DE102017214542A1 - Method, device, computer program and computer program product for verifying a driver assistance system - Google Patents

Abstract

The driver assistance system comprises a driver model which is designed to characterize a predefined property (P) of a driver of the vehicle as a function of at least one predefined input variable. An input reference record (REF) for the driver model is provided. Furthermore, an extended input data record (IN) is provided, which represents a statistical distribution of the at least one input variable, wherein the statistical distribution is dependent on a detection tolerance and / or determination tolerance of the detection sensor system, which is used to detect and / or determine the at least one input variable. The driver's property (P) is determined depending on the driver model and the provided extended input data set (IN). The determined property (P) is classified as a correctly determined property or a wrongly determined property depending on a reference property corresponding to the input reference data set (REF). Alternatively or additionally, a reaction (R) of the driver assistance system is determined as a function of the determined characteristic (P) of the driver.

Description

The invention relates to a method for verifying a driver assistance system. The invention further relates to a device for verifying a driver assistance system. The invention further relates to a computer program and to a computer program product for verifying a driver assistance system.

Partial and highly automated driving means intermediate steps between assisted driving, in which the driver is supported by numerous, often separate driver assistance systems in the driving task, and autonomous driving, in which the vehicle runs automatically and without the driver.

In semi-automated driving, the driver assistance system takes over both the steering function and the acceleration and braking functions in certain applications, for example, on certain road types or in defined speed ranges. In addition, the driver must continuously monitor the car and be able to take control of the vehicle at all times. An example of such a system is the so-called congestion assistant, which takes over with the help of a distance radar and a camera in halting traffic brakes and acceleration and steering.

In highly automated driving, the driver no longer has to be on permanent alert. The system recognizes its performance limits itself and actively asks the driver to take control of the vehicle in the respective borderline cases. The driver must therefore be able to take over the driving tasks within a certain time.

An essential prerequisite of automated or autonomous driving is that the environment detection by the vehicle is highly reliable. Modern vehicle assistance systems use environment models that allow, for example, a prediction of various traffic situations. The verification of the vehicle assistance systems and in particular of the surrounding models takes place predominantly by driving many kilometers with the active vehicle assistance system in the most diverse traffic environments.

As mentioned above, semi-automated and highly automated driving requires assurance that the driver is able to regain control of the vehicle within a certain time.

For modern driver assistance systems for semi-automated or highly automated driving, it is therefore also important to be able to reliably assess a situational awareness of the driver.

The object on which the invention is based is to provide a method and a device which makes it possible to reliably verify a driver assistance system for partially and / or highly automated driving and to reduce the effort for verification.

The object is solved by the features of the independent claims. Advantageous embodiments are characterized in the subclaims.

The invention is characterized according to a first aspect by a method for verifying a driver assistance system of a vehicle. The invention is characterized according to a second aspect by a device for verifying a driver assistance system of a vehicle. The device is designed to carry out the method or an embodiment of the method. In this case, the driver assistance system comprises a driver model that is designed to characterize a predefined property of a driver of the vehicle as a function of at least one predefined input variable. An input reference data set for the driver model is provided, which comprises at least one reference value for the at least one input parameter of the driver model. Here, the at least one reference value satisfies a predetermined detection accuracy requirement. Furthermore, an extended input data record is provided, which represents a statistical distribution of the at least one input variable, the statistical distribution being dependent on a detection tolerance and / or determination tolerance of the detection sensor system, which is used to detect and / or determine the at least one input variable during operation of the driver assistance system , The driver's characteristic is determined depending on the driver model and the provided extended input data set. The determined property is classified as a correctly determined property or a wrongly determined property depending on a reference property corresponding to the input reference data set. Alternatively or additionally, a reaction of the driver assistance system is determined as a function of the ascertained property of the driver and the determined reaction is classified as a correctly determined reaction or incorrectly determined reaction depending on a reference reaction corresponding to the input reference data record.

Advantageously, the driver assistance system can thus be verified with sufficient reliability.

The method described above allows easy protection of the driver assistance system. The characteristic of the driver and, if necessary, the response of the driver assistance system may be simulated based on the driver model, depending on the input data set.

Securing a driver assistance system that relates to driver characteristics requires that the driver assistance system function be verified for a variety of different drivers, different driving situations, and different driving behaviors of the vehicle.

