DE102017219065A1 - METHOD AND DEVICE FOR CONFIGURING AN AUTOMATED VEHICLE GUIDANCE - Google Patents

Abstract

The invention relates to the field of automated driving of vehicles, in particular of motor vehicles. Specifically, the invention relates to a method for configuring an at least highly automated vehicle guidance of a vehicle. The method comprises the following automated steps: (i) detecting, preferably non-contact, user input to a man-machine interface related to an instruction for configuring the automated vehicle guidance as well as environmental information characterizing the current local environment of the vehicle; (ii) deriving a default for the automated vehicle guidance of the vehicle by means of an evaluation of the user input as a function of the surroundings information; (iii) determining a driving strategy related to the local environment of the vehicle in accordance with the specification; and (iv) generating and outputting control signals for driving the vehicle to perform the automated vehicle guidance in accordance with the established driving strategy. The invention further relates to a computer program for carrying out the method, a device set up to carry out the method and a vehicle having such a device.

Description

The present invention relates to the field of automated driving of vehicles, especially motor vehicles. Specifically, the invention relates to a method and apparatus for configuring an at least highly automated vehicle guidance of a vehicle.

In the automation of vehicle guidance, especially of motor vehicles such as cars and trucks or buses or even single-track vehicles, such as motorcycles, different levels of automation, which provides a corresponding control system (system) are regularly differentiated (see German Federal Highway Research Institute, BaSt , 2015).

In conventional "non-automated driving" ("Driver Only"), the driver performs the entire length of the travel (acceleration / deceleration) and the lateral guidance (steering) permanently (during the entire journey). Accordingly, no automated system intervenes in the longitudinal or transverse guidance.

A next level, "assisted driving", is also not yet considered as automated driving or automated vehicle guidance, but can be regarded as its precursor and starting point. The driver permanently performs either the transverse or the longitudinal guidance. The other subtask is executed within certain limits by the system. The driver must monitor the system permanently and be ready to take over the vehicle guidance at all times. Examples include the so-called "Adaptive Cruise Control" (ACC), in which the longitudinal guidance by means of adaptive distance and speed control is automatic, and the parking assistance, where the transverse guidance by the parking assistant (automatic steering in parking spaces) takes place while the driver, the longitudinal guidance controls.

The further stages described below represent various automation levels of automated driving or automated vehicle guidance, with the degree of automation increasing from stage to stage.

In "partially automated driving (TAF)", the system takes over both the lateral and longitudinal guidance of the vehicle for a certain period of time or in specific situations. The driver must continue to monitor the system permanently and be ready to take over the task at all times. Examples include motorway assistants, which provide automatic longitudinal and lateral guidance on highways to an upper speed limit, above which the driver himself must take over these tasks.

In "highly automated driving (HAF)", the system takes over the lateral and longitudinal guidance for a certain period of time or in specific situations. The driver no longer has to monitor the system permanently. He receives sufficient time reserve before he has to take on the driving task himself. The system warns the driver so before. An example of this is a highway chauffeur system that provides automatic longitudinal and lateral guidance on highways up to an upper speed limit where the driver does not need to permanently monitor, but must respond with some time reserve after takeover request.

In "fully automated driving (VAF)", the system completely takes over the lateral and longitudinal guidance in a defined application, for example when driving on the highway. The driver does not have to monitor the system. Before leaving the application, the system prompts the driver with sufficient time reserve to take on the driving task. If this is not done, the system is returned to the lowest risk state. System limits are all recognized by the system and the system is in all situations able to return to the lowest-risk system state.

"Autonomous (" driverless ") driving" represents the highest level of automation, in which the system takes over the entire vehicle guidance from start to finish and none of the persons in the vehicle must act as, at least temporary, driver.

To operate a non-autonomous motor vehicle or its various functions, or to their setting or configuration by the driver, various man-machine interfaces are known. In addition to the steering wheel and pedals, these are, above all, manually operated mechanical or electrical operating elements, such as buttons, buttons, levers, sliding or dials, touch-sensitive surfaces (for example track pads) or screens (touch screens). In particular, if a large number of different choices are to be made available by means of such man-machine interfaces, which is often done by means of a menu control, usually the user-friendliness and the access time suffer.

