Method and apparatus for providing a digital positioning map, computer program, and storage medium

Abstract

A method for providing a digital positioning map, the method having the following steps: a) obtaining road segments (C 11 …C nn ) by analyzing and processing data of a road network (100) to be mapped; b) classifying the road segments (C 11 …C nn ) considering the defining conditions of the road network (100); c) classifying the road segments (C 11 …C nn ) considering environmental sensor data; d) merging the road segments (C 11 …C nn ) into clusters (C1…Cn) considering the classifications performed in steps b) and c); e) jointly processing the clusters (C1…Cn) respectively by computational techniques; f) transmitting the created digital positioning map to a vehicle.

Description

Technical Field

[0001] The present invention relates to a method and an apparatus for providing a digital positioning map for a vehicle. The present invention also relates to a computer program. The present invention also relates to a machine-readable storage medium. Background Art

[0002] EP 3 293 489 A1 discloses a method and an apparatus for providing track bundles (Bahnbündeln) for map data analysis. A scheme for generating trajectory bundles (Trajektorienbündeln) for map data analysis is proposed. The scheme includes receiving data related to a limited geographical area. The data is collected by sensors of a plurality of devices moving in the limited geographical area and includes checkpoints indicating position, route, speed, time, or a combination thereof.

[0003] WO 2018 / 126228 A1 discloses a method for generating signs and lanes for a high-resolution map for an autonomous vehicle.

[0004] A positioning system having a feature-based digital positioning map for determining vehicle position and vehicle orientation is a core system component of an automated driving function.

[0005] A grid-based scheme or a road-segment-based scheme for creating a digital positioning map is known. In the grid-based scheme (Scheme A), a map area is divided into regular rectangles, where the data within each rectangle constitutes a partition, which is also referred to as spatial clustering (spatial cluster). Then, the data is processed independently of the data outside the rectangle.

[0006] Another known scheme (Scheme B) consists in grouping the data recorded along lanes for map creation purposes. In this case, a partition is a set of successive road segments.

[0007] Both of the two schemes A and B known per se allow mapping over a large area. Although grids can be calculated quickly and with low overhead in Scheme A, partitions are not considered for map density. When creating a map for a larger area (such as an entire country), there may be partitions with a large amount of data (such as big cities) and partitions with very little data (such as rural areas), which may cause problems in the map creation process.

[0008] The scalability of the process is known, for example, from "Tectonic SAM: Exact, Out-of-Core, Submap-Based SLAM" (Proceedings of the 2007 IEEE International Conference on Robotics and Automation, 2007).

[0009] Different methods of local clustering are also known. Summary of the Invention

[0010] The object of the present invention is to provide an improved method for providing a digital positioning map for a vehicle.

[0011] According to a first aspect, this object is solved by a method for providing a digital positioning map, the method having the following steps:

[0012] a) Obtaining road segments by analyzing and processing data of the road network to be mapped;

[0013] b) Classifying the road segments while taking into account the defining conditions of the road network;

[0014] c) Classifying the road segments while taking into account environmental sensor data;

[0015] d) Merging the road segments into clusters while taking into account the classifications performed in steps b) and c);

[0016] e) Processing the clusters together computationally;

[0017] f) Transmitting the created digital positioning map to the vehicle.

[0018] Advantageously, a high-precision digital positioning map can be provided based on the data processing performed together. Advantageously, in this way, the computing power for creating the digital positioning map can be scaled or parallelized very well.

[0019] According to a second aspect, this object is solved by a device which is set up to carry out the proposed method for providing a digital positioning map.

[0020] According to a third aspect, this object is solved by a computer program which includes instructions which, when the computer program is executed by a computer, cause the computer to carry out the proposed method.

[0021] According to a fourth aspect, this object is solved by a machine-readable storage medium on which the proposed computer program is stored.

[0022] Advantageous developments of the method are the subject of the development aspects.

