Method for generating a high definition road map for a motor vehicle, as well as an assistance system

Abstract

A method for generating a high definition road map (10) for a motor vehicle (36). A position of the motor vehicle (36) is detected by a global navigation satellite system and at least one sensor data

Description

FIELD OF THE INVENTION

[0001] The present disclosure relates to the field of automobiles. More specifically, the present disclosure relates to a method for generating a high definition road map for a motor vehicle, as well as to a corresponding assistance system. BACKGROUND INFORMATION

[0002] Generating high definition road maps, which may also be referred to as road networks, is already known in the state of the art. According to the state of the art, the road networks for navigation maps are generated from data acquired by mobile mapping vehicles or area photographs. Lane-level information, which is the basis for a high-definition map, is acquired largely manually from high-definition images, laser point clouds, or DGPS / INS trajectories. Inferring such detailed information manually is timeconsuming and labor-intensive.

[0003] US 7,979,172 B2 discloses a method and system for enabling semi-autonomous or autonomous vehicle travel including creating dedicated travel lanes for vehicle travel, determining the location of vehicles in the dedicated travel lanes, and providing instruction to the vehicles in the dedicated level lanes to enable them to travel in a manner which maximizes speed. To determine the location of the vehicles, a vehicle travel management system may be arranged in each vehicle to determine the location of the vehicle, for example, using a GPS-based positioning system or a similar system, and which also optionally receives travel instructions which can be implemented by a vehicle control system to control the direction and speed of the vehicle. The vehicle control system may be designed to assume control of vehicle travel without requirir an occupant of the vehicle and to control vehicle travel based on the provid instructions.

[0004] According to the state of the art, manually inferring detailed information in the road map is time-consuming and labor-intensive. There are no existing methods to infer high-definition maps from sensor data fully automatically. There are existing methods for inferring road networks from vehicular position data automatically and there are map matching techniques to map a low precision vehicle location on a road network. However, the state of the art does not do both at the same time and thus fails to provide a valid basis for further processing to generate high-definition maps automatically. SUMMARY OF THE INVENTION

[0005] It is an object of the present invention to provide a method as well as an assistance system, by which an automatic generating of a high definition road map may be realized.

[0006] This object is solved by a method as well as an assistance system according to the independent claims of the present invention. Advantageous embodiments are presented in the dependent claims.

[0007] One aspect of the invention relates to a method for generating a high definition road map for a motor vehicle by an assistance system, wherein a position of the motor vehicle is detected by a global navigation satellite system of the motor vehicle and wherein at least one sensor data is detected by at least one detection device of the motor vehicle, and wherein the current position of the motor vehicle and the current sensor data are linked to generate a trace point by an electronic computing device of the assistance system.

[0008] It is envisaged that in a first step hypotheses for a plurality of road map nodes are generated by the electronic computing device along a plurality of trace points from the motor vehicle and / or a further motor vehicle and in a second step the plurality of the road map node hypotheses are linked by the electronic computing device in order to generate the high definition road map. The trace points are belonging to at least two traces.

[0009] Therefore, the method according to the invention distinguishes fror the art methods by the fact that the resulting data structure of the road netv may be referred to as the high definition road map, includes linked sensor aaia acquired along the road topology. This feature may be highly valuable for applications in the field of high-definition maps and automated driving. Its underlying data structure in this or modified form is probably the only way to generate the high-definition maps automatically using vehicle probe data and to test road geometries for changes. Another distinctive feature of the method is the usage of sensor data to generate the road network, apart from position and initial sensors.

[0010] Furthermore, a prerequisite of the procedure is that the processed data is of the following data structure. All sensor data acquired by vehicles needs to be linked with a global navigation satellite system position and time of the acquisition. The sensor data, which is available at a higher rate than the global navigation satellite system position, is linked to the latest global navigation satellite system position acquired. Such data point in the following is referred to as the trace point. These trace points need to be linked per vehicle according to the chronology of the acquisition. The resulting data structure is then referred to as a trace. The procedure of the road network extraction is divided into two phases. In phase one, which is referred to as step one, hypotheses for the road network nodes are generated along the trace points of the processed vehicle probe data. In phase two, which is referred to as the second step, the road network node hypotheses are linked.

[0011] In an embodiment, in the first step for generating the hypotheses an iteration is performed, beginning with a determination of a local center point aggregated from the position of the trace points, the center points are then condensed in such a way that a final iteration results in the hypotheses.

