DE102009024131A1 - Traffic lane determining method for driver assistance system of vehicle, involves comparing curvature values of lane curvature with reference values of lane curvature, where reference values are determined from drive dynamics data - Google Patents

Abstract

The method involves taking pictures of surroundings of a vehicle by an image sensor (1) that is arranged at the vehicle. Plausibility of a traffic lane is inspected, where the traffic lane is determined from the taken pictures of the image sensor. Curvature values (Ck) of a lane curvature are determined from the taken pictures of the sensor by a Kalman filter (2), and are compared with reference values (Cf) of the lane curvature, where the reference values are determined from drive dynamics data.

Description

The The invention relates to a method for determining a traffic lane a vehicle according to the preamble of the claim 1 and a driver assistance system for a vehicle with means for determining a lane for the vehicle according to the preamble of claim 7.

From the DE 103 49 631 A1 a driver assistance method on the basis of lane information is known, wherein depending on the lane information driver information and / or an action are triggered. The lane information is derived from at least two lane identifying information.

From the DE 103 54 290 A1 a method for determining the lane of a vehicle is known. In this case, distance data of at least one distance sensor are evaluated, it being possible to determine from the distance data a lateral distance to a reference body which extends at least partially laterally next to the traffic lane. Depending on the distance data, lane data are determined, with noise information being filtered out of the distance data, which is caused in particular by vehicles or discontinuities of a reference body present in the area of the lane.

Of the Invention is based on the object, an improved method for determining a lane of a vehicle from recordings of an environment of the vehicle, by means of an image sensor arranged on the vehicle be recorded. Furthermore, the invention is the task based, a driver assistance system for a vehicle with improved means for determining a lane for the vehicle from recordings of an environment of the vehicle and a on the vehicle can be arranged image sensor for recording the environment of Vehicle to specify.

The The object is with respect to the method by the in claim 1 specified characteristics and with regard to the driver assistance system solved by the features specified in claim 7.

advantageous Embodiments of the invention are the subject of the dependent claims.

at The method according to the invention becomes a traffic lane for a vehicle from recordings of an environment of the vehicle determined by means of an image sensor arranged on the vehicle be recorded. The plausibility of the Recordings of the image sensor determined lane checked, from the images of the image sensor curvature a Lumbar curvature determined and with comparative values of the Lane curvature to be compared, from dynamic driving Data are determined.

By the plausibility check of the images determined lane is tested, whether the determined lane realistic is. As a result, the reliability of a increased such lane averaging. In particular, can advantageous the quality of functions of the vehicle be increased, for the control of such a determined lane is used by implausible detected lanes not to Control of these functions are used. The plausibility check based on a comparison of the image recordings of determined curvature values a lane curvature determined from driving dynamics data Comparative values of the lane curvature has the advantage that thereby captures images based on data obtained from the camera data are independent, checked for plausibility can be. This increases the reliability the plausibility check especially against known methods in which the plausibility check based on the image itself, for example, by image processing this recordings, is performed. By the investigation The comparison values from dynamic data also give these comparison values favorably the actual lane curvature very much exactly again, reducing the quality of the plausibility check is further increased.

Further Advantages, features and details of the invention are described below based on embodiments with reference to drawings described.

there demonstrate:

1 schematically a method for checking the plausibility of a lane determined from image captures in a block diagram, and

2 Curvature values of a lane curvature determined from image recordings and a tolerance band for deviations of the curvature values of comparison values as a function of a vehicle speed.

each other corresponding parts are in all figures with the same reference numerals Mistake.

1 schematically shows a block diagram of a method for checking the plausibility of a lane of a vehicle, which are determined from images of an environment of the vehicle, which by means of an image sensor arranged on the vehicle 1 be recorded.

From the of the image sensor 1 -housed NEN recordings are made using a Kalman filter 2 Curvature values C _{K of} a lane curvature determined.

At the same time, values of a vehicle speed v and a steering wheel angle δ _{LW of} the vehicle, which are detected by corresponding sensors of the vehicle and via a data connection 3 , For example, via a so-called CAN bus (CAN = Controller Area Network), provided by means of a single-track model 4 Comparison values C _{F of} the lane curvature determined.