To address this, more manufacturers have recently introduced voice controls as some other human-machine interfaces that are typically designed to distinguish and recognize a narrowly-restricted cataloged set of predefined voice prompts. In particular, camera-based, detection of individual hand gestures as an additional man-machine interface for controlling selected Vehicle functions introduced, with this possibility is limited to the detection and distinction of very few predefined rudimentary movement sequences, so that accordingly only a very limited range of functions is controllable.

The object of the present invention is to provide a method and a device for at least highly automated vehicle guidance, that is, a vehicle guidance whose degree of automation corresponds at least to highly automated driving (HAF), which enables a vehicle occupant, in particular the driver, of a vehicle. In a simple way to influence the vehicle or driving behavior of the vehicle, while this is controlled by a highly or more highly automated vehicle management and without interrupting it.

A solution to this problem is achieved according to the teaching of the independent claims. Various embodiments and modifications of the invention are subject of the dependent claims.

A first aspect of the invention relates to a method for configuring an at least highly automated vehicle guidance of a vehicle, in particular of a motor vehicle, comprising the following automated steps: (i) Detecting, preferably non-contact, a user input on a human machine relating to an instruction for the configuration of the automated vehicle guidance Interface as well as environmental information characterizing the current local environment of the vehicle; (ii) deriving a default for the automated vehicle guidance of the vehicle by means of an evaluation of the user input as a function of the surroundings information; (iii) determining a driving strategy related to the local environment of the vehicle in accordance with the specification; and (iv) generating and outputting control signals for driving the vehicle to perform the automated vehicle guidance in accordance with the established driving strategy.

The term "vehicle guidance" within the meaning of the invention means the dynamic motion guidance of a vehicle, in particular its transverse and longitudinal guidance. A "configuration of the vehicle guidance" is understood as meaning an adjustment of the vehicle guidance with regard to a direct control of the vehicle for a specific driving maneuver, ie, for a steering maneuver or a braking or acceleration maneuver, superordinate plane. The setting thus defines, in the sense of a boundary condition, the solution space, at least partially, from which the respective at least highly automated vehicle guidance provides a concrete solution, i. Driving strategy, with concrete measures for vehicle control, selects, which is then automatically implemented in the form of corresponding concrete by means of a corresponding control of the longitudinal and lateral control components of the vehicle in appropriate driving maneuvers. The term "at least highly automated vehicle guidance" means that the degree of automation of the vehicle guidance must at least correspond to the highly automated vehicle guidance described above. In particular, therefore, a fully automated as well as an autonomous vehicle management fall under this term.

"Environment information" in the sense of the invention means data that characterize the current local environment of the vehicle, ie. H. Describe, classify or label aspects of it. A "local environment" is understood here to be the immediate spatial environment of the vehicle, ie a room area surrounding the vehicle that is comparable to the space area that the driver typically surveys in conventional non-automated driving and takes into account the current and timely vehicle control or the environment , which is typically detectable by means of vehicle-based sensors (eg cameras, radar, LiDAR, laser) from the vehicle. In particular, objects such as roadside facilities (such as lane markings, traffic signs, traffic lights, etc.), or other vehicles, passers-by or other road users, or infrastructural facilities (such as bridges or tunnels) located in front of, beside, or behind the local environment may be used are located in the vehicle. Typical extents of the local environment are about 1 km or less, i. z. B. at a few hundred meters.

The surroundings information can in particular be detected by the vehicle itself by means of a corresponding sensor system or can be completely or partially received via a communication interface from an external vehicle source, for example from a traffic monitoring system or a navigation system, and recorded in this way. The environment information may in particular characterize the position, orientation, dimension, extent, movement or other characteristics or states of objects in the local environment of the vehicle.

