WO2017178232A1

WO2017178232A1 - Method for operating a driver assistance system of a motor vehicle, computing device, driver assistance system, and motor vehicle

Info

Publication number: WO2017178232A1
Application number: PCT/EP2017/057405
Authority: WO
Inventors: Marco Heimberger; Jean-Francois Bariant; Markus Heimberger
Original assignee: Valeo Schalter Und Sensoren Gmbh
Priority date: 2016-04-15
Filing date: 2017-03-29
Publication date: 2017-10-19
Also published as: DE102016106978A1

Abstract

The invention relates to a method for operating a driver assistance system (2) of a motor vehicle (1) by detecting an object (O) in first sensor data acquired by a vehicle sensor device (S) from a surrounding area (3) of the motor vehicle (1) and identifying an object-specific feature (M) on the object (O) in the first sensor data; determining a first pose of the motor vehicle (1) with respect to the object-specific feature (M) on the basis of the object-specific feature (M), recognizing the object (O) in second sensor data acquired by the sensor device (S) on the basis of the object-specific feature (M), determining a second pose of the motor vehicle (1) with respect to the object-specific feature (M) on the basis of the object-specific feature (M); determining a pose of the motor vehicle (1) with respect to the object-specific feature (M), which results from a movement of the motor vehicle (1) between the acquisition of the first and second sensor data, by means of odometry; and detecting an odometry error if a difference between the resulting pose and the second pose exceeds a predetermined threshold value. The invention also relates to a computing device (9), to a driver assistance system (2) and to a motor vehicle (1).

Description

Method for operating a driver assistance system of a motor vehicle, computing device, driver assistance system and motor vehicle

The invention relates to a method for operating a driver assistance system of a motor vehicle by detecting an object in by a vehicle-side

Sensor device from a surrounding region of the motor vehicle detected first sensor data, identifying an object-specific feature on the object in the first sensor data and determining a first pose of the motor vehicle to the object-specific feature based on the object-specific feature. The invention also relates to a computing device, a driver assistance system and a

Motor vehicle with such a driver assistance system.

A method for operating a driver assistance system of the aforementioned type is known for example from DE 10 2010 056 217 A1. In this case, an orientation of a motor vehicle to an adjacent vehicle is determined on the basis of an image taken by a vehicle-side camera of the adjacent vehicle. The image is processed by a computing device and identified at least one object on the adjacent vehicle. The orientation of the motor vehicle is determined based on an image of the identified object. Such an object may be, for example, an official license plate and / or tires of the other vehicle. The information about the orientation of the neighboring vehicle may be used to assist the driver in performing parking operations.

It is an object of the present invention to provide a solution to provide a versatile driver assistance system, by means of which a motor vehicle can be located particularly reliable.

This object is achieved by a method by a

Computing device, solved by a driver assistance system and by a motor vehicle with the features according to the respective independent claims.

Advantageous embodiments of the invention are the subject of the dependent

Claims, the description and the figures.

In one embodiment of a method according to the invention for operating a driver assistance system, an object may be in by a vehicle-side sensor device detected from an environmental region of the motor vehicle detected first sensor data and an object-specific feature on the object in the first sensor data are identified. In addition, a first pose of the motor vehicle to the object-specific feature can be determined based on the object-specific feature. In addition, the object is in through the sensor device from the

Recognized area detected second sensor data based on the object-specific feature and a second pose of the motor vehicle to the

object-specific feature determined on the basis of the object-specific feature. In addition, in particular, a pose resulting from a movement of the motor vehicle between the detection of the first and the second sensor data becomes

Motor vehicle to the object-specific feature determined by odometry and the Odometriefehler detected if a deviation between the resulting pose and the second pose exceeds a predetermined threshold.

In a preferred embodiment of the method for operating a

Driver assistance system, an object detected in a vehicle-side sensor device from a surrounding area of the motor vehicle first sensor data and identified an object-specific feature on the object in the first sensor data. In addition, a first pose of the motor vehicle to the

object-specific feature determined based on the first sensor data. In addition, the object is detected in the second sensor data detected by the sensor device from the surrounding area on the basis of the object-specific feature, and a second pose of the motor vehicle for the object-specific feature is determined on the basis of the object-specific feature. In addition, a pose of the motor vehicle to the object-specific feature resulting from a movement of the motor vehicle between the detection of the first and the second sensor data is determined by means of odometry and the odometry error is detected if a deviation between the resulting pose and the second pose exceeds a predetermined threshold value.

