Dynamic validation of a function of a sensor unit of a railway vehicle for object localization

The dynamic validation method for rail vehicle sensor units using reference objects and self-localization systems addresses the limitations of static validation, ensuring accurate and continuous obstacle detection and reducing collision risks by enabling real-time, reliable obstacle localization.

EP4467419A1Inactive Publication Date: 2024-11-27SIEMENS MOBILITY GMBH

Patent Information

Application Number
EP2024172971
Authority / Receiving Office
EP · EP
Patent Type
Applications
Current Assignee / Owner
Priority Date
2023-05-25
Filing Date
2024-04-29
Publication Date
2024-11-27
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Current on-board sensor systems for rail vehicles lack dynamic validation capabilities for obstacle localization, leading to potential inaccuracies and reduced reliability, especially during normal operation and varying weather conditions, and require time-consuming static validation methods.

Method used

A method and arrangement for dynamic validation of a sensor unit for rail vehicle object localization using a reference object with known position, measured by a sensor unit like lidar or radar, and self-localization units to determine relative positions and route distances, allowing for real-time validation and detection of obstacles along the rail route.

Benefits of technology

Enables reliable and precise validation of obstacle localization functions during dynamic scenarios, reducing the risk of collisions by providing accurate and continuous monitoring of obstacles, even in challenging conditions, and allowing for automated responses to maintain system reliability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure IMGAF001_ABST
    Figure IMGAF001_ABST
Patent Text Reader

Abstract

A method for the dynamic validation of a function of a sensor unit (7) of a rail vehicle (6) for object localization is described. In this method, a reference object (3), whose reference position (RP) is known, and which is located on or along a track (2) traversed by the rail vehicle (6), which, in addition to at least one reference object (3), has a plurality of reference measurement points (RMP) distributed along the track (2), the positions (P-RMP) of which are also known, is detected by the sensor unit (7). The relative position (RLP) of the reference object (3) is determined by measuring a distance (d) of the reference object (3) to the sensor unit (7) and an angle (α) from which the reference object (3) is detected by the sensor unit (7).Based on the relative position (RLP) and a railway track map (SLK) depicting the route of the railway track (2), a distance (ds) to the reference object (3) along the railway track is determined. An ego-pose (EP) of the rail vehicle (6) is also determined by a self-localization unit (4) mounted on the rail vehicle (6). A first reference measurement point (RMP1) from the multitude of distributed reference measurement points (RMP) is selected, which is closest to an ego position (EPS) encompassed by the ego-pose (EP), and a second reference measurement point (RMP2) from the multitude of distributed reference measurement points (RMP) is selected, which is closest to the reference object (3). Based on the first reference measurement point (RMP1) and the second reference measurement point (RMP2), a reference distance (dsR) to the reference object (3) is determined.Finally, the object localization function of the sensor unit (7) is validated by determining a deviation (AW) of the determined distance (ds) from the reference distance (dsR). An arrangement (1) for the dynamic validation of the object localization function of a sensor unit (7) of a rail vehicle (6) is also described.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] The invention relates to a method for dynamically validating a function of a sensor unit of a rail vehicle for object localization and an arrangement for dynamically validating a function of a sensor unit of a rail vehicle for object localization.

[0002] In rail transport, it occasionally happens that reference objects such as people or road vehicles, shopping trolleys thrown onto the tracks, or even boulders or fallen trees end up on the track and therefore pose a threat to the safety of rail traffic. In the case of people and road vehicles, they are themselves highly endangered due to the possibility of a collision with a moving rail vehicle. Therefore, such obstacles must be detected in good time in order to initiate braking for an approaching rail vehicle so that a collision between the rail vehicle and the obstacle in question can be avoided. Furthermore, rail vehicles that are on the same track must also be detected in order to avoid a collision between these rail vehicles.Therefore, one of the critical tasks of a rail vehicle driver is to constantly monitor the track area for possible obstacles that could collide with the rail vehicle and to decide whether or not a reaction, such as a braking maneuver, needs to be initiated.

[0003] When controlling rail vehicles without a driver, it is particularly important to have a system that can automatically locate potential obstacles correctly and decide whether they could pose a danger to the rail vehicle itself or other road users, since by definition there is no need for a driver to monitor the track.

[0004] Lidar systems and radar systems, which can detect distances and directions in which obstacles are located, are suitable for such tasks.

[0005] Automated detection systems for potential obstacles have already been experimentally investigated and developed. Infrastructure-based solutions, in which multiple sensors are deployed to monitor a specific section of a railway line, are sometimes used to monitor areas where accidents frequently occur, such as level crossings. However, it is not feasible to equip the entire rail network with infrastructure-based obstacle detection systems. Therefore, there is a need for an on-board system to fulfill the driver's task of observing and monitoring the railway line.