A "specification for the automated vehicle guidance of the vehicle" in the sense of the invention means a boundary condition derived from the user input and additionally in dependence on the surroundings information for the named solution space. In particular, this may correspond to a specific instruction associated with a particular executed user action with respect to the man-machine interface, or, as in the case of voice control, the recognized linguistic instruction of the user or occupants. The user input may relate in particular to one or more aspects of the vehicle environment characterized by the environment information, such as a vehicle driving ahead of the vehicle. The default could be, for example, in any case, "follow the immediately preceding vehicle," or in the case of a multi-lane road (which is recognized from the environment information): "reduce (or minimize) the frequency of lane changes".

A "driving strategy" in the sense of the invention means a solution from the solution space defined by means of the specification with concrete intermediate goals or measures for vehicle control, which is thus consistent with the specification. In particular, the driving strategy can be long-term and relate, for example, to an entire trip or even at short notice and, for example, relate only to a specific period, route section or a specific current driving situation. In the example given above with regard to the frequency of the lane changes, a driving strategy could, for example, be analogous to overtaking a further vehicle currently located to the right of the vehicle and then making a lane change in a gap identified before it, and then as long as possible on this lane remain without making any further overtaking maneuvers or more lane changes than specified by a selected threshold set to be "low" or in historical comparison with the already traveled distance. These elements of the driving strategy can, in particular, be understood as "intermediate goals" which, in their interaction with the overriding objective definition (solution space definition), are not necessarily directly determined by them, but must be derived from them. This can be done in particular by means of a solution of at least one corresponding mathematical optimization problem or simply by computer-aided trial and error evaluation of different approaches.

By means of the method, it is thus possible for an occupant of the vehicle to have a specific influence on a highly or even more automated vehicle guidance of the vehicle without, however, interrupting or even having to undertake the vehicle guidance in order to achieve a desired future behavior of the vehicle, in particular his Driving behavior, to effect. This opens the way for a cooperative vehicle management, in which the automated vehicle management interacts with the driver or other occupants, without the fact that at a given time or in a certain driving situation either only the automatic or only the driver determines the vehicle guidance or takes over. Thus, the automated vehicle guidance can respond to wishes of the driver or other vehicle occupants and be configured accordingly, while at the same time their benefits, in particular the concomitant comfort, safety and / or freedom of action of the occupants can be maintained.

In the following, preferred embodiments of the method will be described which, unless expressly excluded, may be combined as desired with one another as well as with the other aspects of the invention described below.

According to some embodiments, deriving the default comprises (i) detecting a driving instruction included in the user input relating to the vehicle itself and / or at least one off-vehicle reference object; (ii) checking the environment information as to whether an object corresponding to the reference object can be identified in the local environment of the vehicle; and (iii)
if so, deriving the default to define the driving instruction in relation to the identified object. In this way, the driving instruction to be implemented as part of the automated vehicle guidance or the corresponding specification can in particular also relate directly to an object external of the vehicle. For example, in the previously mentioned example, the specification "sequence of the immediately preceding vehicle" refers to this preceding vehicle as an external object and defines a boundary condition to be met for the driving strategy to be derived in relation or dependence thereon. In this way, in comparison to known systems, the scope of possibilities for influencing the vehicle guidance can be expanded and the definition of specifications about corresponding user inputs can be simplified on the basis of the thus created possibility of reference to objects in the vehicle environment.

According to some embodiments, the user input is freely selectable and is evaluated in the evaluation by means of a KI-based recognition method for their content recognition. Thus, it becomes possible to provide as user input not only a selection from a predefined catalog of possible user inputs, but instead to make arbitrary user inputs without such a restriction, which are analyzed by the KI-based recognition method to reveal their contents. So can in the case of a speech-based human-machine interface by the user (driver or occupant of the vehicle) made an input in free speech, without previously learned cataloged phrases or terms must be used. AI-based recognition methods for speech recognition are basically known in particular in the context of speech recognition solutions for digital assistants on smartphones and computers (cf., for example, "Siri" from Apple, Inc, "Alexa" from Amazon, Inc., etc.). On the other hand, the specification derived from the speech input can also be selected from a predetermined finite catalog of default options, which, however, can optionally be modifiable, in particular expandable, in particular via corresponding learning functions or software updates.

