DE102021101805A1 - DRIVER MONITORING SYSTEM FOR MOTOR VEHICLES - Google Patents

Abstract

A motor vehicle is provided, wherein the motor vehicle is configured to take over, at least partially, automated transverse and/or longitudinal guidance of the motor vehicle, the motor vehicle has a camera system that is configured to acquire image data of a driver of the motor vehicle, and the motor vehicle a driver monitoring system connected to the camera system, which is configured to receive the image data acquired by the camera system and to determine a degree of attention of the driver based on the received image data. The driver monitoring system is designed to determine the level of alertness of the driver in order to determine a position of a face, a direction of gaze, and a position and an orientation of the driver's arms and hands as input data for an artificial neural network based on the image data received from the camera system Inputting input data into the artificial neural network, which is designed to determine the driver's level of attention based on the input data, and to determine the driver's level of attention using the neural network based on the input data.

Description

FIELD OF THE INVENTION The present invention relates to an at least partially automated motor vehicle with a driver monitoring system that is designed to determine the degree of attentiveness of a driver of the motor vehicle. The invention also relates to a training data record for training an artificial neural network, which is designed to determine the level of attention of the driver of the motor vehicle. The invention also relates to a method for training the artificial neural network. The invention also relates to using the training data set to train the artificial neural network.

Conventional driver monitoring systems or devices have sensors installed in the vehicle interior, by means of which data are detected which allow a control unit to determine whether a driver is monitoring the vehicle and/or the driving situation, even if the vehicle is (partially) automated. i.e. the lateral and/or longitudinal guidance is taken over by the vehicle.

From the U.S. 2019/367038 A1 such a driver monitoring apparatus is known, which is provided with a driver monitoring camera that captures an image of a driver's face, a detection unit that acquires safety check action information related to a safety check action of the driver based on the image of the driver's face captured by the driver monitor camera is detected, and is provided with a determination unit that determines whether or not the driver neglects the safety check action based on the security check action information detected by the detection unit.

A development of driver assistance systems for autonomous driving, especially in urban areas, can differ from the way traditional driver assistance systems are developed. Instead of having development and testing up to a certain start of production, after which customers can buy the developed vehicle together with driver assistance systems built into it, autonomous systems will be brought to market in stages, with safety drivers or operators being used in development who have the correct functionality of the vehicle during test drives.

These safety drivers or operators have the task of intervening during test drives, i.e. taking over vehicle guidance manually as soon as automated vehicle guidance does not work as expected or desired. Only after a certain number of test drives in real traffic will a newly developed driver assistance system be brought onto the market, for example via a software update. Then safety drivers and operators return to the vehicles and begin testing another driver assistance system and a test cycle begins again for the other driver assistance system. In this way, autonomous vehicles gradually penetrate into more complex areas and gain more complex maneuvering capabilities over time.

The difference between safety driver and operator is the level of technical detail that the driver is exposed to. The safety driver has access to deeper insights into the underlying autonomous system, e.g. B. the visualization of sensors, detected objects and obstacles as well as the control intention of the vehicle. This allows the safety driver to intervene more quickly than the operator, e.g. B. if a collision-relevant object is not recognized by the autonomous system. The operator does not have to have such a deep understanding and is therefore closer to the actual customer.

The monitoring of such a technical system, especially when the system to be monitored is becoming more and more sophisticated, can become a monotonous task for the safety driver or operator. It is known that there is a high probability that the safety driver or the operator can lose the attention required to monitor the system if the system to be monitored is functioning correctly over a longer period of time. This harbors the risk that the safety driver or operator will intervene too late or not at all in a critical situation that requires the safety driver or operator to intervene, and an accident or collision cannot be avoided. To counteract this, the duration for which the safety driver or operator carries out a test drive can be reduced. The disadvantage is that it cannot be determined whether the security driver or operator is paying the necessary attention to the motor vehicle to be monitored and/or the driving situation. Visual recordings by the safety driver or operator are also conceivable, so that in the event of an accident, possible inattentiveness on the part of the safety driver or operator can be proven and he/she can be held responsible for the possible accident. However, this does not increase safety for the safety driver or operator or for other road users. In addition, these systems can easily be circumvented, e.g. B. by wearing a mask or covering a camera lens.

Against the background of this prior art, the object of the present invention is to specify a device and a method which are suitable for overcoming at least the above-mentioned disadvantages of the prior art.

The object is solved by the features of the independent claims. The dependent claims relate to preferred developments of the invention.