[0006] Most on-board systems include systems for measuring the position of detected obstacles. Knowing the exact position of a potential obstacle allows an assessment of the hazards posed by the obstacle. In particular, the position of the potential obstacle must be determined with sufficient accuracy to estimate how far away the potential obstacle is from a rail vehicle and whether the potential obstacle is within the rail vehicle's clearance gauge or not, or whether it might be within the rail vehicle's clearance gauge or not.

[0007] In order to meet safety requirements, the precision of the localization of potential obstacles by the on-board systems used must be tested and validated.

[0008] To prepare for such a validation, a test object is first positioned on a test track. Position markers are then defined on the test track. During the validation, a rail vehicle equipped with a sensor device to be validated moves to the positions marked by the defined position markers. The determination of the position markers is usually done individually by hand, making it quite time-consuming. Furthermore, the movement of the rail vehicle to the individual position markers can lead to inaccuracies in the positioning of the rail vehicle, which can negatively impact the reliability of the validation.

[0009] Most previous approaches to validating the localization of potential obstacles are limited to a static determination of the distance between a rail vehicle and the target object to be detected using conventional measurement methods. However, these methods only allow for static evaluation and validation.

[0010] It is also desirable to validate the localization of potential obstacles during normal operation of a rail vehicle, since the quality of the localization may deteriorate during the ongoing operation of a rail vehicle, for example due to weather changes.

[0011] The task is therefore to provide a validation of the function of an on-board sensor system for obstacle localization for rail vehicles, which can also be used in dynamic scenarios.

[0012] This object is achieved by a method for dynamically validating a function of a sensor unit of a rail vehicle for object localization according to patent claim 1 and an arrangement for dynamically validating a function of a sensor unit of a rail vehicle for object localization according to patent claim 11.

[0013] In the method according to the invention for dynamically validating the function of a sensor unit of a rail vehicle for object localization, a reference object is detected by the sensor unit while the rail vehicle is traveling on a rail line. Such a reference object can be arranged next to the rail line at a predetermined reference position, but it can also be arranged directly on the track area of ​​the rail line at a predetermined reference position on or at the rail line as a type of "real obstacle" or "dummy" obstacle in order to test the function for obstacle detection and localization. The sensor unit determines a relative position of the reference object by measuring a distance of the reference object from the sensor unit and an angle from which the reference object is detected by the sensor unit.A sensor unit suitable for distance measurement, preferably an active sensor unit, particularly preferably a lidar unit or a radar unit, is used as the sensor unit. If a "real obstacle" or "dummy" obstacle is detected as a reference object, which is located on the track, i.e., particularly within the clearance profile of the rail vehicle, the obstacle detection function of the rail vehicle can also be directly tested under real-world conditions.

[0014] The reference position of the reference object is known. This means that the exact geographical position of the reference object was determined in advance.

[0015] The reference object is arranged on and at a rail line traveled by the rail vehicle, which rail line has at least one reference object and a plurality of reference measuring points distributed over the rail line, the positions of which are also known. The reference positions of the reference object and the reference measuring points can, for example, be determined in advance by a measuring vehicle traveling along the rail line and by the measuring vehicle locating the reference object and the reference measuring points. In order to locate the reference object and the reference measuring points, the measuring vehicle, preferably a rail vehicle, preferably comprises a self-localization unit, which, as part of the previously carried out journey over the rail line, when passing by orWhen the position of a reference object or a reference measuring point is passed, the position of the rail vehicle and thus also the position of the relevant reference object or reference measuring point is measured.

[0016] The position of a reference measurement point can be determined either by the vehicle's own self-localization unit or by a second localization unit specifically designed to generate the ground truth system comprising the reference measurement points. In the second case, the sensors and measuring units for determining the "ground truth data" for validation are completely separate from the measuring units for object localization or obstacle detection. This allows for particularly reliable validation of object localization due to the independence of the measuring units used for validation from the measuring units used for regular obstacle detection and localization.

[0017] Based on the relative position and a rail map depicting the course of the rail line, the distance between the remaining track and the reference object along the rail line is determined. The "course" of the rail line is understood to be the trajectory formed by the rail line along which the rail vehicle moves.

