Computer-implemented method and device for determining a position and / or speed value of a rail vehicle

Abstract

The invention relates to a computer-implemented method for determining a position and / or speed value of a rail vehicle (1) moving on a track (3), in which the position and / or speed value is determined using a data fusion algorithm, in which data from at least one GNSS receiver (9) and at least one inertial measuring device (10) of the rail vehicle (1) are processed by the data fusion algorithm, and in which a distance (D) traveled by the rail vehicle (1) is used as at least one state variable of the data fusion algorithm.To improve the accuracy of the position and / or speed value, the invention provides that the distance traveled (D) is operatively linked with at least one further state variable, such as position or speed, by means of the data fusion algorithm, at least one distance measurement value is determined when passing over location markers (4), in particular balises, arranged along the driving route (3), and is processed by the data fusion algorithm in such a way that both the distance traveled (D) and the operatively linked further state variables are updated.

Description

[0001] The invention relates to a computer-implemented method for determining a position and / or speed value of a rail vehicle moving on a track, in which the position and / or speed value is determined using a data fusion algorithm, in which data from at least one GNSS receiver and at least one inertial measuring device of the rail vehicle are processed by the data fusion algorithm, and in which a distance traveled by the rail vehicle is used as at least one state variable of the data fusion algorithm.

[0002] The invention further relates to a device for determining a position and / or speed value of a rail vehicle moving on a track, comprising at least one GNSS receiver, at least one inertial measuring device, and at least one data processing device designed to determine the position and / or speed value using a data fusion algorithm, wherein data from the GNSS receiver and the inertial measuring device are processed by the data fusion algorithm and the distance traveled by the rail vehicle is used as at least one state variable of the data fusion algorithm.

[0003] Such methods or devices are described, for example, in EP 4 406 813 A1, which is included here in its entirety. Another prior art device is described in EP 4 339 069 A1.

[0004] Rail vehicles utilize systems that employ both GNSS receivers and inertial measuring devices for time, position, and / or speed measurement. The combination of a GNSS receiver and an inertial measuring device (IMU - Initial Measuring Unit) represents an optimally complementary system. An inertial navigation system calculates the relative change in position, speed, and orientation by integrating the accelerations and angular velocities measured by the inertial measuring device. The GNSS receiver provides absolute and drift-free position, speed, and / or time values ​​(PVT).

[0005] Using a data fusion algorithm, such as a Kalman filter, which processes both the GNSS receiver and inertial measurement device (IMD), has the advantage that sections with GNSS shadowing effects can be bridged by the relative measurements from the IMD. Furthermore, the systematic errors of the IMD during GNSS reception can be estimated, which in turn improves the effectiveness of the IMD during GNSS shadowing. In a so-called strapdown algorithm, for example, a navigation solution consisting of position, velocity, and orientation is propagated over time using measurements from an IMD. Here, accelerations and rotation rates measured by the IMD are integrated, and measurement errors accumulate. As a result, the errors of the navigation solution also increase over time.However, if position and / or velocity measurements are available from the GNSS receiver, these are processed in the data fusion algorithm, leading to a correction of the navigation solution propagated by the strapdown algorithm. As described above, the data fusion algorithm typically used for the simultaneous use of GNSS and inertial measurement devices is a Kalman filter, but other data fusion algorithms would also be possible.

[0006] In railway environments, due to buildings, bridges, mountains, gorges, avalanche barriers, overhead lines, tunnel portals, stations, and station roofs, etc., there are often extended sections of track with GNSS obstruction, meaning no clear line of sight to the GNSS satellites. This can severely limit the availability of measurements from the GNSS receiver. Since the values ​​of the inertial navigation system cannot be corroborated by the GNSS receiver's data in these areas, navigation errors increase over time, and the associated estimated measurement uncertainties (variances) widen. With currently known systems, these measurement uncertainties can sometimes increase so significantly after only about one minute that they are no longer usable for accurate speed and / or position determination.Although higher-quality inertial measuring devices with significantly better performance are already known, which use, for example, fiber-optic gyroscopes or ring laser gyroscopes, these are very expensive and therefore often less attractive.