According to this, the object is achieved by a motor vehicle, the motor vehicle being configured in order to automatically take over transverse and/or longitudinal guidance of the motor vehicle at least at times.

The motor vehicle can be a passenger car. Automated driving can take place in such a way that the motor vehicle moves (largely) autonomously. The motor vehicle can be a vehicle with autonomy level 1, i.e. it can have certain driver assistance systems that support the driver in operating the vehicle, such as the adaptive cruise control (ACC). The motor vehicle can be a vehicle with autonomy level 2, i.e. it can be partially automated in such a way that functions such as automatic parking, lane keeping or lateral guidance, general longitudinal guidance, acceleration and/or braking are taken over by driver assistance systems. The motor vehicle can be a vehicle with autonomy level 3, i.e. conditionally automated in such a way that the driver does not have to continuously monitor the vehicle system. The vehicle independently carries out functions such as triggering the turn signal, changing lanes and/or staying in lane. The driver can turn to other things, but if necessary the system will prompt the system to take over the lead within a warning period. The motor vehicle can be a vehicle with autonomy level 4, i.e. so highly automated that the vehicle system is permanently in charge of driving the vehicle. If the system can no longer handle the driving tasks, the driver can be asked to take control. The motor vehicle can be a vehicle with autonomy level 5, i.e. fully automated in such a way that the driver is not required to perform the driving task. No human intervention is required other than setting the target and starting the system. The vehicle can do without a steering wheel and pedals.

The motor vehicle has a camera system, which is designed to acquire image data of a driver of the motor vehicle, and a driver monitoring system connected to the camera system, which is designed to receive the image data acquired by the camera system and, based on the received image data, assign a degree of attention to the driver determine.

It is conceivable that the camera system has a camera directed into the vehicle interior for acquiring or recording the image data. The image data can be continuously acquired from the camera system. The image data can also be referred to as video image data.

It is conceivable that the camera system is divided into two subsystems. Both subsystems can each have a camera directed into the vehicle interior for acquiring the image data. The image data provided by the subsystems can be used by the driver monitoring system, in particular in the sense of a sensor data fusion, to determine the driver's level of attention.

One of the two subsystems can be a so-called "Driver Camera System" (DCS) and the other of the two subsystems can be an interior camera system (IRK). Conventionally, the DCS is used to determine the attentiveness or the degree of attentiveness of the driver of the motor vehicle, which allows longer hands-off times during automated driving in a vehicle with autonomy level 2 (see above). On its own, however, the DCS is limited to tracking the driver's face and eyes. This proposal goes further in that the output of the DCS is only one of the inputs to the proposed driver monitoring system. Additional input can be obtained from the IRK, more specifically from a gesture recognizer of the IRK. Gesture recognition is based on tracking the driver's hands and arms. It can thus be determined whether the driver's hands are on the steering wheel and/or whether the driver is interacting with another system, such as a smartphone and/or a user input system of the vehicle. While gesture recognition is traditionally used exclusively for driver interaction with the motor vehicle, for example for volume control and/or making phone calls, it is proposed here to use the tracking data from the arms and hands of the IRK as input data for the driver monitoring system. The proposed camera system can therefore be designed to provide both video images of the driver's face and his line of sight based on a position and/or orientation of the driver's face and/or eyes and a position and orientation of the driver's arms and hands.

The driver monitoring system is designed to determine a position of a face, a viewing direction, and a position and an orientation of the driver's arms and hands as input data for an artificial neural network based on the image data received from the camera system in order to determine the driver's degree of attentiveness.

The driver monitoring system is further configured to input the input data into the artificial neural network, which is configured to determine the driver's level of attentiveness based on the input data, and to determine the driver's level of attentiveness based on the input data using the neural network.

In particular, it is conceivable that the driver monitoring system processes the data from the DCS and the IRK in essentially real time in an onboard control unit. This real-time processing can constantly determine the driver's attention level. It is conceivable that the driver monitoring system, and in particular its neural network, relies on training data for the same inputs for different levels of attention, which originate from test studies in a driving simulator and/or from real events that led to an accident and/or a critical situation. It can also be ensured in this way that the driver monitoring system is developed further, i.e. the neural network is continuously trained and thus the accuracy of the determination of the driver's level of attention can be improved. It is also conceivable that the driver monitoring system is designed to detect or detect manipulation of the driver monitoring system based on the input data of the camera system, for example because the driver monitoring system determines that facial recognition of the driver is not possible due to the driver covering his face and /or covering of a camera lens of a camera of the camera system leads to a lack of usable image data.

