EP3388307A2 - Merging of infrastructure-related data, in particular of infrastructure-related data for railway vehicles - Google Patents

Abstract

A method is disclosed for creating at least one new data record (60), in particular comprising infrastructure data for vehicles, wherein
the method comprises the following steps:
Reading (S1, S2) at least one existing data set from at least one source (10, 10a, 10b, 10c, 10d), wherein
the at least one existing record contains data describing an infrastructure section,
Conversion (12, 12a, 12b, 12c, 12d) of the at least one data record to at least one new data record (60) with a uniform format, wherein
the at least one new data record (60) is suitable for enabling the operation of vehicles, in particular by means of assistance from driver assistance systems, wherein
the at least one new data set is suitable for enabling the operation of vehicles, in particular by means of assistance from driver assistance systems.

Description

The present invention relates to a method for merging infrastructure-related data, in particular infrastructure-related data for vehicles, in particular rail vehicles, a computer program product for carrying out this method and an apparatus for carrying out this method.

For the operation of rail vehicles on railway lines and also for the management, maintenance and repair of trackside infrastructure, a data-based overview of the entire infrastructure and distance-related information is required.

Today there are a variety of infrastructure-related data for railway rail networks, which have grown historically, from different sources with varying degrees of accuracy and direction, sometimes with errors and incomplete ones.

Typically, such infrastructure-related data includes information about the route characteristic, such as the grade, information on maximum allowable speeds or the positions of stops, and the like. Also information regarding the maintenance intervals or the general condition of the route are conceivable.

So far, special data sets have been used for certain tasks, but they are also inconsistent, incomplete and sometimes only reflect simplified representations of reality in the respective databases.

These available datasets are for today's and tomorrow's rail vehicle operation and management, maintenance and service requirements Repair inadequate as data is needed consistently, completely, with the appropriate level of detail and data quality.

So far, the separately existing records for a route network for the respective purpose such as management of the route for operation and maintenance were sufficient. However, the application of a driver assistance system now requires the fusion and unification of several previously existing data sets.

These considerations apply not only to modern rail vehicles, but can be transferred in the same way to other track-bound vehicles. Under track-bound vehicles are to be understood hereinafter as vehicles which follow a fixed track and these can not leave, or leave only under certain conditions, such as maglev trains or driverless transport vehicles.

This invention is based on the task of generating new data records from existing infrastructure-related data sets that are consistent, complete, with the corresponding level of detail and with the corresponding data quality.

This object is achieved by means of the independent claims, advantageous developments are the subject of the dependent claims.

According to the invention, a method is provided which creates at least one new data record which contains infrastructure data for vehicles.

The creation of the new data record preferably proceeds essentially by first reading in at least one existing data record from at least one source and subsequently converting it to at least one new data record having a uniform format, wherein the at least one existing data record contains data which includes an infrastructure segment The standardized format is suitable for making the new data record accessible to a driver assistance system of a vehicle, thereby enabling the operation of such a driver assistance system. Thus, under further processing, especially conversions of infrastructure-related coordinates, Understand plausibility checks for data consistency and other processing steps.

Preferably, in a further step, a new data record is created by means of a data fusion of at least two existing data records, whereby the information content of the new data record is at least as high as the information content of the at least two existing data records.

For example, two records are merged so that both records describe only a separate part of an infrastructure section. The merged, new data set now contains the data and information of both data sets, whereby the information content of this data set is higher than that of the two individual data sets.

Preferably, an evaluation of a quality of the data records takes place, whereby certain data are identified on the basis of the quality as trusted data records for the further fusion.

The quality of a data record preferably describes how much the data record is suitable for the description of an infrastructure section and the further processing.

For example, the evaluation of the quality can be done by analyzing a data set for its information content. If there are two plausibly evaluated data sets that describe the course of the same route, which is characterized by strongly varying slope, then the data set with the higher quality is to be evaluated, which has the higher resolution, that has more nodes. As a result, this record describes the track in much greater detail compared to the other record. A high level of detail, ie a high number of interpolation points, can thus be a measure of high data quality.