[0018] Furthermore, during the current journey of the rail vehicle, an ego pose of the rail vehicle and thus also an ego pose of the rail vehicle's sensor unit, whose relative pose to the rail vehicle is known or can be easily determined using sensors, is determined by a self-localization unit mounted on the rail vehicle. In addition to the exact ego position, the ego pose also indicates the orientation of the rail vehicle. The ego position is preferably specified in geographical coordinates. The orientation of the rail vehicle can be described, for example, by an angle relative to a predetermined reference direction.

[0019] Subsequently, a first reference measurement point from the plurality of distributed reference measurement points is determined, which is closest to an ego position encompassed by the ego pose, and a second reference measurement point from the plurality of distributed reference measurement points is determined, which is closest to the reference object. The determination and identification of the reference measurement points is preferably carried out based on a data directory in which the reference positions of all reference measurement points are specified.

[0020] Based on the knowledge of the position of the first reference measuring point and based on the knowledge of the position of the second reference measuring point, a reference route distance to the reference object, more precisely a length of a reference route that the rail vehicle covers between the first reference measuring point and the second reference measuring point, is determined.

[0021] Finally, the function of the sensor unit for object localization is validated by determining the deviation of the measured distance from the reference distance. If the determined deviation exceeds a predetermined threshold, it is determined that the object localization is no longer sufficiently accurate.

[0022] If the deviation exceeds a predetermined threshold, the object localization must be classified as no longer reliable, and preferably a notification to a maintenance unit or an automated response is triggered. An automated response preferably includes switching to a redundant sensor unit for object localization or an automated recalibration of the sensor unit for object localization that may no longer be functioning accurately.

[0023] Advantageously, the track's own reference measurement points provide so-called ground truth data, which can be used for the precise validation of sensor-based object localization. A particularly advantageous feature is the ability to determine a reference track distance completely independent of the shape and topology of the rail line. In particular, reference track distances can be determined with high accuracy along the track at curves and / or hills and valleys, as the distances between the reference objects and the rail vehicle are determined by summing the distances between a "chain" of reference measurement points. For this purpose, the individual reference measurement points are arranged close to one another compared to the variability of the track trajectory, so that any curvature or gradient can either be neglected or approximately taken into account in the distance determination.

[0024] The inventive arrangement for dynamically validating the function of a sensor unit of a rail vehicle for object localization comprises a rail line with a plurality of reference measuring points distributed along the rail line, the positions of which are known. As explained in more detail later, the reference positions of the reference measuring points can be determined once in advance by a measurement by a rail vehicle, which preferably localizes itself, and stored in a data directory.

[0025] The arrangement according to the invention has at least one reference object on or in the rail line, the reference position of which is also known.

[0026] Part of the arrangement is also a rail vehicle whose object localization function is to be validated. The rail vehicle has a sensor unit configured to detect an object, in particular a reference object, wherein a relative position of the reference object is determined by measuring a distance of the reference object from the sensor unit and an angle from which the reference object is detected by the sensor unit. The rail vehicle also includes a self-localization unit for determining an ego pose of the rail vehicle.

[0027] In addition, the arrangement, preferably the rail vehicle, comprises a track distance determination unit for determining a track distance of a track still to be traveled to the reference object along the rail line based on the relative position and on a rail line map that depicts the course of the rail line. Using the rail line map, the course of the trajectory, which describes a section of rail line between the position of the rail vehicle and the position of the reference object, can be determined, and thus the path length of this section can be determined, in particular calculated.

[0028] The arrangement, preferably the rail vehicle, also has a reference unit for selecting a first reference measurement point from the plurality of reference measurement points distributed across the rail line, which is closest to an ego position encompassed by the ego pose, and a second reference measurement point from the plurality of reference measurement points distributed across the rail line, which is closest to the reference object. The first and second reference measurement points are preferably determined based on a comparison of the distances of the individual reference measurement points to the rail vehicle or the reference object, respectively, and a selection of those reference measurement points that are positioned closest to the rail vehicle or the reference object.

[0029] Part of the arrangement according to the invention, preferably part of the rail vehicle, is also a reference distance determination unit for determining a reference distance to the reference object based on the first reference measurement point and the second reference measurement point. The reference distance is understood to be the length of a path or route traveled by the rail vehicle between passing the first reference measurement point and passing the second reference measurement point.

[0030] Finally, the arrangement, preferably the rail vehicle, also comprises a validation unit for validating a function of the sensor unit for object localization by determining a deviation of the determined track distance from the reference track distance. The arrangement according to the invention shares the advantages of the method according to the invention for dynamically validating a function of a sensor unit of a rail vehicle for object localization.

