DE102016213783A1 - A method, apparatus and computer readable storage medium having instructions for determining the lateral position of a vehicle relative to the lanes of a lane - Google Patents

Abstract

A method, apparatus (20, 30), and computer readable storage medium having instructions for determining the lateral position of a vehicle relative to the lanes of a lane. An image processing unit (22) detects (10) lane markings. A position determining unit (23) determines (11) a home position for the vehicle. An evaluation unit (24) determines (12) possible lateral positions of the vehicle by comparing the detected lane markings with lane marking information from a lane geometry map for the starting position determined for the vehicle. In this case, at least a first set of possible lateral positions for a first detected lane marking and a second set of possible lateral positions for a second detected lane marking are determined. Furthermore, the evaluation unit () determines (13) a final lateral position of the vehicle by applying at least one evaluation function to the possible lateral positions.

Description

Field of the invention

The present invention relates to a method, an apparatus and a computer-readable storage medium having instructions for determining the lateral position of a vehicle relative to the lanes of a lane, and in particular for determining a relative lateral position with sub-lane accuracy. The present invention further relates to a vehicle with such a device.

Background of the invention

Modern vehicles are becoming increasingly autonomous, i. The vehicles provide the driver with more and more functions and systems that assist him with control of the vehicle by indicating or taking over parts of the vehicle control. Such functions and systems require a great deal of information about the vehicle and its surroundings.

For example, the function "lane-exact navigation" requires knowledge of the lane on which the vehicle to be navigated, the "ego vehicle", is located. This lane is also referred to as "ego lane". Furthermore, for the automatic driving and car-to-car-based applications in addition to the knowledge of the ego lane also more detailed information regarding the cross-placement of the ego vehicle relative to the ego lane needed. It must be known at all times with sub-lane accuracy where the ego vehicle is laterally in relation to the lane.

The publication US 2014/0358321 A1 discloses a method for recognizing and tracking the boundaries of a traffic lane. The method uses maps with road geometry information, GPS data, historical data, and other vehicle locations to determine current position.

The publication EP 2 899 669 A1 describes a method for determining the lateral position of a vehicle relative to the lane of a road. With the help of a camera, geometric aspects of the lanes are determined, for example lane markings. The determined aspects are classified and used for position determination. The classification requires training of the classification unit.

The publication DE 10 2012 104 786 A1 describes a system for accurately estimating a lane in which a vehicle is traveling. A tracker system provides estimated tracks that are determined in different ways. Examples are lane markers detected by a camera, a leading vehicle or GPS / maps which are accurate down to the track level. The estimated tracks are provided with confidence information. The estimated traces and the corresponding confidence information are merged to give a detected trace.

In summary, essentially three approaches to the determination of the lateral position of a vehicle relative to a roadway are currently being pursued.

A first approach is to use a highly accurate digital lane geometry map with absolute centimeter accuracy combined with a high accuracy dual frequency GPS system. Here, the position in the map with the help of the GPS sensor is determined without further imaging sensors. Due to GPS and map inaccuracies regarding the absolute position, however, it is often impossible to associate the ego vehicle with the correct lane. In addition, a solution with a high-precision map and high-precision GPS is very cost-intensive.

Another approach is to use imaging sensors, e.g. a camera system. In this way, the assignment of the ego vehicle to lanes is made possible based on lanes determined by the sensor system. The use of imaging sensors without the simultaneous use of a digital map, however, means that often only one or two lanes are detected by the sensors. The positioning of the ego vehicle can then take place only relative to the recognized lanes, but not relative to all lanes.

A third approach combines imaging sensors with map information regarding the number and type of lane marking. By using imaging sensors and the information from a digital map, how many lanes are present and with which edge marking they are present (dashed, solid, ...), the ego vehicle can be assigned to all lanes. Due to distance errors of the lane markings detected by the sensors relative to the ego vehicle, however, the accuracy of the cross-placement with respect to the corresponding lane is not sufficient for the above-mentioned applications.

Summary of the invention

It is an object of the invention to provide a method and apparatus for determining a lateral position of a vehicle relative to the lanes of a lane, which allow determination of the relative lateral position with sub-lane accuracy.

This object is achieved by a method having the features of claim 1, by an apparatus having the features of claim 10, and by a computer-readable storage medium having instructions according to claim 11. Preferred embodiments of the invention are the subject of the dependent claims.

• – Detektieren von Fahrbahnmarkierungen;
• – Bestimmen einer Ausgangsposition für das Fahrzeug;
• – Ermitteln von möglichen lateralen Positionen des Fahrzeugs durch Vergleichen der detektierten Fahrbahnmarkierungen mit Fahrbahnmarkierungsinformationen aus einer Fahrstreifengeometriekarte für die für das Fahrzeug bestimmte Ausgangsposition, wobei zumindest eine erste Menge von möglichen lateralen Positionen für eine erste detektierte Fahrbahnmarkierung und eine zweite Menge von möglichen lateralen Positionen für eine zweite detektierte Fahrbahnmarkierung ermittelt werden; und
• – Bestimmen einer abschließenden lateralen Position des Fahrzeugs durch Anwendung zumindest einer Bewertungsfunktion auf die möglichen lateralen Positionen.