According to some variants of these embodiments, the AI method has a learning method, by means of which at least one further user input detected at the human-machine interface is evaluated and used to adapt a process used to derive the specification. This can be done in particular in the sense of so-called "reinforcing learning". The said further user input can in particular represent a response to a previous query of the system, wherein the system preferably then makes such a query, if there is an ambiguity or ambiguity in the evaluation of one or more previous user inputs. Thus, the system is given a learning capability that can be used, in particular, to continuously optimize the reliability and quality of the response of the system to user input.

According to some embodiments, the user input is at least partially as an acoustic input, in particular as a voice input, and it is detected as such. In this way, a non-contact and simple operation of user input at the man-machine interface is possible. In particular, this also makes it possible for other occupants of the vehicle as the driver in a simple way to make user inputs, since their performance does not or not to a similar extent depends on the position of the user in the vehicle, such as in a limited to touch-based human machine Interfaces, such as Buttons or touchscreens, that would be the case.

According to some embodiments, the user input is at least partially performed by one or more gestures performed by a user, in particular hand, arm or head gestures, and is detected as such. The input of user input by means of gestures can be carried out in particular in addition to other input options, such as voice-based user input, or even as sole user input. In this way, in turn can make contactless inputs, which also do not necessarily depend on the position of the user in the vehicle, and in particular can be performed silently. For example, a user could use a pointing gesture to identify an object in the vehicle to which the user input relates, with the desired behavior of the vehicle relating to the object simultaneously using the same or another gesture or another, eg voice-controlled, user input on the or can specify or enter a corresponding man-machine interface.

According to some embodiments, the specification relates to influencing a driving strategy for the automated vehicle guidance with respect to one or more of the following aspects or a combination of at least two thereof: (i) a vehicle guidance and / or route guidance of the vehicle beyond its current local environment; (ii) one or more local driving maneuvers; (iii) one or more vehicle dynamic characteristics of the vehicle.

A non-limiting example of the case (i) is the previously mentioned requirement to follow a preceding vehicle. In this case, although the default can be defined with reference to the immediate vehicle environment - namely, the vehicle ahead in the vicinity of the vehicle as an object - but this specification affects the vehicle over a longer period of time, during which the Vehicle can also leave at the time of user input present environment and often. A non-limiting example of the case (ii) is, for example, a specification according to which the vehicle is to make a lane change in a certain direction or at the next (allowable) opportunity to the side roadside drive to stop. A non-limiting example for case (iii) is a specification according to which the distance to immediately preceding vehicles is to be increased or shortened, more / less lane changes are to be made or an overall faster / slower driving. Overall, a broad range of options for defining the default and thus influencing automated vehicle guidance is thus provided.

According to some embodiments, the determination of the driving strategy is carried out dynamically, to which besides the specification additionally one or more of the following variables or a combination thereof as an input variable or input variables for the Determining the driving strategy are used: (i) one or more parameters identifying a current vehicle condition; (ii) traffic control or traffic monitoring information. In this case, the input variables or their combination are respectively sensory detected and / or received by means of a communication device. In this way, additional boundary conditions for defining the solution space from which a driving strategy is to be determined can be taken into account. The consideration of the current vehicle state can be used in particular, possibly due to this vehicle condition to avoid critical vehicle guidance or related driving maneuvers, so as not to affect the safety of the vehicle and its occupants. For example, if the engine is so heated after a long highway ride that its temperature is at the edge of a permitted operating range, the solution space, and thus the resulting driving strategy, could be defined as operating the engine in a lower power range until a substantial increase in fuel efficiency Lowering its operating temperature can be determined. In a similar manner, traffic control information, such as congestion reports, can be taken into account in order to avoid, for example, time-shifting or delaying, for example, conflicting lane changes or other driving strategies.

According to some embodiments, in case of evaluating the user input depending on the surroundings information, no or no clear result results as default or the driving strategy resulting from a derived specification is at least temporarily not feasible within the automated vehicle guidance, via the man-machine interface. Interface issued a message in this regard and / or requested a new user input. In this way, the driver or inmates are informed about the lack of practicability of the previous user input or of the corresponding specification and optional provision for a subsequent re-user input, which makes it possible to replace the specification accordingly.