[0012] In another embodiment, for the condensation a discrete global grid is used for the trace points.

[0013] In another embodiment, in the second step directional road map link hypotheses are generated based on a membership of traces belonging to road map node candidates.

[0014] Another aspect of the invention relates to an assistance system foi high definition road map for a motor vehicle comprising at least one electroi device, wherein the assistance system is configured to perform a method according io the preceding aspect. In particular, the method is performed by the assistance system.

[0015] Advantageous forms of the method are to be regarded as advantageous forms of the assistance system, wherein the assistance system comprises features allowing to perform the method.

[0016] Further advantages, features, and details of the invention derive from the following description of a preferred embodiment as well as from the drawings. The features and feature combinations previously mentioned in the description as well as the features and feature combinations mentioned in the following description of the figures and / or shown in the figures alone may be employed not only in the respectively indicated combination but also in any other combination or taken alone without leaving the scope of the invention. BRIEF DESCRIPTION OF THE DRAWING

[0017] The novel features and characteristics of the disclosure are set forth in the independent claims. The accompanying drawings, which are incorporated in and constitute a part of this disclosure, illustrate exemplary embodiments and together with the description serve to explain the disclosed principles. The same reference signs are used throughout the figures to refer to identical features and components. Some embodiments of the system and / or the method according to the present subject-matter are now described below, by way of example only and with reference to the accompanying figures.

[0018] Fig 1 shows a schematic top view of a road.

[0019] Fig. 2 shows top view of a road.

[0020] Fig. 3 shows a schematic flow chart of an embodiment of the method.

[0021] Fig 4 shows a schematic top view of a road.

[0022] Fig. 5 shows another schematic top view of a road.

[0023] Fig. 6 shows another schematic top view of a road.

[0024] Fig. 7 shows another top view of a road.

[0025] In the figures same elements or elements having the same function are indicated by the same reference signs. DETAILED DESCRIPTION

[0026] In the present document, the word "exemplary" is used herein to mean "serving as an example, instance, or illustration." Any embodiment or implementation of the present subject matter described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other embodiments.

[0027] While the disclosure is susceptible to various modifications and alternative forms, specific embodiments thereof have been shown by way of example in the drawings and will be described in detail below. It should be understood, however, that it is not intended to limit the disclosure to the particular forms disclosed, but on the contrary, the disclosure is to cover all modifications, equivalents, and alternatives falling within the scope of the disclosure.

[0028] The terms “comprises", “comprising”, or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a setup, device or method that comprises a list of components or steps does not include only those components or steps but may include other components or steps not expressly listed or inherent to such setup or device or method. In other words, one or more elements in a system or apparatus preceded by “comprises” or “comprise” does not or do not, without more constraints, preclude the existence of other elements or additional elements in the system or method.

[0029] In the following detailed description of the embodiments of the disclosure reference is made to the accompanying drawings that form a part hereof and in which are shown by way of illustration specific embodiments in which the disclosure may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the disclosure, and it is to be understood that other embodiments may be utilized and that changes may be made without departing from the scope of the present disclosure. The following description is, therefore, not to be take sense.

[0030] Fig. 1 shows a schematic view of a road map 10. Fig. 1 shows a trace 12, a plurality of center points 14, a center point heading 16, and a radius 18 for aggregation in the current iteration.

[0031] In particular, Fig. 1 shows a method for generating a high-definition map 10 for a motor vehicle 36 by an assistance system 20, wherein a position of the motor vehicle 36 is detected by a global navigation satellite system of the motor vehicle 36 and wherein at least one sensor data is detected by at least one detection device of the motor vehicle 36 and wherein the current position of the motor vehicle 36 and the current sensor data are linked by an electronic computing device of the assistance system 20 to a trace point 22.

[0032] In a first step, hypotheses for a plurality of road map nodes 24 (Fig. 2) are generated by the electronic computing device along a plurality of trace points 22 from the motor vehicle 36 and / or a further motor vehicle, wherein the trace points are belonging to at least two traces 12, and in a second step the plurality of the road map nodes 24 hypotheses are linked by the electronic computing device in order to generate the high definition road map 10.