[0007] Various methods for improving the accuracy of the Kalman filter are also known in the literature. These include, for example: • JL Farrell: “Carrier phase processing without integers”, Proceedings of the Institute of Navigation 57th Annual Meeting 2001, pages 423 - 428, June 11-13, 2001, Albuquerque, NM, USA. • F. van Graas and JL Farrell: “GPS / INS - a very different way”, Proceedings of the Institute of Navigation 57th Annual Meeting 2001, pages 715 - 721, June 11-13, 2001, Albuquerque, NM, USA. • SI Roumeliotis and JW Burdick, “Stochastic cloning: a generalized framework for processing relative state measurements,” Proceedings 2002 IEEE International Conference on Robotics and Automation (Cat. No.02CH37292), Washington, DC, USA, 2002, pp. 1788-1795 vol.2, doi: 10.1109 / ROBOT.2002.1014801.

[0008] However, these methods are rather complex and may result in a problematic increase in computing effort.

[0009] It is therefore the object of the present invention to improve the known methods and devices in order to provide a simple and cost-effective navigation solution with position and / or speed determination for rail vehicles.

[0010] According to the invention, the problem is solved by linking the distance traveled with other state variables, such as position or speed, using the data fusion algorithm, and by determining at least one distance measurement value when passing over location markers, in particular balises, arranged along the route, and processing it by the data fusion algorithm in such a way that both the distance traveled and the linked other state variables are updated.

[0011] Furthermore, the object of the invention is solved by the fact that the data fusion algorithm is designed in such a way that the distance traveled is linked to other state variables, such as position or speed, and the data processing device is designed to determine at least one distance measurement value when passing over location markers arranged along the route, in particular balises, and to process it by means of the data fusion algorithm in such a way that both the distance traveled and the linked other state variables are updated.

[0012] The solution according to the invention has the advantage that it allows for the simple provision of improved position and / or speed values ​​and improves the variance of these values. This navigation solution according to the invention enables, for example, the highly accurate navigation of very long tunnel passages with prolonged GNSS signal loss, as the uncertainty of position and speed determination is reduced. The solution according to the invention allows for the very simple and effortless use of the distance traveled as a measured value in the data fusion algorithm, in particular a Kalman filter.

[0013] Updating the distance traveled and the related state variables could also be called correcting, since the previous values, which have a larger error, are replaced by new, less error-prone values.

[0014] The data fusion algorithm according to the invention determines the position and / or velocity values. As is customary, the data from the GNSS receiver and the inertial measuring device are processed by the data fusion algorithm. Furthermore, the distance traveled by the rail vehicle is used as a state variable of the data fusion algorithm. Here, the distance traveled is the path taken by the rail vehicle along the track. Therefore, the distance traveled is not the shortest distance between two points in space, but rather the path between two points on the track, taking into account the trajectory of the track. A crucial aspect of the solution according to the invention is that the distance traveled is operatively linked to other state variables by the data fusion algorithm.This interrelationship establishes a connection between these state variables, so that if one of the state variables changes, the other state variables automatically and correlate with it.

[0015] The physical relationship between the state variables is represented in the data fusion algorithm according to the invention.

[0016] Furthermore, in the method according to the invention, at least one distance measurement is determined when the vehicle passes over location markers, in particular balises, arranged along the route. When the vehicle passes over a location marker, it can be detected and identified, for example, by a received identification code. For example, the distance measurement can be determined using two location markers. This distance measurement has a much lower variance than a distance value determined by the data fusion algorithm during a GNSS shadowing and is processed by the data fusion algorithm in such a way that the distance traveled and also the other related state variables are updated. "Update" here means that values ​​previously calculated by the data fusion algorithm are corrected. The data fusion algorithm serves to estimate states and can, for example, be a Kalman filter.The determined state variables may exhibit increased variance. Updating further state variables with the distance measurement can reduce this variance. Since the distance and the other state variables are interrelated, the low variance of the determined distance measurement also has a positive effect on the related state variables.

