Methods for evaluating data packets acquired during recording drives

Abstract

Method for evaluating data packets (1) acquired during a large number of award runs (2) to generate an HD map of route for vehicles, comprising the following steps: a) Receiving a large number of data packets relating to recording journeys (2) with recording vehicles, wherein the data packets each contain a trajectory (3) of the recording journey (2) and further landmark data (4) acquired by the recording vehicle during the recording journey (2) along the trajectory (3); b) Selecting a subset (13) of the set of data packets received in step a); c) Performing an alignment of the trajectories (3) in the data packets selected according to step b) to determine an aligned item graph (5); d) Performing an alignment of trajectories (3) from further data packets to the aligned item graph (5) determined according to step b).

Description

State of the art

[0001] The procedure described here concerns the generation of digital high-resolution map data (so-called HD maps) that can be used for applications of highly automated and possibly autonomous driving.

[0002] Digital maps play a crucial role in the implementation of assisted and automated driving functions in vehicles. Digital maps are data sets that describe the situation along a vehicle's route and, as a supplement to real-time environmental data, enable significantly improved implementation of assisted and automated driving functions than would be possible based solely on real-time environmental data.

[0003] Such digital maps (unlike simpler digital navigation maps used by road navigation systems) contain data about the situation in the immediate vicinity of the route. This allows assisted and automated driving functions to, in a sense, see beyond the visibility horizon of the vehicle's environmental sensors and to better assess and evaluate the environmental data acquired by these sensors. For example, if the environmental sensors detect a specific traffic situation in the immediate vicinity of the vehicle, this situation can be better assessed using data from a digital map that describes the further course of the road. For instance, conclusions can be drawn about how the traffic situation is likely to develop as the journey continues along the road.

[0004] Such digital maps regularly contain a significant amount of data describing the roadway and its immediate surroundings. This includes roadway and marking data, data on objects in the vicinity of the roadway, road topology and lane-level geometry, as well as semantic contextual information such as traffic signs and traffic light positions. The accuracy and depth of information provided by such maps regularly exceeds the amount of information that can be acquired using sensors alone. Extracting comprehensive semantic information directly from raw sensor data in real time presents a major challenge, especially in complex traffic scenarios. By using digital maps that incorporate the aforementioned data, the environmental data acquired by sensors can be enriched and analyzed more efficiently.In particular, the use of digital maps in many situations or for many types of information means that it is only necessary to correctly evaluate the information contained in the digital maps using the sensor data.

[0005] Such digital maps (unlike simple digital navigation maps used by road navigation systems) are not typically available in a uniform format for the entire road network. These digital maps are provided by a wide variety of service providers. Unlike simple digital navigation maps, they are also often not available in a fixed, uniform global coordinate system. The data underlying these digital maps is frequently acquired through test drives of vehicles along routes / roads. Highly individualized coordinate systems are often used to describe the position of the data on the map, referring, for example, to the position and, if applicable, the orientation of the vehicle that performed the test drive. In this case, a coordinate system exists that is referenced to the route.Often, it is not possible to precisely convert digital maps, or the positional data contained within them, into a uniform global coordinate system. This is frequently because, during the acquisition of the data underlying the digital map, no precise positions are available that would be necessary for converting positional data into a global coordinate system. This might be because no GNSS signal is available at all, or the GNSS signal quality during acquisition is insufficient to determine an exact position. A typical example would be a digital map describing the surroundings of a vehicle in a tunnel. In such a situation, no GNSS signal is available. Nevertheless, digital maps are required for assisted and automated driving functions.

[0006] The process described here concerns the processing and preparation of swarm-collected data for the creation and updating of high-resolution digital maps—specifically, so-called HD maps. The initial creation (production) of an HD map and its subsequent updating are distinct tasks that nevertheless share many similarities, particularly regarding the initial collection of the necessary data for creation and / or updating. When the creation of an HD map is mentioned below, this generally also refers to its updating, and vice versa.

[0007] Traditionally, high-resolution maps for applications of highly automated and possibly autonomous driving are created by having manufacturer-specific fleets record data for mapping or by using other information sources to obtain data.

[0008] This traditional approach presents classic map providers with two fundamental challenges today: the ability to provide map updates and the difficulty of collecting sufficient data for very high map accuracy (granularity). Both challenges are related to the size of the vehicle fleet that collects the data used for the map. The data collected so far by classic map providers is typically only sufficient for mapping with the granularity common to conventional in-vehicle navigation systems. Such maps are usually referred to as SD cards. For highly automated and potentially even autonomous driving functions, maps are required that contain data for every single lane of a road and also describe the surrounding area in detail.Such cards are usually grouped under the term "HD card" introduced above.