According to a first aspect of the invention, a method for determining a lateral position of a vehicle relative to the lanes of a roadway comprises the steps of:

• - detecting lane markings;
• Determining a starting position for the vehicle;
• Determining possible lateral positions of the vehicle by comparing the detected lane markings with lane marking information from a lane geometry map for the vehicle determined starting position, wherein at least a first set of possible lateral positions for a first detected lane marker and a second set of possible lateral positions for a detected second detected lane marking; and
• Determining a final lateral position of the vehicle by applying at least one evaluation function to the possible lateral positions.

• – Eine Bildverarbeitungseinheit zum Detektieren von Fahrbahnmarkierungen;
• – Eine Positionsbestimmungseinheit zum Bestimmen einer Ausgangsposition für das Fahrzeug; und
• – Eine Auswerteeinheit zum Ermitteln von möglichen lateralen Positionen des Fahrzeugs durch Vergleichen der detektierten Fahrbahnmarkierungen mit Fahrbahnmarkierungsinformationen aus einer Fahrstreifengeometriekarte für die für das Fahrzeug bestimmte Ausgangsposition und zum Bestimmen einer abschließenden lateralen Position des Fahrzeugs durch Anwendung zumindest einer Bewertungsfunktion auf die möglichen lateralen Positionen; wobei die Auswerteeinheit eingerichtet ist, für das Ermitteln von möglichen lateralen Positionen des Fahrzeugs zumindest eine erste Menge von möglichen lateralen Positionen für eine erste detektierte Fahrbahnmarkierung und eine zweite Menge von möglichen lateralen Positionen für eine zweite detektierte Fahrbahnmarkierung zu bestimmen.

According to a further aspect of the invention, a device for determining a lateral position of a vehicle relative to the lanes of a roadway comprises:

• An image processing unit for detecting lane markings;
• A position determination unit for determining a starting position for the vehicle; and
• An evaluation unit for determining possible lateral positions of the vehicle by comparing the detected lane markings with lane marking information from a lane geometry map for the starting position determined for the vehicle and determining a final lateral position of the vehicle by applying at least one evaluation function to the possible lateral positions; wherein the evaluation unit is set up to determine at least a first set of possible lateral positions for a first detected lane marking and a second set of possible lateral positions for a second detected lane marking for determining possible lateral positions of the vehicle.

• – Detektieren von Fahrbahnmarkierungen;
• – Bestimmen einer Ausgangsposition für das Fahrzeug;
• – Ermitteln von möglichen lateralen Positionen des Fahrzeugs durch Vergleichen der detektierten Fahrbahnmarkierungen mit Fahrbahnmarkierungsinformationen aus einer Fahrstreifengeometriekarte für die für das Fahrzeug bestimmte Ausgangsposition, wobei zumindest eine erste Menge von möglichen lateralen Positionen für eine erste detektierte Fahrbahnmarkierung und eine zweite Menge von möglichen lateralen Positionen für eine zweite detektierte Fahrbahnmarkierung ermittelt werden; und
• – Bestimmen einer abschließenden lateralen Position des Fahrzeugs durch Anwendung zumindest einer Bewertungsfunktion auf die möglichen lateralen Positionen.

In accordance with another aspect of the invention, a computer-readable storage medium includes instructions that, when executed by a computer, cause the computer to perform the following steps:

• - detecting lane markings;
• Determining a starting position for the vehicle;
• Determining possible lateral positions of the vehicle by comparing the detected lane markings with lane marking information from a lane geometry map for the vehicle determined starting position, wherein at least a first set of possible lateral positions for a first detected lane marker and a second set of possible lateral positions for a detected second detected lane marking; and
• Determining a final lateral position of the vehicle by applying at least one evaluation function to the possible lateral positions.

The sub-lane-specific determination of the lateral vehicle position relative to the roadway is based on the alignment of lane markings, which were detected by an imaging sensor system, with lane marker geometries from a digital lane geometry map. This is not subject to any special requirements for absolute accuracy, but contains the lane center geometries and lane edge geometries with high accuracy to each other. Based on the comparison, possible lateral positions of the vehicle are determined. By applying at least one weighting function to the determined possible lateral positions, a final lateral position is determined, i. the most likely or best possible position. For determining the possible lateral positions, a plurality of detected lane markings, preferably two lane markings, are used. For each of these lane markings, an attempt is made to determine a set of possible lateral positions. The resulting quantities are subsequently combined.

In this way it is prevented that only poorly positioned possible lateral positions are determined due to an incorrectly detected lane marking. For example A tarmac could be erroneously recognized as a road marking. If the determination of the lateral position was based solely on this supposed lane marking, a faulty result would be expected.

Preferably, the first detected lane marking and the second detected lane marking are those lane markings which have the smallest lateral distance to the starting position of the vehicle. The lateral accuracy of the geometries detected by the camera tends to decrease significantly with greater lateral distance from the camera. By looking at the detected lane markings with the smallest lateral distance to the vehicle, the lane markings are thus selected with presumably the greatest accuracy with respect to this distance measurement.

