[0059] By determining whether a predefined driving maneuver is possible on the basis of the determined driving situation, the driver model and the neural network, a particularly reliable hit accuracy for the correct driving maneuver assistance is possible since both a current situation and a neural network are used for the determination.

[0060] Furthermore, the method can be applied to many possible driving maneuvers provided that appropriate training data for the neural network are available.