[0009] To address the two problems of sufficiently frequent map updates and adequate map granularity, approaches to generating maps based on swarm data are currently being discussed. Swarm data is preferably collected indirectly by the customers of vehicle manufacturers – that is, by every single driver who operates a vehicle. During everyday driving, sensor data from the vehicles is sent to the vehicle manufacturer and made available to the map service provider. The map service provider processes this data to create highly accurate HD maps.

[0010] Creating maps from such swarm data is a complex process that typically involves several steps. The data processing effort required to generate high-definition maps from swarm data is regularly very high. Therefore, measures that reduce this data processing effort are generally desirable.

[0011] An important step is the alignment of data acquired during individual vehicle recording runs. The problems associated with varying degrees of accuracy in position determination and the different coordinate systems in which sensor-acquired data from these recording runs are stored have been described above.

[0012] The purpose of the method described here is to disclose a novel approach for the particularly efficient alignment of individual recording runs relative to one another. This purpose is achieved by the method according to the features of the independent claims. Further advantageous embodiments are specified in the dependent claims, as well as in the description and, in particular, in the description of the figures. It should be noted that a person skilled in the art can combine the individual features in a technologically meaningful way and thereby arrive at further embodiments of the invention.

[0013] This section describes a method for evaluating data packets acquired during numerous award runs to generate an HD map of the route traveled by vehicles, comprising the following steps: a) Receiving a large number of data packets relating to recording journeys with recording vehicles, wherein the data packets each contain a trajectory of the recording journey and further landmark data acquired by the recording vehicle during the recording journey along the trajectory; b) Selecting a subset of the set of data packets received in step a); c) Performing an alignment of the trajectories in the data packages selected according to step b) to determine an aligned item graph; d) Performing an alignment of trajectories from further data packets to the aligned item graph determined according to step c).

[0014] To create a high-definition (HD) map of the environment (environmental map), large amounts of data must be collected regularly. In particular, it is also frequently necessary to combine data from various sources. This includes, in particular, data collected from different recording runs by various recording vehicles along the same route over a period of time. Such recording runs, and especially the data packets collected during them, are commonly referred to as "drives." A "drive" preferably comprises a data packet containing all the data recorded during a specific recording run. A "drive" typically includes data acquired during a segment of a recording vehicle's journey.For example, such a section can have a length between 200 meters and 5 kilometers. Recording drives are typically divided into such sections. Each of these sections constitutes a separate recording drive for processing in subsequent processing steps. The data from different drives or recording drives concerning specific route sections must usually be combined to obtain a sufficient amount of data for a particular route section in order to generate suitable, high-precision map data.

[0015] During these recording drives, raw data is collected using sensors on the recording vehicles and made available in the form of recorded drives or "drives". In a subsequent processing step, highly precise map data is then created using this raw data, which becomes part of the high-precision map.

[0016] Such a recording drive also includes a trajectory that describes the route taken by the recording vehicle during the drive. In preferred implementations, this trajectory describes the drive in global coordinates. However, due to inaccurate vehicle position determinations during the drive, this trajectory is often inaccurate and contains errors. Data and trajectories acquired in different recording drives can exhibit different errors or deviations. These trajectories are aligned according to steps c) and d) of the described procedure.

[0017] For the subsequent processing step, it is crucial that the data acquired during different recording runs can be processed together. This requires aligning (also called mapping or referencing) the recording runs with each other. The terms "alignment," "mapping," and "referencing" are used synonymously below. These terms refer to relating the different recording runs to each other so that joint data processing of the data from the various recording runs is possible for the subsequent processing step. In other words: If a specific object (e.g.,If a specific building in the vicinity of a roadway is detected in a slightly different location during a first recording run due to inaccuracies in the data determination in common coordinates than during a second recording run, then there is indeed a deviation in the common coordinates used during the recording runs, which usually affects all features detected during the recording runs.

[0018] Mapping the various recorded drives to each other typically uses information about vehicle movement in the form of wheel rotations (odometry), rough global vehicle positions (GNSS), and observations of the local environment from sensors such as cameras, laser scanners, radar, etc., which can be referenced or mapped to one another. Due to the inaccuracy of each of these sensor types, and especially due to noise in the sensor signals acquired, this alignment (or mapping or referencing) is usually a complex optimization problem requiring the best possible solution.

