DE102018001970A1 - Method for operating a control system for automated vehicle guidance - Google Patents

Abstract

The invention relates to a method for operating a control system (1) for the automated vehicle guidance of a vehicle (Ego). According to the invention, a driver (FF) of the vehicle (Ego) is in an imminent critical situation during an automated driving operation of the vehicle (Ego) Control system (1) can not handle itself, prompted to take over the vehicle control or to approve a performance of a given maneuver (FM), reaching the critical situation and a takeover period within which the driver (FF) must take over the vehicle or the predetermined driving maneuver (FM) is to be predicted by a situation analysis, wherein a state of the driver (FF) is determined and depending on the determined condition of the driver (FF) a driver response time (Ft) is determined, the determined driver response time (Ft) with the predicted take-over time period and the vehicle (ego) is delayed if the driver response time (Ft) is greater than the predicted takeover period.

Description

The invention relates to a method for operating a control system for automated vehicle guidance according to the features of the preamble of claim 1.

From the prior art, as in the DE 10 2012 112 802 A1 A method for controlling a vehicle is known, which includes a driver assistance system that enables autonomous, partially autonomous and manual driving, an environment detection unit, an evaluation unit evaluating the environmental situation of the vehicle and the surroundings of the environment during the vehicle an autonomous or partially autonomous driving controllable hazard warning device for issuing a warning signal as a takeover request to the driver depending on the assessment of the surrounding situation of the vehicle includes. It is envisaged that by means of a risk estimation unit on the basis of the environment data and vehicle dynamics data during autonomous or semi-autonomous driving the takeover probability is determined with which driver intervention is likely to be required soon. Further, the attentiveness level of the driver is estimated by means of an attention estimating unit, and finally, from the assumption probability depending on the attentiveness level of the driver, a period of time until the generation of the warning signal is determined.

• - das Fahrzeug nähert sich einer hinsichtlich der Streckenführung kritischen Stelle oder befindet sich an einer solchen Stelle,
• - das Fahrzeug nähert sich einer Stelle mit Verkehrsstörungen oder befindet sich an einer solchen Stelle,
• - das Fahrzeug nähert sich einer Stelle mit Sichtbeeinträchtigungen oder befindet sich an einer solchen Stelle,
• - das Fahrzeug wird durch die Abstands- und Geschwindigkeitsregelfunktion beschleunigt,
• - es liegt eine Anomalie des Verkehrsflusses in der Fahrzeugumgebung vor.

In the DE 10 2015 014 139 A1 For example, a method for operating a distance and speed control function of an autonomously or highly automated vehicle and a driver assistance system for carrying out the method are described. The procedure shall initiate at least one safety-enhancing action if the driver is inattentive and if at least one of the following conditions is met:

• the vehicle approaches or is in a position critical to the route,
• - the vehicle approaches or is in a position with traffic congestion,
• - the vehicle is approaching or in a place with visual impairments,
• the vehicle is accelerated by the distance and cruise control function,
• - there is an anomaly of the traffic flow in the vehicle environment.

The invention is based on the object to provide a comparison with the prior art improved method for operating a control system for automated vehicle guidance.

The object is achieved by a method for operating a control system for automated vehicle guidance with the features of claim 1.

Advantageous embodiments of the invention are the subject of the dependent claims.

In a method according to the invention for operating a control system for the automated, in particular highly automated, vehicle control of a vehicle during an automated, in particular highly automated, driving operation of the vehicle in an imminent critical situation, which the control system can not cope automatically, especially there is not designed for their automatic management, prompted to take over the vehicle management or to approve a performance of a predetermined, in particular prescribed by the control system, driving maneuver, reaching the critical situation and a takeover period, within which the driver, in particular to defuse the critical situation that must take on vehicle guidance or approve the given driving maneuver, is predicated by a situation analysis, wherein a state of the vehicle driver determined is detected, in particular, and depending on the determined, in particular recognized, the driver's reaction time is determined, wherein the determined driver reaction time is compared with the predicted takeover time and the vehicle is delayed, in particular, is delayed in advance, when the driver response time is greater than the predicted takeover duration. This delay, in particular anticipatory deceleration, of the vehicle achieves adaptive ride adaptation.

The term automated, in particular highly automated, vehicle guidance or automated, in particular highly automated, driving operation, is understood in particular to mean highly automated driving according to SAE Level 3. Here, the control system monitors its functional limits and gives the vehicle guidance to the driver when the functional limits are reached. Ancillary activities, in particular limited and / or predetermined ancillary activities of the driver are allowed. In particular, the driver can give the responsibility for highly automated driving temporarily to the vehicle and to devote other non-driving tasks.

The predetermined or specific driving maneuver is, in particular, a driving maneuver which is not secured by the system, ie by the control system, for example a lane change.

By decelerating the vehicle is achieved that the critical situation occurs at a later date. This later entry results in an increase in the takeover time remaining to the driver. This deceleration of the vehicle can thus be ensured that the driver of the vehicle in good time, d. H. within its driver reaction time, can take over or the execution of the given maneuver in time, d. H. within its driver response time, may approve.

