DE102014202385A1 - Method for detecting a driver avoidance request in a driver assistance system for motor vehicles - Google Patents

Abstract

A method for detecting a driver avoidance request in a driver assistance system for motor vehicles, in which are evaluated as primary inputs of the steering angle and the steering angular velocity and the vehicle speed (V) of the vehicle, comprising the following steps: - Determining model trajectories (M1, M2, Mn) specify the primary inputs as functions of time, based on additional inputs that characterize the traffic environment and / or state of the host vehicle, determining at least one model trajectory corresponding to an evasive maneuver of the vehicle, and checking whether at least one alternative Model trajectory that does not correspond to an evasive maneuver of the vehicle - if at least one alternative model trajectory exists, selecting one of the determined model trajectories based on the degree of correspondence with measured values of the primary input variables, Decide whether a driver escape request exists based on the selected model trajectory.

Description

State of the art

The invention relates to a method for detecting a driver avoidance request in a driver assistance system for motor vehicles, in which the steering angle and / or the steering angle speed and the vehicle speed of the vehicle are evaluated as primary input variables.

There are driver assistance systems for motor vehicles in development, which support the driver in the event of an imminent collision (especially in longitudinal traffic) in an evasive maneuver (Evasive Steering Support, ESS). To do this, the driver must first initiate the evasive maneuver himself (for example, after a collision warning by the system). If the system recognizes that the driver wants to avoid the obstacle, the vehicle is stabilized by means of a superimposed steering torque or by asymmetrically controlling the brakes on a suitable avoidance trajectory. Performing the challenging evasive maneuver will be much easier and safer for the driver.

Since the backup assistance is activated only after a recognized evasion request, the algorithm for evasion request identification must have the highest possible recognition rate with the lowest possible false trigger rate. The Applicant has disclosed a method in which an environmental sensor, e.g. a radar sensor detects whether an object exists with which a collision is imminent. When this condition is satisfied, the steering angle (e.g., steering wheel angle) and the steering angular velocity (preferably low-pass filtered) are evaluated. If both values exceed a threshold, the backup support is activated.

At high speeds, for example, above 60 km / h, where ESS is previously available, the known method works very reliable. However, if an alternate assistance function is used even at lower speeds, e.g. is to be implemented in city traffic, there is a risk that the error trigger rate increases in the conventional method, because, as long-term tests have shown when driving at low speed strong steering wheel wrench and high steering wheel angular velocities occur frequently, even if no evasive maneuvers take place. Reasons are at very low speeds, where an evasion support would not be useful anyway, such as parking and maneuvering. At somewhat higher speeds, steering angle and steering angular velocity values are particularly high in connection with turn maneuvers. However, in this speed range, evasive maneuvers, e.g. To avoid a collision with a road crossing pedestrian, be quite useful, so that in itself would be desirable to be able to offer evasion support in this speed range can.

Out DE 10 2006 043 676 For example, a driver assistance system is known in which primary input parameters which are characteristic of the lateral dynamics of the vehicle are evaluated in order to be able to evaluate the driver's attention and then decide on this basis whether or not to issue a collision warning to the driver.

Disclosure of the invention

The object of the invention is to provide a method which allows recognition of a driver escape request with high detection rate and low false trigger rate even in the lower speed range.

• – Bestimmen von Modelltrajektorien, die die primären Eingangsgrößen als Funktionen der Zeit angeben, auf der Grundlage zusätzlicher Eingangsgrößen, die das Verkehrsumfeld und/oder den Zustand des eigenen Fahrzeugs charakterisieren, wobei mindestens eine Modelltrajektorie bestimmt wird, die einem Ausweichmanöver des Fahrzeugs entspricht, und geprüft wird, ob mindestens eine alternative Modelltrajektorie existiert, die keinem Ausweichmanöver des Fahrzeugs entspricht,
• – wenn mindestens eine alternative Modelltrajektorie existiert, Auswählen einer der bestimmten Modelltrajektorien anhand des Ausmaßes der Übereinstimmung mit gemessenen Werten der primären Eingangsgrößen, und
• – Entscheiden, ob ein Fahrerausweichwunsch vorliegt, anhand der ausgewählten Modelltrajektorie.

This object is achieved according to the invention by a method with the following steps:

• Determining model trajectories indicating the primary inputs as functions of time, based on additional inputs characterizing the traffic environment and / or the state of the host vehicle, determining at least one model trajectory corresponding to an evasive maneuver of the vehicle, and checked whether there is at least one alternative model trajectory that does not correspond to an evasive maneuver of the vehicle,
• If at least one alternative model trajectory exists, selecting one of the determined model trajectories based on the degree of correspondence with measured values of the primary inputs, and
• Decide whether a driver evasion request is present based on the selected model trajectory.