[0033] In particular, Fig. 1 shows the first step, which may also be referred to as phase one of the method. The phase one includes: the generation of the network node hypotheses, wherein the network node may be referred to as the road map node 24, is an iterative process, beginning with the determination of the local center points 14 aggregated from the global navigation satellite system positions of the trace points 22, wherein the global navigation satellite system may be a GPS. These center points 14, which are referred to as condensation nuclei, are condensed in a way that the final iteration results in hypotheses for road network nodes 24.

[0034] In order to determine the condensation nuclei, a discrete global grid is used for space partitioning of the trace points 22, as shown in Fig. 1. All trace points 22 pertaining to one tile of the grid are aggregated by averaging their weight. The result of the aggregation determines the weighted center point 14 in each grid tile and is used as condensation nucleus for further processing steps.

[0035] The relevant parameters of the algorithm are at least the size of thi grid and the minimum number of trace points 22. If the minimum number of 22 per tile is below a threshold, no nucleus is determined in the relevant tile, i ne parameters to determine the weight of the trace point 22 may be the current GPS position of the trace point 22 and the spread of the adjacent trace points 22 in the trace. Once the condensation nuclei are determined, they iteratively condense to the final road network node hypotheses. This is done by defining the center point 14 at the position of each nucleus and growing a radius 18 around them stepwise as shown in Fig. 2. If in one step a radius 18 around a center point 14 covers an adjacent center point 14, both are fused to one resulting center point 14, as shown.

[0036] Fig. 2 shows another schematic top view of a road map 10. In Fig. 2, the fusion is shown, which is based on an aggregation algorithm. This algorithm determines the position and the weight of the resulting center point 14. The radius 18 around the center points 14 grows with each iteration without particular order among the center points 14. The final result is reached with the maximum radius 18 for the relevant location, which is shown in Fig. 2. The maximum radius 18 is determined from an estimation of the resulting road width for each specific location. This estimation is based on the local trace point 22 distribution. If the approximate local road width is known from a different data source, it may be fed to the algorithm in order to enhance the result.

[0037] The relevant parameter of the algorithm is the estimated road width, which is derived from trace point position spread.

[0038] Fig. 3 shows a flowchart according to an embodiment of the method. Phase two or the second step starts with the road network node candidates (RNNC) 26 determined from the road network node hypotheses, inferred along the traces 12 in phase one, as depicted in Fig. 4. In phase two, the nodes are linked along the course of the inferred road.

[0039] In the first step of phase two, directional road network link hypotheses are generated. This generation is based on the membership of traces 12 belonging to RNNCs. Each trace belonging to two RNNCs generates a directional road network link candidate (RNLC) between them along the driving direction of the relevant trace. This cross-linking (RNLC generation in flow chart Fig. 3) results in correct and incorrect link candidates as depicted in Fig. 5. The incorrect link candidates are identifiec corrected subsequently by a set of heuristics. These heuristics are applied i manner as depicted in the flow chart in Fig. 3. The key aspect here is adaptnig wo। m11y parameters during the iteration process. This adaption aims at minimal corrections on the RNLC network in the first loop, then iterates until no corrections are applied any more with the current parameterization, then adapts the working parameters and starts over. This approach prevents corruption of the network and makes the algorithm converge at the same time. After the final step of the procedure, road map node candidates 26 and road network link candidates are treated as final, thus forming the final road network.

[0040] In the accompanying drawings reference sign 26 describes road map node candidates, reference sign 28 refers to correct road network link candidates, reference sign 30 refers to incorrect road network link candidates, reference sign 32 refers to incorrect road map node candidates, and reference sign 34 refers to traces linked to the wrong road network section.

[0041] According to Fig. 3 in a first block S1 resort traces on branches are shown. A distinction is made between sections of the road network, wherein each section begins and ends at a branch. At such branches in the road network each trace point 22 linked to an RNNC is checked regarding its associations to the right section behind the branching. If a trace segment between trace points 22 links two trace points 22 belonging to different sections behind a branch (as depicted in Fig. 7), this subsequent trace point 22 is considered to be moved to an RNNC belonging to the correct section. The decision if movement is to be performed involves checking the agreement of the origin RNNC and the target RNNC after the movement is performed. The agreement is determined using all sensor data in both RNNCs (see also “Merge intermediate nodes”).