A second aspect of the invention relates to a device for configuring an at least highly automated vehicle guidance of a vehicle, in particular of a motor vehicle. In this case, the device is set up to carry out the method according to the first aspect of the invention, preferably according to one of its embodiments described herein.

A third aspect of the invention relates to a vehicle, in particular a motor vehicle, with a device according to the second aspect of the invention.

A fourth aspect of the invention relates to a computer program which contains instructions which, when executed on one or more processors of a device according to the second aspect, cause them to carry out the method according to the first aspect, in particular according to one of its embodiments described herein.

The above, in each case for the method according to the first aspect and its embodiments thus applies equally to the device, the vehicle and the computer program according to the further aspects of the invention.

Further advantages, features and applications of the present invention will become apparent from the following detailed description in conjunction with the figures, in which the same reference numerals for the same or corresponding elements of the invention are used throughout.

• 1 ein Fahrzeug gemäß einer bevorzugten Ausführungsform der Erfindung;
• 2 ein Flussdiagramm zur Veranschaulichung einer bevorzugten Ausführungsform des erfindungsgemäßen Verfahrens; und
• 3 schematisch ein Fahrzeug, welches insbesondere das Fahrzeug aus 1, sein kann, einschließlich eines beispielhaften Fahrzeugumfelds, wobei das Fahrzeug mit einer Vorrichtung zur Ausführung des Verfahrens gemäß einer bevorzugten Ausführungsform der Erfindung ausgestattet ist;

Showing:

• 1 a vehicle according to a preferred embodiment of the invention;
• 2 a flow diagram illustrating a preferred embodiment of the method according to the invention; and
• 3 schematically a vehicle, which in particular the vehicle 1 , may be, including an exemplary vehicle environment, the vehicle being equipped with an apparatus for carrying out the method according to a preferred embodiment of the invention;

First, it will open 1 Referenced. The vehicle schematically illustrated therein 1 According to one embodiment of the invention, to its dynamic vehicle guidance on the one hand vehicle components 2 to its longitudinal control, in particular a drive train and an automatic braking system, and on the other vehicle components 3 to its transverse control, in particular a steering on. Furthermore, the vehicle points 1 a control device 4 for vehicle guidance, which has a processor unit having at least one processor 4a as well as a corresponding memory associated signal 4b contains. In the store 4b One or more computer programs that are configured are stored on the processor unit 4a to run and thereby cause the control device, a inventive method, for example according to the in 2 illustrated embodiment.

Furthermore, the vehicle points 1 a man-machine interface 5 on that with the control device 4 is signal-connected and a gesture recognition unit 5a , a speech recognition unit 5b and an output unit 5c having. The gesture recognition unit 5a In particular, it may have one or more cameras that are set up and so in the vehicle 1 are arranged so that they can detect sensory gestures performed by a driver or other occupants of the vehicle, in particular hand gestures, and can provide corresponding signals characterizing recognized gestures in the form of image data. The gesture recognition unit 5a can be trained or initialized by the driver in advance by increasing the recognition accuracy and reliability in particular by means of a one-time learning process for each gesture. The output unit 5c may in particular output acoustic and / or visual messages or signaling and in particular be implemented by means of a, preferably touch-sensitive, display device (touch screen) including audio function.

The speech recognition unit 5b has one or more microphones and corresponding signal transducers to the recorded by means of the microphones acoustic signals, in particular - if available - voice-based user input of a driver or other occupants of the vehicle 1 to convert to voice data. The processing of the image data and / or the voice data for the purpose of deriving a specification for the automated vehicle guidance of the vehicle 1 In particular, depending on the embodiment, either already in the gesture recognition unit 5a or the speech recognition unit 5b or in the control device 4 respectively.

Furthermore, the vehicle points 1 a vehicle sensor system having at least one sensor, preferably a plurality of different sensors 6 which is set up to sense one or more parameters for describing a current vehicle state and the vehicle environment and, in the form of corresponding sensor data, via a corresponding signal connection of the control device 4 to provide. In addition, the vehicle points 1 a communication device 7 with a radio antenna coupled to it 8th by means of which the vehicle can wirelessly receive messages and other information from an external information source, such as a traffic monitoring system or the Internet, and optionally also to send to them.