By means of the method, a driver assistance system can be implemented, which is designed to determine poses with respect to at least one object in the surrounding area of the motor vehicle and also during the pose determination for detecting an odometry error. The pose includes a position or a distance of the motor vehicle relative to the object and an orientation of the motor vehicle to the object. The determination of the pose, that is to say the determination of a spatial position between the motor vehicle and the object, becomes a particular one so-called SLAM (Simultaneous Localization and Mapping, Simultaneous Localization and Map Generation). In the SLAM process,

For example, by a computing device of the driver assistance system, at the same time creates an environmental map describing the surrounding area and estimated the pose of the motor vehicle within this map. The environment map is in particular a two-dimensional map, in which the objects detected by the vehicle-side sensor device can be registered and thus spatial positions of the objects are indicated to the motor vehicle. This can be the

Motor vehicle, for example, in an origin of the area map

descriptive coordinate system. With knowledge of an absolute position of the motor vehicle, which is indicated for example in GPS coordinates, the absolute position of the object can also be determined on the basis of the specific pose of an object with respect to the motor vehicle.

Such per se known SLAM methods are usually very computationally intensive, since the environment map is usually constructed incrementally and thus objects that were detected, for example, in a first measurement must be recognized particularly reliable and accurate in a second measurement. Especially in the object recognition of, for example, other vehicles using

Ultrasonic sensors, the positions of the vehicles can be determined only very inaccurate, since the ultrasonic sensors measure depending on the pose of the motor vehicle, ie orientation or angle and distance of the motor vehicle to the other vehicle, different contour points on the other vehicle. In order to obtain more reliable results, the data of several measuring cycles and / or sensors are thus usually fused, whereby the computing power increases significantly.

In the method according to the invention, it is now provided in particular, the

Posenbestimmung the motor vehicle based on the object-specific feature, which is located on the object to perform. For this purpose, the object is first recognized based on the first sensor data. In particular, a first image of at least one camera of the motor vehicle is detected as the first sensor data and / or first laser scanner data of at least one laser scanner of the first sensor data

Motor vehicle recorded. Preferably, the first sensor data is a first image of the surrounding area of the motor vehicle captured by the camera. In this first image of the surrounding area, which was acquired at a first measurement time, the object is identified and the at least one object-specific feature is identified on the object. The object-specific feature is in particular a feature of the object, by means of which the object can be uniquely identified and recognized. Then, based on the object-specific feature, the first pose of the motor vehicle is determined to the object-specific feature. Based on the object, which is clearly identifiable, and based on the first pose then the

Area map can be created by the motor vehicle and the object in the

Map to be entered and thereby have a predetermined by the first pose spatial position to each other.

Then, second sensor data, for example a second camera image and / or second laser scanner data, are acquired by the sensor device at a second measurement time. In particular, a second camera image of the surrounding area is detected as the second sensor data. In particular, between the detection of the first and the second sensor data, ie between the first and the second measurement time, several minutes or even several hours or days lie. The first and the second sensor data are thus in particular no sensor data, which are recorded directly in chronological succession. The first and the second sensor data can be recorded, for example, when the motor vehicle is on a parking space search and passes several times the same location and thus the same object in the surrounding area. It can also be provided that the first and the second sensor data are detected on a route section which is traveled regularly by the motor vehicle, for example in the daily drive to a workstation and / or to a driver's residence

Motor vehicle. The second sensor data are searched for the object and can be clearly recognized based on the object-specific feature.

As soon as the object has been recognized by the computing device of the driver assistance system on the basis of the object-specific feature, the second pose of the

Motor vehicle to the object determined based on the object-specific feature.

In addition, the pose of the motor vehicle at the second detection time is additionally determined by means of odometry. By odometry, the motor vehicle can estimate its location based on data of its propulsion system. Such data of

Propulsion system, for example, a number of Radumdrehungen and steering angle of the motor vehicle. Odometry can thus be done by capturing the data of the

Propulsion system as well as knowledge of vehicle-specific geometry data, such as a known wheel diameter, a distance are determined, which the motor vehicle from the first pose to the second measurement time

has covered. An endpoint of the distance traveled from the first pose it determines as the pose, which holds the motor vehicle in the detection of the second sensor data. If there is no deviation between the second pose and the pose determined by odometry, it is assumed that there is no odometry error. But if the second pose deviates from the pose determined by odometry, it is assumed that there is an odometry error in the pose determination. The Odometriefehler can, for example, from certain road conditions, changing or incorrect information about the

Wheel geometry, etc. result.