By generating and providing the extended input data set, it is possible to test the driver model for a plurality of input data values and thus to ensure a verification for a representative coverage of drivers, so that possible errors or deviations in the input variables that are caused by different drivers, not have to be secured by actually driven kilometers with different drivers. The input variables are dependent in particular on characteristics of the respective driver.

Furthermore, a multiplicity of different driving situations and / or driving behavior can be taken into account, so that possible errors or deviations in the input variables, which are caused by driving situations and / or the driving behavior of the vehicle, do not have to be protected by actually driven kilometers.

Furthermore, the simulation on the basis of the extended input data set and the driver model facilitates a repeatability of the verification or a resimulation, in particular after an adaptation of the driver model.

The extended input data set is preferably determined on the basis of the input reference data set and a quality statistic that represents the detection tolerance and / or the detection tolerance of the detection sensor system, for the detection and / or determination of the values for the at least one input variable in the vehicle during operation of the driver assistance system, in particular in a normal operation outside of a test operation, is used.

In an advantageous embodiment according to the first and second aspects, the classification is dependent on a comparison of the determined property with the reference property. Preferably, the determined property is classified as a correctly-determined property if it is equal to the reference property or deviates from the reference property within a predetermined tolerance range, and classified as a wrong-determined property if not equal to the reference property or more than the predetermined tolerance range of the reference property Reference property deviates. This allows a simple decision as to whether the evaluation of the driver's characteristic by the driver assistance system is sufficiently reliable. In particular, it is very easy to determine key figures which enable an assessment of a degree of fulfillment or improvement of the driver model.

In a further advantageous embodiment according to the first and second aspects, the reference property is equal to a determined property in which the input reference data record is applied to the driver model. This allows easy provision of the reference property.

In a further advantageous embodiment according to the first and second aspects, during a real driving operation of the driver assistance system, the values of the input variables are detected with a reference sensor system that has a higher detection accuracy than the detection sensor system provided in the vehicle for the detection of the values. At least a portion of the stored values of the inputs are provided as an input reference data set and / or included in the input reference data set.

In order to determine the characteristic of the driver during a journey with the vehicle, for example a driver's attention or fatigue, a driver's line of sight can be determined by means of an eye tracker. Here, devices and systems are referred to as eye trackers, make the eye through a recording eye movements and allow an analysis of the eye movements. To provide the input reference data set, the viewing direction can be determined on the basis of supplementary camera images which have a higher resolution and / or better illumination and / or mechanical stability and so on. In particular, it is thus also possible to verify the viewing direction and its detection accuracy determined by the eye tracker. Analogously, the determination of the reference values for a position of the driver and / or a head pose of the driver can be used. Optionally, the values of these input variables, which are detected with the provided detection sensor system of the vehicle, can be verified in this way as well.

In a further advantageous embodiment according to the first and second aspects, the values of the at least one input variables for different driving situations and / or different driving behavior of the vehicle and / or recorded different driver behavior. In this case, the driver's behavior particularly includes a driving style of the driver. Advantageously, this makes it possible to provide an input reference data set for a large number of different boundary conditions with regard to driving situations, driver behavior and driving behavior of the vehicle. Preferably, during the actual driving operation, data relating to the respective driving situation and / or the respective driver behavior and / or the respective driving behavior of the vehicle are also recorded and stored. The input reference data are therefore so-called "ground truth data". That is, the input reference data set represents a data record in which it has been determined, in particular by a comparison, that the data acquired during the real driving operation coincide with the determined data of the driver model or of the driver assistance system and / or lead to a similar result. The input reference data thus meet the requirements of the respective driving situation, the respective driving behavior of the vehicle and a real driver behavior. The respective reference values of the at least one input variable of the input reference data set are preferably to be regarded as the respective reference for a specific driving situation with a specific driver behavior, in particular a specific driving style of the driver, and a driving behavior of the vehicle.

In a further advantageous embodiment according to the first and second aspects, the characteristic of the driver is determined as a function of a predefined driving situation and / or a predetermined driving behavior of the vehicle. In particular, information about the respective driving situation and the respective driving behavior of the vehicle can be included in the driver model for determining the characteristics.

Depending on which driver's characteristic is to be evaluated, the assessment may also depend on a driving situation or traffic situation and / or driving behavior. In an attention evaluation, for example, certain areas and / or objects in the vicinity of the vehicle for which the attention is to be assessed can be determined. Alternatively or additionally, a change of a direction of travel in the attention evaluation can be taken into account.