The method according to the invention thus makes possible an adaptive adaptation of the highly automated driving operation taking into account in particular continuously determined system limits of the control system, the condition of the vehicle driver and a respective situation criticality. Thus, the method allows for individual cognitive assistance with empathic, d. H. understanding and attentive, intelligence, adapted to the driver's reaction time of the driver.

In particular, the vehicle is delayed as long as the vehicle driver has not followed the request to take over the vehicle guidance or to approve the driving maneuver and / or as long as the critical situation exists. If the deceleration sufficiently reduces the criticality of the situation and / or the situation changes during the deceleration, further deceleration is no longer necessary and, for example, the automated driving mode can then be continued. If the vehicle driver has followed the request to take over the vehicle guidance or to approve the maneuver, the driver will continue to drive the vehicle or the approved driving maneuver will be carried out on the system side.

The situation analysis is performed, for example, based on a sensory environment observation of an environment of the vehicle. By means of such an environmental detection by means of an environmental detection sensor system, a respective critical situation can be detected and its criticality can be determined.

For example, driving maneuvers of other, in particular surrounding, road users are predicted and taken into account for the situation analysis, and / or, for example, limits of the control system and / or uncertainties with regard to the limits of the control system are taken into account. A critical situation which can not be mastered by the control system is present, for example, if the situation analysis, in particular taking into account predicted driving maneuvers of the surrounding road users, shows that the vehicle must, at least with high probability, make a lane change, but a secured automatic lane change is not possible is because other vehicles approaching on a neighboring lane may not be timely detected due to system limitations of the control system, in particular due to a limited range for observing the sensors of the vehicle approaching other vehicles. Uncertainties of the system boundaries, which are based for example on measurement errors of the sensors used, are advantageously taken into account in the situation analysis.

The system boundaries or the critical situation are also achieved, for example, when the situation requires a regulation to a speed or acceleration that is outside a predetermined range of values defined as permissible, or if a complex situation, for example a construction site, a roundabout, a confusing intersection and / or a congestion and / or other complex situations for which the control system is not designed to cope.
The condition of the vehicle driver is determined, for example, by a driver's observation, in particular detected. Such a driver's observation can for example be done by at least one camera of the vehicle. As a result, a multiplicity of different states can be reliably detected and a corresponding driver response time can be determined.

In particular, a minimum period of time is determined as the driver response time that the vehicle driver is expected to require due to his condition for taking over the vehicle guidance or for approving the given driving maneuver. This enables a secure takeover of the vehicle guidance or an informed decision regarding the approval.

When determining the vehicle driver reaction time, for example, a respective secondary activity and / or at least one vital parameter of the vehicle driver and / or at least one preceding measurement or several such measurements of the driver's reaction time of this vehicle driver, in particular during one or more previous journeys, are taken into account. This allows the most accurate possible determination of the driver response time.

The request to take over the vehicle guidance or to approve the execution of the predefined driving maneuver takes place, in particular, by means of a voice output, for example by a voice output in the sense of "Take over the vehicle guidance!". The voice output is advantageously, in particular in time-critical situations in which the takeover time is shorter than or equal to the same time as the driver reaction time, initiated by an audible warning signal and / or optical signal. The audible warning signal is output, for example, in the form of a single or multiple beeps. The visual warning signal is output, for example, as a pictogram showing a steering wheel with two hands enclosing the steering wheel. The output of the icon is preferably on a display on the vehicle cockpit and on a display of a device on which the driver performs a secondary activity. By the acoustic / visual warning signal followed by the voice output, a quick and intuitive detection of the respective request by the driver is ensured in urgent cases. The driver can respond or respond, in particular granting or not granting approval. Thus, an action request or a voice dialogue between the control system and driver is performed.

The predefined driving maneuver, for the implementation of which it requires the approval of the driver, for example, a system side, d. H. in particular by the control system, unsecured maneuver, in particular a maneuver that can not be performed safely due to a limited sensor range of the vehicle's environment detection sensors. Such a maneuver is, for example, a lane change on a highway when approaching a slow other vehicle or at a stationary obstacle or when approaching a lane narrowing or at a lane end. In such cases, the driver, advantageously by the combination of audible / visual warning and voice dialogue, is asked to approve the execution of the driving maneuver. The driving maneuver is thus secured by the approval of the driver.

Embodiments of the invention are explained in more detail below with reference to drawings.

• 1 schematisch einen Ablauf einer Durchführung eines autonomen Spurwechsels mit Genehmigung durch einen Fahrzeugführer,
• 2 schematisch eine adaptive hochautomatisierte Fahrzeugführung,
• 3 schematisch ein Diagramm mit kumulierten Reaktionszeiten einer Probandengruppe bei verschiedenen Nebentätigkeiten,
• 4 schematisch eine Verkehrssituation während einer adaptiven hochautomatisierte Fahrzeugführung, und
• 5 schematisch eine weitere Verkehrssituation während einer adaptiven hochautomatisierte Fahrzeugführung.