[0023] One advantageous extension of the method provides for assigning road segments to clusters in step d). In this way, a defined assignment from road segments to clusters is carried out, so that parallel processing of road segments is advantageously not performed. In this way, a high degree of systematization during the creation of a digital positioning map is advantageously supported.

[0024] Another advantageous extension of the method provides for aggregating clusters with road segments in step d). In this way, the aggregation of clusters can be carried out starting from the origin until these clusters have a defined data size. In this way, good scalability during the creation of a digital positioning map is advantageously supported.

[0025] Another advantageous extension of the method provides for constructing the clusters to be of as similar a size as possible in terms of data extent in step d). In this way, good parallelization of the computational technical processing for the clusters can be achieved.

[0026] Another advantageous extension of the method provides that in step d), when a defined size is reached, the aggregation of clusters with road segments is no longer continued. In this way, the computing power for processing the individual clusters can be very well divided among the available computers.

[0027] Another advantageous extension of the method provides for merging road segments of the same type into clusters in step d). In this way, for example, objects of traffic infrastructure (such as tunnels, all roads with / without intersections, etc.) can be merged, which advantageously simplifies the processing of the digital map.

[0028] Another advantageous extension of the method provides for assigning defined numerical values to the road segments in steps b) and c). This advantageously supports the simple classification of road segments.

[0029] Another advantageous extension of the method provides for matching the data volume of one or more clusters to the computing power of the map creation device. In this way, good scalability of the computer used for creating a digital positioning map is advantageously supported.

[0030] Another advantageous extension of the method provides that the defined features already determined during the classification of the road segments in steps b) and c) can be used in the vehicle for the digital positioning map of the defined road segments. This advantageously supports providing regionally varying degrees of good map quality, whereby, for example, in a defined area, due to reduced quality, fully autonomous driving cannot be achieved with the aid of the digital positioning map.

[0031] Further measures for improving the invention are shown in detail below together with the description of preferred exemplary embodiments of the invention based on the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In the accompanying drawings:

[0033] Figure 1 An overview diagram is shown with an explanation of the working method of the proposed method;

[0034] Figure 2 A schematic diagram of the proposed method for providing a digital positioning map for a vehicle is shown. DETAILED DESCRIPTION

[0035] Figure 1 A schematic diagram with an explanation of the working method of the proposed method is shown. It can be seen that there are a series of small road sections C 1n …C nn A road network 100 of chunks of road segments is to be formed from the data of these road segments into a digital positioning map. It can also be seen that these road segments are combined into so-called clusters C1 ... Cn. Figure 1 For example, the road segment C of cluster C3 31 …C 3n It is shown with dashed lines in order to indicate that the road sections contained therein are guided in one or more tunnels.

[0036] It is proposed that the corresponding data of the individual road sections, which are detected by sensor technology, are processed in their entirety in clusters C1 ... Cn. In this way, the computational effort for creating or providing a digital positioning map can be scaled or parallelized very well. The computing power of different computers provided for this purpose can thus be very well balanced and used in a balanced manner.

[0037] Preferably, the road segments are gathered starting from a so-called origin until a certain, defined amount of data is reached. Such a starting point can be, for example, a road junction, a tunnel or another infrastructure element of the same type.

[0038] Figure 1 A total of three clusters C1, C2 and C3 are shown, wherein the road section C is shown in dashed lines in cluster C3. 31 …C 3n Cluster C1 represents a junction scene, and cluster C2 represents an approach area on a highway. All clusters C1, C2, C3 are shown only qualitatively, exemplarily and not to scale, and can obviously also have more or fewer road sections C than shown. 11 …C nn .

[0039] The proposed solution supports an efficient and balanced map creation process, which results in a high-quality localization map. As a result, balanced map creation is supported, which leads to higher efficiency in creating a digital localization map.