In a further advantageous embodiment according to the first and second aspects, the characteristic of the driver is determined in each case for a multiplicity of predetermined driving situations and / or a multiplicity of predetermined driving behavior of the vehicle.

In a further advantageous embodiment according to the first and second aspects, the characteristic of the driver includes a condition and / or a behavior of the driver.

In a further advantageous embodiment according to the first and second aspects, the property of the driver represents a driver's attention. A reliable evaluation of the driver's attention can be used advantageously for driver assistance systems for partially and highly automated driving. This makes it possible to increase the safety of the driver assistance functions and avoid misuse of the driver assistance functions, since it is always possible to reliably monitor during a journey, in particular during a partially autonomous journey, whether the driver is correspondingly attentive to the situation. A true and proper situational awareness of the driver requires that the driver perceive the objects in the environment, understands their meaning and accurately predicts the changes in the environment and the future condition of the objects for a sufficient amount of time. Such an assessment of situational awareness is therefore often based on an assessment of what the driver is attentive to, and driver intent recognition depending on current behavior of the driver.

In a further advantageous embodiment according to the first and second aspects, the characteristic of the driver represents a tiredness of the driver. The tiredness can be determined depending on a head pose and / or a lid closure frequency.

In a further advantageous embodiment according to the first and second aspects, the driver model comprises a first input which is representative of a driver's line of sight, and / or a second input which is representative of a driver's head pose, and / or a third input variable which is representative of a driver's head position.

The head pose may include a tilt and / or yaw angle of the head. A viewing direction may include a gaze vector and / or nose vector.

According to a third aspect, the invention is characterized by a computer program, wherein the computer program is designed to carry out the method for verifying a driver assistance system or an optional embodiment of the method.

According to a fourth aspect, the invention is characterized by a A computer program product comprising executable program code, wherein the program code, when executed by a data processing device, executes the method for verifying a driver assistance system or an optional embodiment of the method.

In particular, the computer program product comprises a medium which can be read by the data processing device and on which the program code is stored.

• 1 ein Ablaufdiagram zur Verifikation eines Fahrerassistenzsystems.

Embodiments of the invention are explained in more detail below with reference to the schematic drawing. It shows:

• 1 a flow chart for verification of a driver assistance system.

The 1 shows a flowchart of a program for verification of a driver assistance system.

The program can be processed by a device for the verification of a driver assistance system.

For this purpose, the device has in particular a computing unit, a program and data memory and, for example, one or more communication interfaces. The program and data memory and / or the arithmetic unit and / or the communication interfaces can be formed in a structural unit and / or distributed over several structural units.

In particular, the program for verifying a driver assistance system is stored on the program and data memory of the device for this purpose.

The driver assistance system to be verified has at least one driver model which is designed to characterize at least one predetermined characteristic P of the driver depending on predetermined input variables. The driver assistance system is a driver assistance system that evaluates a state and / or a behavior of the driver of the vehicle in order to carry out one or more driver assistance functions. For example, the driver assistance system is designed to execute a driver assistance function depending on the determined driver state and / or the determined driver behavior, for example output a warning and / or actively intervene in a vehicle control, for example to initiate braking.

The driver assistance system has, for example, a driver model for evaluating a driver's attention and / or fatigue. For example, the driver assistance system for controlling a driver assistance function may require information as to whether the driver has been looking at a particular object within the last second. Furthermore, it is also possible to distinguish between, for example, "certainly not seen", "possibly seen" or "with a high probability" instead of a pure yes / no function.

The program can thus be used, for example, to verify whether the driver assistance system performs an attention evaluation and / or fatigue evaluation sufficiently reliably for the driver.

The program will be in one step S1 started in which variables can be initialized if necessary. The program will be in one step S3 continued.

In the step S3 becomes an input reference record REF for the at least one driver model provided, the reference values for the respective input variables of the driver model includes. The providing may include, for example, retrieving the input reference record REF from a database. The input reference data record REF is preferably determined in advance of an actual verification simulation and stored in the database.

In the case of attention evaluation and / or fatigue assessment, the input variables may include, for example, a driver's line of sight and / or a head pose and / or a head position of the driver.

The head pose may include a tilt and / or yaw angle of the head. A viewing direction may include a gaze vector and / or nose vector.

The program will be in one step S5 continued.