The motor vehicle can have a system connected to the driver monitoring system for detecting a user input, the system for detecting the user input being designed to output information about a detected user input to the driver monitoring system. The driver monitoring system may be configured to further determine the input data for the artificial neural network based on the information received from the system for detecting the user input.

This means that it is particularly conceivable that an interaction between the driver and the motor vehicle is also integrated into the proposed driver monitoring system, e.g. by bus messages being triggered by operating a radio and/or an air conditioning system and being entered into the driver monitoring system. The proposed system can therefore be designed to receive video images of the driver's face and his line of sight from the DCS, the position and orientation of arms and hands from the IRK and the driver's interactions with vehicle operating systems from a vehicle bus, in particular an entertainment Bus to obtain as data for determining input data for the artificial neural Netzt, by means of which the driver's level of attention can be determined.

The driver monitoring system may be configured to output a control signal to the motor vehicle based on the driver's determined level of alertness.

The driver monitoring system can be configured in particular to use the control signal, depending on a state of the motor vehicle, to prevent the motor vehicle from taking over the lateral and/or longitudinal guidance of the motor vehicle, or to end the automated lateral and/or longitudinal guidance of the motor vehicle. In the case of the termination of the automated lateral and/or longitudinal guidance, it is particularly conceivable that the driver monitoring system is designed to control an output device in such a way that it outputs an acoustic, haptic and/or visual warning signal to the driver, the automated lateral and/or or to end longitudinal guidance after the warning signal has been issued and to stop the motor vehicle depending on the driver's reaction to the warning signal which has been issued, or to give the driver control of a transverse and/or longitudinal guidance of the motor vehicle.

Specifically, this can mean that the driver monitoring system is designed to prevent the autonomous system from becoming active if it is in a deactivated state and the driver's attention or level of attention is classified as too low and/or due to missing input data no input data for the artificial neural network can be determined on the driver monitoring system. If a loss of attention or input data to the driver monitoring system is detected by the driver monitoring system while autonomous driving is active, ie when the driver or operator is tired and/or distracted or tampering with the system, the driver monitoring system may be configured to alert the driver through a combination of audible (e.g. beeps), visual (instrument cluster flashing, lights on) and/or haptic (e.g. steering wheel and/or seat vibration) feedback and then either turn on the control handing the driver back or bringing the vehicle to a safe stop before the driver monitoring system eventually disables autonomous driving.

The driver monitoring system can be designed to record the image data received from the camera system and/or the control signal output to the motor vehicle on a particularly crash-proof and/or redundant memory and/or by streaming to a backend.

In other words, inputs and outputs of the system can be recorded on a crash-proof memory, a redundant memory and/or streamed to a backend, for example to be available or to be used for training the artificial neural network.

It is conceivable that an output device connected to the driver monitoring system, which is designed to output information received from the driver monitoring system to the driver, and an input device connected to the driver monitoring system, which is designed to detect an input from the driver and one for input from the driver output corresponding information to the driver monitoring system is provided.

The driver monitoring system can be configured to receive information about the surroundings of the motor vehicle, based on the received information about the surroundings of the motor vehicle to control the output device in such a way that it outputs information to the driver, which prompts the driver to respond to the output information about the surroundings of the motor vehicle to make an input via the input device of the motor vehicle, and based on the information received via the input device, which corresponds to the input of the driver detected by means of the input device, and/or a period of time between the output of the information about the environment of the motor vehicle and the input of the driver is to output a control signal to the motor vehicle.

The output device can have a head-up display and the driver monitoring system can be designed to control the head-up display based on the received information about the surroundings of the motor vehicle in such a way that it outputs the information to the driver, which prompts the driver to Response to the output information about the environment of the motor vehicle to make the input via the input device of the motor vehicle.

This means that the driver monitoring system can be designed to determine the driver's level of alertness in the manner described above and also to actively involve the driver in the driving process and increase his level of alertness, to output information about the driver in terms of a task that one Driver reaction required. It is conceivable that information from the autonomous system is used to formulate questions for the driver about environmental conditions in the motor vehicle environment. For example, the driver monitoring system may be configured to use the output device to display a graphical representation of an upcoming traffic light and ask the driver to confirm that the status of the real traffic light matches that displayed. Another example is that the driver monitoring system uses the output device to display a visualization of an upcoming intersection and asks the driver whether the assumptions made by the motor vehicle about having the right of way at the intersection are correct or not. To minimize distraction, these displays can be in the head-up display. In addition, these questions can also be asked via the vehicle's audio system, i. H. the system can ask through the speakers: "Please press volume + if the upcoming traffic light is green, or volume - if it is not!". The driver must then press the button on the motor vehicle's steering wheel to answer the question. In addition to the driver's inputs and the processing described above, the system can also analyze a reaction time and whether the question was answered correctly or not. It should be noted that the system may be designed to intentionally generate false assumptions about the environment such that the driver cannot always easily confirm the assumptions made by the system.

