Methods for locating a vehicle on a lane and vehicle

Abstract

The invention relates to a method for locating a vehicle (1) on a lane (LS) of a multi-lane road (2) taking into account the GNSS position of the vehicle (1), information read from a digital road map (3), and the preceding lanes (LS) detected by the vehicle (1) using vehicle sensors (4) based on environmental features. The method according to the invention is characterized by the following process steps: - Determining the GNSS position of the vehicle (1) and reading the number of lanes for the road segment corresponding to the GNSS position from the digital road map (3); - Determining at least one left and / or right lateral distance (LA-L, LA-R) of the vehicle (1) to a left and / or right road edge boundary (5) from the sensor data supplied by the vehicle sensors (4); - Determining the number of lanes (FS) to the left and right of the Ego lane (FS-Ego) driven on by the vehicle (1) from the lateral distance (LA-L, LR-R) and lane width information; - Generation of a lane model (6), representing a virtual representation of the lane alignment of the road (2); and - Feeding the lane model (6) to a downstream driver assistance system as input data.

Description

[0001] The invention relates to a method for locating a vehicle on a lane of a multi-lane road according to the type defined in more detail in the preamble of claim 1, and to a vehicle for carrying out the method.

[0002] Modern vehicles are equipped with a wide variety of driver assistance systems to enhance driving safety and comfort. For example, adaptive cruise control allows the vehicle to automatically maintain a specific distance to the vehicle ahead by taking over longitudinal control. Lateral control can also be assisted, for instance, by lane keeping assist or lane change assist. Some driver assistance systems require a virtual representation of the vehicle's surroundings, including the lane markings of the road, to function correctly.

[0003] It is known to determine the lane layout ahead using machine vision. For this purpose, the section of road ahead of the vehicle is captured by a forward-facing camera. The camera images are processed, and the respective lanes are identified, taking into account any detected lane markings. Typically, the vehicle's own lane, as well as the adjacent lanes to its left and right, if present, are recognized. However, such systems reach their limits with more than three lanes, increasing the risk of not correctly detecting all lanes. Consequently, it is not possible to determine which of the four or more lanes the vehicle is traveling in.

[0004] Modern vehicles can be equipped with an integrated navigation system that determines their position by comparing a GNSS position with a digital road map, for example, using GPS. This digital road map can assign a number of lanes to each road segment. While high-resolution maps, also known as HD maps, store extensive information such as the precise lane layout—that is, the underlying geometry of each lane—this is not the case with low-resolution maps, also known as SD maps. However, such HD maps are not yet widely available. It is expected that in the near to medium term, such HD maps will only be available for critical infrastructure areas.Furthermore, a reliable and particularly accurate GNSS signal is required to accurately determine the lane being traveled based on HD maps.

[0005] German patent DE 10 2016 112 913 A1 discloses a method and a device for determining a vehicle's self-position. The method involves identifying environmental features in the vicinity of a vehicle. Coordinates of these features are determined from a digital road map within a reference coordinate system. Taking into account the respective viewing angles and object sizes of the environmental features, the vehicle's position is then determined within the reference coordinate system, similar to position determination using a compass and a paper map. This allows the vehicle to be located within its lane and this information to be provided to downstream driver assistance systems for deriving automated driving maneuvers. A disadvantage, however, is the requirement to include corresponding coordinates of environmental features in the map data. Such information is often unavailable or...The underlying file size of the digital road map increases disproportionately. Furthermore, locating the vehicle relative to the recognized environmental features is technically complex due to the need to take a bearing.

[0006] From DE 10 2015 001 386 A1, a method for determining the lateral position information of a motor vehicle on a roadway is known, wherein radar data describing at least a portion of the roadway is acquired with at least one radar sensor of the motor vehicle. By evaluating the radar data, environmental features describing the position of a roadway boundary are detected and located. From these environmental features, the course of the roadway boundaries and the lateral distances of the motor vehicle to the roadway boundaries are determined. The lateral position information is then determined as a function of the lateral distances of the motor vehicle to the roadway boundaries.