With this method it is possible, in situations where the driver makes heavy steering actions, to distinguish whether the most plausible cause of these steering actions is that the driver intends to evade maneuvers, or whether it is more likely that the driver will perform these steering actions other reasons.

For the mathematical description of the model trajectories, the steering angle and the steering angle velocity are particularly useful. However, it is not excluded within the scope of the invention that other input variables are measured and evaluated, for example, the yaw rate and / or the yaw acceleration, in which the information about the steering angle and the steering angular velocity is included only implicitly. Likewise, if necessary, further input variables can be measured and evaluated.

Advantageous developments and refinements of the invention are specified in the subclaims.

• – die Historie der Fahrzeuggeschwindigkeit in den Sekunden vor dem Manöver,
• – der Zustand des Fahrtrichtungsanzeigers (Blinker)
• – Informationen über vorausliegende Kurven, Kreuzungen oder Abzweigungen, insbesondere anhand Daten einer digitalen Karte,
• – Bewegungsdaten vorausfahrender Fahrzeuge,
• – das Vorhandensein einer im Navigationssystem programmierten Fahrtroute,
• – Positon, Geschwindigkeit und Klasse (z.B. Fußgänger, Pkw, Lkw, ...) des kollisionsträchtigen Objekts.

In particular, additional input variables are:

• The history of vehicle speed in the seconds before the maneuver,
• - the condition of the direction indicator (turn signal)
• Information about the curves, intersections or branches ahead, in particular using data from a digital map,
• - movement data of vehicles in front,
• The presence of a route programmed in the navigation system,
• - Position, speed and class (eg pedestrians, cars, trucks, ...) of the collision-critical object.

In order to determine the model trajectory, in addition to the above-mentioned input variables (or a suitable selection thereof), a knowledge base can also be used in which information or rules about typical behavior of human drivers are stored, which indicate which vehicle trajectories are to be expected in which traffic situations. This knowledge base can be formed for example by empirical data and / or by a corresponding collection of rules and laws (expert knowledge).

In the following, embodiments of the invention will be explained in more detail with reference to the drawing.

Show it:

1 a block diagram of a driver assistance system, with which the invention is executable;

2 a sketch of a traffic situation to explain the method according to the invention;

3 a schematic representation of model trajectories in the sketch according to 2 ; and

4 and 5 Flowcharts for embodiments of the method according to the invention.

This in 1 a driver assistance system shown as a block diagram comprises an electronic data processing system 10 on, the data from an environment sensor 12 which is represented here by a radar sensor. The radar sensor is installed, for example, in front of the vehicle and serves to locate vehicles in front and other objects ahead of their own vehicle.

Further input variables are provided by the data processing system 10 from a vehicle speed sensor 14 which measures the vehicle speed V of the own vehicle and a steering wheel sensor 16 which measures the steering angle φ (turning angle of the steering wheel) and the steering angular velocity dφ / dt. Furthermore, the data processing system communicates with a navigation system 18 the vehicle, which provides information about roadways, intersections and the like in the form of an electronic map and, if a route guidance system is activated, also provides information about the route currently being followed.

A flasher switch 20 provides information about the state of the direction indicator of the own vehicle.

The data of the environment sensor 12 be from a detection module 22 evaluated, which detects in particular the distances, relative velocities and azimuth angle of located in front of the own vehicle objects.

The data of the vehicle speed sensor 14 and the steering wheel sensor 16 and in the example shown also the data of the navigation system 18 (Own vehicle position) and turn signal switch 20 be from a recording module 24 continuously, each recorded for the duration of a sliding time window of predetermined length.

Based on the data of the acquisition module 22 and the recording module 24 evaluates a collision warning module 26 continuously the traffic situation and checked on the basis of the location data and the dynamic data of the own vehicle if among the located objects at least one object is, which represents a potential obstacle, thus an object, from which a danger of collision emanates. Based on this assessment, the collision warning module decides 24 whether an (optical or acoustic) collision warning is to be output to the driver. If a collision warning is issued, the recording module will be sent at the same time 24 a signal which "freezes" the record window, ie, the data stored at that time is no longer cleared, but the new incoming data will still be recorded continuously.