The control device 4 is through that in his store 4b stored at least one computer program configured based on a man-machine interface 5 as a voice and / or gesture input made user input and by means of the vehicle sensor 6 and / or via the communication device 7 received vehicle state information and environment information, which is the local environment of the vehicle 1 characterize a specification for the automated vehicle guidance of the vehicle 1 deduce a driving strategy consistent therewith, and control signals to drive the vehicle components 2 for longitudinal control and the vehicle components 3 for lateral control of the vehicle 1 to generate and these control signals for performing an automated vehicle guidance in accordance with the established driving strategy to the vehicle components 2 and 3 issue.

To explain the implemented by means of the computer program method is now on 2 Referenced. In the process, first in one step S1 as previously stated with reference to 1 described at the man-machine interface 5 recorded a user input for the configuration of the automated vehicle guidance. The activated degree of automation of the vehicle guidance corresponds to highly automated driving or even an even higher degree of automation, such as fully automated driving.

In a further step S2 who also simultaneously or before the step S1 can be done by means of vehicle sensors 6 detects the surrounding environment of the vehicle characterizing environment information and / or by means of the communication device 7 received and also the local vehicle environment characterizing information from an off-vehicle data source, such as a traffic control or monitoring system receive and the control device 4 made available.

There will be in one step S3 the user input is evaluated in dependence on the detected environment information by means of a method based on artificial intelligence (AI) in order to derive an instruction expressed by means of the user input as an unambiguous specification for the configuration of the automated vehicle guidance. The use of a method based on artificial intelligence can in particular be advantageously used to be able to process free speech inputs beyond predetermined cataloged voice commands for deriving a corresponding specification.

Can such a unique specification be derived successfully (step S4 - yes), so in a further step S5 a vehicle strategy related to the local vehicle environment for automated vehicle guidance determined, in particular calculated, which is in accordance with this requirement. In other words, it defines one or more planned maneuvers to be executed, which are selected in their combination of a possible solutions to fulfill the specification containing solution space. Its boundary conditions can except by in the step S3 optionally derived by further restrictive criteria, in particular based on the detected current vehicle state and / or permitted operating ranges of the vehicle 1 be determined. Also in this step, in particular AI methods can be used.

Is doing in step S5 an immediately realizable driving strategy successfully determined (step S6 - yes), so be in one step S7 the already mentioned control signals for controlling the automated longitudinal and lateral guidance of the vehicle 1 generated in accordance with this driving strategy and to the corresponding vehicle components 2 and 3 output. The process then loops to the step S1 back to be ready for a new run. Otherwise ( S6 - no) will omit the step S7 to the step S1 returned. Preference is given in addition to the human-machine interface 5 a corresponding signaling output (not shown), the driver or the occupants of the vehicle 1 informed that the instruction made by the user input could not be successfully implemented.

According to an alternative thereto (shown by dashed arrow) can in one step S8 besides such a message at the man-machine interface 5 also requesting another user input in response to it, and in a further step S9 the AI procedure is trained on the basis of the feedback made by the further user input (eg by deep learning) before the procedure for step S1 returns. In this case, the message can in particular provide two or more implementable ways for the approximate implementation of the specification, which as such can not be directly implemented, for selection, and a corresponding selection of the user at the man-machine interface 5 Request. In step S9 then the selection made to train the AI process may be used to step to a same or similar future user input S1 or one in the step S3 derived default on this basis to calculate an actually implementable driving strategy, and so reduce the rate of unsuccessful configuration attempts with respect to the vehicle management over time. In particular, optimizations with regard to the user inputs of a specific user and their peculiarities can thus take place.

The same applies accordingly, if in step S4 it is determined that there is no clear specification for the configuration of the automated driver in the step S3 could be derived (S4 - no). Even then becomes the step S8 branches and in the subsequent step S9 accordingly, the AI procedure to step S3 to evaluate the user input based on the in step S8 recorded feedback trained and thus optimized.