[0007] German patent application DE 10 2010 033 729 A1 discloses a method for determining the position of a vehicle on a roadway. According to the method, data from a satellite signal sensor, a digital map, a line detection sensor, a vehicle dynamics sensor, and / or an environment sensor are combined in such a way that the vehicle's position on the roadway is determined with at least lane-specific accuracy. In particular, the method can also be used to determine the number of lanes.

[0008] The present invention is based on the objective of providing an improved method for locating a vehicle on a lane of a multi-lane road, which allows a reliable determination of the lane travelled by the vehicle on the respective road using simple map material and can be implemented with minimal technical effort.

[0009] According to the invention, this problem is solved by a method for locating a vehicle on a lane of a multi-lane road with the features of claim 1. Advantageous embodiments and further developments, as well as a vehicle for carrying out the method, are described in the dependent claims.

[0010] A generic method for locating a vehicle on a lane of a multi-lane road, taking into account the vehicle's GNSS position, information read from a digital road map, and the preceding lanes detected by the vehicle using vehicle sensors based on environmental features, is further developed by the following process steps: - Determining the vehicle's GNSS position and reading the number of lanes for the road segment corresponding to the GNSS position from the digital road map; - Determine at least one left and / or right lateral distance of the vehicle to a left and / or right edge of the roadway from the sensor data supplied by the vehicle's sensors; - Determining the number of lanes to the left and right of the vehicle's ego lane from the lateral distance and lane width information; - Generating a lane model representing a virtual map of the lane layout of the road, in particular including the lanes detected by the vehicle sensors and those extending beyond the lanes detected by the vehicle sensors; and - Feeding the lane model to a downstream driver assistance system as input data.

[0011] The method according to the invention makes it possible to reliably determine the ego lane traveled by the vehicle on a multi-lane road. This is also possible using low-resolution map data, i.e., SD cards. The technical effort required to implement the method according to the invention is minimal.

[0012] A lane profile is determined using the vehicle's sensors to detect the lanes ahead. The vehicle's sensors are designed to detect its own lane and, at most, one lane to its right and left, but not, for example, four, five, six, or even more lanes. For a complete lane model, it is essential to also know the lanes not detected by the vehicle's sensors. In other words, the lane profile can be determined based on the lanes detected by the vehicle's sensors, and a complete model of the road ahead can be generated based on the number of lanes determined using this method. In a sense, the lanes detected by the vehicle's sensors are supplemented by virtual lanes derived from the number of additional lanes to the left and right of the vehicle, which are not detected by the vehicle's sensors.The additional lanes adapt to the course of the lane detected by the vehicle's sensors. For example, the vehicle's sensors detect three lanes: the ego lane and one lane to its left and one to its right. However, the process then identifies four lanes to the left and right of the ego lane; that is, three additional lanes are added to the lane model on each side, in addition to the lanes detected by the vehicle's sensors.

[0013] The GNSS position can be a geoposition determined by GPS, Galileo, Beidou, GLONASS, or similar systems. To determine the GNSS position, the vehicle is equipped with suitable positioning devices, such as a GPS receiver. By comparing the GNSS position to a map, the road segment traveled by the vehicle can be determined. Various pieces of information are assigned to this road segment in the digital road map, including at least the number of lanes.

[0014] Using suitable vehicle sensors, the vehicle determines the distance to the left and / or right edge of the roadway. Taking this lateral distance(s) and the lane width information into account, it is possible to determine which lane of the multi-lane road the vehicle is traveling in. The lane width information indicates the width of each lane. The lanes can all have the same width or vary in width. Lane width information can be measured, particularly using the vehicle sensors. Alternatively, a typical standard lane width can be assumed, or the respective lane width information can be read from the digital road map, if known.

[0015] The road or section of road being traveled on by the vehicle comprises at least one lane, or possibly more, such as two, three, four, five, or even more lanes. Depending on which lane the vehicle is in, there may be a corresponding number of lanes to the left or right of the vehicle's lane. These lanes can include not only regular traffic lanes, but also, for example, highway on-ramps or off-ramps, or hard shoulders.

[0016] Furthermore, a lane model is generated, depicting the lane layout. This lane model describes not only the geometric arrangement of the lanes in the respective road segment, but also which of these lanes is the ego lane. The lane model is then fed as input data to downstream driver assistance systems, ensuring their reliable operation.