From the determined comparison values C _F , and the values of the driving speed v and the steering wheel angle δ _LW is determined by means of a tolerance band map 5 a tolerable deviation ΔC of the curvature values C _{K is determined} from the comparison values C _F.

• – Die Krümmungswerte C_K dürfen von den Vergleichswerten C_F innerhalb der Spurbreite der Fahrspur abweichen, wenn anzunehmen ist, dass das Fahrzeug in der Fahrspur fährt. Mit Zunahme der Fahrgeschwindigkeit v wird die dabei tolerierte Abweichung kleiner.
• – Die Krümmungswerte C_K können von den Vergleichswerten C_F aufgrund eines Modellfehlers des Kalmanfilters 2 und/oder des Einspurmodells 4 abweichen. Diese Modellfehler können fahrgeschwindigkeitsabhängig sein; so kann mit zunehmendem Schlupf beispielsweise der Modellfehler des Einspurmodells 4 zunehmen. Bei Geschwindigkeiten von 0 km/h ist kein Vergleich möglich.
• – Aufgrund eines verspäteten Lenkeinschlags kann sich ein zeitlicher Versatz zwischen einer Änderung des Lenkradwinkels δ_LW und der Bilddateninformation ergeben.
• – Die tolerierbare Abweichung ΔC kann ferner von einer Assistenzfunktion abhängig sein, der die Ermittelung der Fahrspur dient. So ist beispielsweise bei längsdynamischen Assistenzfunktionen die tolerierbare Abweichung ΔC enger zu wählen als bei querdynamischen Assistenzfunktionen.

The tolerance band characteristic _diagram dependent on the values of the vehicle speed v and the steering angle δ _LW preferably takes into account at least one of the following blurs:

• The curvature values C _K may deviate from the comparison values C _F within the lane width of the lane if it is to be assumed that the vehicle is driving in the lane. As the vehicle speed v increases, the deviation tolerated thereby becomes smaller.
• The curvature values C _K can be obtained from the comparison values C _F due to a model error of the Kalman filter 2 and / or the single track model 4 differ. These model errors may be driving speed dependent; For example, with increasing slip, the model error of the one-track model may increase 4 increase. At speeds of 0 km / h no comparison is possible.
• - Due to a delayed steering angle, a time lag between a change of the steering wheel angle δ _LW and the image data information may result.
• The tolerable deviation .DELTA.C can also be dependent on an assistance function, which serves to determine the lane. Thus, for example, the tolerable deviation ΔC is to be selected to be narrower in longitudinally dynamic assistance functions than in transverse dynamic assistance functions.

The tolerable deviations ΔC are preferably determined in the experiment and as a tolerance band characteristic map 5 saved.

In a plausibility check 6 Finally, it checks to see if the images from the image sensor 1 determined curvature values C _K differ by more than the tolerable deviation ΔC from the comparison values C _F determined from the values of the vehicle speed v and the steering wheel angle δ _LW . If this is not the case, the pictures from the image sensor 1 determined lane considered plausible, otherwise as implausible. The result of the plausibility check 6 is output in a status variable Sta_C.

2 schematically shows a result of such a plausibility check 6 , Shown are the pictures from the image sensor 1 determined curvature values C _K and the comparison values C _F determined from the values of the vehicle speed v and the steering wheel angle δ _LW in dependence on the vehicle speed v, as well as a tolerance band B, within which the curvature values C _K are classified as plausible. The boundary lines of the tolerance band B therefore extend at a distance from the respective tolerable deviation ΔC (generally dependent on the vehicle speed v) above and below the comparison values C _F. In three regions I, II, III, the curvature values C _{K run} outside the tolerance band B. In these regions I, II, III, the curvature values C _K and those from the images of the camera 1 therefore, they are considered as plausible in the other areas.

The method described above can be modified in various ways. For example, by means of the Kalman filter 2 In addition to the curvature values C _K further track parameters, such as a lane curvature change and / or an angle between a longitudinal axis of the vehicle and the lane, are determined. These further track parameters can then be used analogously to the curvature values C _K for checking the plausibility of the images taken by the camera 1 determined lane are used.