On the basis of the newly arriving dynamic data, in particular on the basis of the current values of the steering angle and the steering angle speed as well as in the recording module 24 recorded history then assesses a decision module 28 the driver's behavior to determine if a driver's intention is to initiate an evasive maneuver (be it on his own initiative or in response to the collision notice). The decision module takes effect 28 also on a knowledge base 30 in which empirical knowledge about typical behaviors of human motor vehicle drivers is stored. In the knowledge base 30 Supplementary information about the traffic environment is also evaluated by the entry module 22 For example, information about the classes of the located objects, in particular the potential obstacle (pedestrians, cars, trucks, ...), as well as current movement data of the potential obstacle and possibly other relevant objects.

If the decision module 28 decides that an evasion request of the driver is actually recognizable, it activates an evasion assistant 32 (ESS), who then actively intervenes in the steering and / or braking system of the vehicle to actively assist the driver in performing the evasive maneuver.

The operation of the decision module 28 will now be explained using the example of a traffic situation, the sketchy in 2 is shown.

A vehicle 34 equipped with the driver assistance system (the "own vehicle"), approaches an intersection 36 on which it is just at this moment a collision of two more vehicles 38 . 40 comes. The vehicle 40 blocks part of your own vehicle 34 traveled lane and thus represents an obstacle that poses a risk of collision. The collision warning module 26 then issues a collision warning. In 2 is still a possible evasion trajectory 42 for your own vehicle 34 shown. When the driver of the vehicle 34 originally had the intention at the intersection 36 go straight ahead, he would now probably try the obstacle (vehicle 40 ) on this evasion trajectory 42 to drive around.

However, it is also conceivable that the driver of the vehicle 34 originally intended to turn right at the intersection and, accordingly, a turn-off trajectory 44 to follow by first on a turning lane 46 changes and then at the crossroads into a crossroads 48 einbiegt. In that case it would be irritating and disturbing to this driver, and possibly even dangerous, if the Dodge Assistant 32 nevertheless, the vehicle would try on the avoidance trajectory 42 to keep.

In the decision module 28 Therefore, an algorithm is implemented, with which it is possible to distinguish with low error rate between the two possibilities presented above.

The decision module calculates this 28 first a model trajectory 42 ' , in the 3 is represented as a sequence of circles and that of the evasion trajectory 42 in 2 equivalent. The diameter of the circles in 3 gives the respective instantaneous speed of the own vehicle 34 , that is, the distance that the vehicle lies within a time interval dt of fixed length. The curvature of the model trajectory and the derivative of this curvature correspond to the steering angle φ and its derivative dφ / dt. The mathematical representation of the model trajectory 42 ' is thus equivalent to an indication of the steering angle φ and the steering angular velocity dφ / dt as functions of time (or as functions of the location, which can be converted at zeitantee functions at known driving speed V).

In the in 3 For the sake of simplification, it is assumed that the driver passes through the evasion trajectory at constant speed. In general, however, the driving speed is time-dependent. In the determination of the model trajectory 42 ' Movements can also be considered that the vehicle 40 possibly still executes, as well as sizes that the presumed dimensions of the vehicle 40 and the like. Also the dynamic data of your own vehicle 34 , in particular the initial speed when detecting the obstacle, have an influence on the course of the model trajectory. Algoritms for the calculation of such model trajectories are known as such and are also used, for example, in the calculation of target trajectories for the evasive maneuver in the evasive assistant 32 used.

Furthermore, the decision module calculates 28 at least one alternative model trajectory 44 ' that may be different from the driver of the vehicle 34 desired route corresponds, in the example shown the turn trajectory 44 , In determining these and possibly further alternative model trajectories, the decision module takes effect 28 on the navigation system 18 provided digital map, the information about the presence of the intersection 36 and the cross street 48 contains and possibly also information about the presence of the turn lane 46 , Optionally or additionally, supplementary information of the environmental sensor system may also be used in this context 12 be resorted to, for example, a video camera that recognizes lane and lane courses.

With the help of theoretical models or empirical data from the knowledge base 30 In addition, the estimated speed course on the model trajectory 44 ' modeled. For example, the driver will reduce the speed just before the actual turn, which is in 3 It can be seen from this that the diameters of the circles in this area of the trajectory decrease.

Typically, the driver is already on the turn lane before the change 46 reduce his speed.

In 3 A line T0 marks the time at which the obstacle is detected and the collision warning is issued. The model trajectories 42 ' and 44 ' However, they also include a period of time prior to this time T0. For this period, the measured data of the vehicle speed V, the steering angle φ, and the steering angular velocity dφ / dt are in the recording module 24 available.

By comparing the model data with the measured data, one obtains a criterion for distinguishing which of the two model trajectories the driver was more likely to choose. The decrease in the vehicle speed V even before the time T0, ie even before the time at which the obstacle was even recognizable, indicates that the driver anyway had the intention to turn at the intersection.