In the step S5 an extended input data set IN is provided, which represents a statistical distribution of the at least one input variable, wherein the statistical distribution is dependent on a detection tolerance and / or determination tolerance of the detection sensor system, which is used to detect and / or determine the at least one input variable. The providing may include, for example, retrieving the extended input record IN from the or another database. Preferably, the extended input data set IN is determined in advance of the actual verification simulation and stored in the corresponding database.

The extended input data record IN includes, for example, a multiplicity of one-dimensional input data arrays, wherein the number of elements of the input data array is predetermined by the number of input variables and the respective one-dimensional input data array in each case comprises a value for the respective input variable. In particular, the statistical distribution of the values of the various input variables is mapped with the aid of the multiplicity of input data arrays.

The program will be in one step S7 continued. In the step S7 the driver's property P is determined depending on the driver model and the provided extended input data set IN. In the case of attention evaluation, for example, a field of view area and / or a primary field of vision and / or peripheral field of view of the driver is determined as a function of the input variables specifically mentioned above.

The property P of the driver is in the step S7 For example, also determined depending on a driving situation or a traffic situation and / or a driving behavior of the vehicle.

For example, the attention and / or fatigue of the driver is determined depending on a driving situation. For this purpose, for example, areas and / or objects are determined, to which the driver should focus his attention and which not. The attention evaluation thus makes possible at least a statement with which probability an object to be viewed and / or an area to be considered lie in a field of view area, in particular in a primary field of vision, of the driver.

Depending on the given areas or depending on the objects to be viewed, it takes longer for the human to fully perceive the object. The human perceives an object at the latest after 1 to 10 seconds. The attention evaluation therefore also takes into account, for example, how long and / or how often a driver directs his primary field of view to a particular object or area.

Alternatively or additionally, the driver's attention and / or tiredness is determined as a function of a driving behavior of the vehicle.

Today's assistance systems for detecting the driver's attention include, for example, a camera sensor that determines the head pose, viewing direction and / or position of the driver, for example based on facial features. The camera sensor can use a head model of a driver for this purpose. The head model can hereby be based on specific and anatomical or physiological features such as the corner of the eye, the corner of the mouth, the tip of the nose and their relation to one another. The driver's attention estimation may then be based on a detected head movement within a fixed vehicle model. The problem here, however, is that the driver's attention is often assessed incorrectly due to the high dynamics in traffic.

For determining the input quantities, a camera sensor can be used, which determines the input variables on the basis of various facial features.

The dynamics of the road traffic or the driving behavior of the vehicle is determined, for example, as a function of road data that is representative of a road course in the direction of travel of the vehicle, and / or vehicle data, for example steering angle, yaw rate and so forth.

The attention and / or fatigue is therefore also determined, for example, depending on road data and / or vehicle data.

The program will be in one step S9 or S11 continued. In the step S9 the determined property P is classified as a correctly determined property or false-determined property depending on a reference property corresponding to the input reference data set REF. Preferably, the classification is dependent on a comparison of the determined property P with the reference property. The reference property is in particular equal to a determined property in which the input reference data record REF is applied to the driver model.

In the alternative or additional step S11 a reaction R of the driver assistance system is determined as a function of the determined characteristic P of the driver and the determined reaction R is classified as a correctly determined reaction or incorrectly determined reaction depending on a reference reaction corresponding to the input reference data record REF.

The property P the driver is, for example, determined for at least a portion of the input data arrays of the extended input data set IN and the program is after the step S9 or the step S11 in the step 7 continued until a sufficient number of the input data has been evaluated. After evaluation of the sufficient number of input data, preferably after evaluation of the entire extended input data set IN, the program is in one step S13 completed.

The determination of the input reference data set REF preferably takes place independently of the execution of the program for the verification of the driver assistance system.

In order to provide the input reference data set REF, the values of the input variables are detected and stored, for example, during a real driving operation of the driver assistance system with a reference sensor system that has a higher detection accuracy than the arranged detection sensor system provided in the vehicle for the acquisition of the values. At least a portion of the stored values of the inputs are provided as an input reference data set and / or included in the input reference data set.

Belonging to the input variables of the driver model, for example, the driving situation and / or the driving behavior of the vehicle and the property P of the driver, in particular the driver state and / or the driver behavior, and the reaction R of the driver assistance system are detected and stored. Preferably, these data are detected by means of additional sensors and not taken over by the driver assistance system data. Thus, it can be verified whether the input reference data, when applied to the driver model, lead to the same result that was also detected or observed in real driving.