According to one aspect of the invention, for determining a set of possible lateral positions for a selected detected lane marker, this lane marker is successively assigned to different lane markers of the lane geometry map. An attempt is then made to determine a possible lateral position for each of these assignments. Clearly, it is assumed here that the selected detected lane marking would correspond to the respective lane marking of the lane geometry map. In this way, it is possible to reliably determine a plurality of possible lateral vehicle positions from the comparison of the detected lane markings with the lane markings of the lane geometry map.

According to one aspect of the invention, intersections of a line orthogonal to the vehicle movement with the detected lane markings are determined for assigning a selected detected lane marker to a selected lane marker of the lane geometry map. Likewise, intersections of the orthogonal line are determined with the lane markings of the lane geometry map. The intersection points determined for the detected lane markings are then shifted along the orthogonal line such that the intersection of the selected detected lane marker comes to rest on the intersection of the selected lane marker of the lane geometry map and the relative distances of the detected lane markings are preserved. By reducing the situation analysis to intersections, the computational effort required to implement the solution according to the invention can be significantly reduced.

According to one aspect of the invention, the shifted intersection point of at least one further detected lane marking is compared with the intersection points of the lane markings of the lane geometry map nearest to its position, and a similarity of the intersections is assessed. The further lane marking is then assigned to the lane marking of the lane geometry map with the best rated similarity of the intersections. After the assignment of the selected detected pavement marker to the selected pavement marker of the lane geometry map, an attempt is made to determine associations between the further detected pavement markings and the further lane markings of the lane geometry map. The respective assignments are preferably evaluated individually. From the totality of the evaluations of all assignments, a rating for the considered assignment of the selected detected roadway marking to the selected roadway marking of the lane geometry map can then be derived. This can also be used to evaluate the possibly determined possible lateral position.

According to one aspect of the invention, for evaluating the similarity of the intersections, at least one of the following aspects is considered: If the lane marker of the lane geometry map has already been assigned to another detected lane marker, the order of the lane markers is maintained, what is the geometric distance of the intersections, is that correct Line type of the lane markings match, does the line direction of the associated lane markings match? Each of these aspects allows conclusions as to whether a match of the points of intersection is plausible and probable. At the same time, the individual aspects can be checked without much computational effort, so that the comparison of the intersections can take place virtually in real time. Preferably, several aspects are considered and the resulting ratings combined.

According to one aspect of the invention, for each assignment of the selected detected lane marker to a selected lane marker of the lane geometry map, an overall score is determined based on the scores of the intersections. An assignment is discarded if the associated overall score falls below a quality measure. The comparison of the overall rating with a quality measure makes it possible to discard obviously implausible allocations and not to be included in the further evaluation. The quality measure is in particular undershot, if a larger proportion of the detected lane markings no lane markings of the digital map could be assigned.

According to one aspect of the invention, a provisional possible lateral position of the vehicle is determined by shifting the home position determined by the displacement of the points of intersection parallel to the orthogonal line. With this approach, a preliminary possible lateral position can be determined with little effort, which can serve as a basis for a later optimization.

In accordance with one aspect of the invention, a provisional possible lateral position score function is determined based on weighted distances between the intersections of the further detected lane markers and the intersections of the respective lane geometry map associated lane markers. The provisional possible lateral position is then shifted parallel to the orthogonal line such that the evaluation function has a functional minimum. The resulting shifted provisional possible lateral position is finally determined as a possible lateral position of the assignment. In this way, on the basis of the provisional possible lateral position, an optimized lateral position is determined, which takes into account the results of the associations of the detected lane markings to the lane markings of the lane geometry map.

Preferably, a method according to the invention or a device according to the invention are used in an autonomously or manually controlled vehicle, in particular a motor vehicle.

Further features of the present invention will become apparent from the following description and the appended claims, taken in conjunction with the figures.

Brief description of the figures

1 schematically shows a method for determining a lateral position of a vehicle relative to the lanes of a roadway;

2 shows a first embodiment of a device for determining a lateral position of a vehicle relative to the lanes of a roadway;

3 shows a second embodiment of a device for determining a lateral position of a vehicle relative to the lanes of a roadway;

4 shows an approach for associating a selected detected pavement marker with a selected pavement marker of a lane geometry map;

5 shows an approach for assigning further detected lane markings;

6 shows a preferred embodiment of the in 1 represented method;

7 shows an example of detected map intersections;

8th shows an example of determined BV line intersections;

9 shows various routing assignments and routing vehicle locations for a first detected pavement marker; and

10 shows various routing assignments and routing vehicle locations for a second detected pavement marker.

Detailed description of the embodiments

For a better understanding of the principles of the present invention, embodiments of the invention will be explained in more detail below with reference to the figures. It is understood that the invention is not limited to these embodiments, and that the features described can also be combined or modified without departing from the scope of the invention as defined in the appended claims.