With additional reference to 3 Certain aspects of the method for selected exemplary applications will now be explained. 3 shows a four-lane road (in 3 vertical) with two lanes for each direction of travel coming from a two-lane road (in 3 horizontally) is crossed. On the four-lane road are located in each direction several vehicles. In particular, there is a vehicle according to the invention 1 , about the vehicle 1 according to 1 on the in 3 from bottom to top running direction road. There is another white vehicle in front of this vehicle W , which goes straight, a black further vehicle S which also drives straight ahead and has already reached the intersection, as well as another blue vehicle, which has already entered the intersection B , which has already set to turn into the crossroads to the left and a corresponding direction indicator (eg turn signal) has activated. On the opposite lane come the vehicle 1 a second blue vehicle B as well as a red further vehicle R opposite.

In a first exemplary application, the driver of the vehicle takes 1 due to the left-turning vehicle B and the right of the vehicle 1 located less traveled right lane by voice input "lane change to the right" a user input at the man-machine interface 5 in front. By means of vehicle sensors 6 of the vehicle 1 the vehicle has recognized that next to it on the right side another lane is available. It is an environment information that the local environment of the vehicle 1 characterized and to which the user input of the driver specifically refers. By evaluating the user input, the specification of a lane change is derived to the right (step S3 ) and then considering the relative position and speed of the other vehicles W . B and S calculated in the same direction of travel a concrete driving strategy (step S5 ) based on it corresponding control signals for controlling the vehicle components responsible for the longitudinal and transverse guidance 2 and 3 to generate (step S7 ). The driving strategy can be used as "intermediate goals" in particular an adaptation of the speed of the vehicle 1 and define the timing and the spatial course of the lane change.

In another exemplary application, the left-turn vehicle is B a vehicle that is the driver of the vehicle 1 would like to follow (for example, when traveling by train). An appropriate voice input could be "Follow the blue vehicle in front of us". Additionally, the driver may make a corresponding hand gesture, such as pointing at the vehicle B perform to its user input at the man-machine interface 5 to specify or to provide a corresponding redundancy and amplification of the input. In step S3 On the one hand, the speech and gesture input is analyzed in order to recognize its content, and on the other hand it has to be deduced which of the two in the local environment of the vehicle 1 existing and sensory recorded blue vehicles B that is to be followed. Here, in particular if the voice input is not completely recognized, the redundancy effected by means of the gesture input can be utilized in front of the vehicle 1 moving and just turning left vehicle B to recognize as the right vehicle as well as on the oncoming lane oncoming another blue vehicle B to ignore. Accordingly, based on the default, the now preceding vehicle being identified B to follow, calculated a corresponding driving strategy (step S5 in 2 ), which includes in particular intermediate targets for braking and turning left at the intersection.

In another exemplary application, the driver could voice input to the man-machine interface 5 driving to and stopping the vehicle 1 Request a "next" parking space. In this example, there are both the fee-based parking garage P1 as well as the toll-free parking lot P2 in the local environment of the vehicle 1 and about the same distance away. As part of the step S2 receives the vehicle 1 by means of the communication device 7 and an Internet connection created thereby, surrounding information, including information regarding the location and fees for parking P1 and P2 , Since both parking facilities are located at a similar distance from the vehicle and the parking P1 Covered but subject to charge and parking P2 is not covered but free of charge, the derivation of a clear specification from the detected user input may fail initially for lack of clarity, so in the step S4 to the step S8 is branched, in which the driver in the step S8 will be asked to make a choice between the two parking facilities P1 and P2 based on him at the man-machine interface 5 to display received received environment information. The driver then chooses, for example, the covered parking P1 despite their fee, so may in the step S9 the AI method for evaluating user input is trained in such a way that it considers this selection as a fundamental preference for later evaluations.

Further exemplary use cases or user inputs and corresponding specifications can be, in any case, "for example:" Please turn right at the next intersection "," leave the motorway at the next exit "," please overtake the vehicle before us "," in front of us stop next hotel / ATM "," please less lane change "," please keep less distance ".