Furthermore, a method for training an artificial neural network, which is designed to determine a degree of attention of a driver of a motor vehicle based on input data and a predetermined number of degrees of attention.

The method involves acquiring a training data set for the artificial neural network, which has at least one set of input data for each attention level, with each set of input data having one information about a position of a face, a viewing direction, and a position and an orientation of arms and hands of a driver, and training the artificial neural network with the acquired training data set.

What was described above with reference to the motor vehicle also applies analogously to the method and vice versa.

The invention also relates to the use of a training data set for training an artificial neural network, which is designed to determine a degree of attention of a driver of a motor vehicle based on input data and a predetermined number of degrees of attention.

The training data set has at least one set of input data for each level of attentiveness, each set of input data including information about a position of a face, a viewing direction, and a position and an orientation of the driver's arms and hands.

What has been described above with reference to the motor vehicle and the method also applies analogously to the use and vice versa.

Furthermore, a training data set for training an artificial neural network, which is designed to determine a degree of attention of a driver of a motor vehicle based on input data and a predetermined number of degrees of attention, is provided.

The training data set has at least one set of input data for each level of attentiveness, each set of input data including information about a position of a face, a viewing direction, and a position and an orientation of arms and hands of a driver.

What has been described above with reference to the motor vehicle, the method and the use also applies analogously to the training data set and vice versa.

The invention described above therefore proposes a technical monitoring system for a driver, in particular a safety driver or operator, which can continuously determine the driver's level of alertness. For this purpose, technical systems can be used in combination that are already installed in the motor vehicle but were originally intended for other applications. An extension of the proposed system is conceivable, which involves the driver directly in the driving task and thus ensures that the driver is informed about the traffic situation in which the vehicle is located.

• 1 zeigt schematisch ein Fahrerüberwachungssystem für ein Kraftfahrzeug, und
• 2 zeigt schematisch ein Ablaufdiagramm eines Verfahrens zum Trainieren eines künstlichen neuronalen Netzes.

The following is an embodiment related to 1 and 2 described.

• 1 shows schematically a driver monitoring system for a motor vehicle, and
• 2 FIG. 1 schematically shows a flow chart of a method for training an artificial neural network.

This in 1 Schematically illustrated motor vehicle 1 has a driver assistance system 2, which is designed to at least temporarily take over automated transverse and/or longitudinal guidance of motor vehicle 1.

Motor vehicle 1 has a camera system 3, a driver monitoring system 4 connected to camera system 3, a system for detecting a user input or a user input detection system 5, which is connected to driver monitoring system 4, and an output device 6, which is also connected to driver monitoring system 4.

The camera system 3 is designed to acquire image data 31 of a driver (not shown) of the motor vehicle 1 and to output this to the driver monitoring system 4 .

The user input detection system 5 is designed to detect a user input and to output information 51 about the detected user input to the driver monitoring system 4 .

The output device 6 is designed to output information received from the driver monitoring system 4 to the driver.

The driver monitoring system 4 is configured to receive the image data 31 acquired by the camera system 4 and the information 51 about the detected user input from the user input detection system 5 and based on the received image data and the information 51 received from the user input detection system 5 to a degree of attention 42 of the driver determine.

To determine the level of attention 42 of the driver, the driver monitoring system 4 is designed to, in a first step S1, a position of a face, a viewing direction, and a position and an orientation of the driver's arms and hands as input data 41 for to determine an artificial neural network based on the image data 31 received from the camera system and the information 51 received from the user input detection system 5 .

In order to determine the level of attention 42 of the driver, the driver monitoring system 4 is designed to enter the input data 41 determined in the first step S1 into the artificial neural network in a second step S2, which is designed to determine the level of attention of the driver based on the input data 41 , and to determine the level of attention 42 of the driver by means of the neural network based on the input data 41 .

Driver monitoring system 4 is also designed to determine a first control signal 43 in a third step S3 based on driver attention level 42 determined in second step S2 and to output it to motor vehicle 1 , in particular to driver assistance system 2 .