1 schematically shows a method for determining a lateral position of a vehicle relative to the lanes of a roadway. In a first step, lane markings are detected 10 , Likewise, a starting position for the vehicle is determined 11 , By comparing the detected lane markings with lane marking information from a lane geometry map for the vehicle's initial position, possible lateral positions of the vehicle are determined 12 , In this case, at least a first set of possible lateral positions for a first detected lane marking and a second set of possible lateral positions for a second detected lane marking are determined.

For example, the first detected lane marking and the second detected lane marking are those lane markings which have the smallest lateral distance to the starting position of the vehicle. A final lateral position of the vehicle is then determined by applying at least one evaluation function to the possible lateral positions 13 ,

2 shows a simplified schematic representation of a first embodiment of a device 20 for determining a lateral position of a vehicle relative to the lanes of a roadway. The device 20 has an image processing unit 22 to detect 10 of road markings. For this purpose, the image processing unit uses 22 For example, image information of a camera unit 25 that have an entrance 21 the device 20 be received. The device 20 also has a position determination unit 23 for determining 11 a starting position for the vehicle. The home position is based, for example, on reception data of a GPS receiver 26 Certainly, the same over the entrance 21 can be received.

An evaluation unit 24 determined 12 possible lateral positions of the vehicle by comparing the detected lane markings with lane marking information from a lane geometry map for the starting position determined for the vehicle. For this purpose, the lane geometry map contains lane center geometries and lane edge geometries with high accuracy to each other. In this case, at least a first set of possible lateral positions for a first detected lane marking and a second set of possible lateral positions for a second detected lane marking are determined. For example, the first detected lane marking and the second detected lane marking are those lane markings which have the smallest lateral distance to the starting position of the vehicle. Further determined 13 the evaluation unit 24 a final lateral position of the vehicle by applying at least one evaluation function to the possible lateral positions.

The from the evaluation unit 24 certain lateral position of the vehicle is preferably via an exit 27 the device 20 made available for further processing, for example for processing in a guidance system. It can also be stored in a memory 28 the device 20 be stored, for example, for a later evaluation. The entrance 21 and the exit 27 may be implemented as separate interfaces or as a combined bidirectional interface.

The image processing unit 22 , the position determination unit 23 as well as the evaluation unit 24 can be implemented as dedicated hardware, such as integrated circuits. Of course, they can also be partially or fully combined or implemented as software running on a suitable processor.

3 shows a simplified schematic representation of a second embodiment of a device 30 for determining a lateral position of a vehicle relative to the lanes of a roadway. The device 30 has a processor 32 and a memory 31 on. For example, the device is 30 to a computer or a controller. In the storage room 31 are filed instructions that the device 30 when executed by the processor 32 to perform the steps according to one of the described methods. The in the store 31 filed instructions thus embody one by the processor 32 executable program that realizes the inventive method. The device has an entrance 33 for receiving information. From the processor 32 generated data will be via an output 34 provided. In addition, they can be stored in memory 31 be filed. The entrance 33 and the exit 34 can be combined to a bidirectional interface.

The processor 32 may include one or more processing units, such as microprocessors, digital signal processors, or combinations thereof.

The stores 28 . 31 The embodiments described may have volatile and / or non-volatile memory areas and include a wide variety of memory devices and storage media, for example hard disks, optical storage media or semiconductor memories.

For determining a set of possible lateral positions for a selected detected lane marking, this lane marking can be successively assigned to different lane markings of the lane geometry map. An attempt is then made to determine a possible lateral position for each of these assignments. 4 shows an approach for associating a selected detected pavement marker with a selected pavement marker of the lane geometry map. First, intersections of a line orthogonal to the vehicle movement are determined with the detected lane markings 14 , Similarly, intersections of the orthogonal line are determined with the lane markings of the lane geometry map.

The intersection points determined for the detected lane markings are then shifted along the orthogonal line 16 in that the point of intersection of the selected detected lane marking comes to lie on the intersection of the selected lane marking of the lane geometry map and the relative distances of the detected lane markings are preserved.

For the determination of a possible lateral position for an assignment, preferably further detected lane markings are assigned. One approach shows 5 , The shifted intersection of at least one further detected lane marking is compared with the nearest to its position intersections of the lane markings of the lane geometry map 17 and evaluated a similarity of intersections 18 , The further lane marking is then assigned to the lane marking of the lane geometry map with the best rated similarity of the intersections 19 , For evaluating the similarity of the points of intersection, for example, consideration of the following aspects comes into question: If the lane marking of the lane geometry map has already been assigned to another detected lane marking, the sequence of lane markings is retained, what is the geometric distance of the intersection points, is the line type of the lane markings identical, and does the line direction of the associated lane markings match?

Description of a preferred embodiment

Hereinafter, a preferred embodiment of the invention will be described in detail. The method is based on a series of input data. What is needed first are the geometry and property information of the visible lane markings determined by the imaging sensor system. These are referred to below as BV lines (BV for image processing). Also required is absolute position information with direction and speed information. This can be provided for example by the vehicle GPS. For the absolute position information, in the present embodiment, a distinction is made between absolute position information representing the direct result of GNSS (GNSS) position finding (GNSS position data) and those based on past GNSS Location were interpolated by dead-reckoning (absolute position data). Optionally, relative, jump-free position information can also be used, e.g. is determined by motion estimation. In addition, map data with high relative accuracy regarding lanes and lane markings are used.