Showing:

• 1 schematically a process of performing an autonomous lane change with the permission of a driver,
• 2 schematically an adaptive highly automated vehicle guidance,
• 3 a schematic diagram with cumulated reaction times of a group of subjects in different secondary activities,
• 4 schematically a traffic situation during an adaptive highly automated vehicle management, and
• 5 schematically another traffic situation during an adaptive highly automated vehicle management.

Corresponding parts are provided in all figures with the same reference numerals.
Based on 1 to 5 The following is a method for operating a control system 1 for highly automated vehicle guidance described in which a driver FF of a vehicle ego during an automated driving operation in an imminent critical situation, to automatically handle the control system 1 is not designed, prompted to take over the vehicle or a performance of a particular, not guaranteed in particular system side, maneuver FM To approve, for example, the implementation of a lane change. The respective driving maneuver FM is thereby the rule system 1 determined, ie predetermined.

In the method, reaching the critical situation and a takeover period, within which the driver FF to defuse the critical situation, the vehicle must take control or the specific driving maneuver FM must be approved by a situation analysis, in particular based on a sensory environment observation. Furthermore, a state of the vehicle driver FF , in particular by a driver observation, detected and depending on this becomes a driver response time Ft determined, which represents in particular a minimum period of time, the driver FF due to its condition for taking over the vehicle management or for the approval G the particular driving maneuver FM is expected to need. The determined driver response time Ft is compared to the predicted takeover duration and the vehicle ego is delayed, in particular anticipatory, when the driver response time Ft is greater than the takeover period, in particular if and as long as the driver FF the call AU to take over the vehicle guidance or for approval G of the driving maneuver FM not followed. Due to this delay, especially forward-looking Delay of the vehicle ego Thus, an adaptive ride adjustment is achieved.

For example, if there is a situation where the driver is FF for example, ten seconds of driver response time due to its condition Ft for the assumption of the driving task, ie the vehicle guidance required, but he must take over within eight seconds to avoid the occurrence of the critical situation, in particular to defuse a dangerous situation, then the anticipatory delay has the consequence that the control system 1 not automatically manageable critical situation occurs later. The later entry has an enlargement of the driver FF remaining takeover period result. If the vehicle ego delayed in the present example, that the takeover period of eight seconds increases to ten seconds or more, then by the anticipatory deceleration of the vehicle ego made sure that the driver FF the driving task, ie the vehicle guidance, on time, ie within its driver response time Ft , can take over.

An unmanageable critical situation is present, for example, if the situation analysis, in particular taking into account predicted driving maneuvers surrounding road users, that the vehicle ego , in particular with high probability, must perform a lane change, but a secured automatic lane change is not possible because on a neighboring lane approaching other vehicles AF due to system limitations, in particular due to a limited range of for observing the approaching other vehicles AF used sensors, may not be detected in time. Uncertainties of the system boundaries, which are based for example on measurement errors of the sensors used, are advantageously taken into account in the situation analysis.

The system boundaries or the critical situation are also achieved, for example, when the situation requires a regulation to a speed or acceleration that is outside a predetermined range of values defined as permissible, or if a complex situation, for example a construction site, a roundabout, a confusing intersection , a congestion and / or other complex situation exists, for coping with the regulatory system 1 not designed.

The driver response time Ft in particular depending on secondary activities and / or vital parameters of the driver FF and / or determined from measurements during previous trips.

The calls AU , in particular takeover requests, are issued in particular as voice issues. Speech outputs are preceded by an audible and / or visual warning signal, especially in time-critical situations where the driver must react quickly. The output of this acoustic or visual warning signal is advantageously dispensed with in non-critical situations. The acoustic signal is output, for example, as a single or multiple beep. The audible warning signal is output, for example, as a pictogram on a display of the vehicle and / or a display of a device on which the driver performs a secondary activity, wherein the pictogram shows, for example, a steering wheel that is encompassed by two red hands.

In one possible embodiment, this is the specific driving maneuver FM which it is authorized to carry out G of the driver FF requires a system unsafe driving maneuver FM ie a driving maneuver FM , which can not be performed safely due to the limited sensor range. Such a maneuver FM is, for example, a lane change on a highway when approaching a slow other vehicle AF or at a stationary obstacle or when approaching a lane narrowing or at a lane end. In such cases, the driver becomes FF advantageously by a voice dialogue SD asked to carry out the driving maneuver FM to approve. The driving maneuver FM is thus approved G of the driver FF hedged.

The method thus enables an adaptive adaptation of the highly automated driving operation, taking into account in particular continuously determined system limits of the control system 1 , the condition of the driver FF and a respective situation criticality. Under Highly Automated Driving (HAF) - Conditionally Automated Drive 3 is understood in particular that the control system 1 monitors its functional limits and to the driver FF when these functional limits are reached. Limited defined ancillary activities of the driver FF are allowed.

In particular, the method represents an individual cognitive assistance with empathic, ie understanding and attentive, intelligence. An identification of the respective vehicle driver FF can for example be based on his personal profile, also referred to as "shared mobility" profile. This profile can for example be referred to by a central server, in particular of the vehicle manufacturer, as a cloud platform, and / or from a personal mobile terminal, for example, a mobile phone, the driver FF and / or from the shared mobility card.