Preferably, the assessment of the quality can also be made or influenced by the user by incorporating his own assessment of the quality in the process.

Preferably, an identification of data of both data sets takes place, which describe the same infrastructure section in order to avoid duplication of the data in the later fused data set.

Preferably, based on the evaluation of the quality of the data records, certain trusted data sets are determined which serve as a data basis in the further procedure, if other data sets provide conflicting data.

Preferably, such a trusted dataset is a dataset which has been previously read from the sources, or it has been specially created for this purpose and provided for further processing of the method.

In a further advantageous embodiment of the invention, connecting points of the at least two data records are also identified in order to fuse them, corresponding to the infrastructure.

Preferably, the at least two data sets are checked before and / or after the merger with regard to existing data gaps and / or further inconsistencies, these being identified for further processing.

The source from which the data records are read in may preferably be existing data records which have a static character, that is to say are immutable.

Preferably, dynamically generated data records can be read in from other sources, which are generated, for example, during operation of the vehicles or by the driver. These dynamically generated data records are preferably used to supplement the already existing static data records.

Preferably, when the at least two data records are merged into at least one new data record, a plausibility check of the existing data records and / or the at least one new data record takes place.

Preferably, at least one existing data record is defined as a trusted data record as a reference for the plausibility check. This trusted dataset serves as the sole data source when multiple datasets describe one and the same leg of the route, but provide different or contradictory information.

The selection of the trustworthy data source can preferably take place via a weighting or evaluation of the quality between the individual data records which serve as data sources.

Preferably, a plausibility check involves checking for data consistency. Preferably, a review of the elevation data is carried out primarily at the transition points between individual data sets, whereby discontinuities are avoided.

Preferably, the driver assistance systems located in the vehicles are designed to carry out the autonomous or semi-autonomous operation of the vehicles on the basis of the newly generated data records.

Preferably, the newly created datasets have infrastructure-related information about a road network in which the vehicles move.

In a special embodiment of the invention, these are rail vehicles which move on a rail network.

In a further advantageous embodiment of the invention, the invention is provided as a computer program product with program code stored on a machine-readable carrier for carrying out the method described above, wherein the carrier can be, for example, a USB stick or the like, or the computer program product via a server by means of a suitable data connection can be made available.

In a further advantageous embodiment of the invention, the invention is in the form of a device which processes the method described above.

Such devices are preferably devices which are designed for installation in vehicles, for example in the form of an on-board unit, or path-side devices, such as signal boxes or control stations.

When installed in a vehicle, the device preferably serves to process the data records present on the vehicle, so that no new data records have to be generated.

Preferably, the device is provided in a planning office, which undertakes long-term care of the data, which serve as the basis for the operation of the vehicles on a road network, which is formed from several infrastructure sections.

Preferably, the device has at least one suitable interface to the vehicle or the path-side device, which is adapted to determine the required data for the method described above.

As a result, the device has the option of developing the required data for carrying out the method, for example via data carriers or data connections.

The device preferably has at least one user-accessible interface so that the user can transmit additional information and / or static or dynamic data to the device via this interface.

For example, the user can transfer data from a test drive, which was undertaken to supplement existing data records, to the device via this interface. There is preferably also the possibility to transmit unforeseen events directly to the device during a journey.

Preferably, the device is designed to be installed in a track-bound vehicle, in particular a rail vehicle, and to interact with it.

The invention is not limited only to the embodiments described above. Rather, further embodiments are conceivable, which results from a combination of at least two of the embodiments described above.

In the following, the description of the invention with reference to preferred embodiments with reference to the accompanying drawings.

Fig. 1

ein Flussdiagramm einer Ausführungsform des erfindungsgemäßen Verfahrens mit den wesentlichen Schritten des Verfahrens.