These are for example through a map data server 110 made available. Map information is referred to as DLM lanes (DLM: Detailed Lane Model) and DLM Lanemarkings. Preferably, results from the previous procedure or previous procedures of the method are included as a history, except, of course, in the first iteration.

The basic procedure of the procedure is divided into several parts, which in 6 are partially summarized in dashed blocks.

As part of a treatment 40 The input data is used to determine an approximate vehicle position 41 , This serves as a starting point for the alignment of lane markings with map data. In addition, there is a conversion and aggregation of the geometries of the BV lines.

The geometries of the lane markings detected by the camera system are typically described by clothoids in a vehicle-relative Cartesian coordinate system. The descriptions of the clothoid are translated into polylines, which closely approximate the clothoid geometries. Conversion to polylines is performed because the algorithms operating subsequently on the pavement marking geometries are so much easier to implement. In the present embodiment, the coordinates of the polylines are transferred from the vehicle-relative Cartesian coordinate system to the WGS84 coordinate system 42 , The following algorithms operate in the WGS84 coordinate system because the map data and vehicle position and motion information is also typically present in this coordinate system.

The geometries of the lane markings recognized by the camera system always start just in front of the vehicle and have an extension of a few meters approximately in the direction of travel / camera viewing direction. Occasionally, road markings are first recognized and transmitted correctly by the camera system, but are no longer detected shortly thereafter. The BV lines present in an algorithm iteration are therefore always buffered and compared with the new geometries transmitted by the camera system during the next iteration 43 ,

In the following, essential geometry aspects are extracted 50 , Points of BV-lines and DLM-Lanemarkings are calculated for the longitudinal vehicle position.

To do this, first an orthogonal line of configurable length is constructed at a configurable distance from the approximate vehicle position 51 , Subsequently, intersections of the DLM Lanemarkings and the DLM Lanes are determined with the orthogonal route 52 , both together referred to as card intersections. An example of such map intersections is in 7 shown. To see is a roadway FB with four lanes FS, which are delimited by road markings FBM from each other, and the associated lane centers FSM. Also shown are the home position AP, the constructed orthogonal line OL and the map intersections, separated by intersections SPM with the DLM Lanemarkings (Marking intersection) and Intersections SPL with the DLM lanes (Lane intersection). For each card intersection preferably its lateral position relative to the vehicle, for example as distance from the center of the orthogonal route, and in DLM Lanemarkings the information of the type of associated lane marking (dashed / solid line, guardrail, roadside, ...) recorded. In a further step, the intersections of the orthogonal path with the BV lines are formed 53 , referred to as BV line intersections. An example of such BV line intersections is in 8th shown. To be seen in addition to those already out 7 known elements, the detected BV-lines BVL and the intersections BVS of the orthogonal line OL with the recognized BV-lines BVL. Again, for each intersection BVS preferably its lateral position relative to the vehicle and the information of the type of recognized lane marking (line dashed / solid, guardrail, roadway edge, ...) recorded.

If sufficiently informative information on BV lines is available, then a number of possible vehicle positions is determined below 60 , Otherwise it will be an error handling 90 initiated. For example, it may be determined that sufficient information is present when at least two BV line intersections are found at a distance apart of, for example, 0.5 lane widths and a distance from the approximate vehicle position of less than, eg, 1.5 lane widths.

To determine the quantity of possible vehicle positions, a configurable number of BV lines is used, for example two BV lines. For this purpose, the process described below for determining a number of possible vehicle positions based on a selected BV line is carried out several times in succession, always based on a different number of BV lines of the determination. Preferably, those BV lines are used which have the smallest distance between the vehicle position and its associated intersection with the orthogonal line and at the same time have a probability of existence above a configurable threshold value. The underlying BV line is referred to below as the lead BV line.

To generate possible vehicle positions based on a lead BV line, an attempt is made to generate a possible vehicle position at each intersection with a DLM Lanemarking. It is intuitively assumed that the lead BV line corresponds to the DLM Lanemarking belonging to the respective marking intersection. The lane marking identified in each case with the guide BV line is referred to below as the guide mark.

Thus, the procedure considers in turn each BV-line used as the lead BV line and for each of these lead BV-lines the DLM-Lanemarking of each existing marking-intersection as a guide marker.

The currently considered assignment of the roles of the leading BV line and the guide marking is referred to below as a guide assignment. For each individual routing assignment, an attempt is made to determine a possible vehicle position according to the method described below. It is not certain that this provision is successful. Thus, at most exactly one possible vehicle position is determined for each guidance assignment.

For each routing assignment, the BV line intersections are placed on the orthogonal path so that the intersection of the guidance BV line is on the guide marker and the distances of all BV lines to each other are maintained.