While at least one exemplary embodiment has been described above, it should be understood that a large number of variations exist. It should also be understood that the described exemplary embodiments are nonlimiting examples only and are not intended to thereby limit the scope, applicability, or configuration of the devices and methods described herein. Rather, the foregoing description will provide those skilled in the art with guidance for implementing at least one example embodiment, it being understood that various changes in the operation and arrangement of the elements described in an exemplary embodiment may be made without departing from the scope of the appended claims deviated from, and its legal equivalents.

LIST OF REFERENCE NUMBERS

1

vehicle

2

Vehicle components for longitudinal control of the vehicle

3

Vehicle components for the lateral control of the vehicle

4

Control device for vehicle guidance

4a

processor unit

4b

Storage

5

Human-machine interface

5a

Gesture recognition unit

5b

Voice recognition unit

5c

output unit

6

vehicle sensor

7

communication device

8th

antenna

B

blue additional vehicles

R

red another vehicle

S

black further vehicle

W

white further vehicle

P1

fee parking garage

P2

free parking

Claims (12)

Method for configuring an at least highly automated vehicle guidance of a vehicle (1), in particular of a motor vehicle, comprising the following automated steps: Detecting (S1, S2) a user input on a man-machine interface (5) concerning an instruction for configuring the automated vehicle guidance as well as environmental information characterizing the current local environment of the vehicle (1); Deriving (S3) a specification for the automated vehicle guidance of the vehicle (1) by means of an evaluation of the user input as a function of the surroundings information; Determining (S5) a driving strategy related to the local environment of the vehicle (1) in accordance with the specification; and Generating and outputting (S7) of control signals for controlling the vehicle (1) for carrying out the automated vehicle guidance in accordance with the established driving strategy. Method according to Claim 1 wherein the deriving (S3) comprises the specification of: detecting a driving instruction contained in the user input, which relates to the vehicle (1) itself and / or at least one vehicle-external reference object; Checking the surroundings information as to whether an object (B) corresponding to the reference object can be identified in the local environment of the vehicle (1); and if so, deriving the specification to define the driving instruction in relation to the identified object (B). Method according to one of the preceding claims, wherein the user input is freely selectable and is evaluated in the evaluation by means of a KI-based recognition method for their content recognition. Method according to Claim 3 , wherein the AI method comprises a learning method by means of which at least one detected at the man-machine interface (5) (S8) further user input is evaluated and used to adapt a process used to derive the default (S9). Method according to one of the preceding claims, wherein the user input at least partially takes place as an acoustic input and is detected as such. Method according to one of the preceding claims, wherein the user input is at least partially carried out by means of one or more gestures carried out by a user and is detected as such. Method according to one of the preceding claims, wherein the specification relates to influencing a driving strategy for the automated vehicle guidance with regard to one or more of the following aspects or a combination of at least two of these aspects: a vehicle guidance and / or route guidance of the vehicle (1) extending beyond its current local environment; - one or more local driving maneuvers; - One or more dynamic characteristics of the vehicle (1). Method according to one of the preceding claims, wherein the determination (S5) of the driving strategy is carried out dynamically, for which, in addition to the specification, additionally one or more of the following variables or a combination thereof are used as input variables or input variables for the definition of the driving strategy: one or more parameters characterizing a current vehicle state; - traffic control or traffic monitoring information; wherein the input variables or their combination are respectively detected by sensors and / or received by means of a communication device (7). Method according to one of the preceding claims, wherein (S4) in the event that no or no clear result results as a result when evaluating the user input as a function of the surroundings information or the driving strategy resulting from a derived specification is at least temporarily not feasible within the scope of the automated vehicle guidance (S6), via the man-machine interface (5) a message in this regard is issued and / or a re-user input is requested (S8). Device (4 to 7) for configuring an at least highly automated vehicle guidance a self-propelled vehicle (1), in particular a motor vehicle, the device being arranged to carry out the method according to one of the preceding claims. Vehicle (1), in particular motor vehicle, with a device according to Claim 10 , Computer program containing instructions which, when executed on one or more processors (4a) of a device according to Claim 10 they cause the method according to one of Claims 1 to 9 perform.

Images