In particular, a stationary object is detected as the object. The invention is based on the finding that the odometry adjustment can only be carried out reliably and correctly if the global position of the object has not changed.

The object-specific feature can be recognized in the images, for example by means of pattern recognition. To determine the first and second pose from the feature, edges of the object-specific feature may be identified and an angle between the edges recognized. The detection of the edges can be done by means of a known edge detection method, according to which in the recorded image for color transitions - for example, from black to white - is searched. Depending on the angle, the orientation of the motor vehicle can then be determined, since the angle represents a measure of an orientation angle of the static object with respect to the motor vehicle.

Based on the SLAM method, which determines the poses based on the

performs object-specific characteristics, with very little data and

Computing power needed for the localization of the motor vehicle, in addition an odometry adjustment can be performed and thus also a reliable

Pozzo determination by Odoemtrie be provided. This can advantageously be provided a reliable and versatile driver assistance system.

Preferably, in the presence of the odometry error, a correction value for compensating the odometry error is determined based on the second pose. So it's about it

assumed that the detected from the second sensor data pose is the correct current pose of the motor vehicle. The correction value is then determined so that the resulting pose determined by odometry corresponds to the second pose. In particular, continuous detection of the object-specific features of objects in the surrounding area and mapping of their positions in the environment

Ambient map or two-dimensional plane can thus be corrected by means of the SLAM method continuously the Odometriefehler and thus a particularly reliable driver assistance system can be provided.

In an advantageous embodiment of the invention, after identifying the object-specific feature, the object is assigned an identification number which is stored in a vehicle-side storage device. Based on this

object-specific identification number or ID, which the

If the object-specific characteristic is assigned uniquely, the object can be uniquely identified as soon as the object-specific characteristic has been recaptured. In this case, the ID is stored together with information about the object-specific feature in the storage device. In particular, the ID and the information are stored at least for so long that the odometry adjustment can also be carried out if, for example, the motor vehicle returns to the global position after several days, in which the object with the object-specific feature is located.

In a further development of the invention, the second sensor data for

Recognition of the object only then searched for the object-specific feature, if it was detected based on data of a navigation system of the motor vehicle, that the surrounding area from which the first sensor data are detected, and the surrounding area from which the second sensor data has been detected, at least partially overlap. To navigate the motor vehicle by means of the navigation system, a global position of the motor vehicle in GPS coordinates can be determined by means of GPS (Global Positioning System). The

Motor vehicle can thus be located globally by GPS. This global position of the motor vehicle can by means of the navigation system to the first

Measuring time are detected. If it has been detected by the navigation system that the motor vehicle is at the second measuring time at least in the vicinity of the global position, which held the motor vehicle at the first measuring time, for example in the same street, the second sensor data for identifying the object after the object-specific feature. Thus, computing time and computing power can be saved in an advantageous manner, since the

Sensor data will only be searched if a recognition of the object is likely. Particularly preferably, a parked other vehicle is recognized as the object and identified as the object-specific feature is a motor vehicle license plate, in particular at least one alphanumeric character of the motor vehicle license plate. In other words, this means that the license plates or license plates are detected by parked vehicles. These number plates can be detected continuously by means of the camera and their positions can be imaged in the two-dimensional plane or environment map. So it will be the exact positions or spatial positions of the

Number plates used to correct the continuous odometry errors using the SLAM method. This embodiment takes advantage of the fact that almost every vehicle has such an official license plate, so that it can be used reliably for the identification of the vehicle. These license plates can be identified particularly easily, for example, using pattern recognition. As these license plates are usually a

have predetermined orientation on the vehicle, the license plates can be used advantageously for reliable determination of the pose. For this purpose, in particular longitudinal edges of the license plate are recognized, and determines an angle between the longitudinal edges as a measure of the orientation angle of the motor vehicle to the other vehicle. In addition, in particular, the at least one alphanumeric character representing information about the vehicle is identified and the ID assigned to this alphanumeric character, so that the parked vehicle can be uniquely identified.