[0040] Similar to the above-mentioned, itself-known solution B, the proposed method proposes to assign road segments in important relevant areas in the form of clusters C1…Cn. However, additionally, qualitative parameters and complex critical scenarios, such as intersections, tunnels, etc., are considered, and these scenarios can thus be mapped consistently.

[0041] As a result, in this way and method, a data-driven clustering solution is implemented for the creation of a digital localization map.

[0042] The proposed solution is based on HAD mapping considering localization-important relevant parameters, such as road infrastructure (e.g., tunnels, critical road leadings, intersections, highway exits, etc.), connection quality, and localization quality. As a result, expected probability values are assigned to road segments of the road network to be mapped. These probability parameters are used to classify the road segments.

[0043] Based on the classification, clusters C1…Cn are formed, and these clusters determine the data to be processed together when creating a digital localization map. In this way, balanced map creation can be achieved, which has complex scenarios important for HAD map creation.

[0044] One advantage of the proposed solution is that it combines a data-driven, classifying or categorizing solution with the clustering process.

[0045] For the map creation of a digital localization map over a large area (HAD-Mapping, highly autonomous driving), it is very important to efficiently process a large amount of data and have a high map quality. The proposed solution allows scalable map creation, which results in a high-quality map.

[0046] Scaling can be achieved by grouping the road network to be mapped into individual road segments C 11 …C nn and processing these road segments in a data-technical manner in the form of groups (clusters). High quality of the digital localization map can be achieved through the merging of road segments, which is based on the qualitative classification of these road segments.

[0047] One advantage of this "soft clustering" (using "soft" criteria) is that the process of creating a digital positioning map can achieve good scalability, which is necessary for map creation over a large area. As a result, the creation of a digital positioning map can be performed in a computationally balanced creation program. In addition, after considering the important relevant influences in map creation through clustering, the digital positioning map created in this way advantageously has high quality.

[0048] The process of creating or providing a feature-based digital positioning map proposed above is further elaborated in detail below:

[0049] First, road segments C are calculated along the road network that is important and relevant for the digital positioning map 11 …C nn Here, the road segments C 11 …C nn represent an approximation of the possible trajectories of vehicles traveling in the road segments of the road network, where the size of the road segments C 11 …C nn is configurable.

[0050] A first classification or categorization of the road segments C 11 …C nn is performed for the assignment of features specific to the road network. These features encode the complexity and partitioning trends of the road segments C 11 …C nn For example, a low number can be assigned to separable road segments C 11 …C nn (e.g., due to exceeding data limits), and conversely, a high number can be assigned to road segments C that are not suitable for separation 11 …C nn For example, a low number can be assigned to separable road segments C (e.g., due to exceeding data limits), and conversely, a high number can be assigned to road segments C that are not suitable for separation.

[0051] Scenarios that are challenging for creating a digital positioning map (such as tunnels, intersections, highway exits, etc.) typically receive higher numbers, while rural roads are more likely to receive lower numbers.

[0052] The second classification performed is based on environmental sensor data detected and provided by vehicles through sensor technology. Here, the availability and quality of the environmental sensor data can be considered. For example, a high number can be assigned to the following road segments C 11 …C nn : On this road segment, it is only difficult or completely impossible to achieve GPS positioning (e.g., this may be the case in a tunnel). Conversely, a low number can be assigned to the following road segments: On this road segment, there is high-quality environmental sensor data and high positioning quality can be achieved.