[0017] The basis of the solution according to the invention is the extension of a state vector of the data fusion algorithm to include the state of the distance traveled. The state vector comprises the various state variables. In a system model underlying the data fusion algorithm, a relationship between this distance and the velocity, which is also included in the state vector, is represented according to the invention.

[0018] The solution according to the invention allows the measured values ​​for distance traveled, which are frequently available in the railway sector, for example from linear encoders or location markers such as balises, to be used directly by the data fusion algorithm. This distance measurement can be processed directly in the data fusion algorithm because the added state now establishes a direct relationship between the distance measurement and the other elements of the state vector of the data fusion algorithm.

[0019] During the journey of the rail vehicle, position and / or speed values ​​are determined essentially continuously, even between points in time when distance measurements are available, such as when passing landmarks like balises. This is necessary because, for example, current train control systems require constantly updated position and / or speed values. A value for the distance traveled, continuously determined by the data fusion algorithm, is also updated according to the continuously re-estimated speed. This is because, according to the invention, the distance traveled is present as a state variable in the data fusion algorithm. This is due to the relationship between distance and speed introduced in the system model, through which both appear as state variables in the data fusion algorithm.The system model can be viewed as the physical model underlying the data fusion algorithm.

[0020] Due to the relationship established according to the invention between distance traveled and speed, correlations between this distance and the other state variables also form during the propagation phases. This means that, in the processing of the distance measurement value according to the invention, which was determined using the detected location marker, not only is the distance propagated based on the speed corrected, but also all other state variables, such as position, speed, and the like, are corrected due to the correlations.

[0021] All state variables of the data fusion algorithm together can also be referred to as a state vector. The correction of all state variables by the distance measurement is a significant advantage of the invention.

[0022] The distance estimated according to the invention using the data fusion algorithm has rather little practical significance immediately after processing the distance measurement and can, for example, be reset to an initialization value, which is described in more detail below. Of greater importance is that, as already described above, the other state variables are continued with the corrected values.

[0023] By using the distances between the location markers according to the invention, the states in the data processing algorithm are updated, thereby reducing the variances in the state estimates that have widened during free propagation between the location markers. Speed, position, distance, and their confidence intervals are used, for example, by current train control systems. The solution according to the invention results in smaller speed confidence intervals than in the prior art, allowing the train to operate closer to the permissible line speed and to calculated braking curves. This can advantageously reduce travel times. Smaller position and distance confidence intervals allow braking before stopping points to occur later, or the target braking at the stopping points to be performed with greater accuracy.This allows the train, for example, to stop more precisely in front of platform doors, thus facilitating boarding and alighting. In modern train control systems such as CBTC, landmarks like balises (kilometer markers) are used for absolute positioning. The average distance between landmarks in CBTC is less than approximately 500 meters. Thanks to the regular support provided by the distance measurements between the landmarks, as described in the invention, long GNSS shadows can be traversed with high accuracy.

[0024] The solution according to the invention can be further developed by advantageous embodiments, which are described below.

[0025] The distance traveled can be reset to a predetermined fixed initial value, particularly zero, upon passing a first known landmark located along the route. This has the advantage that resetting to the fixed initial value allows for simple calculation and subsequent updating of the distance traveled.

[0026] Furthermore, upon reaching at least a second known landmark located along the route, the distance traveled can first be set to the determined distance measurement and then to the initialization value, where the distance measurement represents the distance traveled between the first and second landmarks. This has the advantage that a distance value can be determined at each subsequent landmark, and the distance measurement and other related state variables can be updated. This, of course, presupposes that further information about the landmark is known, from which the distance measurement can be determined. This information about landmarks is usually available in databases to which rail vehicles have access.For example, if a first location marker in the form of a balise is passed over, the distance traveled and its associated variance are set to the initial value of zero. If the second location marker in the form of a balise is then passed over, the distance between these balises, which can be read from a database, for example, can be directly processed as a distance measurement in the data fusion algorithm. This is possible because the distance state variable added according to the invention now establishes a direct relationship between the distance measurement and a state variable or an element of the state vector. After processing the distance measurement, the distance traveled state variable and its associated variance are again set to the initial value of zero to enable processing of the distance to the next balise. This process is repeated continuously.