Driver monitoring system 4 is designed to output first control signal 43 depending on the state of motor vehicle 1 . In other words, using first control signal 43, driver monitoring system 4 can, in a state in which driver assistance system 2 is not executing the lateral and/or longitudinal guidance of motor vehicle 1, prevent driver assistance system 2 from performing the lateral and/or longitudinal guidance of the vehicle Motor vehicle 1 takes over. However, in a state of the motor vehicle 1 in which the driver assistance system 2 is currently executing the automated lateral and/or longitudinal guidance of the motor vehicle 1, the driver monitoring system 4 is configured to carry out the automated lateral and/or longitudinal guidance of the motor vehicle 1 by means of the first control signal 43 to end.

In the latter case it is conceivable that the driver monitoring system 4 controls the output device 6 by means of a second control signal 44 determined in the third step S3 such that the output device 6 outputs an acoustic, haptic and/or visual warning signal to the driver. Driver monitoring system 4 can then end the automated lateral and/or longitudinal guidance by means of first control signal 43, and depending on the driver's reaction to the warning signal that is issued, motor vehicle 1 can be stopped or the driver can control a lateral and/or longitudinal guidance of the motor vehicle 1 handed over. The driver's reaction can be detected, for example, by means of the user input detection system 5 connected to the driver monitoring system 4 .

The motor vehicle 1 can also have a memory 7, which is connected to the driver monitoring system 4, the driver monitoring system 4 being designed to store the image data 31 received from the camera system 3, the control signals 43, 44, 45, 46 and/or output to the motor vehicle 1. or to store the information received from the user input detection system 5 on the memory 7. The memory 7 can in particular be a crash-proof and/or redundant memory. It is also conceivable to send this information 31, 43, 44, 51 additionally or alternatively by streaming to a backend from the driver monitoring system 4 and to record it there.

In the present case, the driver monitoring system 4 is also designed to receive information 21 about the surroundings of the motor vehicle 1 from the driver assistance system 2 . Based on the received information 21 about the surroundings of the motor vehicle 1, the driver monitoring system 4 controls the output device 6 by means of a third control signal 45 in such a way that it outputs information to the driver, which prompts the driver to respond to the output information about the surroundings of the motor vehicle 1 to make an input via the input device or the user input detection system 5 of the motor vehicle. Based on the information received via the input device 5, which corresponds to the driver's input detected by means of the input device 5, and/or a period of time that lies between the output of the information about the surroundings of the motor vehicle 1 by means of the output device 6 and the driver's input , the driver monitoring system 4 outputs a fourth control signal 46 to the driver assistance system 2 . In a manner analogous to second control signal 44, fourth control signal 46 can be used to prevent driver assistance system 2 takes over the lateral and/or longitudinal guidance of motor vehicle 1, or an automated lateral and/or longitudinal guidance of motor vehicle 1 that is currently being carried out by means of driver assistance system 2 is terminated.

To output the information to the driver, the output device 6 can have a head-up display (not shown).

The following is related to 2 a method for training the artificial neural network of the driver monitoring system 4 is described, the artificial neural network being configured in the manner described above to determine the degree of attention of the driver of the motor vehicle 1 based on input data 42 and a predetermined number of degrees of attention

As 2 can be seen, the method essentially has two steps S11 and S12.

In the first step S11 of the method, a training data set for the artificial neural network is acquired, which has at least one set of input data for each degree of attention. Each set of input data includes information about a position of a face, a viewing direction, and a position and an orientation of arms and hands of a driver.

In the second step S12 of the method, the artificial neural network is trained with the training data record acquired in the first step S11 of the method.

Reference List

1

motor vehicle

2

driver assistance system

21

Information about an environment of the motor vehicle

3

camera system

31

image data

4

driver monitoring system

41

Input data for the artificial neural network

42

level of attention of the driver

43 - 46

control signals

5

User input detection system or input device

51

Information corresponding to user input

6

dispenser

7

Storage

S1-S3

Steps to determine driver alertness level

S11, S12

Steps to train the artificial neural network

QUOTES INCLUDED IN DESCRIPTION

This list of documents cited by the applicant was generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent Literature Cited

• US 2019367038 A1 [0003]

Claims (10)