In particular, by the above-mentioned speech dialogue SD between the driver FF and the rule system 1 to take over the vehicle management and in particular for approval G the determined, ie predetermined, driving maneuver FM By means of the method an empathic assistance system is realized. In particular, the application of the speech dialogue SD Missing, especially non-time critical, information supplement and empathic impression of this assistance function with the driver FF produce. In 1 this is exemplified by an autonomous lane change, with missing information due to a too short sensor range to the rear by a voice input of the driver FF be compensated.

1 shows a sequence of the method on the example of the lane change along a timeline t , The control system 1 recognizes a specific maneuver to be performed as required FM the lane change. Then the request is made as a voice output AU for approval G the implementation of this particular, ie of the control system 1 predetermined, driving maneuvers FM For example, a voice output that a lane change is required, associated with the question of whether an area is left behind the vehicle ego is free, ie if there is no other vehicle AF approaches.

Shown in 1 are the two possible reactions of the driver FF , For example, a rejection occurs AL the permit G by the driver FF ie he does not approve the lane change, for example because he is on the other lane FS to which to change, an approaching other vehicle AF recognizes. This rejection AL takes the form of a voice response from the driver FF ie between the control system 1 and the driver FF a dialogue takes place. For example, the driver replies FF with "no, wait, not free" or something similar. As the driver FF approval G has not granted, the lane change is not performed. Alternatively, the driver recognizes FF that no other vehicle AF approaching, and granted the permit G , again in the form of a voice response, for example, "Los, links is free" or similar, wherein the beginning of this voice response may additionally contain a predetermined formulation by which the control system 1 recognizes that the driver FF just with the control system 1 communicated.

There is now a translation Ü this voice response of the driver FF in one for the rule system 1 usable form and an evaluation A this language answer to complete the situation analysis. This results in the result e that the driver FF approval G to carry out the particular maneuver, here the lane change granted. The result of this communication K between the control system 1 and the driver FF is the driving maneuver execution FD this particular driving maneuver FM , here the lane change, by the control system 1 ,

An alternative to this speech dialogue SD For example, would be a requested lane change, activated for example by the driver FF by turning on a respective direction indicator of the vehicle ego , By means of the speech dialogue SD In contrast to this alternative, it creates an empathic impression of a personalized intelligent driver assistance system. This will be the driver FF actively held "in the loop", ie he is involved in the driving event, if his intervention and / or his help and / or his input, ie, for example, his information, and / or his decision is needed. This increases his attention. Furthermore, it is easier for the driver in time-critical situations FF to undertake the driving task or to provide assistance and / or information if the control system 1 is close to its functional limit, for example due to lack of sensor technology to the rear or because of limited range or temporarily increased uncertainty of the sensor.

To develop dialogues between the driver FF and the rule system 1 or the vehicle ego For example, already existing speech dialogue systems, in particular from the computer and mobile telephone sector, can be used. These known systems process the information from such speech dialogues SD especially cloud-based over the internet. For vehicle applications which require the predictive situation evaluation and driver state while driving, a local implementation is more advantageous, ie the algorithms, in particular speech recognition algorithms, locally in the vehicle ego directly to ensure the integrity and availability of the connection.

2 shows the adaptive adaptation of the highly automated driving operation that can be realized by the method with continuous determination of the system limits of an autonomous drive and comparison with the context of driver status and situation criticality.

The adaptive request AU to take over the vehicle guidance or for approval G of the particular, ie of the control system 1 predetermined, driving maneuvers FM during a partially autonomous or highly automated journey, depending on the determined system boundaries and in the context of the criticality of a respective traffic situation and taking into account the current driver response time Ft of the driver FF adapted and executed. Between the call AU and the takeover by the driver FF or the permit G the particular driving maneuver FM by the driver FF fits the control system 1 a longitudinal and transverse control of the vehicle ego such that a critical situation is defused and, for example, a takeover by the driver FF or the specific driving maneuver FM is no longer required or this takeover of vehicle guidance or approval G the particular driving maneuver FM by the driver FF is possible in time.

The realization consists of a combination of several process steps S1 to S7 , as in 2 shown, the method steps S1 to S7 for example, successively or more than one method step S1 to S7 can take place simultaneously. Ie. the sequence of method steps described below S1 to S7 does not have the order of execution of these procedural steps S1 to S7 correspond.

A first process step S1 is a determination of the condition of the driver FF , Depending on the condition of the driver FF then becomes its driver response time Ft determined. For example, there may be driver response times Ft from one second to seven seconds.

input variables ES1 for this first process step S1 are, for example, a type of a respective secondary activity of the vehicle driver FF , a cognitive condition of the driver FF and / or a vital state of the vehicle driver FF to determine one or more vital parameters of the driver FF be determined. For example, in the first process step S1 Driver response times Ft measured and / or detected, for example by means of reaction times such as in computer games, by means of Internet programs and / or by contact-free detection in particular, for example by means of cameras and / or in particular non-contact sensors. This is advantageously with the vital condition of the driver FF combined.