[0031] A large portion of the aforementioned components of the arrangement according to the invention can be implemented entirely or partially in the form of software modules in a processor of a corresponding computing system, e.g., in a control device or computing system of a rail vehicle or an externally arranged computing system. A largely software-based implementation has the advantage that even previously used computing systems can be easily retrofitted with a software update to operate in the manner according to the invention.In this respect, the object is also achieved by a corresponding computer program product with a computer program that can be loaded directly into a computer system, with program sections for executing the steps of the method according to the invention, at least the steps executable by a computer, in particular the steps of determining a relative position of the reference object to the sensor unit, determining a distance of a route still to be traveled, selecting a first reference measuring point and a second reference measuring point, determining a reference distance, and validating a function of the sensor unit for object localization when the program is executed in the computer system. Such a computer program product can, in addition to the computer program, optionally include additional components such as documentation and / or additional components, including hardware components such as hardware keys (dongles, etc.).) to use the software.

[0032] A computer-readable medium, e.g., a memory stick, a hard disk, or another portable or permanently installed data storage device, can be used for transport to the computer system or control device and / or for storage on or in the computer system or control device, on which the program sections of the computer program that can be read and executed by a computer system are stored. For this purpose, the computer system can, for example, have one or more cooperating microprocessors or the like.

[0033] The dependent claims and the following description each contain particularly advantageous embodiments and developments of the invention. In particular, the claims of one claim category can also be developed analogously to the dependent claims of another claim category and their description sections. Furthermore, within the scope of the invention, the various features of different embodiments and claims can also be combined to form new embodiments.

[0034] In a variant of the method according to the invention for the dynamic validation of a function of a sensor unit of a rail vehicle for object localization, the reference measuring points are located in tunnels and are measured with a high-precision self-localization unit.

[0035] The position of a reference measurement point can be determined either by the vehicle's own self-localization unit, preferably a GNSS / IMU system, or by a second localization unit, which is specifically designed to generate the "ground truth data" comprising the reference measurement points and also preferably comprises a GNSS / IMU system. In the second case, the sensors and measuring units for determining the "ground truth data" for validation are completely separated from the measuring units for object localization, allowing for particularly reliable validation of the object localization.

[0036] In one variant of the method according to the invention for dynamically validating the function of a sensor unit of a rail vehicle for object localization, the positions of the reference measuring points and the reference object(s) are determined by a reference measurement during a preliminary measurement run. A measuring vehicle, preferably a rail vehicle equipped with a self-localization unit, travels along the rail line, passing the reference measuring points and the reference object(s) as it travels along the rail line, and their geographical positions are determined using the self-localization unit. Advantageously, the preliminary measurement run establishes a database in the form of a so-called "ground truth," which serves as a basis for comparison for the subsequent dynamic validation. The ground truth data enables the determination of an ideally expected result.Reference result of an object detection or object localization. By comparing measurement data with ground truth data, a validation of an object localization function can be performed.

[0037] In a variant of the method according to the invention for dynamically validating a function of a sensor unit of a rail vehicle for object localization, the self-localization unit comprises at least one of the following types of self-localization units: a navigation unit for receiving a satellite navigation signal, an inertial measurement unit, a self-localization unit for determining the ego pose of the rail vehicle based on track markings.

[0038] A navigation unit enables the determination of a position of a rail vehicle on the basis of satellite navigation signals, but also the determination of an orientation of a rail vehicle on the basis of a trajectory of the rail vehicle, which is recorded on the basis of satellite navigation signals, and possibly also on the basis of map data which represent the course of the rail route travelled by the rail vehicle.

[0039] An inertial measuring unit, also known as an IMU measuring unit, can also determine the orientation of a rail vehicle with high precision.

[0040] Preferably, the self-localization unit is designed to determine the ego pose of the rail vehicle based on track markings that include the reference measurement points. In this variant, the reference measurement points are used not only to determine the reference track distance, but also for the distance measurement to be validated. This variant can be used if the entire rail line is equipped with reference measurement points and the rail vehicle is to be used only on this rail line, so that self-localization of the rail vehicle can always be carried out based on the reference measurement points, even during regular operation.

[0041] Particularly preferably, the known reference position is assigned to the reference object by identifying the reference object and determining the known reference position based on a look-up table in which a geographical position is specified for each reference object. The reference object preferably comprises an identification function for its identification, preferably a feature that is easily usable for identification. Advantageously, the position of the reference object can be inferred based on its identity.

[0042] The reference position of the reference object is also preferably determined by determining a nearest reference measurement point to the reference object and by assigning the geographical position of the nearest reference measurement point to the reference object. In this variant, the reference object does not need to be identified directly. Rather, an "approximate" position of the reference object can be determined based on the nearest reference measurement point. By comparing the determined "approximate" position with the positions of reference measurement points stored in a look-up table, a reference object can be identified and its exact position determined.