Fig. 2

ein weiteres Flussdiagramm einer Ausführungsform des erfindungsgemäßen Verfahrens mit einer detaillierten Darstellung einzelner Verarbeitungsschritte bis hin zur Fertigstellung des fusionierten Datensatzes.

In detail shows:

Fig. 1

a flowchart of an embodiment of the method according to the invention with the essential steps of the method.

Fig. 2

a further flowchart of an embodiment of the method according to the invention with a detailed representation of individual processing steps to the completion of the fused data set.

Fig. 1 shows a flowchart of an embodiment of the method according to the invention with the essential steps of the method.

After the start of the data fusion, the first step is to read S1 static data. Static data is to be understood as meaning all infrastructure-related data which are essentially static, that is unchangeable. An example of static data is information about the slope of a link as it does not change further during normal operation.

In a further step, reading S2 of dynamic data is subsequently carried out. Dynamic data is preferably understood to mean data that was generated during the operation of the vehicles or through special test drives. These dynamic data represent a possibility to supplement or close existing data gaps of the static data, but they can also be used for enriching the static data with further information, such as the position of a temporary construction site.

In a third step, a plausibility check S3 of the read-in data takes place. Obviously non-plausible data, such as negative maximum speeds, or height information that does not correspond to the typical terrain course, sorted out, or at least marked as faulty for further processing. For example, the plausibility check S3 can be compared with other data classified as correct or plausible, such as, for example, map data.

In a further step, the evaluation S4 of a quality of the data records that were read in beforehand takes place.

The quality of a data record describes how much the data set is suitable for the description of an infrastructure segment and further processing.

In this case, the quality of the data is weighted such that at least one data source is defined or recognized as more trustworthy or more accurate than the other data sources and whose data parameters are adopted in case of doubt or in the case of several data sources which supply the same data parameters. a majority vote can be taken. It is also conceivable to define different trustworthy data for different route sections. Thus, a first record for a first stretch can be defined as trusted. However, this turns out to be bad or unfit for a subsequent second stretch. Instead, a second record is trusted and used as a possible data foundation.

This weighting happens on a case-by-case basis not only on a data source-specific basis, but can also be changed by data-parameter-specific exemptions. For example, in case of doubt the data of a data source should be used, if especially the data parameter "Direction Indicator" is contradictory, it should be taken over by another data source.

The data records defined as trustworthy are also referred to as nodes or anchor points. They serve, as described above, the possibility of retreating to these data records, if other data sets contain plausible or erroneous information.

The now prepared data records are merged into a further step by means of data fusion S5 to form a data record in unit format.

In the data fusion S5 itself, filters and compensation calculations (FFT analysis, high / low pass filter, averaging, moving average, extra / interpolation, etc.) are applied to data gaps or implausible locations which are not determined by the anchor points could be resolved.

In addition, the compensation calculations can also be used to correct the resolution of the data records in order to obtain a desired node resolution of the new data record.

For example, it is checked whether a height jump, ie a difference in the value of height, closely spaced route points is even possible because there is a maximum allowable slope for rail vehicles. If this maximum allowable slope is exceeded in the test, such a height jump is not recognized as plausible. Such intrinsic contradictions can to some extent be expected to be plausible.

If these measures do not lead to plausible data sets, a routing mechanism can be used to use closely adjacent routes (in this way) In one case, for example, the GPS measured route data is prioritized higher than the timetable data for a route section).

If all these measures are not used and if such a position can not be extrapolated or interpolated, it is marked and can then be manually edited by a user. Alternatively, the processing takes place via manually defined rules, which are used either specifically for single or multiple locations, or generally for an entire route network. These rules can be used to correct or supplement plausible places or data gaps in the data sources so that higher network coverage can be achieved.

Result of the data fusion S5 is a data record in the unit format, which is subjected in a further step to a further plausibility check S6 to ensure that the transition points of this data record to a next, subsequent data record, which already exists in the unit format, are consistent and plausible.

This plausibility check S6 essentially follows the same steps as those during the data fusion S5, except that now a check is made against adjacent data records.