This determination of the vital state and the one or more vital parameters required for this purpose can be carried out, for example, as in US Pat DE 10 2011 110 486 A1 whose entire contents are hereby incorporated by reference. Here a non-contact, camera-based measurement of vital parameters for driver assistance and vehicle guidance systems, determined from only one signal of the facial area, is described. Alternatively or additionally, the condition of the vehicle driver FF For example, be determined by means of a non-contact pulse measurement, especially for driver assistance and vehicle guidance systems, such as in the DE 10 2011 016 772 A1 whose entire contents are hereby incorporated by reference. The determined condition of the vehicle driver FF , in particular its vital state, is advantageously in the determination of the driver response time Ft considered.

The driver response time Ft of the driver FF depends in particular on the type of secondary activity, depending on the time of day, ie for example in the morning, in the evening, at night, during working hours, more relaxed on the way to / from a holiday destination or stressed in a traffic jam, strained because of too many appointments in a calendar, depending on the Vitality and / or a state of health, depending on fatigue and / or depending on a number of people in the vehicle and their age, ie whether they are young or old and whether, for example, a noise level is loud or quiet.

The current driver response times Ft For example, they may be measured just prior to departure to determine how to classify a vehicle driver response for the particular ride. For example, there are websites that measure reaction times with full attention and focus on the driver's task. Typical reaction times, based on average measurements of 2000 test persons in old age 18 to 90 Years, for example, are 300 ms to 1100 ms. This shows an increase in reaction times with increasing age.

This increase is essentially linear or in a slightly upwardly curved curve. For example, an average reaction time, with full attention and focus on the driver's task, is 429 ms at the age of 35 years and 1011 ms at the age of 85 years.

Alternatively or additionally, measured driver response times Ft from previous trips to determine the current driver response time Ft be used.

As driver response times Ft of the driver FF in particular depending on the nature of the secondary activity, for example, driver response times Ft used which were determined in volunteer trials, especially in driving simulator studies, for example, with more than 400 subjects. Examples of possible sideline activities include reading news and / or books on a tablet computer, watching movies, looking for something in your own pocket.
In such driving simulator studies, for example, it was determined by experiments that the driver response time Ft varies between one second and two seconds as a function of the non-driving activity performed outside. The determination of these driver reaction times Ft is advantageously based on own driving simulator experiments.

For example, driver response times have been Ft determined from a driving simulator experiment with 96 subjects, with a takeover of highly automated driving in the semi-automated driving was examined in a missing lane marking in a left turn with crosswind from the left, in the case of non-intervention of the driver FF a drift to the right of the roadway takes place. Driver response times Ft Here are glances from the tablet computer with the corresponding side tasks on the road (measured with the headmounted eye tracker system Dikablis). A control group had no secondary tasks. A search for an item called a "treasure hunt" was done in a passenger seat in a box, ie, it required turning the subject to the right, often with both hands, but that was up to the test subjects.

This resulted in the in 3 illustrated courses of the cumulative reaction times, wherein on the abscissa axis the driver response time Ft and on the ordinate axis the percentage PA the subject is removed. Shown are the progressions of the cumulative reaction times for reading VL , for movie watching VF , for the treasure hunt mentioned above VS and a course of cumulative reaction times for the control group VK , The treasure hunt should in particular represent the case that one searches in the pocket in the passenger seat a ringing mobile phone.

From this a limit point can be determined. The cut-off point calculates the number of more than 95% of the subjects who had the eye-to-eye shift from the tablet to the street within 1.26 seconds during movie viewing (4.57% of the subjects were slower) ). For reading, and especially the treasure hunt, this number of subjects who made the change of vision within 1.26 seconds was less. However, this applies to the control group.

The control system 1 Fits autonomous or highly automated driving, depending on the estimated uncertainty of the system boundaries, on the criticality of the situation as well as on the driver reaction times Ft on. With a driver FF his looking at the road is the driver response time Ft For example, one to two seconds, he has his hands on the steering wheel, it is for example one to three seconds, with a distracted behavior, for example, something in a footwell of the vehicle ego For example, it is two to seven seconds. Takeover periods, particularly from autonomous or highly automated driving to manual driving, are important for comparison with the criticality of the situation prediction and the subsequent adaptive adaptation of the highly automated driving with adaptation of the speed.

The vehicle ego advantageously determines continuously during the journey, for example by means of sensors, the probable driver response time Ft of the driver FF that is typical of the particular sub-task being performed, and takes into account, for example, an expected extended driver response time Ft in the prompting process, for example by a timely request AU and / or by delaying the vehicle ego after the request AU , The following values of the driver reaction times Ft are based on the results of the driving simulator test.

If a steering torque of zero Nm is registered during highly automated driving, it means that no power transmission of the driver FF to the steering wheel, ie no hand on the steering wheel. Ie. an additional driver response time Ft must be included in the takeover period. If the sensors, for example a driver observation camera, are an upright driver FF detected with a head direction directed straight ahead, the additional driver response time is Ft for example, about one second. Alternatively, the individual driver response times Ft be taken into account, which were determined in the manner described above directly before the start of the journey, for example via a website.