[0019] Especially when large quantities of individual recording drives are to be processed together in the subsequent processing step, the mapping is a very complex process.

[0020] Here, a method for the efficient and highly parallelizable processing of a large number of "drives" or recording drives for specific route sections during alignment is proposed.

[0021] It is proposed that the alignment of different recording runs not be solved in a single large optimization problem, but rather that a subset of the received data packets relating to recording runs be selected in step b) and pre-aligned in step c). Further recording runs, or rather their associated data packets, are then aligned subsequently in step d). This is described as determining a common pose graph for a plurality of recording runs, which contains a kind of computational rule for how the individual recording runs and their trajectories must be processed together so that the data collected in the data packets can be combined to build high-definition maps. The "common pose graph" can be understood as a kind of location-based computational rule that applies to the individual features (trajectories, landmark data, etc.).) can be applied to the data packets to process them together to create an HD map. The common pose graph is preferably determined in step c). Due to the reduced subset of recording trips, the optimization problem to be solved here is significantly smaller than it would be if all recording trips were processed together directly.

[0022] The fact that less data is used to determine the pre-defined pose graph according to step c) is usually less relevant to the accuracy of the determined pose graph. The deviations that would arise if all available recording data were processed together are small. The pose graph is preferably not corrected by performing step d). The alignment of the subsequent trajectories is preferably carried out according to the pre-defined pose graph determined in step c) without further correction of the pose graph. In fact, this does not cause any major difficulties for processing the data to produce HD maps, because HD maps are needed to better assess the vehicle's surroundings. Exact global coordinates are not usually required for this purpose.The additional data packages and recording journeys processed according to step d) are used in particular to collect and integrate statistical information about the behavior of road users in lanes (speeds and other data about driving behavior).

[0023] The method is particularly preferred if a SLAM algorithm is used in step c).

[0024] Another approach for step c), which can be used in combination and / or as an alternative to a SLAM algorithm, is the use of a Kalman filter, which describes a kind of “alignment step” for combining GNSS sensor signals as the first data stream and IMU data as the second data stream.

[0025] "SLAM" stands for "Simultaneous Localization and Mapping." This refers to the simultaneous localization and mapping of a robot. SLAM algorithms are particularly well-known from the field of robots that move through an unknown environment while simultaneously creating a map of that environment and continuously determining their position within that map.

[0026] The SLAM method compares all available sensor-acquired additional information about a vehicle's surroundings with information already available in the form of a map, and the map is then improved or corrected using this additional information. This involves a continuous supplementation or correction of map data based on additional sensor data. The SLAM approach is comparable to the way the human brain typically perceives its surroundings. A person, while observing a room, knows what the room looks like and what is in it. A kind of virtual map of the room exists in the person's mind. Through spatial perception, the person automatically supplements their mental image of their surroundings, or rather, the virtual map of the environment they have in mind.

[0027] Regarding the processing of swarm data for the production of highly accurate maps, the SLAM method is preferably used to obtain aligned driving trajectories or route profiles of the individual journeys from which the swarm data was collected. The sensor measurements of all sensors of a vehicle are recorded relative to the vehicle's position when determining the respective data. Errors in determining the vehicle's position are reflected in the driving trajectory or route profile of the vehicle's journey. Aligning the driving trajectories or route profiles of the individual recording journeys from the data collection process can typically be achieved using the SLAM method. Preferably, landmarks detected by the vehicle's sensors from different recording journeys are matched and used to reference / align the recording journeys relative to each other.In particular, such a step makes it possible to use swarm data to create HD maps.

[0028] A special case of the SLAM algorithm is the so-called Full-SLAM algorithm, which is particularly suitable for carrying out step c) of the described procedure.

[0029] A full SLAM algorithm (also known as offline SLAM or batch SLAM) is a variant of the SLAM problem in which the entire trajectory of a vehicle and the map of the environment are optimized simultaneously, based on all sensor data collected from the beginning to the end of the movement.

[0030] The procedure is particularly advantageous if, in step c), an optimization problem is created based on landmark data contained in different data packages and which can be assigned to each other, and which is then solved to carry out the alignment.

[0031] Landmark data is preferentially integrated into the data packets of different recording runs in such a way that landmarks present in one recording run can be found again in another. These landmarks have a unique position. Landmarks present in the data packets of two (or more) recording runs can be used to align these recording runs. Even if the exact position of the recording vehicle is not known in both recording runs, a relative position of the recording vehicles or the trajectories of the recording runs to each other is still possible using such landmarks.