In particular, based on the first and / or the second pose of the motor vehicle with respect to the parked vehicle and based on predetermined geometric data describing the parked motor vehicle, at least one geometric dimension of a parking space adjacent to the parked vehicle is determined for the motor vehicle. In particular, transverse parking gaps can be detected on the basis of the pose of the motor vehicle with respect to the vehicle registration number. If the longitudinal edges of the license plate, for example, extend parallel to a vehicle longitudinal direction of the motor vehicle, the transverse parking space can be detected in a passing of the motor vehicle to the parked vehicles. Based on given

Geometry data, which in particular a length of the parked vehicle

Describe can be determined as the geometric dimension of the parking space, in particular a depth of the parking space thus. The particular pose can therefore be used advantageously in many ways by the driver assistance system. Alternatively or additionally, a milestone may be recognized as the object and a kilometer indication on the milestone may be identified as the object-specific feature. The milestones are usually attached to lane edges of lanes at regular intervals so that they can be used to determine the distance traveled and to determine location. These

Embodiment is particularly advantageous if the motor vehicle is regularly, for example daily, on a particular road section and thus regularly passes by the same milestone. The odometry data are thus recorded over a particularly long period of time, for example, between passing the milestone on a first day and passing the milestone on the following day. The odometry data are then recorded over the entire period, even if the motor vehicle has been turned off in the meantime.

It proves to be advantageous if in third sensor data, another milestone is detected and a further mileage is recorded on the further milestone, based on the mileage on the milestones a distance between the milestones is determined, one between the milestones of the

Vehicle covered route is detected by odometry and the

Odometer error is detected if a further deviation between the distance and the distance traveled exceeds a further predetermined threshold. This embodiment is based on the finding that the milestones are usually set up at regular intervals. By the mileage on the milestones in the form of numbers, the distance between two

Milestones are determined. In addition, using odometry,

in particular on the basis of a number of wheel revolutions which are determined by the motor vehicle between the milestones covered route. If there is no odometry error, the distance covered should correspond to the distance between the milestones. However, if the distance covered by odometry deviates from the distance between the milestones, then the odometry error is also detected. This is particularly advantageous when the motor vehicle passes every milestone only once and thus can not perform the Odometrievgleich the same milestone.

It can also be provided that a house is recognized as the object and, as the object-specific feature, a house number plate, in particular at least one digit on the house number plate, is identified. In particular, in this

Embodiment, the detection of the object, so here the house, linked to the performed by the navigation system of the motor vehicle position determination. For example, if it was detected by the navigation system that the

Motor vehicle repeatedly enters a certain street, so a specific house number can be recognized at a house in this street, and be performed on the basis of this house number odometry adjustment.

The invention also relates to a computing device for a driver assistance system of a motor vehicle, which is designed to carry out a method according to the invention. The computing device can be integrated, for example, in a vehicle-side control unit.

An inventive driver assistance system comprises an inventive

Computing device and a sensor device for detecting a

Surrounding area of the motor vehicle. The sensor device may comprise, for example, at least one laser scanner and / or at least one camera. The

Driver assistance system is designed in particular to perform a determination of the pose of the motor vehicle by means of a SLAM method and by means of odometry and thus to determine an environment map of the motor vehicle. Based on

Ambient map and the corrected Odometriedaten can be made possible in particular an at least semi-autonomous maneuvering of the motor vehicle.

For example, an at least semi-autonomous parking process for the motor vehicle can be carried out by the driver assistance system.

Moreover, the invention relates to a motor vehicle with a driver assistance system according to the invention. The motor vehicle is designed in particular as a passenger car.

The preferred embodiments presented with reference to the process according to the invention

Embodiments and their advantages apply correspondingly to the computing device according to the invention, the driver assistance system according to the invention and to the motor vehicle according to the invention.

Further features of the invention will become apparent from the claims, the figures and the description of the figures. The features and feature combinations mentioned above in the description, as well as those mentioned below in the description of the figures and / or features and combinations of features shown alone in the figures can be used not only in the particular combination specified, but also in other combinations or in isolation, without departing from the scope of the invention. Thus, embodiments of the invention are to be regarded as encompassed and disclosed, which are not explicitly shown and explained in the figures, however, emerge and can be produced by separated combinations of features from the embodiments explained. Embodiments and combinations of features are also to be regarded as disclosed, which thus do not have all the features of an originally formulated independent claim. Moreover, embodiments and combinations of features, in particular by the embodiments set out above, are to be regarded as disclosed, which go beyond or deviate from the combinations of features set out in the back references of the claims.