[0017] For example, the lane width information indicates that all lanes are 2.5 meters wide. The digital road map shows that there are four lanes. The left lateral distance is measured and yields a value of 6.75 meters. This allows the vehicle to be assumed to be in the third lane from the left. This lane is then designated as the ego lane.

[0018] According to the invention, the at least one lateral distance is determined based on camera images generated by at least one side camera, wherein the detection area of ​​the side camera is oriented at least partially in the transverse direction of the vehicle. With the aid of corresponding side cameras, the left or right edge of the roadway, viewed in the transverse direction of the vehicle, can thus be visually detected. The visual determination of the lateral distance can also be fused with distance determination based on depth information. This allows for a particularly reliable distance determination. The respective side cameras can, for example, be integrated into a side mirror of the vehicle or at another suitable location on the vehicle's exterior. It is thus possible to determine the lateral distance by processing camera images generated by a respective side camera, which will be discussed in more detail below.

[0019] The system envisages identifying a recurring feature of the road edge boundary in the camera images, based on the direction of travel. The lateral distance is then determined by considering a predefined standard longitudinal distance of the feature in the direction of travel, the vehicle's speed, and a detection frequency for the feature. Road networks often contain infrastructure objects used as road edge boundaries, such as guardrails, concrete walls, warning beacons, pylons, traffic cones, and the like. These objects are characterized by specific dimensions and are installed at defined intervals, depending on the region. The frequency with which these characteristic visual features appear in camera images varies depending on the vehicle's lateral distance from the objects and its speed.For different combinations, corresponding characteristic frequencies can then be stored in a data memory in the vehicle as a reference. This makes it possible to determine the lateral distance.

[0020] Each side camera has a defined detection range relative to its surroundings. If the vehicle is relatively close to the edge of the road, then relatively few features, such as one to four, are visible in a single camera image. As the vehicle speeds up, the time it takes for each feature to move across the camera image decreases. The detection frequency thus increases with the vehicle's speed. Conversely, if the vehicle is farther from the edge of the road, more features are visible in the camera image due to the distance, for example, five to ten features. Here, too, the time it takes for each feature to move across the camera image decreases with increasing speed.Taking into account the fixed relationship between lateral distance, locomotion speed, standard longitudinal distance and detection frequency of the feature, the lateral distance can be determined if locomotion speed, standard longitudinal distance and detection frequency are known.

[0021] An advantageous further development of the method according to the invention provides that the at least one lateral distance is determined by processing depth information, in particular depth information generated by an ultrasonic sensor system, a radar sensor system, and / or a laser scanner. Various sensor types are known that allow the determination of depth information. This makes it possible to determine a relative distance between the vehicle and the edge of the road. Taking into account the installation position of the respective sensor on the vehicle and its orientation relative to the surroundings, the distance to the edge of the road extending orthogonally to the direction of travel in the transverse direction of the vehicle can be determined from this relative distance. This allows the vehicle to be located relatively accurately perpendicular to the road.

[0022] Guardrails are particularly preferred as road edge boundaries, and the posts supporting the guardrail are considered a recurring feature. Guardrails are installed especially on well-developed road sections such as highways, expressways, federal roads, dangerous urban sections, and the like. Thus, guardrails are found on a high proportion of the roads in the road network. Furthermore, the guardrail posts are relatively easy to identify as a visual feature. This allows for reliable feature detection regardless of visibility conditions influenced by, for example, ambient brightness or weather. The reliable implementation of the method according to the invention is therefore ensured. The frequency at which the respective cameras generate camera images is relatively high and is in the range of, for example, 30 to 120 Hertz.The high sampling rate reduces the risk of overlooking relevant features.

[0023] A further advantageous embodiment of the method according to the invention further provides that: the camera images captured by the side camera are mapped into the frequency domain, in particular using Fast Fourier Transform; Depending on the speed of movement, a search frequency interval is set in the frequency range; The search frequency interval is examined for a dominant frequency signal, which is characterized by the fact that the dominant frequency signal falls below a defined frequency uncertainty threshold; and The lateral distance is determined from the height of the dominant frequency signal, the speed of movement, and the standard longitudinal distance.