[0032] It is particularly advantageous if, in step d), trajectories from each additional data package are individually aligned to the pre-directed pose graph determined in step c).

[0033] Preferably, the aligned position graph from step c) is assumed to be final for the further execution of step c). All subsequent recording runs and their trajectories are aligned to this position graph. Preferably, in steps c) and d), transformation data is created for each recording run and each trajectory of a recording run, enabling the transformation of data collected during the recording run, starting from the trajectory, onto the (common) position graph. It is particularly preferred that this (common) position graph is defined in step c). Each subsequent trajectory or recording run to be aligned is aligned to this position graph. This can be done individually, making it possible to perform the process in parallel.The aligned pose graph is fixed after step c) and is preferably used as the final reference for the further alignment of any other trajectory / recording run according to step d).

[0034] Furthermore, it is preferred if the landmark data in the data packets were acquired using at least one sensor of one of the following sensor types on the recording vehicles: - Camera sensors, - Lidar sensors, - Radar sensors, and - Ultrasonic sensor.

[0035] These are common sensor types that can be used to collect data during recording drives with normal customer vehicles.

[0036] Furthermore, it is preferred if the recording journeys each concern a segment of a journey of a recording vehicle which has a minimum length, wherein the minimum length is greater than 500 meters.

[0037] Preferably, the average segment length is not significantly higher than the minimum length. In special situations, considerably longer segments can be used in certain areas. A uniform segment length greatly simplifies data processing.

[0038] Furthermore, it is preferred if the recording vehicles used to carry out the recording runs are customer vehicles, the recording runs being carried out in the regular operation of the customer vehicles and data packages relating to the recording runs according to step a) being transmitted via an online interface to a central location for carrying out the described procedure.

[0039] This describes the swarm-based execution of recording drives in the field.

[0040] It is particularly preferred if, according to step d), map data is created both based on data aligned according to step c) and based on data aligned according to step d).

[0041] Preferably, in the further processing of data for the production of HD maps, no distinction is made as to whether the data originates from recording drives aligned according to step c) or from recording drives aligned according to step d).

[0042] Furthermore, it is preferred if data concerning trajectories contained in the data packets were determined using GNSS systems installed in the recording vehicles.

[0043] Such GNSS systems are not usually highly precise. However, by aligning them according to steps c) and d), it is still possible to use this data to create HD maps.

[0044] This section will also describe a mapping system for producing HD maps with - a data collection facility for collecting data, wherein segments of recorded journeys are collected in the data collection facility, - a central data storage device for storing segments of recorded journeys, wherein the central data storage device is designed in such a way that new segments can be inserted into the central data storage device and existing segments can be removed from the central data storage device, and - a data alignment device for generating an aligned pose graph according to a described procedure with steps a) to d) on the basis of the segments of recording journeys stored in the central data storage.

[0045] It should be noted that the special advantages and design features described above are also applicable and transferable to the mapping system.

[0046] The mapping system is preferably operated at a central instance and serves to provide users with constantly up-to-date HD maps.

[0047] The invention and its technical context are explained in more detail below with reference to the figures. The figures show preferred embodiments, to which the invention is not limited. It should be noted in particular that the figures, and especially the size relationships shown in the figures, are only schematic. They show: Fig. 1: schematically illustrates the difference between raw data and aligned raw data; Fig. 2: schematically an approach to performing a raw data alignment in a single alignment step; Fig. 3: schematically, a staged approach to performing raw data alignment with an initial pre-alignment and subsequent data alignments; and Fig. 4: a flowchart and a device for creating HD maps using the method described herein.

[0048] In Fig. Figure 1 illustrates what happens when data contained in data packets relating to recording journeys is aligned according to the procedure described here.

[0049] In the left part of Fig. Figure 1 shows examples of unaligned raw data. The raw data 1 here consists, for example, of landmark data 4, which describes specific objects along a road or lane. These can be, for example, traffic signs, buildings, lane markings, or similar structures. Landmark data 4, represented by stars, is part of a first data package collected during an initial recording run. Landmark data 4, represented by a star and a circle, is part of a second data package collected during a second recording run. Both recording runs were conducted on the same road or lane. Each recording run, or data package, also includes a trajectory 3, which describes the route traveled by the recording vehicle during the recording run.However, due to inaccuracies in the position determination of the respective recording vehicle, there are deviations between the trajectories 3, which also lead to inaccuracies in the recorded positions of the landmark data 4 in the individual data packages.