[0043] The determination of a reference distance to the reference object preferably comprises the addition of the distances between adjacent reference measuring points, which the first reference measuring point and the second reference measuring point as well as all reference measuring points lying between the first reference measuring point and the second reference measuring point.

[0044] Advantageously, a length of a trajectory between the first and the second reference measuring point can be determined exactly as ground truth with the aid of the reference measuring points lying between the first reference measuring point and the second reference measuring point.

[0045] Preferably, the reference measurement points and their reference positions are stored in a look-up table, and the look-up table is preferably accessed when selecting the first reference measurement point and the second reference measurement point and when determining the reference track distance in order to determine the reference positions of the reference measurement points and their distances from one another. Advantageously, the look-up table provides exact position and distance values ​​that are independent of the function of the sensor unit or the self-localization unit of the rail vehicle, so that they can be used as ground truth for validating the function of said units of the rail vehicle.

[0046] In one variant of the method according to the invention for dynamically validating the function of a sensor unit of a rail vehicle for object localization, the reference object is movable and can assume different reference positions. Its current reference position is determined by a self-localization unit arranged on the reference object. Advantageously, the number of reference objects on a rail line can be reduced because the reference object can move to different positions. In an advantageous embodiment of the method according to the invention, the rail line map, which depicts the course of the rail line, includes the reference measurement points and their reference positions.

[0047] In one embodiment of the arrangement according to the invention, the rail vehicle has at least one of the following units, preferably all of the following units: the localization unit, the distance determination unit, the reference unit, the reference distance determination unit, the validation unit.

[0048] Advantageously, the validation of the rail vehicle can be carried out independently of communication with a stationary evaluation device.

[0049] Alternatively, the aforementioned units can also be arranged separately or stationary. For example, the aforementioned units can be part of a separate, preferably stationary evaluation device. The evaluation device additionally comprises a receiving unit for receiving the distance of the reference object from the sensor unit determined by the sensor unit and the ego pose of the rail vehicle determined by the self-localization unit.

[0050] In this variant, the rail vehicle comprises a transmitting unit for transmitting the distance of the reference object to the sensor unit determined by the sensor unit and the ego pose of the rail vehicle determined by the self-localization unit.

[0051] This variant has the advantage that the evaluation device can now be used flexibly to validate the function of different rail vehicles, whereby only a single validation unit is required for the aforementioned application cases. This way, resources can be saved. The evaluation device can be permanently installed on a specific test track, or it can be designed as a mobile or transportable device for validation on different test tracks. For transport, the evaluation device can be designed with a chassis and a traction unit for transport between different test tracks. In the latter variant, the application area of ​​the evaluation device is extended to different test rail lines, so that resource efficiency is further increased through the versatile application possibilities of one and the same validation unit.

[0052] The self-localization unit in or on the rail vehicle of the arrangement according to the invention acquires measurement data for position estimation and orientation estimation for the position and, if applicable, the pose of the rail vehicle. This measurement data can be acquired using technically very different types of measuring devices. The aforementioned measurement data can include the following types of measurement data: Satellite navigation data, IMU data (IMU = inertial measurement unit), speed sensor data, odometry data.

[0053] Satellite navigation can achieve high levels of accuracy for position determination or direction of movement determination within the framework of differential satellite navigation by supplementing reference stations and evaluating the phase shift of the carrier wave. An inertial measurement unit can be used for orientation measurements. Speed ​​data and odometry data can also be used for position determination.

[0054] Preferably, the rail line has a curved course. By measuring the reference line spacing based on the reference measurement points, which were measured very precisely in advance, distances between reference objects and rail vehicles can also be precisely measured over curved rail lines and used as a basis for comparison or ground truth for validation.

[0055] The invention is explained in more detail below with reference to exemplary embodiments in the accompanying figures. They show: FIG 1 a schematic representation of an arrangement for dynamic validation of a function of a sensor unit of a rail vehicle for object localization according to an embodiment of the invention, FIG 2 a flowchart illustrating a method for dynamically validating a function of a sensor unit of a rail vehicle for object localization according to an embodiment of the invention, FIG 3 a circular distance on a sphere, FIG 4 a representation of angular relationships in a circular arc, FIG 5 a schematic representation of an arrangement for dynamic validation of a function of a sensor unit of a rail vehicle for object localization according to an alternative embodiment of the invention.