In a corresponding manner, characteristic signatures in the course of the steering angle and / or the steering angle velocity can also be taken into account.

Further indications are provided by the status of the turn-signal switch 20 (in case of a turn request, the driver will set the turn signal before or at the latest shortly after time T0) as well as in the navigation system 18 saved route.

In 4 a possible process flow in the form of a flow chart is shown. If the collision warning module 26 has detected an acute risk of collision and has possibly issued a collision warning, a number of model trajectories M1, M2, ..., Mn is determined in step S1, which come in the current traffic situation as possible trajectories of the own vehicle in question, Among them is at least one avoidance trajectory on which the obstacle can be avoided. Naturally, further (alterantive) model trajectories can only be calculated if the infrastructure (course of the lanes and lanes, turn-offs, etc.) permits such alternative trajectories at all.

Input variables in step S1 are, in particular, the location data about the object from which the risk of collision originates, possibly location data of other objects which are to be considered as possible further obstacles, the position of the own vehicle known from the navigation system and the map data from the navigation system , Each of the model trajectories determined in this way contains information that indicates the associated time profile of the steering angle and / or the steering angle speed (and preferably the travel speed V).

In a further step S2 then the model data with the of the Aufzeichnunsmodul 24 adjusted measured data indicating the history of the steering angle, the steering angle speed and the vehicle speed. For the Trajektorienmatching probalistic methods are known, which allow each model trajectory M1, M2, Mn assign a probability value P1, P2, Pn, indicating how large in view of the history of the measured data is the probability that the enthralling model trajectory of the driver actually corresponds to the desired trajectory. Other known methods for trajectory matching work with fuzzy logic. In that case, corresponding membership measures are obtained instead of the probability values P1, P2, Pn.

In the in 4 In the example shown, the step S2 is followed by a further step S3, in which the preliminary results obtained by the trajectory matching are fused with further features, in particular with the turn signal, the route stored in the navigation system, and the like. Through this merger, the probability values or membership measures may still be changed. The effects of the further features on the certainty values can be modeled, for example, using a Bayesian network.

The model trajectory with the highest probability or the highest membership measure is then selected as the probably desired trajectory. If the selected trajectory is an evasion trajectory, it is checked in step S4 whether (as before) an actual collision risk exists. If so, in step S5 the alternate assistance is initiated based on the selected avoidance trajectory.

5 illustrates an alternative method in which the fusion with the further features such as turn signal, Navi route and the like according to step S3 in 4 already integrated in step S1. In this case, the probability values P1, P2, Pn obtained in step S2 immediately result Selection of the most probable trajectory and thus possibly the detection of an alternative request.

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 102006043676 [0005]

Claims (9)

Method for detecting a driver avoidance request in a driver assistance system for motor vehicles ( 34 ), in which as the primary input variables the steering angle (φ) and the steering angular velocity (dφ / dt) and the vehicle speed (V) of the vehicle are evaluated, characterized by the following steps: - determining model trajectories ( 42 ' . 44 ' ), which indicate the primary input variables as functions of the time, on the basis of additional input variables, the traffic environment and / or the state of the own vehicle ( 34 ), wherein at least one model trajectory ( 42 ' ), which corresponds to an evasive maneuver of the vehicle, and it is checked whether at least one alternative model trajectory ( 44 ' ) that does not correspond to an evasive maneuver of the vehicle - if at least one alternative model trajectory ( 44 ' ), selecting one of the determined model trajectories based on the degrees of correspondence with measured values of the primary input quantities, - deciding whether a driver escape request exists based on the selected model trajectory. Method according to Claim 1, in which the additional input variables comprise information about the traffic infrastructure on the basis of a digital map. Method according to Claim 1 or 2, in which the additional input variables are a state of a blinker switch ( 20 ). Method according to one of the preceding claims, in which the additional input variables contain information about a route guidance in a navigation system ( 18 ) comprise selected route. The method of any one of the preceding claims, wherein at least one of the primary inputs is continuously recorded during a sliding time window and the additional inputs comprise the history of the recorded quantities. Method according to one of the preceding claims, in which the selection of one of the determined model trajectories is carried out by probabilistic trajectory matching. Method according to one of claims 1 to 5, wherein the selection of one of the determined model trajectories by trajectory matching by means of fuzzy logic. Method according to Claim 6 or 7, in which the results of the trajectory matching are subsequently fused with at least one of the additional input variables. Dodge assistant for motor vehicles, characterized by an electronic data processing system ( 10 ) in which a method according to any one of claims 1 to 8 is implemented.

Images