[0050] In the right part of Fig. Figure 1 illustrates what is to be achieved through alignment according to the procedure described here. The goal is to assign the recognized landmark data 4 in the two data packets to each other as accurately as possible. Inaccuracies in the position determination of the respective recording vehicle and deviations between the trajectories 3 in the two data packets are compensated for. The aim is not to map the trajectories 3 to each other exactly, as different recording vehicles may not have driven in exactly the same way during the recording runs and may, for example, have driven in different lanes. A uniform pose graph 5 will be determined that relates the raw data (landmark data) existing in the data packets to each other as closely as possible, or with which aligned raw data 14 can be generated.The pose graph 5 particularly preferentially establishes a mathematical relationship between the trajectories 3 of the different recording drives or the respective associated data, which can be used to combine the different recording drives to produce and / or update an HD map.

[0051] According to the Fig. 2 and Fig. Section 3 now describes the multi-stage alignment of raw data relating to recording drives, as carried out using the procedure described here according to steps a) to d). To perform the alignment, an optimization problem is solved in which the landmark data contained in the individual data packets of the recording drives are related to each other and an optimal solution is found to align the trajectories of the individual recording drives to each other in such a way that a common pose graph for the recording drives is found.

[0052] According to Fig. 2 and Fig. Only the trajectories of the recorded journeys are shown in each of the 3. The points on the trajectories according to the Fig. 2 and Fig. 3 represent individual vehicle poses 18. The arrows represent the relative differences of the vehicle poses 18 during the respective recording runs.

[0053] According to Fig. 2. This is considered a unified optimization problem, which must be set up and solved for alignment using all available recording runs. The arrows between the individual vehicle positions 18 indicate this.

[0054] According to Fig. Figure 3 illustrates the approach underlying the procedure described here. A subset 13 is selected from the total set of recording journeys (represented here only as trajectories 3). Initially, only these trajectories 3, or recording journeys, are aligned to determine a forward-directed pose graph 5. This is represented here by arrows between the individual vehicle poses 18, which are assigned to process step b) (selection) and process step c) (determining the forward-directed pose graph). Subsequently, the remaining recording journeys, or trajectories 3, are assigned to the forward-directed pose graph determined according to step c). Arrows assigned to step d) are shown to illustrate this.

[0055] It is evident that this in Fig. The optimization problem shown in step c) is smaller or comprises less data than the one in Fig. The optimization problem shown in section 2 involves considering all recorded journeys or trajectories in a single optimization problem. This advantage increases significantly as soon as the number of journeys optimized in section d) considerably exceeds the number used in section c).

[0056] According to Fig. 4 The procedure for evaluating data packets acquired during recording drives according to steps a) and b) in the context of creating and / or updating HD maps is schematically illustrated.

[0057] The procedure is executed using a data processing system 15, which is schematically divided into components here. This division is not mandatory. The data processing system 15 can also be structured differently. The representation is only exemplary. The components of the data processing system 22 shown here are a data collection unit 6, a data storage unit 10, which is further divided into a data update unit 7 and a data reduction unit 8, and a data processing unit 9.

[0058] The data collection device 6 serves primarily to collect raw data that can be used to create HD maps. Raw data is collected, for example, from recording trips 2 with fleet vehicles 16 in the field. Such vehicles preferably record raw data during their regular operation (swarm-based), which can then be used with the data processing system 15 described here to create HD maps. The collected data is preferably transmitted from the fleet vehicles 16 in the field to the data processing system 15, operated by a central instance, via a transmission interface 17 (e.g., a mobile network connection). The raw data initially contains recording trips in unsegmented form. In the raw data, the collected data is typically simply represented along the routes that the vehicles regularly traveled in the field. This could, for example, be the commute to work and / or shopping trips of the drivers of the vehicles in the field.In data collection facility 6, the raw data is processed to make it particularly suitable for further processing in the creation and / or updating of HD maps. This preferably includes, in particular, a division of the recorded journeys captured in the raw data, namely a so-called segmentation, in which more easily processable segments of recorded journeys are created from the raw data.