[0027] To use a commonly used algorithm, a Kalman filter, especially an extended Kalman filter, can be employed as the data fusion algorithm. The Kalman filter can be, for example, a Total State Filter or, alternatively, an Error State Filter.

[0028] The invention also relates to a data processing device with means for carrying out the aforementioned method according to the invention and its advantageous embodiments. Such a data processing device can, for example, be a computer.

[0029] Furthermore, a computer program product with program commands for carrying out the said inventive method and / or its embodiments is claimed, wherein the inventive method and / or its embodiments can be carried out by means of the computer program product.

[0030] Furthermore, a provisioning device for storing and / or providing the computer program product is claimed. The provisioning device is, for example, a data carrier that stores and / or provides the computer program product. Alternatively and / or additionally, the provisioning device is, for example, a network service, a computer system, a server system, in particular a distributed computer system, a cloud-based computer system, and / or a virtual computer system, which preferably stores and / or provides the computer program product in the form of a data stream.

[0031] Finally, the invention also relates to a rail vehicle which, according to the invention, comprises a device for determining a distance traveled by a rail vehicle according to one of the aforementioned embodiments.

[0032] The invention will now be explained with reference to the accompanying drawings. These show: Fig. 1 a schematic representation of an exemplary embodiment of a rail vehicle according to the invention traveling on a track; Fig. 2 a schematic representation of a positional error of the rail vehicle Fig. 1 compared to the state of the art; Fig. 3 a schematic representation of a flowchart of an exemplary embodiment of a method according to the invention for determining a position and / or speed value for the rail vehicle Fig. 1.

[0033] First, the invention is described with reference to the exemplary embodiment in Fig. 1 explained.

[0034] Fig. Figure 1 shows an exemplary embodiment of a rail vehicle 1 according to the invention, which moves in a direction of travel 2 along a track 3. The rail vehicle 1 can be, for example, a train, a locomotive, a subway, a tram, or the like. Several location markers 4 are arranged along the track 3, which in the exemplary embodiment are Fig. 1 are designed as balises. The route 3 also has a hatched section 5 in which GNSS signal is blocked upwards, for example by a tunnel 6, and therefore no connection to a GNSS satellite 7 is possible. The tunnel 6 is in Fig. Figure 1 is shown schematically only and should be considered merely an example. The obstruction to the GNSS satellite 7 can also be caused by other structures, such as houses, bridges, avalanche barriers, overhead power lines, tunnel portals, station buildings, or similar. Furthermore, only one GNSS satellite 7 is shown as an example, although of course many such GNSS satellites 7 exist for GNSS navigation.

[0035] The rail vehicle 1 in the exemplary embodiment in Fig. 1 comprises at least one device 8 according to the invention for determining a position and / or speed value. This device 8 in turn comprises at least one GNSS receiver 9, at least one inertial measuring device 10, at least one location marker detection device 11 and at least one data processing device 12.

[0036] The GNSS receiver 9 generates position and / or velocity values ​​by means of a connection to several GNSS satellites 7. These position and / or velocity values ​​are transmitted to the data processing unit 12 for further processing. Alternatively, line-of-sight measurements from the GNSS sensor 9 can also be used.

[0037] The inertial measuring device 10 is designed in a known manner as an inertial measuring device and also provides data on the spatial motion of the rail vehicle 1, such as accelerations and rotation rates or velocity and angle increments. These values ​​are also transmitted to the data processing device 12.

[0038] The location marker detection device 11 detects the location markers 4 when they are passed by the rail vehicle 1. In the exemplary embodiment in Fig. In the embodiment 11, the location marker detection device is configured as a balise detection device which, upon passage, receives a balise telegram containing an identification code of the location marker 4 from the location marker 4. The determined location marker identification code is transmitted to the data processing device 12 for further processing. Additional information for each location marker 4 is stored in a database 13 and can be retrieved using the location marker identification code. This additional information is stored in the embodiment in Fig. 1 at least one distance D traveled to the respective previous location marker 4.