[0024] According to the invention, the repeating feature of the road edge is mapped into the frequency domain. This allows for particularly reliable detection of the repeating feature and the determination of the lateral distance, taking the relevant parameters into account. In particular, using a Fast Fourier Transform (FFT), a computationally very efficient calculation with a short execution time is possible. The standard longitudinal distance is known for each object representing the road edge. Considering different regions, the respective standard longitudinal distances can be stored for different objects. The vehicle's speed can be read via a bus system and determined, for example, based on wheel rotation speed. The dominant frequency signal is derived accordingly from the camera images converted by FFT.Various relationships between respective frequency signals depending on the speed of movement and the standard longitudinal distance to respective lateral distances are stored, for example in the form of a database or table, which reliably allows the lateral distance to be determined.

[0025] According to a further advantageous embodiment of the method according to the invention, a side camera with a fisheye lens is used. This increases the camera's field of view, thereby increasing the probability of detecting relevant features of road edge boundaries. This improves the reliability of determining the lateral distance based on camera images and ultimately further enhances the implementation of the method according to the invention. Particularly in situations where the view of the road edge boundary is partially obstructed by adjacent vehicles, determining the lateral distance based on the detection frequency of the feature allows for a more reliable and accurate distance determination than a distance determination based on the number of recognizable features, especially when using a fisheye lens to enlarge the field of view.

[0026] A further advantageous embodiment of the method according to the invention provides that the road section ahead of the vehicle is captured with a front camera, lane markings are identified in the camera images generated by the front camera using machine vision methods, and at least the ego lane is determined based on the identified lane markings. This allows sensor fusion of the information read from the digital road map based on the GNSS position and the lanes determined by machine vision. In this way, it can be determined with particular reliability in which ego lane the vehicle is located. As already mentioned at the outset, the detection capability of lanes based on machine vision reaches its limits with an increasing number of lanes. However, at least the ego lane traveled by the vehicle can be identified.Furthermore, the left and right lanes immediately adjacent to the Ego lane are identifiable, if present. However, if the road section has more than three lanes, machine vision alone cannot accurately determine which lane is the Ego lane. This is possible, however, by combining it with the previously described steps of the inventive method. Considering the lanes identified based on machine vision thus increases the reliability of identifying the correct lane as the Ego lane.

[0027] For example, machine vision can determine that the vehicle is in the middle of three adjacent lanes. Taking the lateral distance into account, it can be determined that the vehicle is in the second lane from the right on a four-lane road. This allows the system to determine that the left and right adjacent lanes detected in the camera images generated by the front camera are the first and third lanes, counting from right to left across the roadway. Objects detected in these lanes can be integrated into the lane model, providing further information to downstream driver assistance systems.

[0028] It is further preferred that, if the number of lanes determined from the digital road map for the road segment is equal to or less than the maximum number of lanes identifiable using machine vision methods, the lane model is generated solely based on the lanes determined using machine vision methods. This further increases the computational efficiency in determining the ego lane. As described above, the detection capability of lanes based on machine vision of the camera images generated by the front camera is limited. Typically, however, the ego lane and the lanes immediately to its left and right are recognizable.If the digital road map at the GNSS position indicates that the road segment has only one to three lanes, the ego lane can be reliably determined based solely on machine vision, relative to the other lanes of the road segment. The additional determination of the ego lane taking lateral distance into account is then unnecessary, but can be performed optionally to increase reliability.

[0029] A further advantageous embodiment of the method according to the invention provides that the following is used as a downstream driver assistance system: - an environment representation device is used, wherein the environment representation device provides for the display of a virtual replica of the vehicle environment in the form of the lane model on a display device in which at least the vehicle and the lanes of the road section are displayed, wherein the vehicle is shown driving on the lane on which it has been located; - a navigation system is used that calculates driving instructions based on the lane model; and / or - a system is used to derive at least partially automated control commands for the vehicle based on the lane model.