[0056] In FIG 1 A schematic representation or a top view of an arrangement 1 for dynamically validating the function of a sensor unit 7 of a rail vehicle 6 for object localization is shown. The arrangement 1 comprises a curved rail section 2, next to which a reference object 3 is located.

[0057] The reference object 3 is in FIG 1 on the right side is only shown schematically. Furthermore, on the track 2 there is a rail vehicle 6, which is FIG 1 on the left side is also shown only schematically. The reference object 3 is located at a distance d from the sensor unit 7 of the rail vehicle 6.

[0058] The rail vehicle 6 has a self-localization unit 4 configured to determine an ego pose EP and, in particular, an ego position EPS of the rail vehicle 6. The self-localization unit 4 comprises a satellite navigation unit for position determination. To determine the orientation of the rail vehicle 6, the self-localization unit 4 comprises an inertial measurement unit and / or processes rail route map data on which the orientation of a rail line can be seen as a function of the position of the rail vehicle 6.

[0059] During a preparatory run (not shown), the rail vehicle 6 travels along the rail line 2 and, when passing a respective reference measuring point RMP, determines its respective position or reference position with the aid of its self-localization unit 4.

[0060] In the FIG 1 In the scenario shown, the sensor unit 7 of the rail vehicle 6 measures the distance d to the reference object 3 located in the side area of ​​the rail line 2, as well as an angle α from which the reference object 3 is detected, during a regular journey. The data mentioned provide information about the relative position RLP of the reference object 3 to the rail vehicle 6.

[0061] The rail vehicle 6 also comprises a track distance determination unit 5. The track distance determination unit 5 is configured to determine a track distance ds of a track still to be traveled to the reference object 3 on the basis of the relative position RLP of the reference object 3 to the sensor unit 7. For this purpose, a rail route map SLK (see also step 2.V in FIG 2 ), which depicts the course of rail line 2. The rail line map SLK can be obtained, for example, from a data storage device (not shown).

[0062] The rail vehicle 6 also comprises a reference unit 8, which is configured to determine a first reference measuring point RMP1 of the plurality of distributed reference measuring points RMP, which is closest to an ego position EPS encompassed by the current ego pose EP of the rail vehicle 6, and to determine a second reference measuring point RMP2 of the plurality of distributed reference measuring points RMP, which is closest to the reference object 3. The first reference measuring point RMP1 is in FIG 1 obscured by rail vehicle 6.

[0063] The rail vehicle 6 also comprises a reference distance determination unit 9 for assigning the current ego pose EP of the rail vehicle 6 to the next reference measuring point RMP1 (also referred to as the first reference measuring point RMP1) to an ego position EPS encompassed by the ego pose EP and for determining a reference track distance ds R and on the basis of the reference measuring point RMP assigned to the reference object 3.

[0064] Finally, the rail vehicle 6 also comprises a validation unit 10 for validating an object localization of the sensor unit 7 by determining a deviation AW of the determined track distance ds from the reference track distance ds R .

[0065] In FIG 2 a flowchart 200 is shown which illustrates a method for dynamically validating a function of a sensor unit 7 of a rail vehicle 6 for object localization according to an embodiment of the invention.

[0066] In step 2.1, a reference measurement is taken over at least part of a rail line 2. The reference measurement is performed by a rail vehicle 6, which travels this part of the rail line 2 once. It continuously measures its own position using a self-localization unit 4 and, with knowledge of this ego position EPS, establishes a reference measuring point RMP at regular intervals, for example, every 10 cm. The reference measuring point RMP is then determined in each case. Furthermore, all reference objects 3 arranged along the rail line 2 are also measured, i.e., their respective geographical positions GPO are determined.

[0067] In step 2.II, a regular journey of a rail vehicle 6 is started on the measured rail line 2. In the FIG 2 In the example illustrated, the rail vehicle 6 is the same rail vehicle 6 that was also used to determine the ground truth data in step 2.1. It should be expressly mentioned at this point that the invention is not limited to this specific embodiment, and the surveying of the rail line 2 and the regular travel can be carried out by different rail vehicles.

[0068] In step 2.III, the rail vehicle 6 detects and localizes a reference object 3 arranged next to the rail line 2 with the aid of a sensor unit 7 arranged on the rail vehicle 6, which is configured for distance measurement. By self-localizing the rail vehicle 6 with a determination of a current ego pose EP of the rail vehicle 6 and by estimating the position of the reference object 3 based on the determined current ego position EPS of the rail vehicle 6, the geographical position GPO of the reference object 3 can be at least roughly estimated. Based on this rough estimate and on a look-up table in which all exact geographical positions GPO of the reference objects 3 are specified, the detected reference object 3 is clearly identified and its exact geographical position data GPO or its reference position RP is determined.