Alternatively or additionally, instead of the single track model 4 a multidimensional map for determining the comparison values C _F are used, whereby the computational effort required for this determination is reduced, or the comparison values C _F can be determined from already existing CAN signals of a driven curve curvature, for example from a brake control device and / or a vehicle dynamics control ,

Furthermore, as input variables of the tolerance band characteristic map 5 In addition or as an alternative to the values of the driving speed v and the steering wheel angle δ _LW , static and / or at least one further driving dynamic variable, for example a vehicle lateral acceleration and / or a yaw rate of the vehicle, can be used to determine the tolerable deviation ΔC.

An inventive driver assistance system for a vehicle has means for determining a lane for the vehicle from recordings of a Environment of the vehicle and an on-vehicle image sensor 1 for picking up the surroundings of the vehicle, and an evaluation unit for determining and comparing curvature values C _K from the images of the image sensor 1 and of comparison values C _F from driving dynamics data, in particular from values of the driving speed v and the steering angle δ _LW . By means of the evaluation unit are preferably the determination of the curvature values C _K from the images of the image sensor 1 by means of a Kalman filter 2 , the comparison value C _F by means of a one-track model 4 or multi-dimensional characteristic map, the tolerable deviation ΔC by means of a tolerance band map 5 and a plausibility check 6 as each described above feasible.

1

image sensor

2

Kalman filter

3

Data Connection

4

single-track

5

Tolerance band map

6

plausibility check

C _K

curvature value

C _F

comparison value

v

driving speed

δ _LW

steering wheel angle

Sta_C

state variable

B

tolerance band

I

first Area

II

second Area

III

third Area

QUOTES INCLUDE IN THE DESCRIPTION

This list The documents listed by the applicant have been automated generated and is solely for better information recorded by the reader. The list is not part of the German Patent or utility model application. The DPMA takes over no liability for any errors or omissions.

Cited patent literature

• - DE 10349631 A1 [0002]
• - DE 10354290 A1 [0003]

Claims (7)

Method for determining a traffic lane for a vehicle from images of an environment of the vehicle, which is detected by means of an image sensor ( 1 ), characterized in that the plausibility of the images of the image sensor ( 1 ) is checked by taking from the images of the image sensor ( 1 ) Curvature values (C _K ) of a lane curvature are determined and compared with reference values (C _F ) of the lane curvature, which are determined from vehicle dynamics data. Method according to Claim 1, characterized in that values of a driving speed (v) and a steering wheel angle (δ _LW ) of the vehicle are determined and used as vehicle dynamics data from which the comparison values (C _F ) of the lane curvature are determined. A method according to claim 1 or 2, characterized in that the curvature values (C _K ) from the images of the image sensor ( 1 ) by means of a Kalman filter ( 2 ) be determined. Method according to one of the preceding claims, characterized in that for the comparison of the images of the image sensor ( 1 ) determined curvature values (C _K ) from a tolerance band map ( 5 ) a tolerable deviation (ΔC) of the curvature values (C _K ) from the comparison values (C _F ) is determined as a function of values of driving dynamics variables. A method according to claim 4, characterized in that as driving dynamics variables of the tolerance band characteristic map ( 5 ) a vehicle speed (v) and / or a steering wheel angle (δ _LW ) and / or a vehicle lateral acceleration and / or a yaw rate of the vehicle are determined and used. A method according to claim 4 or 5, characterized in that the from the images of the image sensor ( 1 ) determined lane is currently considered to be plausible if the deviation of the instantaneous curvature value (C _K ) from the current comparison value (C _F ) is smaller than the currently determined tolerable deviation (.DELTA.C). Driver assistance system for a vehicle with means for determining a traffic lane for the vehicle from images of an environment of the vehicle and an image sensor which can be arranged on the vehicle ( 1 ) for recording the environment of the vehicle, characterized by an evaluation unit for determining and comparing curvature values (C _K ) of a lane curvature from the images of the image sensor ( 1 ) and comparative values (C _F ) of the lane curvature from vehicle dynamics data.