[0042] In one embodiment, block S2 describes a mesh removal. If two or more RNLCs starting at a single RNNC are joint to a single RNNC after a small number of intermediate RNNCs and each of these intermediate RNNCs has only one outgoing and one incoming RNLC, a mesh is present as depicted in Fig. 5. The mesh is removed by finding pairs of RNNCs in the mesh and merging them using their positions and weights.

[0043] In one embodiment, block S3 describes a directional cross-linking removal. If an RNNC contains traces driven in different directions, this RNNC is split. Directions are determined by weighting and merging the measured heading at a trace point 22 with the calculated heading from adjacent trace point 22 positions. The resulting val determined for all headings belonging to an RNNC. These headings are clu a threshold, which is a parameter of the algorithm. If the clustering results in unieiein directions, the RNNC is split as follows.

[0044] In one embodiment, block S4 describes the merging of solitary links. RNNCs having only one RNLC to the rest of the network are merged to surrounding RNNCs.

[0045] In one embodiment, block S5 describes the merging of loops. Mutually linked RNNCs in different directions are merged by position and weight if they have both additional ingoing and outgoing links in the same direction. Otherwise they are left for directional cross-linking removal with adapted working parameters in subsequent iterations.

[0046] In one embodiment, block S6 describes the merging of close nodes. Linked RNNCs with a distance below a threshold are merged by position and weight.

[0047] In one embodiment, block S7 describes the merging of intermediate nodes. An RNNC which is located between two adjacent RNNCs and has a weight below a threshold compared with the adjacent ones is distributed to the two adjacent nodes.

[0048] In one embodiment, block S8 describes the merging of false splits. Branches in the road topology which are represented by branches in the RNNC network but are branched too early (as depicted in Fig. 6) are merged. The algorithm for identification of such early branches is a key asset of this method. The algorithm uses all sensor data in the relevant RNNCs to be merged in order to determine their agreement. If this agreement is sufficiently high (depending on the working parameter of the current iteration) merging is applied. The relevant sensor data among others includes road border distance. Further parameters are among others the angle between ingoing and outgoing RNLCs.

[0049] In one embodiment, block S9 describes the splitting of inconsistent nodes. If the agreement of an RNNC is low (see “Merge false splits”) the node is split by finding two or more clusters of trace points 22 in the RNNC which have a high agreement. An additional test to gain confidence if splitting is required is based on the angle of the outgoing RNLCs at branches in the road network.

[0050] In one embodiment, block S10 describes an adaptation of local wo parameters for the loop removal.

[0051] In one embodiment, block S11 describes the regeneration of the links. Depending on the working parameters all links at an RNNC are removed and regenerated according to “RNLC generation”.

[0052] In one embodiment, block S12 describes the removal of separated nodes. RNNC that have no RNLC are removed.

[0053] In one embodiment, block S13 shows the removal of dead ends. RNNCs having ingoing but no outgoing RNLCs, which are linked to a number of trace points 22 below a certain threshold, which belongs to the working parameters, are removed.

[0054] In one embodiment, block S14 shows the removal of separated structures. Network fractions (structures of RNNCs linked by RNLCs) which have no connection to the rest of the network and which are linked to a number of trace points 22 below a certain threshold, which belongs to the working parameters, are removed.

[0055] In one embodiment, block S15 shows the removal of low frequented links. RNLCs linked to a number of traces 12 below a certain threshold, which belongs to the working parameters, are removed.

[0056] The last block S16 shows the removal of U-turn links similarly to the removal of low frequented links, however, with a lower threshold of linked traces 12 and an additional check if the angle between RNCLs is above a certain threshold, which belongs to the working parameters.

[0057] Figs. 1 to 7 show an automated road network extraction. Reference Signs 10 road map 12 trace 14 center point 16 center point heading 18 radius 20 assistance system 22 trace point 24 road map node 26 road map node candidate 28 road network link candidate correct 30 road network link candidate incorrect 32 road map node candidate incorrect 34 traces linked to the wrong road network section 36 motor vehicle S1 block S2 block S3 block S4 block S5 block S6 block S7 block S8 block S9 block S10 block S11 block S12 block S13 block S14 block S15 block S16 block S17 block

[0058] Also disclosed herein is a method for generating a high definition road map for a motor vehicle by an assistance system, wherein a position of the motor vehicle is detected by a global navigation satellite system of the motor vehicle and wherein at least one sensor data is detected by at least one detection device of the motor vehicle and wherein the current position of the motor vehicle and the current sensor data are linked to generate a trace point by an electronic computing device of the assistance system, characterized in that in a first step hypotheses for a plurality of road map nodes are generated by the electronic computing device along a plurality of trace points from the motor vehicle and / or a further motor vehicle and in a second step the plurality of the road map node hypotheses are linked by the electronic computing device in order to generate the high definition road map.