[0059] The data collection process is initially independent of the subsequent creation and / or updating of the HD map. The segments of recorded journeys generated by the data collection unit 6 are stored in the data storage unit 10 to be used later for the creation and / or updating of HD maps. Here, the data storage unit 10 is simplified to include a data update unit 7 and a data reduction unit 8. The purpose of the data storage unit 10 is to store and provide only data, segments, or recorded journeys that offer relevant added value for the creation and / or updating of an HD map. For example, if several thousand segments of recorded journeys along a main road are collected by the data collection unit 6, it is generally not practical to store all of these segments in the data storage unit for later use in creating HD maps.The information in the individual segments is then largely redundant. Rather, the task of the data storage facility 10 is to provide the most up-to-date segments possible, ensuring uniform coverage of the roads and routes to be mapped. Data update 7 prioritizes the addition of relevant, newly added segments. Data reduction 8 prioritizes the removal of segments that are outdated or have been replaced by better, more current segments. The data storage facility 10 is therefore primarily configured to always provide the most up-to-date raw data set possible, exhibiting the most consistent quality possible across all roads and routes to be mapped.

[0060] Through appropriate segmentation according to steps a) and b), a granularity of data stored in data storage facility 10 is achieved. This granularity increases the efficiency of data storage facility 10. If complete real-world vehicle journeys in the field (e.g., a vehicle user's commute to work or shopping) were stored in data storage facility 10, these complete real-world journeys would also have to be added or removed. Segmentation makes it possible to add only particularly relevant segments of a real-world journey, which offer added value for the creation and / or updating of the HD map, to the data set provided in data storage facility 10.

[0061] The data processing unit 9 accesses the data stored in the data storage unit 10 and is responsible for the actual generation and provision of the HD map based on the raw data. The data processing unit 9 can preferably be used both for generating a (new) HD map and for updating map sections of an existing HD map.

[0062] The described procedure, comprising steps a) to d), also takes place in data processing unit 9. This procedure is a preparatory step by which the segmented recording runs provided by data storage unit 10 are preprocessed for the creation and / or updating of HD maps. The segmented recording runs are regularly not aligned with each other. As described above, deviations and inaccuracies in the trajectories of the recording runs must be corrected, or the recording runs must be aligned with each other, so that the data provided by data storage unit 10 can be used for the creation and / or updating of HD maps.

Claims

[1] Method for evaluating data packets (1) acquired during a large number of award runs (2) to generate an HD map of route for vehicles, comprising the following steps: a) Receiving a large number of data packets relating to recording journeys (2) with recording vehicles, wherein the data packets each contain a trajectory (3) of the recording journey (2) and further landmark data (4) acquired by the recording vehicle during the recording journey (2) along the trajectory (3); b) Selecting a subset (13) of the set of data packets received in step a); c) Performing an alignment of the trajectories (3) in the data packets selected according to step b) to determine an aligned item graph (5); d) Performing an alignment of trajectories (3) from further data packets to the aligned item graph (5) determined according to step b). [2] Method according to claim 1, wherein in step c) a SLAM algorithm is used. [3] Method according to claim 1 or 2, wherein in step c) an optimization problem is created based on landmark data (4) contained in different data packages and which can be associated with each other, and which is solved to carry out the alignment. [4] Method according to any of the preceding claims, wherein in step d) trajectories (3) from each further data packet are individually aligned to the pre-directed pose graph determined in step c). [5] Method according to any of the preceding claims, wherein the landmark data (4) in the data packets were acquired using at least one sensor of one of the following sensor types on the recording vehicles: - Camera sensors, - Lidar sensors, - Radar sensors, and - Ultrasonic sensor. [6] Method according to one of the preceding claims, wherein the recording journeys (2) each relate to a segment of a journey of a recording vehicle which has a minimum length, wherein the minimum length is greater than 500 meters. [7] Method according to one of the preceding claims, wherein the recording vehicles with which the recording journeys (2) are carried out are customer vehicles, wherein the recording journeys (2) are carried out in a regular operation of the customer vehicles and data packets relating to the recording journeys (2) according to step a) are transmitted via an online interface to a central location for carrying out the described method. [8] Method according to any of the preceding claims, wherein the map data according to step d) are created based both on data aligned according to step c) and on data aligned according to step d). [9] Method according to one of the preceding claims, wherein data relating to trajectories (3) contained in the data packets were determined using GNSS systems installed in the recording vehicles. [10] Mapping system for producing HD maps with - a data collection device (14) for collecting data, wherein segments (3) of recording journeys (2) are collected in the data collection device (14), - a central data storage device for storing segments (3) of recording journeys (2), wherein the central data storage device is designed such that new segments (3) can be inserted into the central data storage device and existing segments (3) can be removed from the central data storage device, and - a data alignment device (17) for generating an aligned pose graph (5) according to a method according to one of claims 1 to 9 on the basis of the segments (3) of recording journeys stored in the central data storage.

Images