The invention will now be described with reference to preferred embodiments and with reference to the accompanying drawings.

Showing:

Fig. 1 is a schematic representation of an embodiment of a

motor vehicle according to the invention; and

Fig. 2 is a schematic representation of a detected by the motor vehicle

Camera image for determining poses.

In the figures are the same as well as functionally identical elements with the same

Provided with reference numerals.

Fig. 1 shows a motor vehicle 1 according to the present invention. The motor vehicle 1 is designed in particular as a passenger car. The motor vehicle 1 comprises a driver assistance system 2, which is designed to carry out a so-called simultaneous localization and mapping (SLAM) or "simultaneous localization and mapping" process

Driver assistance system 2 is designed to create an environmental map describing a surrounding area 3 of the motor vehicle 1 and at the same time estimate or determine a pose of the motor vehicle 1 in the surroundings map. The driver assistance system 2 has for this purpose a sensor device S, which in addition is designed to detect sensor data from the surrounding area 3. In the present case, the sensor device S has at least one camera 4. The

Sensor device S may also have at least one laser scanner. The cameras 4 are designed to capture the surrounding area 3 of the motor vehicle 1 in images. The cameras 4 form here an all-round camera system or surround-view camera system, wherein a first camera 4 is arranged in a front area 5 of the motor vehicle 1, a second camera 4 is arranged in a rear area 6 of the motor vehicle 1, a third camera 4 on a driver's side 7 of the

Motor vehicle 1 is arranged and a fourth camera 4 is arranged on a passenger side 8 of the motor vehicle 1. By means of the all-round camera system, in particular, the entire surrounding area 3 around the motor vehicle 1 can be captured in images.

The driver assistance system 2 here also comprises a computing device 9, which is designed to detect the pose of the motor vehicle 1, that is to say an orientation and a position of the motor vehicle, on the basis of the sensor data acquired by the sensor device S

I, with respect to an object O (see FIG. 2) in the surrounding area 3 of FIG

Motor vehicle 1 to determine. In addition, the driver assistance system 2 is designed to determine the pose of the motor vehicle 1 in the surrounding area 3 by odometry. For this purpose, the driver assistance system 2 can record data of a propulsion system of the motor vehicle 1, for example a number of revolutions of wheels 10 of the motor vehicle 1 and / or steering angle of the motor vehicle 1. Due to vehicle parameters, for example, due to changing and / or incorrectly given wheel geometries and Fahrwerkgeometrien, it may happen that the pose determination by means of odometry is faulty.

First of the sensor device S first sensor data from the

Surrounding area 3 of the motor vehicle 1 detected. Such first sensor data are shown in FIG. 2 in the form of a first camera image 11 of the camera 4 arranged in the front region 5 of the motor vehicle 1. The motor vehicle 1 is located on a roadway 12 and detects the first camera image during a journey at a first measurement time

I i. The camera 4 may, for example, have a fisheye lens, resulting in the perspective distortion shown in FIG. 2 in the camera image 1 1.

In the first camera image 1 1 are images of objects O, the objects O are formed in this case as parked, other vehicles 13 and as a house 14. The objects O are in particular at least temporarily stationary or immovable objects. The computing device 9 is designed to recognize the objects O in the camera image 1 1 detected by the camera 4 and to identify at least one object-specific feature M on the objects O. The respective object-specific feature can be clearly assigned to an object O, in particular. In the present case, a license plate 15 or license plate is identified as the object-specific feature M of the parked vehicles 13 and a house number plate 16 as the object-specific feature M of the house 14

identified.