[0039] Of course, further information such as precise position values ​​can also be present in database 13, but for the solution according to the invention, only the distance traveled D is necessary. If precise position values ​​were present in database 13, these should be used in the data fusion algorithm instead of the distance traveled D, as they offer greater information content; however, this is not an application of the invention described here. In the exemplary embodiment, database 13 is in Fig. Database 13 is located on rail vehicle 1. Alternatively, database 13 can also be located externally from rail vehicle 1 and accessed remotely, for example via mobile network or similar. Fig. For simplicity, the distance between location markers 1 is always referred to as distance D. However, this distance D can have different measured values ​​between the various location markers 4. Alternatively, the measured distance can also be received directly from location marker 4, as is done, for example, in a so-called linking of balises. Linking (also known as chaining) refers, for example, in the European Train Control System (ETCS), to the logical connection of one balise group with a subsequent balise group. The corresponding distance D is then transmitted to the rail vehicle 1 as it passes. In this case, no database 13 would be necessary.

[0040] The data processing unit 12 receives the various data and is designed to determine a position and / or speed value for the rail vehicle 1. The data processing unit 12 is a computer running software that includes a data fusion algorithm which uses the various data. In the exemplary embodiment in Fig. 1. The data fusion algorithm is designed as an extended Kalman filter.

[0041] The position and / or speed value determined by the data processing unit 12 is transmitted to a train control unit or ATO (Automatic Train Operation) unit of the rail vehicle 1 (not shown) and used to control the rail vehicle 1. The rail vehicle 1 may, for example, be equipped for autonomous driving according to GoA 3 or GoA 4, or for driver-assisted driving according to GoA 2. For this type of train control, a current position and / or speed value of the rail vehicle 1 with the lowest possible error is required at all times.

[0042] When the rail vehicle 1 travels on track 3 without GNSS shadowing, this is unproblematic even with known rail vehicles from the prior art. Problems arise when the rail vehicle 1 is in a section 5 with GNSS shadowing. Here, the solution according to the invention offers a significant improvement over known systems, as described in more detail below.

[0043] In Fig. Figure 2 shows the error of the position and / or speed value during a journey of the rail vehicle 1, as determined according to the present invention or according to the prior art. The three superimposed diagrams in Fig. Figure 2 shows the error of the determined position in the north (top diagram), in the east (middle diagram), and vertically in the direction of altitude (bottom diagram). Black triangles in Fig. 2 marks when a location marker 4 was crossed. The gray shading indicates the time interval during which section 5 with GNSS shadowing was traversed and no GNSS measurements were available.

[0044] As can be seen, the error increases in the region of section 5 whenever GNSS shadowing occurs and, as a result, no position and / or velocity values ​​are delivered from the GNSS receiver 9 to the data processing unit 12. Curve 14 represents three times the standard deviation of the estimation error provided by the data fusion algorithm in a prior art method, i.e., when using a data fusion algorithm but without distance measurements. Accordingly, curve 14 shows the typical behavior for purely inertial navigation, namely a continuous increase in the uncertainty of the state estimation over time. Curve 15 represents the actual navigation errors, which, like the associated uncertainties, also increase over time.

[0045] Curve 16 represents the navigation errors when using the method according to the invention, for example in the rail vehicle 1. Fig. 1. Curve 17 shows the corresponding three-times standard deviations provided by the data fusion algorithm. When processing landmark distance measurements, the position error in the easterly direction (middle diagram) is initially reduced for the second landmark 4 in the GNSS failure interval in section 5, followed by the error in the northerly direction (top diagram) for the subsequent distance measurement, as shown in curve 16. This is also reflected in the corresponding uncertainties of the estimation errors, as shown in curve 17. The altitude error (bottom diagram) is practically unaffected. This behavior is due to the fact that distance measurements only allow partial observation of position, velocity, and other states. A distance measurement only provides information in the direction of the velocity vector of the travel path 3, but not perpendicular to it.Nevertheless, it can be concluded overall that the navigation solution according to the invention, represented as curve 16, clearly benefits from the processing of the distance measurements. As in . Fig. The error in section 5 with GNSS shadowing is significantly reduced by the solution according to the invention, as is clearly visible in section 2.