The vehicle position can accordingly be shifted parallel to the orthogonal route so that the relative positions of the BV lines to the vehicle position are maintained. The vehicle position thus shifted is referred to below as the lead vehicle position. An example of various master assignments and master vehicle positions LFP is in 9 and 10 shown. In this example, three BV lines were detected. In 9 the mean detected BV line is used as lead BV line LBL, in 10 the left detected BV line. The respective control BV line LBL is successively assigned to three different guide markings LM, ie first the dashed marking between the two left lanes, then the one-sided lane boundary (solid and dashed double line) and finally the solid mark on the right side.

The lead vehicle position corresponds to the position provisionally assumed for the possible vehicle position to be determined. However, the final position of the possible vehicle position is determined only in a later step.

Next, assignments between BV lines and DLM Lanemarkings are determined. The respective assignments are evaluated individually. From the totality of the evaluations of all assignments, an evaluation results for the considered guide assignment and the possibly resulting possible vehicle position.

First, the lead BV line is assigned to the lead mark. Thereafter, the other BV lines are considered and possibly assigned, for example in the order of increasing distance of their intersections with the orthogonal of the vehicle position. Each BV line intersection is compared with the nearest to its position marking intersections, the similarity of the intersections evaluated and created an association of the BV line to DLM Lanemarking the best rated intersection.

In determining the similarity of BV line intersections and marking intersections, for example, the criteria described below may be considered. The concrete numerical values and weightings of the individual criteria are preferably configurable.

DLM Lanemarking unoccupied: A first criterion is the question of whether the DLM Lanemarking is unoccupied. If the DLM Lanemarking is already linked to another BV line, the result is a poor evaluation of the similarity.

Preservation of order: A second DLM Lanemarking on the left of a first DLM Lanemarkings, to which a BV line is already assigned, should not be assigned to a BV line on the right of this assigned BV line. If this is the case, this results in a poor evaluation of similarity.

Distance of the intersections: How far are the intersection points apart? A larger distance leads to a worse rating. However, for larger lateral distances of the considered intersections to the lead vehicle position, a greater inaccuracy of the camera system with respect to the estimation of the lateral position of the BV lines may be assumed. Correspondingly, distances of the intersections to one another are less badly evaluated with increasing distance of the intersections to the lead vehicle position.

Line type of lane marking: Does the type of lane marking (dashed / solid line, guardrail, roadway edge, ...) fit together in accordance with DLM Lanemarking and BV Line? For this evaluation, a configurable matrix is preferably provided which assigns a specific value for each combination of the BV-line and the ticket-lane marking type. Thus, common misallocations of the camera, e.g. the detection of a solid as a dashed line, with only a slightly poor rating, improbable misallocations of the camera, e.g. the detection of lane edge as a guardrail, associated with a significantly poor rating.

Line direction: Does the direction of the BV line (at the intersection with the orthogonal line) match the direction of the DLM Lanemarking (at the intersection with the orthogonal)? It should be noted that the lines should be geometrically similar, but an implicit direction is irrelevant.

Gap-free detection: According to the map, there are lane markings, i. Marking intersections, between the marking point to be evaluated and the nearest marking intersection, whose DLM Lanemarking is already assigned to a BV line? Any such unassigned DLM Lanemarking degrades the rating.

In this case, an exception may be provided in the event that the distance of the marking point to be evaluated from the point of intersection of the unassigned DLM Lanemarking is below a configurable threshold value. Two closely spaced DLM Lanemarkings, e.g. a solid next to a dashed line, are u.U. only recognized as a single line. Although such lines are usually also in the DLM data set as only a line with the type dashed-and-drawn (or similar) deposited, but this is not always the case. Likewise, for example, two converging lines may produce two closely spaced marking intersections, but the camera will most likely only detect one line at some point in time. The proposed exception takes into account such situations.

A further exception can be provided for the case in which the distance of the marking intersection point to be evaluated from the guide vehicle position is below a configurable threshold value. Some camera systems often do not detect lane marking (anymore) when the vehicle crosses it.

BV lines with large lateral proximity to the camera / vehicle are deliberately ignored by the configuration of the camera system, for example, to filter out BV lines from road surface tread seams which may be falsely recognized as lane markers. The consideration of the exception regulation in the vicinity of the lead vehicle position is preferably configurable switched on and off. In this way, the method is adaptable to both camera systems requiring filtering vehicle-related BV lines and to those who do not.

Instead of the evaluation depending on the exceeding of a threshold value, a value calculated as a function of the distance or a combination of the two can also be used for the evaluation.

Occasionally, the camera systems used incorrectly detect BV lines, although in fact there are no road markings at the positions determined. If no DLM Lanemarking is found in the search of a DLM Lanemarking for assignment to a BV line for a given BV line and given Leit marker, so that the evaluation of a corresponding assignment reaches or exceeds a configurable quality value, it is assumed that this BV line is a wrongly recognized line by the camera and no assignment is made.

For any BV line that can not be assigned DLM Lanemarking with sufficient quality, the overall rating of the lead assignment deteriorates. An appropriate allocation should always lead to a better overall rating of the lead allocation than would a non-assignment of the BV line.