Motor vehicle (1), wherein the motor vehicle (1): - is designed to at least temporarily take over automated transverse and/or longitudinal guidance of the motor vehicle (1), - has a camera system (3) which is designed to record image data (31) of a driver of the motor vehicle (1) to acquire, and - a driver monitoring system (4) connected to the camera system (3), which is designed to receive the image data (31) acquired from the camera system (3) and based on the received image data (31) to determine a degree of attentiveness of the driver, characterized in that - the driver monitoring system (4) for determining the degree of attentiveness of the driver is designed to: - a position of a face, a viewing direction, and a position and an orientation of arms and to determine the driver's hands as input data (42) for an artificial neural network based on the image data (31) received from the camera system (3), - the input input data (42) into the artificial neural network, which is designed to determine the driver's level of attention based on the input data (42), and - to determine the driver's level of attention using the neural network based on the input data (42). Motor vehicle (1) after claim 1 , characterized in that - the motor vehicle (1) has a system (5) connected to the driver monitoring system (4) for detecting a user input, wherein the system (5) is designed for detecting the user input, information (51) about a detected user input to the driver monitoring system (4), and - the driver monitoring system (4) is designed to determine the input data (42) for the artificial neural network based on the information (51) received from the system (5) for detecting the user input . Motor vehicle (1) after claim 1 or 2 , characterized in that - the driver monitoring system (4) is designed to output a control signal (43, 44, 45, 46) to the motor vehicle (1) based on the determined degree of attention of the driver. Motor vehicle (1) after claim 3 , characterized in that - the driver monitoring system (4) is designed to use the control signal (43) as a function of a state of the motor vehicle (1): - to prevent the motor vehicle (1) from changing the transverse and/or longitudinal guidance of the motor vehicle (1), or - to terminate automated lateral and/or longitudinal guidance of the motor vehicle (1), in particular to control an output device (6) in such a way that it emits an acoustic, haptic and/or visual warning signal to the driver, to end the automated lateral and/or longitudinal guidance after the warning signal has been issued and to stop the motor vehicle (1) depending on the driver's reaction to the warning signal that has been issued, or to give the driver control over a lateral and/or longitudinal guidance of the motor vehicle (1 ) to hand over. Motor vehicle (1) according to one of Claims 1 until 4 , characterized in that - the driver monitoring system (4) is designed to store the image data (31) received from the camera system (3) and/or the control signal (43, 44, 45, 46) output to the motor vehicle on a memory (7) and/or streamed to a backend. Motor vehicle (1) according to one of Claims 1 until 5 , characterized in that the motor vehicle (1) has: - an output device (6) connected to the driver monitoring system (4), which is designed to output information received from the driver monitoring system (4) to the driver, and - a driver monitoring system ( 4) connected input device (5), which is designed to detect an input from the driver and to output information (51) corresponding to the driver's input to the driver monitoring system (4), - the driver monitoring system (4) being designed to: - to receive information (21) about the surroundings of the motor vehicle (1), - to control the output device (6) based on the received information (21) about the surroundings of the motor vehicle (1) in such a way that it outputs information to the driver , which prompts the driver, in response to the output information about the environment of the motor vehicle (1) an input via the input device (5) de s motor vehicle (1), and - based on the information (51) received via the input device (5), which corresponds to the input of the driver detected by means of the input device (5), and/or a period of time between the output of the Information about the environment of the motor vehicle and the driver's input is to output a control signal (46) to the motor vehicle (1). motor vehicle after claim 6 , characterized in that the output device (6) has a head-up display and the driver monitoring system (4) is designed to control the head-up display based on the received information about the surroundings of the motor vehicle (1) so that this Outputs information to the driver, which prompts the driver to make the input via the input device (5) in response to the output information about the environment of the motor vehicle (1). Method for training an artificial neural network, which is designed to determine a degree of attention of a driver of a motor vehicle (1) based on input data (42) and a predetermined number of degrees of attention, characterized in that the method comprises: - Acquiring a training data set for the artificial neural network, which has at least one set of input data for each degree of attention, each set of input data comprising information about a position of a face, a viewing direction, and a position and an orientation of arms and hands of a driver, and Training the artificial neural network with the acquired training data set. Use of a training data set for training an artificial neural network, which is designed to determine a degree of attention of a driver of a motor vehicle (1) based on input data (42) and a predetermined number of degrees of attention, characterized in that the training data set for each degree of attention at least has a set of input data, each set of input data comprising information about a position of a face, a viewing direction, and a position and an orientation of arms and hands of a driver. Training data set for training an artificial neural network, which is designed to determine a degree of attention of a driver of a motor vehicle (1) based on input data (42) and a predetermined number of degrees of attention, characterized in that the training data set for each degree of attention at least one sentence of input data, each set of input data comprising information about a position of a face, a viewing direction, and a position and an orientation of arms and hands of a driver.

Images