[0046] The following describes the computer-implemented method according to the invention for determining a position and / or speed value, which is carried out by the data processing device 12 according to the invention on the rail vehicle 1, with reference to the flowchart in Fig. 3 explained.

[0047] In step 301, the GNSS receiver 9 determines a position and / or velocity value and transmits it to the data processing unit 12. In step 302, the inertial measurement unit 10 determines acceleration and rotation rate measurements and also transmits these to the data processing unit 12. Typically, an inertial measurement unit provides data at a data rate of 100–1000 Hz, while a GNSS receiver provides data at a data rate of 1–10 Hz. Therefore, significantly more measurements are transmitted to the data processing unit 12 in step 302 than in step 301.

[0048] In step 303, the location marker detection unit 11 determines the identification code of a traversed location marker 4. The identification code is transmitted to the data processing unit 12. In the subsequent step 304, the data processing unit 12 accesses the database 13 to determine the data stored for the respective location marker 4. This data stored in the database 13 includes at least the distance traveled to a previous location marker 4. The distance traveled here refers to the path taken by the rail vehicle 1 along the track 3. Since the rail vehicle 1 cannot move freely in three-dimensional space, but only travels one-dimensionally along the track 3 in the direction of travel 2, the exact position of the rail vehicle 1 on the track 3 can be determined from the distance traveled, provided the route of the track 3 is known.This position of the rail vehicle 1 along the track 3 is what today's train control systems or ATO devices need to control the rail vehicle 1.

[0049] In step 305, the data from the GNSS receiver 9 and the inertial measurement device 10 are processed by the data fusion algorithm, which continuously determines a position and / or velocity value. The distance traveled is also processed by the data fusion algorithm. The data fusion algorithm of the data processing device 12 comprises several state variables. In the embodiment shown in the figures, these include, in addition to the distance, the velocity vector and attitude angle such as roll, pitch, and yaw, as well as systematic errors of the inertial measurement device such as bias. In the exemplary embodiment of the invention shown in the figures, the data fusion algorithm is a Kalman filter. The data fusion algorithm processes the various state variables in a predetermined manner that represents reality as accurately as possible.For this purpose, the various state variables in the data fusion algorithm are interconnected. This means that if a measured value is available and used for one or more state variables, the data fusion algorithm corrects all state variables. The measured values ​​originate from the GNSS receiver 9 or from the recognition of the location markers 4 in the form of distance measurements. The data fusion algorithm can thus also use the distance traveled as a distance measurement, which was determined in step 304 using database 13 for the time of crossing location marker 4. The measured values ​​from the inertial measuring device 10 are used to propagate the navigation solution; this navigation solution is corrected when the measured values ​​from the GNSS receiver 9 or the distance measurements are processed.In a so-called strapdown algorithm, the navigation solution, consisting of position, velocity, and orientation, is propagated over time using measurements from the inertial measurement device 10. Accelerations and rotation rates measured by the inertial measurement device 10 are integrated, and measurement errors accumulate. As a result, the errors in the navigation solution also increase over time. However, if position and / or velocity measurements are available from the GNSS receiver 9, these are processed in the data fusion algorithm, leading to a correction of the navigation solution propagated by the strapdown algorithm.

[0050] Since the state variables in the data fusion algorithm are interconnected, using the distance measurement not only updates the distance traveled in the data fusion algorithm, but also the other state variables interconnected with the distance. This results in a significant reduction in the error of the determined position and / or speed value each time a location marker 4 is passed. This is illustrated in the diagram. Fig. 2 clearly, in which the times of crossing the landmarks 4 are also shown.

[0051] In step 306, the distance traveled in the data fusion algorithm is set to a predetermined fixed initial value at the time the location marker is passed. In the exemplary embodiment shown in the figures, this initial value is zero. Since the distance traveled between the location markers 4 in database 13 is always given as a relative value, the initial value normalizes the data, resulting in good comparability when the next location marker 4 is passed. This allows the distance measurement for the next location marker 4 to be used immediately for correction in the data fusion algorithm without requiring any further conversion.