If the assignment of DLM Lanemarkings to all BV lines is completed, it is preferably checked whether the overall rating of the guide assignment falls below a configurable quality measure. If this is the case, the lead assignment is discarded as an implausible assumption and for this purpose no possible vehicle position is output. The further treatment of the lead assignment is dispensed with and the method continues immediately with the next lead assignment. In particular, the quality standard is undercut if a larger proportion of BV lines could not be assigned DLM Lanemarkings.

As described above, the guidance vehicle position has been determined according to the assignment of the guidance mark to the guidance BV line based on the relative position of the guidance BV line to the vehicle.

The location of the possible vehicle position should also take into account the assignments of DLM Lanemarkings to the other BV lines. For this purpose, for example, the squares of the distances of the BV line intersections to the respectively assigned marking intersection points are summed up according to their distance to the vehicle position. There is a smaller weighting at greater distance, e.g. a weight with the inverse of the squared distance. The resulting sum is considered as a function of a shift of all BV line intersections along the orthogonal and determines the functional minimum.

The functional minimum occurs when the BV line intersections are shifted by the weighted arithmetic mean of the distances between the BV line intersections and the respective marking intersections.

The position resulting from shifting the master vehicle position by the value of the functional minimum is the final vehicle position determined from that combination of master BV line and lead mark.

The total determined possible vehicle positions are subsequently evaluated by a sequence of corresponding evaluation functions 70 , Penalty points are awarded with the help of the evaluation functions. A first evaluation function 71 considers the assignment of the line types captured by the camera to the line types stored in the map. For this evaluation, the configurable matrix already described above is preferably used, which assigns a specific value for each combination of BV line and ticket road marking type. A second evaluation function 72 takes into account the history of vehicle positions. Possible vehicle positions that deviate strongly from the history are characterized, for example, by high penalty points. In the present embodiment, a third evaluation function evaluates 73 the lane type. The vehicle is suspected on a regular lane of the road. Possible vehicle positions on non-driving lanes (side lanes, "unknown lanes" of the DLM, and hard shoulder, ...) are therefore poorly evaluated, possible vehicle positions on passable lanes are rated neutral. As another example, significantly higher penalty points are awarded for possible vehicle positions on the oncoming lane than for positions in the direction of travel. The setting of the penalty points depends on the sensor system used and the digital map used. Thus, a specific adaptation for the systems used can be done very easily. As a result of the evaluation process, finally the best rated possible vehicle position is selected 80 ,

During the course of the process, it may occur at various points that due to the lack of input data or an insufficient quality of the input data, a position determination is not possible and the process can not be continued. For such cases, it is possible to leave the modular process at the appropriate location and error handling 90 to initiate, eg a result without To generate position information or with a position information approximately determined by other means. In 6 the paths of the regular sequence are indicated by solid arrows, the paths in case of deviation from the regular sequence by dashed arrows.

For example, in the event of unavailability of map data for the current vehicle position according to absolute position information, no determination of the possible vehicle positions is carried out. In the absence of absolute position information (eg no GNSS reception by structures) the absolute position information can be replaced by an earlier absolute position information and relative position information when relative position information is available. If no relative position information is available, no determination of the possible vehicle positions is carried out if absolute position information is missing. In such cases, a corresponding error condition is output 91 , If, on the other hand, too few BV lines have been detected, it is not possible to find a sufficiently good assignment of DLM Lanemarkings to BV lines or if no single possible vehicle positions have been determined, using a map-based procedure 92 an approximately determined position information can be determined. One possibility is, for example, to assume a continuation of the vehicle movement along the center of the lane according to the map. In a final step 100 the results are prepared, made available for further processing and incorporated into the history.

LIST OF REFERENCE NUMBERS

10

 Detecting road markings

11

 Determining a starting position for the vehicle

12

 Determining possible lateral vehicle positions by comparison with lane geometry map

13

 Determine a final lateral vehicle position by using evaluation function

14

 Determining intersections for detected lane markings

15

 Determining intersections for lane markings of the lane geometry map

16

 Moving the intersections of the detected lane markings orthogonal to the vehicle movement

17

 Compare a shifted intersection with nearest intersections

18

 Evaluate a similarity of intersections

19

 Assigning a lane marking according to the similarity of the intersections

20

 contraption

21

 entrance

22

 Image processing unit

23

 Position Determination Entity

24

 evaluation

25

 camera unit

26

 GPS receiver

27

 output

28

 Storage

30

 contraption

31

 Storage

32

 processor

33

 entrance

34

 output

40

 Preparation of the input data

41

 Determining an approximate vehicle position

42

 Transfer to WGS84 coordinate system

43

 Adjustment of newly transmitted geometries

50

 Extract geometry aspects

51

 Construct an orthogonal line

52

 Determine map intersections

53

 Determining points of intersection with BV lines

60

 Determining possible vehicle positions

70

 Evaluate the possible vehicle positions

71

 First evaluation function

72

 Second evaluation function

73

 Third valuation function

80

 Selection of the best rated vehicle position

90

 error handling

91

 Issue an error condition

92

 Card-based procedure

100

 Processing the results

110

 Map data server

FB

 roadway

FS

 lane

FSM

 lane center

FBM

 road marking

OIL

 orthogonal line

SPM

 Marking cut Punk

SPL

 Lane intersection

AP

 starting position

BVL

 BV line

BVS

 Intersection with BV line

LBL

 Leit-BV line

LM

 Routing label

LFP

 Leit-vehicle position

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited patent literature

• US 2014/0358321 A1 [0004]
• EP 2899669 A1 [0005]
• DE 102012104786 A1 [0006]