[0030] This ensures the reliable operation of the driver assistance systems described above. A corresponding environmental representation allows the driver to better perceive the vehicle's surroundings, even in hectic, confusing, or, in short, "difficult situations." Road users detected by the vehicle can also be highlighted in this environmental representation. In particular, this facilitates the detection of road users in the blind spot, thus reducing the potential for accidents. Furthermore, a corresponding environmental representation supports the driver when transferring control between manual and semi-automated or even autonomous driving modes.

[0031] Furthermore, navigation guidance can be active in the vehicle. Taking the lane model into account, corresponding driving instructions can then be displayed particularly reliably. This reduces the risk of the driver accidentally driving in the wrong lane, especially when a turn is required that involves driving down from the current section of road.

[0032] Since the ego lane is determined with particular reliability according to the invention, reliable control commands for the vehicle can also be generated. In particular, this allows the vehicle to reliably perform a lane change, at least semi-automatically or even autonomously.

[0033] In a vehicle of this type, comprising positioning means, vehicle sensors, and a processing unit, the positioning means, the vehicle sensors, and the processing unit are configured, according to the invention, to carry out a method described above. The vehicle can be any road vehicle, such as a car, truck, van, bus, or the like. As positioning means, the vehicle has at least one suitable GNSS receiver, which allows the determination of the vehicle's position in the form of geocoordinates by processing a signal transmitted by a global navigation satellite system. This makes it possible to locate the vehicle on a digital road map stored on the processing unit. Thus, the road segment traveled by the vehicle can be determined on the digital road map.Vehicle sensors can include a wide variety of sensor systems, such as cameras, laser scanners like LiDAR, radar sensors, ultrasonic sensors and the like.

[0034] The computing unit has at least read access to a computer-readable storage medium, containing machine-interpretable instructions which, when executed by a processor of the computing unit, cause it to implement the individual process steps of the procedure described above.

[0035] Further advantageous embodiments of the inventive method for locating a vehicle on a lane of a multi-lane road and of the inventive vehicle are also evident from the exemplary embodiments, which are described in more detail below with reference to the figures.

[0036] This shows: Fig. 1 a schematic representation of a vehicle traveling on a multi-lane section of road; Fig. 2 a flowchart of a method according to the invention for locating the vehicle on the road; Fig. 3 a schematic detailed representation of the various in Fig. 2 process steps shown; and Fig. 4 a schematic representation of an application of the ego lane traversed by the vehicle, determined by the method according to the invention, by a driver assistance system.

[0037] Fig. Figure 1 shows a vehicle 1 in a top view, traveling on a road 2. The road 2 has several lanes FS. The vehicle 1 also has vehicle sensors 4 and a processing unit 15 for processing the corresponding sensor data. The vehicle sensors 4 include a wide variety of sensor systems, such as a front camera, side cameras, laser scanners, ultrasonic sensors, radar sensors, and the like. The processing unit 15 is capable of recognizing static and / or dynamic objects in the environment. This can be used to provide various driver assistance functions.

[0038] Typically, such vehicles 1 are equipped with a front camera to capture the road ahead. These systems are usually able to recognize lane markings by processing the camera images using machine vision methods and, based on this, to determine the corresponding lane paths. Usually, the ego lane (FS-Ego) driven by the vehicle 1 itself, as well as the adjacent lanes to its left and right, are identifiable. These three lanes are in Fig. 1 shown in white.

[0039] In the illustrated embodiment, however, road 2 has four lanes FS. The in Fig. The leftmost lane FS, highlighted by hatching, is therefore not recognizable by vehicle 1 in the embodiment shown.

[0040] However, for correct functioning, corresponding driver assistance systems require information about the lanes FS actually present on the road segment being traveled and information about which of these lanes FS-Ego the vehicle 1 is currently located in. Using a method according to the invention, this information can be provided reliably and with minimal technical effort.

[0041] The procedure is described using a combination of Fig. 2 and Fig. 3 explained.

[0042] In step 201, vehicle 1 uses its front camera to record the road section ahead. Fig. 3a) shows a corresponding camera image in the upper area 12. As described above, corresponding lane paths are recognized using machine vision, which are in Fig. 3a) are shown below camera image 12. This display also includes detected vehicles ahead. As shown in Fig. As indicated by the arrow pointing from step 202 to step 201, if three or fewer lanes FS are visually detected, it may be sufficient to identify the lanes FS solely by machine vision.