However, if a body posture deviating from the straight alignment, for example by sensors in the seat and / or in the seat back and / or head direction, for example by means of the driver observation camera, is detected, as occurs, for example, when searching in the footwell or on a back seat an additional driver response time Ft of several seconds, for example from about five seconds to seven seconds.

If a steering torque of two to four Nm occurs, it means that the driver FF exerts a measurable force on the steering wheel, ie the additional reaction component for the movement of the hand or hands to the steering wheel is eliminated. An exception is an accidental or intentional operation of the steering wheel with the knee, in addition, if the posture and / or head posture is rotated away sideways, for example, by sensors or otherwise detectable. Here is the additional driver response time Ft as described above with a steering torque of zero Nm.

Alternatively or additionally, a cognitive state and / or vital state of the vehicle driver FF determinable, for example by means of a speech analysis of the intonation and / or the sound of conversations with other vehicle occupants, based on a mood and an emotional state of the vehicle driver FF while driving, as well as based on an intensity of communication with the other vehicle occupants or on the mobile phone or a distraction or engagement with other electronic devices, and at the driver response time Ft to take into account.

In a second process step S2 a determination of critical events from the situation analysis takes place, for example as in DE 10 2012 020 901 A1 whose entire contents are hereby incorporated by reference. In it is a use of apps (including vital data, occupant observation, environment detection and situation analysis) from one / more in the vehicle ego docked smartphones of the driver / vehicle occupant described for driver assistance and vehicle guidance systems. In this second process step S2 Thus, the situation analysis of the current driving condition is performed. From this, the criticality and a prediction, for example over the following one to five seconds, of the driving maneuvers of the surrounding other vehicles AF determined, for example, as in the DE 10 2014 003 343 A1 (Determination of a lane change requirement), the full contents of which are hereby incorporated by reference. input variables ES2 for this second process step S2 are for example fused data of the vehicle ego and the surrounding other vehicles AF ,

In a third process step S3 a monitoring and / or control of the system uncertainties of the control system takes place 1 , This results in a trend of approaching the system limits of the control system 1 determined. The uncertainties of the system boundaries are taken into account, for example, based on the error distribution of the sensors in the context of the observed circumstances of the driving situation. input variables ES3 for this third process step S3 are, for example, sensor data, in particular an environment detection sensor system of the vehicle ego , with a self-diagnostics and uncertainties of the system boundaries.

The analysis of the uncertainties is carried out, for example, using artificial intelligence (AI) techniques, for example Bayesian networks and / or alternative probabilistic techniques that combine the range, error distribution and / or uncertainties of various sensors used for data collection. The size of the uncertainties determines the degree of reliability of available sensor data and the conclusions about the system reliability determined from it.

The sensor diagnostics, in particular the self-diagnostics, and / or the calculation of the error quantities and uncertainties is based on various error propagation algorithms, on statistical methods and / or on methods of system identification.

Then, as already described, an analysis is used to determine the trend in the approach of the uncertainties to the system boundaries. This trend approach to the system boundaries is determined, for example, with the use of Bayesian networks (BN) and / or alternative AI methods. The algorithms use the input features, i. H. especially the input characteristics, of the observed system. These features and their uncertainties are combined by means of Bayesian networks (BN) and / or alternative KI methods and used to analyze the trend of the uncertainties of the system boundaries. The analysis is performed in the context of the continuously determined system boundaries, the driving condition, the driver status, the reaction times and / or the situation criticality.

When a threshold of system boundaries is reached, the corresponding probabilistic states become input, ie input parameters, for decision making to the control system 1 for the adaptive adaptation of highly automated driving.

In a fourth process step S4 a decision is made taking into account the contextual information of the situation analysis and the criticality situation of the current driving situation. There is a comparison of the driver reaction times Ft with a time horizon of the driving maneuver forecast and with the uncertainties of the system boundaries. As context information thus the driver response times Ft and the situation criticality as well as the approach to the system boundaries taken into account. The decision-making in this Context is performed, for example, with Bayesian networks. The probabilistic evaluation becomes after exceeding of a limit value to the control system 1 forwarded.

The takeover periods, in particular from autonomous or semi-automated to manual driving, are relevant for the comparison with the criticality of the situation prediction and relevant for the subsequent adaptive adaptation of highly automated driving, in particular adaptation of the speed and steering of the autonomous or highly automated moving vehicle ego ,

In a fifth process step S5 the adaptive adaptation of highly automated driving takes place. In particular, if the time horizon of the driving maneuver forecast of surrounding other vehicles AF is longer than the driver response time Ft of the driver FF can be achieved by the adaptive adaptation of highly automated driving, in particular by decelerating the vehicle ego , ie by adjusting the driving speed and driving continuation in highly automated driving depending on the first step S1 determined condition of the driver FF For example, a gain of one second to three seconds can be achieved by the driver FF has more time to take over the vehicle guidance or the specific, ie through the control system 1 predetermined, driving maneuvers FM to approve. Ie. it is done as by the connecting arrow P between the first method step S1 and the fifth process step S5 indicated, the adjustment of the vehicle speed and the ride continuation in highly automated driving depending on the uncertainties regarding the highly automated continuation of the ride. By the delay of the vehicle ego and continuation of the highly automated driving operation can be achieved that the predicted takeover period is at least as long as the driver response time Ft or longer than the driver response time Ft ,