[0069] In step 2.IV, a relative position RLP of the reference object 3 to the sensor unit 7 is also determined by measuring a distance d of the reference object 3 to the sensor unit 7 and on the basis of the current ego pose EP of the rail vehicle 6 determined in step 2.III.

[0070] In step 2.V, a distance ds of a route still to be traveled to the reference object 3 is determined on the basis of the relative position RLP of the reference object 3 determined in step 2.IV and on the basis of a rail route map SLK, which shows the course of the rail route 2.

[0071] In step 2.VI, the known current ego position EPS of the rail vehicle 6 is assigned to the next reference measuring point RMP1, which can also be found in a look-up table.

[0072] In step 2.VII, a reference distance ds R . to the reference object 3 is determined based on the reference measuring point RMP1 assigned to the ego position EPS and on the basis of the reference measuring point RMP2 assigned to the reference object 3. For this purpose, the distances between the reference measuring points RMP, which lie between the first reference measuring point RMP1 assigned to the ego position EPS and the second reference measuring point RMP2 assigned to the reference object 3, as well as the distance of the first reference measuring point RMP1 to a neighboring reference point RMP and the distance of the second reference measuring point RMP2 to a neighboring reference measuring point RMP are simply summed.

[0073] Finally, in step 2.VIII, the object localization of the sensor unit 7 is validated by determining a deviation AW between the distance ds determined in step 2.V based on the current distance measurement by the sensor unit 7 and the reference distance ds R determined in step 2.VI. If the deviation AW exceeds a predetermined threshold, the object localization must be classified as no longer reliable, and a notification to a maintenance unit or an automated response must be triggered. For example, an automated response may include switching to a redundant sensor unit or an automated recalibration of the sensor unit 7 that may no longer be functioning accurately.

[0074] In FIG 3 A circular distance on a sphere between a point P and P 0 is shown. Using the so-called Haversine formula, the distance between two given points on a sphere can be calculated. Thus, the distance between two points whose geocoordinates are known can be easily calculated. In particular, exact distances between reference measurement points (RMP) can be determined in this way.

[0075] In FIG 4 A representation of angular relationships in a unit circle is shown. The points A, B and the center Q' form an isosceles triangle whose legs have the length 1. One of the legs, the altitude h of the isosceles triangle, and half the base of the isosceles triangle, form a right-angled triangle with the leg with the unit length 1 as the hypotenuse and the angle θ between the leg of the isosceles triangle and the altitude h. The base of the isosceles triangle is divided into two equal halves at the foot point C. If the altitude h is extended in the x-direction, the extended half-line intersects the unit circle at the point D. The function value sinθ is formed by half the base of the isosceles triangle. The function value cosθ is formed by the altitude h of the isosceles triangle and the function value havθ is formed by the distance between the points C and D. Now, the FIG 4 If the arc distance d 0 shown is calculated from the coordinates (lat 1 , lon 1 ), (lat 2 , lon 2 ) of the two points P, P 0 , where "lat" stands for latitude and "Ion" for longitude and "1", "2" symbolise the two different points P, P 0 , the distance between the points P and P 0 , ie the distance d 0 , is: d 0 = r archav hav θ , where hav θ = hav lat 2 − lat 1 + cos lat 1 ⋅ cos lat 2 ⋅ hav lon 2 − lon 1 . and equation (2) inserted into equation (1) finally gives the distance from d 0 , where r represents the radius of the Earth.

[0076] In FIG 5 is a schematic representation of an arrangement 1 for validating a function of a sensor unit 7 of a rail vehicle 6 for object localization according to a second embodiment of the invention.

[0077] In the second embodiment, the arrangement 1 has a stationary evaluation device 12 separate from the rail vehicle 6 and the reference object 3. In this case, the rail vehicle 6 comprises only the sensor unit 7 for detecting and localizing a reference object 3, the self-localization unit 4 for determining an ego pose EP of the rail vehicle 6, and a transmission unit 5a for transmitting the determined measurement data RLP, EP. FIG 5 a "real" obstacle 3' is also shown in the area of ​​the rail line 2, which is also to be detected or localized by the sensor unit 7 of the rail vehicle 6.

[0078] In this case, the stationary evaluation device 12 comprises a receiving unit 5b for receiving the ego pose EP of the rail vehicle 6 and for receiving the relative position RLP of the reference object 3 or the real obstacle 3' to the rail vehicle 6. Also part of the evaluation device 12 is a route distance determination unit 5, which is configured to determine a route distance ds of a route still to be traveled to the reference object 3 or to the real obstacle 3', also referred to as a dummy obstacle, based on the relative position RLP of the reference object 3 or the real obstacle 3' to the rail vehicle 6. For this purpose, a rail route map SLK, which depicts the course of the rail route 2, is used. The rail route map SLK can be obtained, for example, from a data storage device (not shown).