[0043] In a Fig. In step 202 (3b), as shown, vehicle 1 determines a GNSS position, for example in the form of geocoordinates, using suitable positioning equipment. By comparing this GNSS position with a digital road map 3, vehicle 1 is able to determine the road 2 of a road network it is traveling on. The digital road map 3 contains information such as the number of lanes (FS) assigned to each road segment.

[0044] In step 203, vehicle 1 uses a left and right side camera to record the Fig. 3c) shown camera images 7. In the illustrated embodiment, the side cameras are each equipped with a fisheye lens. In the camera images 7, the processing unit 15 searches for a corresponding road edge boundary 5, here in the form of a guardrail. The detection of the presence of the road edge boundary 5 takes place in the Fig. 3d) shown step 204. In addition, the presence of characteristic features 8 of the road edge 5, which repeat in the direction VR of the road 2, is investigated, here in the form of the posts with which the guardrail is erected. For this purpose, respective camera images 7 are transformed into the frequency domain 9, in particular by means of FFT.

[0045] In step 205, a search frequency interval 10 in the frequency range 9 is determined based on the speed of movement of vehicle 1. Fig. Figure 3e) shows two exemplary representations of the frequency range 9, in which corresponding frequency signals are shown for a lateral distance of the vehicle 1 to the road edge 5 of five meters and thirty meters, respectively. The closer the vehicle 1 is to the road edge 5, the fewer posts are present in the respective camera image 7, and thus correspondingly fewer signal peaks are included.

[0046] In the Fig. In step 206 (3f), the search frequency interval 10 is now examined for a dominant frequency signal 11. The dominant frequency signal 11 is characterized by the fact that its amplitude falls below a defined frequency uncertainty threshold.

[0047] How Fig. As shown in Figure 3g), in step 207 the lateral distance of vehicle 1 to the road edge boundary 5 is calculated as a function of the standard longitudinal distance of features 8 in the direction of travel VR, the speed of movement of vehicle 1 and the corresponding detection frequency of feature 8 in frequency range 9 or the search frequency interval 10. In this case, the left lateral distance LA-L is calculated.

[0048] The lateral distance LA-L and LA-R is in Fig. 3h) is shown for step 208. Here, vehicle 1 is positioned relative to road 2 in a transverse direction. The lanes FS-Ego and FS-N, detected based on machine vision, are indicated by hatching.

[0049] In step 209, the lateral distance LA-R, LA-L determined in this way is fused with the lane detection resulting from machine vision. This shows Fig. 3i) Various scenarios where road 2 has a different number of lanes. Using machine vision, the vehicle 1's own ego lane FS-Ego and the immediately adjacent lanes FS-N to the left and right are always highlighted. Taking into account the number of lanes read from the digital road map 3, the lateral distances LA-R, LA-L, and the lane path recognized by machine vision, a Fig. 4. Generate the lane model 6 shown. Information is now available indicating that the lanes FS highlighted by horizontal hatching are also present.

[0050] Fig. Figure 4 shows the use of the lane model 6 to generate a virtual replica 13 of the environment. This virtual replica 13 can be displayed on an environment representation device. In particular, route navigation can be performed in vehicle 1, with corresponding driving instructions 14 being displayed contact-analogously on the lane FS that vehicle 1 is actually to take. This reduces the risk of the driver of vehicle 1 choosing the wrong lane. Taking the lane model 6 into account, the computing unit 15 could also derive at least semi-automated or even autonomous control commands for vehicle 1, for example, to perform an automated lane change. In the lower part of Fig.Figure 4 shows that the lanes FS detected by the method according to the invention are also highlighted by horizontal hatching. This indicates that in the driving situation shown, all lanes FS of road 2 were correctly detected, and the Ego lane FS-Ego was correctly determined.

[0051] The vehicle 1 and the method according to the invention are characterized by a reliable determination of the ego lane FS-Ego being traveled by the vehicle 1 on a multi-lane road 2. The method can be implemented with minimal technical effort. Most vehicles 1 already possess the necessary hardware components, allowing for easy integration even into existing vehicles. The execution of the method according to the invention is particularly independent of the prevailing weather and lighting conditions. Thanks to sensor fusion, the method according to the invention can be combined with other known methods for lane identification, thus creating system redundancy. Instead of guardrails, other standardized objects, such as concrete barriers, can be used as references.