Claims (12)

Method for determining a lateral position of a vehicle relative to the lanes of a carriageway, comprising the steps of: - detecting ( 10 ) of lane markings; - Determine ( 11 ) an initial position for the vehicle; - Determine ( 12 ) of possible lateral positions of the vehicle by comparing the detected lane markings with lane marking information from a lane geometry map for the home position determined by the vehicle; and - determining ( 13 ) a final lateral position of the vehicle by applying at least one evaluation function to the possible lateral positions; characterized in that for determining ( 12 ) of possible lateral positions of the vehicle at least a first set of possible lateral positions for a first detected lane marker and a second set of possible lateral positions for a second detected lane marker are determined. The method of claim 1, wherein the first detected lane marker and the second detected lane marker are the lane markers having the smallest lateral distance to the home position of the vehicle. A method according to claim 1 or 2, wherein for determining a set of possible lateral positions for a selected detected lane marker, said lane marker is successively assigned to different lane markers of the lane geometry map and an attempt is made to determine a possible lateral position for each of these mappings. The method of claim 3, wherein associating a selected detected pavement marker with a selected pavement marker of the lane geometry map comprises the steps of: determining (FIG. 14 ) of intersections of a line orthogonal to the vehicle movement with the detected lane markings; - Determine ( 15 ) of intersections of the orthogonal line with the lane markings of the lane geometry map; and - move ( 16 ) of the intersections determined for the detected lane markings along the orthogonal line such that the intersection of the selected detected lane marker comes to rest on the intersection of the selected lane marker of the lane geometry map and the relative distances of the detected lane markers are preserved. Method according to claim 4, wherein the method comprises the steps of: - comparing ( 17 ) of the shifted intersection of at least one further detected lane marking with the nearest to its position intersections of the lane markings of the lane geometry map; - Rate ( 18 ) a similarity of the intersections; and - assign ( 19 ) of the further lane marking for lane marking of the lane geometry map with the best rated similarity of the intersections. A method according to claim 5, wherein for evaluating the similarity of the intersections at least one of the following aspects is considered: If the lane marker of the lane geometry map has already been assigned to another detected lane marker, the lane marker order is maintained, what is the geometric distance of the intersections, is that correct Line type of the lane markings match, and agrees the line direction of the associated lane markers. A method according to claim 5 or 6, wherein the method comprises the steps of: Determining an overall score for each mapping of the selected detected pavement marker to a selected pavement marker of the lane geometry map based on the scores of the intersections; and - discarding an assignment if the associated overall rating falls below a quality standard. Method according to one of claims 5 to 7, wherein a provisional possible lateral position of the vehicle is determined by the starting position determined for the vehicle is shifted parallel to the orthogonal line according to the displacement of the intersections. The method of claim 8, wherein the method comprises the steps of: Determining an evaluation function for the provisional possible lateral position based on weighted distances between the intersections of the further detected lane markings and the intersections of the respective lane markings of the lane geometry map; - shifting the provisional possible lateral position parallel to the orthogonal line such that the evaluation function has a functional minimum; and - Defining the shifted provisional possible lateral position as a possible lateral position of the assignment. Contraption ( 20 ) for determining a lateral position of a vehicle relative to the lanes of a roadway, the device ( 20 ) comprises: - an image processing unit ( 22 ) for detecting ( 10 ) of lane markings; A position determination unit ( 23 ) for determining ( 11 ) an initial position for the vehicle; and - an evaluation unit ( 24 ) for determining ( 12 ) of possible lateral positions of the vehicle by comparing the detected lane markings with lane marking information from a lane geometry map for the starting position determined for the vehicle and determining ( 13 ) a final lateral position of the vehicle by applying at least one evaluation function to the possible lateral positions; characterized in that the evaluation unit ( 24 ) is set up for determining ( 12 ) of possible lateral positions of the vehicle to determine at least a first set of possible lateral positions for a first detected lane marker and a second set of possible lateral positions for a second detected lane marker. A computer-readable storage medium having instructions that, when executed by a computer, cause the computer to perform the following steps: - detecting ( 10 ) of lane markings; - Determine ( 11 ) an initial position for the vehicle; - Determine ( 12 ) of possible lateral positions of the vehicle by comparing the detected lane markings with lane marking information from a lane geometry map for the home position determined by the vehicle; and - determining ( 13 ) a final lateral position of the vehicle by applying at least one evaluation function to the possible lateral positions; characterized in that for determining ( 12 ) of possible lateral positions of the vehicle at least a first set of possible lateral positions for a first detected lane marker and a second set of possible lateral positions for a second detected lane marker are determined. Autonomous or manually controlled vehicle, characterized in that it comprises a device according to claim 10 or is adapted to carry out a method according to one of claims 1 to 9.

Images