In a sixth process step S6 a timely request is made AU to the driver FF , In particular by voice output to take over the vehicle or the implementation of the given driving maneuver FM to be approved, ie it will be timely information, especially in the language dialogue SD , from the driver FF regarding the feasibility of the given driving maneuver FM called or the transfer request is made to the driver FF , This is done depending on the situation analysis and the probabilities of the trend approximation to the system boundaries. The speech dialogue SD thus serves to the control system 1 missing information for the execution of the given driving maneuver FM by the driver FF to make available or to induce this to take over the vehicle guidance, with the voice dialogue SD an emphatic impression of this driver assistance with the driver FF is produced.

In a seventh process step S7 leads the rule system 1 in the absence of reaction of the driver FF the vehicle ego For example, safely to a roadside, if the environment detection and free space determination allows. If the free space determination is not reliable, then the vehicle is slow, for example, with 2 - 3 m / s ² in the current lane FS braked to a standstill, so that the following vehicles have enough reaction time.

Furthermore, for example, in particular by the control system 1 Help can be requested and / or it can, especially if the driver FF is again reactive, will be offered a support for a safe onward journey. If the condition of the driver FF is sufficiently good, for example, highly automated, the vehicle ego be moved back onto the road and be safely threaded into the traffic and driven on.

The following is a detailed description of the fifth, sixth and seventh process step S5 . S6 . S7 , which is the adaptive adaptation of highly automated driving in dialogue, especially speech dialogue SD , with the driver FF allows. This adaptation takes place in the context of the missing information, advantageously by means of the dialogue, in particular speech dialogue SD , between the control system 1 and the driver FF can be won. As a result, for example, in the example of the lane change described above, the lane change possibility is confirmed in the case of a recognized critical situation by the control system 1 which, in particular, as already described above, the highly automated driving in the FAS level 3 enabled by an acoustic feedback from the driver FF to the control system 1 , This corresponds to a transfer of responsibility to the driver FF which then takes over the vehicle guidance or the specific, ie by the control system 1 predetermined, driving maneuvers FM approved.
For example, the situation exists that a lane change demand, ie the need for a lane change, by the control system 1 is recognized. However, a highly automated lane change can not be performed because the vehicle ego No sensors or reliable sensors to detect the environment behind the vehicle ego having. For example, the vehicle points ego Although a sensor for environmental detection behind the vehicle ego on, but their range is not sufficient to a particularly fast approaching other vehicle AF on the lane FS to be changed, to recognize in time.

In a first step, the driver becomes FF then on the critical situation, advantageously acoustically and in plain text, ie by a voice output, informed and prompted to take over the vehicle management itself or the specific, ie from the control system 1 predetermined, driving maneuvers FM , here the lane change, to approve.

Such a situation is exemplary in 4 shown schematically. While the own vehicle ego the lane FS follows, the approaching other vehicle approaches AF its preceding other vehicle AF on, which drives slower, whereby at the own vehicle ego preceding vehicle AF a lane change demand arises, such as a lane change direction arrow FP indicated schematically. This also creates for your own vehicle ego a lane change demand, as also by such a lane change direction arrow FP indicated schematically to keep the set vehicle speed, for example, 130 km / h. An autonomous or highly automated follow-on drive of your own vehicle ego ie following the other vehicle ahead AF , associated with an autonomous or highly automated lane change, is not possible if the sensors for the out-of-range environmental monitoring are missing or, for example, their range is too low to estimate the safety of the lane change design. Because of the lack of sensor information, the lane change can therefore not be hedged and not performed.

Therefore, at the same time as the above-described information of the vehicle driver FF , a braking maneuver, starting, for example, with a predetermined comfort delay, with a possible increase up to a maximum delay, introduced, it is a corner frequency of the lateral control so switched that a dynamic evasive maneuver is possible, and a value of the limitation of a maximum allowable steering speed is so increases that a dynamic evasive maneuver is possible.

In a second step, the driver becomes FF , advantageously acoustic / optical and in plain text, ie with warning signal and by voice output, asked to confirm that the specified target lane is backward free, as in 4 shown, and thus of the control system 1 predetermined lane change, ie an evasive maneuver to the specified page, is possible. The driver FF can the confirmation and thus the approval G to carry out this driving maneuver FM , in this example of the lane change, for example, by giving a "yes" or by an alternative positive acknowledgment.

If the lane FS is occupied or will be occupied soon, for example, due to a approaching from behind other vehicle AF , the driver stops FF the particular, ie from the control system 1 predetermined, driving maneuvers FM for example, by a "no" or by an alternative negative statement, ie he does not issue a permit G to carry out this particular maneuver FM , Such a traffic situation is exemplary in 5 shown schematically. Here is the original vehicle driving ahead AF already on another lane FS ie on the left lane FS , changed, so that their own vehicle ego now the slow still driving another vehicle AF approaches, whereby the lane change demand exists. This lane change to the left lane FS However, it is not possible because on this left lane FS own vehicle ego from the left behind another vehicle AF approaches.