[0053] Merge road section C using the standard clustering scheme according to the mentioned classification 11 …C nn wherein starting from the road section C with the highest classification 11 …C nn and then iteratively adding or aggregating adjacent road sections C 11 …C nn to clusters C1…Cn until another separate road section C 11 …C nn has the highest classification, thereby generating additional clusters C1…Cn. It is desired that the clusters have as similar data sizes as possible so that the computer load for calculating the digital location map is uniform Here, specific numerical values cannot be described, and the specific numerical values depend on various parameters such as the number of available computers, the size of the road network to be mapped, etc. Characterized or same - type scenarios are thus advantageously kept unified to a certain extent in one cluster and thereby simplify their processing

[0054] In the case of tracking the cluster size, in this way, a mathematical algorithm can be "run (geführt)" for similar cluster sizes, where the probability of further adding road sections to the cluster in the case of "satisfied size" is reduced

[0055] Figure 2 Shows the principle flow chart of the illustration with the proposed method

[0056] In step 200, the obtaining of road section C 11 …C nn is performed by analyzing and processing the data of the road network 100 to be mapped

[0057] In step 210, the classification of road section C 11 …C nn is performed considering the defined conditions of the road network 100

[0058] In step 220, the classification of road section C 11 …C nn is performed considering the environmental sensor data

[0059] In step 230, considering the classifications performed in steps b) and c), the merging of road section C 11 …C nn into clusters C1…Cn is performed

[0060] In step 240, the clusters C1…Cn are processed together using computing technology respectively

[0061] Finally, in step 250, the created digital positioning map is passed to the vehicle.

[0062] The map creation device for creating the digital positioning map is preferably a server device in the cloud, on which the proposed method runs as software. This supports the simple adaptability of the method. Advantageously, the respective clustering matches the computing power of the individual computers of the map creation device in terms of data scale, thereby supporting the balanced occupancy rate of the computer infrastructure for creating the digital positioning map in this way.

[0063] When implementing the present invention, those skilled in the art can also implement embodiments not previously described.

Claims

1. A method for providing a digital positioning map, the method having the following steps: a) Obtaining road sections (C 11 …C nn ) by analyzing and processing data of the road network (100) of the to-be-drawn map b) classify the road segment (C 11 …C nn ) under consideration of the defined conditions of the road network (100); c) classify the road section (C 11 … C nn ) considering environmental sensor data; d) Merge the road segments (C 11 …C nn ) into clusters (C1…Cn), taking into account the classification performed in steps b) and c); e) jointly processing the clusters (C1…Cn) with computing techniques respectively; f) transmitting the created digital positioning map to a vehicle; Among them, The digital positioning map is a HAD map; wherein the data volume of one or more clusters (C1…Cn) is matched to the computing power of the map creation device; wherein an expected probability value is assigned to a road section of a road network to be mapped, and wherein the probability value is used for classifying the road section; Among them, the road section (C 11 …C nn ) represents an approximation of the possible trajectory of a vehicle traveling in the road sections of the road network; wherein the road sections converge from a starting point until a defined data volume is reached, and the starting point includes intersections, tunnels or other infrastructure elements of the same type; Among them, in step d), the convergence of the clusters (C1…Cn) and the road segments (C 11 …C nn ) is performed until these clusters have a defined data size in order to support scalability when creating a digital positioning map. Among them, when the defined data size is reached, the clusters (C1…Cn) no longer continue to converge with the road segments (C 11 …C nn ); wherein the clusters (C1…Cn) are constructed to be as similar in size as possible in terms of data scale in order to enable parallelization of the computing technique processing for the clusters.

2. The method according to claim 1, wherein In step d), the road segments (C 11 … C nn ) are assigned to the clusters (C1…Cn).

3. The method according to any one of the above claims, wherein In step d), road segments of the same type (C 11 …C nn ) are merged into clusters (C1…Cn).

4. The method according to any one of the above claims, wherein, In steps b) and c), assign a defined numerical value to the road section (C 11 …C nn ).

5. The method according to any one of the above claims, wherein The defined features obtained during the classification of the road sections in steps b) and c) can be used in the vehicle for the digital positioning map of the defined road sections.

6. A device, the device being arranged to implement the method according to any one of claims 1 to 5.

7. A computer program, the computer program comprising instructions which, when the computer program is implemented by a computer, cause the computer to implement the method according to any one of claims 1 to 5.

8. A machine-readable storage medium, on which the computer program according to claim 7 is stored.

Images