To identify the license plate 15 and / or the house number plate 16, the captured image 1 1 of the arithmetic unit 9 of a pattern recognition in terms of corners and edges of the license plate 15 and / or the house number plate 16 are subjected. It can thus be searched in the recorded image 1 1 for a rectangle or a trapezoid. This method of detecting one of the license plate 15 and / or the house number plate 16 is on the one hand particularly reliable and on the other hand also relatively easy to implement. When recognizing the license plate 15 and / or the house number plate 16, first its edges can be detected, and based on the edges then the corners of the license plate 15 and / or the house number plate 16 can be identified. The detection of the edges can be done by means of a known edge detection method, according to which in the recorded image 1 1 for color transitions - for example, from black to white - is searched. If the edges have been detected, then the corners of the license plate 15 and / or the house number plate 16. After recognition of the license plate 15 and / or the house number plate 16, the detection can be made plausible by a length of the image of the license plate 15 and / or House number plate 16 is set in relation to its width, and it is checked whether this ratio is within a predetermined range of values. Thus, the license plate 15 and / or the house number plate 16 can be distinguished from other rectangles, since both the length and the width are fixed in particular for license plates and thus represent a standard size.

In addition, the computing device 9 is designed to use the object-specific features M to determine a respective first pose of the motor vehicle 1 relative to the object-specific features M. The first pose corresponds to the pose which the motor vehicle 1 has when capturing the first camera image 11 to the object-specific feature M. For determining the orientation on the basis of the image of License plate 15 and / or the house number plate 16 longitudinal edges of the license plate 15 and / or the house number plate 16 in Figure 1 1 can be identified and an angle between the longitudinal edges are determined. Depending on the angle, the orientation can then be determined, since the angle represents a measure of an orientation angle of the adjacent vehicle 13 and / or the house 14 with respect to the motor vehicle 1. Based on the

Posenbestimmung with respect to the object-specific features M, the SLAM method can be performed by the objects O and the motor vehicle are mapped in the particular two-dimensional map of the environment. In this case, the motor vehicle 1, for example, in the origin of a two-dimensional

Coordinate system, in which the environment map is determined to be positioned, resulting in positions of the objects O in the two-dimensional map of the environment by the carried out on the basis of the camera image 1 1 Posenbestimmung.

In addition, the computing device 9 each object O based on the

Object-specific feature M assign a unique identification number or ID and deposit them in a vehicle-mounted storage device. It may be advantageous to recognize the numbers or numbers and letters of the license plate 15 and based on which a unique

Identification number or ID matching the parked vehicle 13 is created. When the license plate 15 is detected and recognized again, the information that this is the same license plate 15 of the same vehicle 13 is provided through this ID. This information is especially important if only parts of the number plate were recognized and thus results in a scattering, if the same number plate 15 is recognized and detected again and again, but at different positions. This dispersion can thus be contained in order to converge faster with the SLAM. Furthermore, for example, only one digit can be picked out, which is to be detected to calculate the SLAM, whereby fewer detection points are necessary to calculate a SLAM, and thus run time can be saved.

Also, the computing device 9 can recognize a number on the house number plate 16 and assign it, in particular together with a street in which the house 14 is located, a specific identification number. The road can

For example, be determined by a navigation system of the motor vehicle 1, by which the motor vehicle 1 is located globally. Then, at a second measuring time, a further second camera image is taken by the camera 4 as second sensor data of the surrounding area 3. Now, if the motor vehicle 1 again passes by the objects O, for example because the motor vehicle 1 is in search of a parking space, the images of the objects O are also located in the further camera image. The computing device 9 can now search the further second image for the objects O by the computing device 9 searches the further image for the object-specific features M. If the

Computing 9, the object-specific feature M, for example, one of

License plates 15, has recognized in the further picture, so the object O, so the associated vehicle 13, clearly recognized and a second pose of the

Motor vehicle 1 to the object O determined. It can be provided that the computing device 9 searches the further camera image for the object-specific features M only if it was detected by means of the navigation system of the motor vehicle 1 that the motor vehicle 1 during the recording of the second

Camera image is located approximately in about the same surrounding area 3 as in the recording of the first camera image 1 1.

This second pose is compared with a pose which is determined by the computing device 9 by means of odometry. The pose determined by means of odometry results from the route which the motor vehicle 1 has traveled since the acquisition of the first image 11 until the acquisition of the further image. In other words, starting from the first pose using odometry data or data of the

Propulsion system of the motor vehicle 1, for example, a number of wheel revolutions and / or a steering angle, determines the distance covered and determines the end of the route as the pose with which the second pose is compared. So it is carried out an odometry adjustment. If there is no odometry error, ie if the underlying and recorded odometry data is the current, real data

would match the pose determined by odometry with the second pose. If the two poses do not match, then one becomes

Odometry error detected. Based on the size of the deviation can be a

Correction factor can be determined by means of which the odometry error can be corrected during the position determination of the motor vehicle by means of odometry.