The input reference data are thus so-called "ground truth data", ie the input reference data record REF thus represents a data record in which the requirements for the respective driving situation, the respective driving behavior of the vehicle and a real behavior of the driver, in particular a driving style of the driver Driver, be met.

The detection accuracy of the values of the input variables of the driver model are verified, for example, depending on the associated stored driving situation and / or the associated stored driving behavior and / or driving style of the driver and the associated stored property of the driver.

In particular, the input reference data record REF for the respective driving situation and / or the respective driving behavior and / or the driving style of the driver may include a partial reference data record.

The determination of the extended input data record IN preferably also takes place independently of the execution of the program for the verification of the driver assistance system.

The extended input data record IN includes, for example, values of the at least one input variable, which represents a statistical distribution of a detection tolerance and / or determination tolerance of the detection sensor system, with which the at least one input variable is detected and / or determined.

The tolerances may be different for different driving situations and / or different driving behavior of the vehicle and / or different driving styles of the driver. The extended input data record IN therefore has, for example, partial data sets for the different driving situations and / or driving behavior of the vehicle and / or driving styles of the driver.

LIST OF REFERENCE NUMBERS

REF

Input reference data set

IN

extended input data set

P

Property of the driver

R

Reaction of the driver assistance system

S1 ... S13

program steps

Claims (14)

A method for verifying a driver assistance system of a vehicle, wherein the driver assistance system has a driver model, which is designed to characterize a predetermined property (P) of a driver of the vehicle depending on at least one predetermined input quantity, wherein the method comprises the steps of: - providing an input reference data set (REF) for the driver model comprising at least one reference value for the at least one input parameter of the driver model, the at least one reference value satisfying a predetermined detection accuracy requirement, - providing an extended input data set (IN) representing a statistical distribution of the at least one input parameter, the statistical distribution is dependent on a detection tolerance and / or determination tolerance of the detection sensor system, which is used to detect and / or determine the at least one input variable in the vehicle during operation d driver assistance system is used, determining the property (P) of the driver as a function of the driver model and the provided extended input data record (IN) and classification of the determined property (P) as a correctly determined property or incorrectly determined property as a function of the input reference data record (FIG. REF) corresponding reference property and / or Determination of a reaction (R) of the driver assistance system as a function of the determined characteristic (P) of the driver and classification of the determined reaction (R) as correctly determined reaction or incorrectly determined reaction depending on a reference reaction corresponding to the input reference data record (REF). Method according to Claim 1 in which the classification takes place as a function of a comparison of the determined property (P) with the reference property. Method according to Claim 1 or 2 in which the reference characteristic is equal to a determined characteristic in which the input reference data set (REF) is applied to the driver model. Method according to one of the preceding claims, wherein, during a real driving operation of the driver assistance system, the values of the at least one input variables are detected with a reference sensor system having a higher detection accuracy than the detection sensors provided in the vehicle for detecting the values of the at least one input variable, and at least a portion of the stored values of the inputs are provided as an input reference data set (REF) and / or included in the input reference data set (REF). Method according to Claim 4 in which the values of the input variables for different driving situations and / or different driving behavior of the vehicle and / or different driver behavior are detected. Method according to one of the preceding claims, in which the characteristic (P) of the driver is determined as a function of a given driving situation and / or a predetermined driving behavior of the vehicle. Method according to one of the preceding claims, in which the characteristic (P) of the driver is determined in each case for a multiplicity of predetermined driving situations and / or a plurality of predetermined driving behavior of the vehicle. Method according to one of the preceding claims, wherein the property (P) of the driver includes a condition and / or a behavior of the driver. Method according to one of the preceding claims, wherein the property (P) of the driver represents a driver's attention. Method according to one of the preceding claims, wherein the characteristic (P) of the driver represents a tiredness of the driver. Method according to one of the preceding claims, wherein the driver model, a first input, which is representative of a line of sight of the driver, and / or a second input, which is representative of a head pose of the driver, and / or a third input, which is representative for a head position of the driver. Device for verifying a driver assistance system, wherein the driver assistance system has a driver model which is designed to characterize a predefined property (P) of the driver as a function of at least one predefined input variable, and the device is designed to implement a method according to one of the Claims 1 to 11 perform. Computer program for verifying a driver assistance system, wherein the computer program is designed, a method according to one of Claims 1 to 11 when executed on a data processing device. Computer program product comprising executable program code, wherein the program code when executed by a data processing device, the method according to one of Claims 1 to 11 performs.

Images