Subsequently, when the END of the procedure is completed, the new record is made available for further processing.

This method has the advantage of generating, in large part, automatically a large number of consistent, complete data records which have at least the same or a higher degree of detail and at least the same or a higher data quality.

Fig. 2 shows a further flowchart of an embodiment of the method according to the invention with a detailed representation of individual processing steps up to the completion of the merged new data set 60th

From a plurality of data sources 10, a standardization of the format of individual data records 10a, 10b, 10c, 10d is carried out via a conversion 12. The individual data sets 10a, 10b, 10c, 10d are in various formats before conversion by individually adapted converters 12a, 12b, 12c, 12d. That they describe certain sections of the route in different ways. For example, the data set 10a describes a distance based on absolute distance kilometers, whereas data set 10b only describes distances between two stops. The converters 12a, 12b, 12c, 12d fulfill the task of converting all data records to a desired, uniform format.

The thus unified data sets 10a, 10b, 10c, 10d are subsequently fused by means of the data fusion 14 into a data record in unit format, whereby a data pool 16 in the unit format is obtained.

The following describes the further processing and verification of the merged data.

First of all, the consolidation 18 of the data pool 16 takes place, wherein redundancies, ie overlaps and doublings of data of the data pool 16 are determined and / or corrected.

The data pool 20 thus obtained is then subjected to a general plausibility check 22, which basically checks whether the data records 10a, 10b, 10c, 10d are now linked in the correct format and in a suitable form to a data pool 20. If necessary, corrections of the data pool 20 are made at this point and / or process steps are carried out again and / or messages are output to a user.

The data pool 24 thus obtained is now linked by means of an application-dependent link 26 with further application-dependent data, such as, for example, vehicle or track-specific data, or specific routes.

This ensures that further existing application-dependent data required for the operation of rail vehicles are enriched and supplemented with the information of the data pool 24, so that the ultimately new data record 60 with additional, important for the operation of information, such as maximum allowable Speeds, or existing stops, is provided.

The data pool 28 thus linked is subsequently subjected to a number of checks in order finally to obtain a finished, error-free new data record 60.

In steps 30, 34, 38, 42, track characteristics are checked to ensure that particular information has not been altered or deleted by data fusion 14 in unit format.

Thus, the determination and correction of existing switches 30, and the determination of bases 34 for further processing steps. In addition, a height check 38 and a gradient determination 42 to ensure the consistency of the existing data pool 32, 36, 40, 44 and correct if necessary.

If, for example, unrealistic height jumps between two adjacent data points are detected, then the correction of these height jumps takes place, for example, by the user or the replacement of the height values by data classified as trustworthy.

In the same way, the gradient determination 42 checks whether an existing gradient is plausible, or whether the merger has given rise to a gradient value which is too high for the operation of a rail vehicle.

Here, too, the correction of these gradient jumps takes place, for example, by the user, or the replacement of the altitude values by data classified as trustworthy, as a result of which plausible gradients should occur.

The data pool 44 in unit format thus obtained is subsequently examined in further checks with regard to special features of the infrastructure.

In steps 46, 50, 54, starting from the data pool 44 in the unified format, a check of the speed limits 46 is first made, so that it is ensured that the data pool 48 in the unit format contains the necessary and safety-critical speed restrictions.

If, for example, unrealistic or missing speed limits are detected, they are corrected or supplemented by the user, for example, or by the automatic setting of the speed limits by data classified as trustworthy.

In the same way, the check 54 of route crossings or prohibited route crossings is also proceeding.

If, for example, a crossing point of two intersecting sections in the data pool 52 is determined in unit format, it can be ascertained from the height profiles of both sections whether these two sections actually intersect, or if an underpass is present through which both sections run independently of each other ,

In addition, level crossings or bridges are detected or corrected in this step.

In addition, there is a check 50 of the stops and other signals and infrastructure-related facilities, so that is present especially in autonomous driving every hold option as information in the data pool 52 in unit format.