[0059] In the first step for generating the hypotheses an iteration may be performed, beginning with a determination of a local center point aggregated from the position of the trace points and the center points are condensed in such a way that a final iteration results in the hypotheses.

[0060] For the condensation, a discrete global grid may be used for space partitioning of the trace points.

[0061] In the second step, directional road map link hypotheses may be generated based on a membership of traces belonging to road map node candidates.

[0062] Also disclosed herein is an assistance system for generating a high definition road map for a motor vehicle, wherein the assistance system comprises at least one electronic computing device and is configured to perform a method according to the above.

Claims

1. A method for generating a high definition road map (10) for a motor vehicle (36) by an assistance system (20), wherein a position of the motor vehicle (36) is detected by a global navigation satellite system of the motor vehicle (36) and wherein at least one sensor data is detected by at least one detection device of the motor vehicle (36) and wherein the current position of the motor vehicle (36) and the current sensor data are linked to generate a trace point (22) by an electronic computing device of the assistance system (20), a chronological acquisition of trace points linked to the motor vehicle (36) comprising a trace (12);characterized in thatin a first step, hypotheses for a plurality of road map nodes (24) are generated by the electronic computing device by aggregating a plurality of trace points (22) from the motor vehicle (36) and / or from a further motor vehicle, and determining local centre points of the aggregated trace points (22); andin a second step the hypotheses for the road map nodes (24) are linked by the electronic computing device by identifying road map nodes (24) that belong to the same traces (12) in order to generate the high definition road map (10);the method further comprising:identifying a point in the road network at which the road network splits into two or more branches;identifying a first road map node that lies on a first one of the branches;determining, for each of the trace points (22) from which the first road map node is aggregated, whether to move the respective trace point (22) to a second road map node that lies on a second one of the branches, the determination being carried out by using the sensor data associated with the trace points (22) from which the first road map node and the second road map node are aggregated.

2. The method according to claim 2, characterized in thata discrete global grid is used for space partitioning of the trace points (22) in order to aggregate the plurality of trace points and determine the local centre points of the aggregated trace points.

3. An assistance system (20) for generating a high definition road map (10) for a motor vehicle (36), wherein the assistance system (20) comprises at least one electronic computing device and is configured to perform a method according to claim 1 or 2.01 08 25CLAIMS1. A method for generating a high definition road map (10) for a motor vehicle (36) by an assistance system (20), wherein a position of the motor vehicle (36) is detected by a global navigation satellite system of the motor vehicle (36) and wherein at least one sensor data is detected by at least one detection device of the motor vehicle (36) and wherein the current position of the motor vehicle (36) and the current sensor data are linked to generate a trace point (22) by an electronic computing device of the assistance system (20), a chronological acquisition of trace points linked to the motor vehicle (36) comprising a trace (12);characterized in thatin a first step, hypotheses for a plurality of road map nodes (24) are generated by the electronic computing device by aggregating a plurality of trace points (22) from the motor vehicle (36) and / or from a further motor vehicle, and determining local centre points of the aggregated trace points (22); andin a second step the hypotheses for the road map nodes (24) are linked by the electronic computing device by identifying road map nodes (24) that belong to the same traces (12) in order to generate the high definition road map (10);the method further comprising:identifying a point in the road map at which a road splits into two or more sections; identifying a first road map node that lies on a first one of the sections;determining, for each of the trace points (22) from which the first road map node is aggregated, whether to move the respective trace point (22) to a second road map node that lies on a second one of the sections, the determination being carried out by using the sensor data associated with the trace points (22) from which the first road map node and the second road map node are aggregated.

2. The method according to claim 2, characterized in thata discrete global grid is used for space partitioning of the trace points (22) in order to aggregate the plurality of trace points and determine the local centre points of the aggregated trace points.

3. An assistance system (20) for generating a high definition road map (10) for a motor vehicle (36), wherein the assistance system (20) comprises at least one electronic computing device and is configured to perform a method according to claim 1 or 2.

Images