Advantageously, in the method both yes and no answers as well as alternative positive or negative answers for the speech dialogue SD takes into account an intuitive communication K of the driver FF with the control system 1 to enable. Through the speech dialogue SD ie by system-side language request, ie request AU by voice, and by the voice feedback of the driver FF , advantageously its driver response time Ft be shortened.

In a third step, if the driver FF , in particular acoustically, for example with a "yes" or with another positive feedback release, ie the approval G to carry out the driving maneuver FM , a dynamic evasive maneuver appropriate to the situation is performed to move to the other lane FS switch.

If the driver FF the driving maneuver FM by a "no" or by another negative statement stops, ie the approval G to carry out the driving maneuver FM not granted, a braking maneuver is performed.

If the current lane FS by slower moving or standing other vehicles AF or blocked by other obstacles and if so no, release, ie approval G of the driving maneuver FM by the driver FF done and the driver FF also does not take over manually, ie does not even take over the vehicle, for example, an evasive maneuver is performed on a stationary lane, if it has been previously recognized that such a lane with a sufficient space for the evasive maneuver space is available. Otherwise, the vehicle will be in the current lane FS slowed down to a standstill adjusted to the situation ahead.

LIST OF REFERENCE NUMBERS

1

control system

A

evaluation

AF

another vehicle

AL

rejection

AU

invitation

e

Result

ego

vehicle

FD

Maneuver execution

FF

driver

FM

maneuvers

FP

Lane change direction arrow

FS

lane

Ft

Driver reaction time

G

approval

K

communication

P

connection arrow

PA

percentage of subjects

SD

voice response

t

time beam

Ü

translation

VF

History movie preview

VK

Course control group

VL

Course reading

VS

History treasure hunt

ES1

Input variable for the first method step

ES2

Input variable for the second process step

ES3

Input variable for the third method step

S1

first process step

S2

second process step

S3

third process step

S4

fourth process step

S5

fifth process step

S6

sixth process step

S7

seventh process step

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been 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.

Cited patent literature

• DE 102012112802 A1 [0002]
• DE 102015014139 A1 [0003]
• DE 102011110486 A1 [0043]
• DE 102011016772 A1 [0043]
• DE 102012020901 A1 [0058]
• DE 102014003343 A1 [0058]

Claims (10)

Method for operating a control system (1) for automated vehicle guidance of a vehicle (ego), characterized in that a vehicle driver (Ego) during an automated driving operation of the vehicle (Ego) in an imminent critical situation, which the control system (Ego) 1) is not able to handle itself, is asked to take over the vehicle or to approve the execution of a given maneuver (FM), - reaching the critical situation and a takeover period within which the driver (FF) must take over the vehicle or the predetermined driving maneuvers (FM) must be predicted by a situation analysis, - a state of the vehicle driver (FF) is determined and a driver response time (Ft) is determined depending on the determined state of the driver (FF), - wherein the determined driver reaction time ( Ft) with the predicate ied transfer period is compared and the vehicle (FF) is delayed if the driver response time (Ft) is greater than the predicted acquisition period. Method according to Claim 1 , characterized in that the vehicle (Ego) is delayed as long as the driver (FF) the request (AU) to take over the vehicle guidance or approval (G) of the maneuver (FM) has not followed and / or as long as the critical situation exists , Method according to one of the preceding claims, characterized in that the situation analysis based on a sensory environmental observation of an environment of the vehicle (Ego) is performed. Method according to one of the preceding claims, characterized in that for the situation analysis driving maneuvers other road users are predicted and taken into account. Method according to one of the preceding claims, characterized in that for the situation analysis limits of the control system (1) and / or uncertainties regarding the limits of the control system (1) are taken into account. Method according to one of the preceding claims, characterized in that the state of the vehicle driver (FF) is determined by a driver's observation. Method according to one of the preceding claims, characterized in that the vehicle driver reaction time (Ft) a minimum time period is determined, the driver (FF) due to its state for the assumption of vehicle control or for the approval (G) of the given driving maneuver (FM) expected to require becomes. Method according to one of the preceding claims, characterized in that when determining the vehicle driver reaction time (Ft) a respective secondary activity and / or at least one vital parameter of the vehicle driver (FF) is taken into account and / or at least one preceding measurement of the driver's reaction time (Ft) of this vehicle driver ( FF) is taken into account. Method according to one of the preceding claims, characterized in that the request (AU) for taking over the vehicle guidance or for approval (G) of the implementation of the predetermined driving maneuver (FM) by means of a voice output, which is initiated by an audible and / or visual warning signal , Method according to one of the preceding claims, characterized in that the predetermined driving maneuver (FM) is a driving maneuver (FM) not secured by the control system (1), in particular a driving maneuver (FM) which, due to a limited sensor range, of an environment detection sensor system of the vehicle (Ego ) can not be performed safely.

Images