The correction or supplementation of this information takes place, for example, by the user, or by the automatic setting of the transitions, intersections or stops by the, as trusted classified data.

Subsequently, the generation 58 of application data from the previously created data pool 56 takes place in unit format. In this case, the general information of the data pool 56 is transferred into vehicle-specific new data records 60 in order to be able to transfer these to any vehicles.

In the Fig. 1 and Fig. 2 described method steps can be processed in other, not shown embodiments in other sequence, or even skipped if necessary.

Other embodiments can be formed by dispensing with the execution of special, previously described method steps.

LIST OF REFERENCE NUMBERS

10

data sources

10a

first data source

10b

second data source

10c

third data source

10d

fourth data source

12

conversion

12a

first converter

12b

second converter

12c

third converter

12d

fourth converter

14

data Fusion

16

data pool

18

Consolidate

20

data pool

22

plausibility check

24

data pool

26

application-dependent linkage

28

data pool

30

Determination of switches

32

data pool

34

Create vertices

36

data pool

38

Height-check

40

data pool

42

gradient determination

44

data pool

46

Checking speed limit

48

data pool

50

Checking stops

52

data pool

54

Checking transitions

56

data pool

58

Generation of application data

60

new data set

BEGIN

Start of the data fusion

S1

Reading in static data

S2

Reading dynamic data

S3

plausibility check

S4

rating

S5

data Fusion

S6

plausibility check

THE END

Completion of the data fusion

Claims (14)

A method for creating at least one new data record (60), in particular comprising infrastructure data for vehicles, wherein
the method comprises the following steps: Reading (S1, S2) at least one existing data set from at least one source (10, 10a, 10b, 10c, 10d), wherein the at least one existing record contains data describing an infrastructure section, Conversion (12, 12a, 12b, 12c, 12d) of the at least one data record to at least one new data record (60) with a uniform format, wherein the at least one new data record (60) is suitable for enabling the operation of vehicles, in particular by means of assistance from driver assistance systems. The method of claim 1, wherein
a data fusion (S5, 14) of at least two existing data records is performed to at least one new data record. A method according to claim 2, wherein
a rating (S4) of a quality of the records is performed, wherein
identify certain data by quality as trusted records for further merger. A method according to claim 2 or 3, wherein
Data gaps before and / or after the data fusion (S5, 14) of at least two existing data records or of the at least one new data record (60) are identified, and / or
be marked for further processing, and / or
subsequently eliminated by compensation calculation. Method according to at least one of the preceding claims, wherein the at least one source (10, 10a, 10b, 10c, 10d)
an existing static record, and / or
is a dynamically generated record. A method according to any one of claims 3 to 5, wherein
a plausibility check (S3) of at least two existing data records or a plausibility check (S6) of the at least one new data record (60) is carried out, wherein
the check is also performed against at least one trusted record, wherein
the at least one trusted data set originates from a source (10, 10a, 10b, 10c, 10d) of the read-in data records, and / or
specially made available for the purpose of this test. Method according to at least one of the preceding claims, wherein
the assistance of driver assistance systems includes the autonomous operation of the vehicles. Method according to at least one of the preceding claims, wherein
the infrastructure data includes information about a road network. A method according to claim 8, wherein
the vehicles are track-bound vehicles. A method according to claim 8 or 9, wherein
the vehicles are rail vehicles and
the road network is a railway network. Computer program product with program code stored on a machine-readable carrier for carrying out the method according to one of Claims 1 to 10. Device for installation in a vehicle or a path-side device, which is designed to carry out the method according to one of claims 1 to 10. Apparatus according to claim 12, wherein
the device has at least one suitable interface to the vehicle or the wayside device, which is adapted to determine the required data for the method according to claims 1 to 10. Apparatus according to claim 12 or 13, wherein
the device has at least one user-accessible interface, so that
the user can transmit additional information and / or static or dynamic data to the device via this interface.

Images