Claims

[1] Method for locating a vehicle (1) on a lane (LS) of a multi-lane road (2) taking into account a GNSS position of the vehicle (1), information read from a digital road map (3) and the preceding lanes (LS) detected by the vehicle (1) using vehicle sensors (4) based on environmental features, by the following process steps: - Determining the GNSS position of the vehicle (1) and reading the number of lanes for the road segment corresponding to the GNSS position from the digital road map (3); - Determining at least one left and / or right lateral distance (LA-L, LA-R) of the vehicle (1) to a left and / or right road edge boundary (5) from the sensor data supplied by the vehicle sensors (4); - Determining the number of lanes (FS) to the left and right of the Ego lane (FS-Ego) driven on by the vehicle (1) from the lateral distance (LA-L, LR-R) and lane width information; - Generation of a lane model (6), representing a virtual representation of the lane alignment of the road (2); and - Feeding the lane model (6) to a downstream driver assistance system as input data, characterized in that the at least one lateral distance (LA-L, LA-R) is determined based on the camera images (7) generated by at least one side camera, wherein the detection area of ​​the side camera is oriented at least partially in the transverse direction of the vehicle, wherein a feature (8) of the lane edge boundary (5) that repeats in the direction of travel (VR) of the roadway is identified in the camera images (7), and the lateral distance (LA-L, LA-R) is determined taking into account a predetermined standard longitudinal distance of the feature (8) in the direction of travel (VR), the speed of travel of the vehicle (1) and a detection frequency of the feature (8). [2] Method according to claim 1, characterized by, that at least one lateral distance (LA-L, LA-R) is determined by processing depth information, in particular depth information generated by an ultrasonic sensor system, a radar sensor system and / or a laser scanner. [3] Method according to claim 1 or 2, characterized by , that a guardrail is taken into account as the edge of the roadway (5) and the posts with which the guardrail is erected are taken into account as a repeating feature (8). [4] Method according to any one of claims 1 to 3, characterized by , that the camera images (7) recorded by the side camera are mapped into the frequency domain (9), in particular by means of Fast Fourier Transform; a search frequency interval (10) is set in the frequency domain (9) depending on the speed of movement; the search frequency interval (10) is examined for a dominant frequency signal (11) which is characterized by the fact that the dominant frequency signal (11) falls below a defined frequency uncertainty threshold; and the lateral distance (LA-L, LA-R) is determined from the height of the dominant frequency signal (11), the speed of movement and the standard longitudinal distance. [5] Method according to any one of claims 1 to 4, characterized by that a side camera is used which has a fisheye lens. [6] Method according to any one of claims 1 to 5, characterized by , that the section of road ahead of the vehicle (1) is recorded with a front camera, lane markings are identified in the camera images (12) generated by the front camera using machine vision methods and at least the Ego lane (FS-Ego) is determined based on the identified lane markings. [7] Method according to claim 6, characterized by , that if the number of lanes determined from the digital road map (3) for the road section is equal to or less than a maximum number of lanes identifiable by machine vision methods, the lane model (6) is generated solely on the basis of the lanes (FS) determined by machine vision methods. [8] Method according to any one of claims 1 to 7, characterized by , that as a downstream driver assistance system: - an environment representation device is used, wherein the environment representation device provides for the display of a virtual replica (13) of the vehicle environment in the form of the lane model (6) on a display device in which at least the vehicle (1) and the lanes (FS) of the road section are displayed, wherein the vehicle (1) is shown driving on the lane (FS-Ego) on which it has been located; - a navigation system is used that calculates driving instructions (14) based on the lane model (6); and / or - a system for deriving at least partially automated control commands for the vehicle (1) based on the lane model (6) is used. [9] Vehicle (1) comprising positioning means, vehicle sensors (4) and a computing unit (15), characterized by , that the position determination means, the vehicle sensor system (4) and the computing unit (15) are configured to carry out a method according to one of claims 1 to 8.

Images