In addition, the computing device 9 here based on the specific pose of the

Motor vehicle 1 to the license plate 15 of one of the vehicles 13 and based on at least one predetermined geometric dimension of the parked vehicle thirteenth determine a geometric dimension 17 of a parking space 18 of the motor vehicle 1. Based on a length of the vehicle 13, the computing device 9 can recognize, in particular, a depth of the parking space 18 as the geometric dimension 17. Based on the environment map and the corrected odometry data, the driver assistance system 2 can then provide or perform, for example, an at least semi-autonomous parking process for the motor vehicle 1 in the parking space 18.

Claims

claims

1 . Method for operating a driver assistance system (2) of a motor vehicle (1)

Detecting an object (O) in first sensor data detected by a vehicle-side sensor device (S) from a surrounding area (3) of the motor vehicle (1) and identifying an object-specific feature (M) on the object (0) in the first sensor data;

Determining a first pose of the motor vehicle (1) to the object-specific feature (M) on the basis of the object-specific feature (M);

marked by

Recognizing the object (O) in second sensor data acquired by the sensor device (S) on the basis of the object-specific feature (M);

- Determining a second pose of the motor vehicle (1) to the

object-specific feature (M) based on the object-specific feature (M);

Determining a pose of the motor vehicle (1) resulting from a movement of the motor vehicle (1) between the detection of the first and the second sensor data to the object-specific feature (M) by means of odometry;

Detecting an odometry error if a deviation between the resulting pose and the second pose exceeds a predetermined threshold.

2. The method according to claim 1,

characterized in that

in the presence of the odometry error, a correction value for compensating the odometry error is determined based on the second pose.

3. The method according to claim 1 or 2,

characterized in that

as the first and second sensor data, a first and a second image (11) of at least one camera (4) of the motor vehicle (1) are detected and / or as the first and second sensor data first and second laser scanner data from at least one laser scanner of the motor vehicle ( 1).

4. The method according to any one of the preceding claims,

characterized in that

as the object (O) a stationary object is detected.

5. The method according to any one of the preceding claims,

characterized in that

after identifying the object-specific feature (M) the object (O) is assigned an identification number which in a vehicle-side

Storage device is deposited.

6. The method according to any one of the preceding claims,

characterized in that

the second sensor data for recognizing the object (O) are only searched for the object-specific feature (M), if it was detected on the basis of data of a navigation system of the motor vehicle (1) that the

Surrounding area (3) from which the first sensor data was detected, and the surrounding area (3) from which the second sensor data was acquired; at least partially overlap.

7. The method according to any one of the preceding claims,

characterized in that

as the object (O) a parked other vehicle (13) is detected and as the object-specific feature (M) a motor vehicle license plate (15), in particular at least one alphanumeric character of the motor vehicle license plate (15) is identified.

8. The method according to claim 7,

characterized in that

at least one geometric dimension (17) of a parking space adjacent to the parked vehicle (13) on the basis of the first and / or the second pose of the motor vehicle (1) with respect to the motor vehicle license plate (15) and based on predetermined geometric data describing the parked vehicle (13) (18) for the motor vehicle (1) is determined.

9. The method according to any one of the preceding claims,

characterized in that

as the object (O) a milestone is recognized and identified as the object-specific feature (M) is a kilometer on the milestone.

10. The method according to claim 9,

characterized in that

in third sensor data a further milestone is detected and a further mileage is recorded on the further milestone, based on the mileage on the milestones a distance between the

Milestones is determined, a covered between the two milestones of the motor vehicle (1) distance is detected by odometry and the odometer error is detected if a further deviation between the distance and the distance traveled exceeds a further predetermined threshold.

1 1. Method according to one of the preceding claims,

characterized in that

as the object (O) a house (14) is recognized and as the object-specific feature (M) a house number plate (16), in particular at least one digit on the house number plate (16) is identified.

12. computing device (9) for a driver assistance system (2) of a motor vehicle (1), which is adapted to a method according to one of the preceding

Perform claims.

13. Driver assistance system (2) for a motor vehicle (1) with a sensor device (S) for detecting an environmental region (3) of the motor vehicle (1) and with a computing device (9) according to claim 12.

14. Motor vehicle (1) with a driver assistance system (2) according to claim 13.