[0052] In step 307, when the next location marker 4 is passed, the distance traveled to the previous location marker 4 is again determined using database 13 and used again to update the related state variables. This update again results in the Fig. 2. The error is reduced as shown, and then the initialization value is reset as described in step 306.

[0053] Steps 306 and 307 are then repeated continuously. As a result, the error is reduced again at each location marker 4, as shown in the middle diagram in Fig. 2 is clearly visible from curve 16. QUOTES INCLUDED IN THE DESCRIPTION

[0000] This list of documents cited by the applicant was automatically generated and is included solely for the reader's convenience. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions. Cited patent literature

[0000] EP 4 406 813 A1

[0003] EP 4 339 069 A1

[0003] Cited non-patent literature

[0000] JL Farrell: “Carrier phase processing without integers”, Proceedings of the Institute of Navigation 57th Annual Meeting 2001, pages 423 - 428, June 11-13, 2001, Albuquerque, NM, USA

[0007] F. van Graas and JL Farrell: “GPS / INS - a very different way”, Proceedings of the Institute of Navigation 57th Annual Meeting 2001, pages 715 - 721, June 11-13, 2001, Albuquerque, NM, USA

[0007] S. I. Roumeliotis and J. W. Burdick, „Stochastic cloning: a generalized framework for processing relative state measurements“, Proceedings 2002 IEEE International Conference on Robotics and Automation (Cat. No.02CH37292), Washington, DC, USA, 2002, pp. 1788-1795 vol.2, doi: 10.1109 / ROBOT.2002.1014801

[0007]

Claims

[1] Computer-implemented method for determining a position and / or speed value of a rail vehicle (1) moving on a track (3), where the position and / or velocity value is determined using a data fusion algorithm, in which data from at least one GNSS receiver (9) and at least one inertial measuring device (10) of the rail vehicle (1) are processed by the data fusion algorithm, and in which a distance traveled (D) of the rail vehicle (1) is used as at least one state variable of the data fusion algorithm, characterized by , that The distance traveled (D) is linked to other state variables, such as position or speed, using the data fusion algorithm. at least one distance measurement is determined when passing over location markers (4), in particular balises, arranged along the route (3) and is processed by the data fusion algorithm in such a way that both the distance traveled (D) and the related other state variables are updated. [2] Method according to claim 1, characterized by , that the distance traveled (D) is set to a predetermined fixed initialization value, in particular to zero, upon passing a first known location marker (4) arranged along the route (3). [3] Method according to claim 2, characterized by , that The distance traveled (D) is first set to the determined distance measurement value and then to the initialization value upon reaching at least a second known location marker (4) arranged along the route (3). where the distance measurement represents a distance between the first location marker and the second location marker. [4] Method according to any of the above claims, characterized by , that a Kalman filter, in particular an extended Kalman filter, is used as the data fusion algorithm. [5] Data processing device (12) with means for carrying out the method according to any one of claims 1 to 4. [6] Computer program product comprising program instructions for carrying out the method according to any one of claims 1 to 4. [7] Provisioning device for the computer program product according to claim 6, wherein the provisioning device stores and / or provides the computer program product. [8] Device (8) for determining a position and / or speed value of a rail vehicle (1) moving on a track, with at least one GNSS receiver (9), with at least one inertial measuring device (10), with at least one data processing device (12) designed to determine the position and / or speed value using a data fusion algorithm, wherein data from the GNSS receiver (9) and the inertial measuring device (10) are processed by the data fusion algorithm and the distance traveled by the rail vehicle (1) is used as at least one state variable of the data fusion algorithm, characterized by , that the data fusion algorithm is designed such that the distance traveled is linked to other state variables, such as position or speed, and the data processing device (12) is designed to determine at least one distance measurement value when passing over location markers (4), in particular balises, arranged along the route (3) and to process it by means of the data fusion algorithm in such a way that both the distance traveled and the linked other state variables are updated. [9] Rail vehicle (1) characterized by , that the rail vehicle (1) comprises at least one device (8) according to claim 8.

Images