[0079] The evaluation device 12 also comprises a reference unit 8, which is configured to determine a first reference measuring point RMP1 of the plurality of distributed reference measuring points RMP, which is closest to an ego position EPS encompassed by the current ego pose EP of the rail vehicle 6, and to determine a second reference measuring point RMP2 of the plurality of distributed reference measuring points RMP, which is closest to the reference object 3 or the real obstacle 3'. The first reference measuring point RMP1 is determined based on the current ego pose EP of the rail vehicle 6 and the position data of the reference measuring points RMP, which are stored in a database, which is, for example, part of the reference unit 8. The second reference measuring point RMP2 is determined by comparing the known reference position RP of the reference object 3 or the real obstacle 3' with the position data of the reference measuring points RMP from the database.

[0080] The evaluation device 12 also has a reference distance determination unit 9 for determining a reference distance ds R on the basis of the first reference measuring point RMP1 and the second reference measuring point RMP2 and on the basis of the reference measuring points RMP lying between the first and the second reference measuring point RMP1, RMP2.

[0081] Finally, the rail vehicle 6 also comprises a validation unit 10 for validating an object localization of the sensor unit 7 by determining a deviation AW of the determined track distance ds from the reference track distance ds R .

[0082] Finally, it is emphasized once again that the methods and devices described above are merely preferred embodiments of the invention and that the invention may be varied by those skilled in the art without departing from the scope of the invention, as defined by the claims. For the sake of completeness, it is also emphasized that the use of the indefinite articles "a" or "an" does not exclude the possibility that the respective features may be present in multiple instances. Likewise, the term "unit" does not exclude the possibility that it may consist of multiple components, which may also be spatially distributed.

[0083] Regardless of the grammatical gender of a particular term, persons with male, female or other gender identities are included.

Claims

1. A method for dynamically validating a function of a sensor unit (7) of a rail vehicle (6) for object localization, comprising the steps of: - detecting a reference object (3), the reference position (RP) of which is known and which is arranged on or along a rail line (2) traveled by the rail vehicle (6), which, in addition to at least one reference object (3), has a plurality of reference measuring points (RMP) distributed over the rail line (2), the positions (P-RMP) of which are also known, by the sensor unit (7), wherein a relative position (RLP) of the reference object (3) is determined by measuring a distance (d) of the reference object (3) to the sensor unit (7) and an angle (α) from which the reference object (3) is detected by the sensor unit (7),- determining a distance (ds) of a route still to be traveled up to the reference object (3) along the rail line (2) on the basis of the relative position (RLP) and on the basis of a rail line map (SLK) which shows the course of the rail line (2), - determining an ego pose (EP) of the rail vehicle (6) by a self-localization unit (4) arranged on the rail vehicle (6), - selecting a first reference measuring point (RMP1) from the plurality of distributed reference measuring points (RMP) which is closest to an ego position (EPS) of the rail vehicle included in the ego pose (EP), and a second reference measuring point (RMP2) from the plurality of distributed reference measuring points (RMP) which is closest to the reference object (3), - determining a reference distance (ds, R) a distance still to be traveled to the reference object (3) on the basis of the first reference measuring point (RMP1) and on the basis of the second reference measuring point (RMP2), - validating a function of the sensor unit (7) for object localization by determining a deviation (AW) of the determined distance (ds) from the reference distance (ds R ).

2. Method according to claim 1, wherein the positions (P-RMP, RP) of the reference measuring points (RMP1) and of the reference object (3) or of the reference objects (3) are determined by a reference measurement during a previous measuring run, wherein the rail line (2) is traveled by a rail vehicle (6) which has a self-localization unit (4) and when traveling along the rail line (2) the reference measuring points (RMP) and the reference object (3) or the reference objects (3) are passed and their geographical positions (P-RMP, GPO) are determined using the self-localization unit (4).

3. Method according to one of the preceding claims, wherein the self-localization unit (4) comprises at least one of the following types of self-localization units: - a navigation unit for receiving a satellite navigation signal, - an inertial measuring unit, - a self-localization unit for determining the ego pose (EP) of the rail vehicle (6) on the basis of route markings.

4. The method according to claim 3, wherein the self-localization unit (4) is designed to determine the ego pose (EP) of the rail vehicle (6) on the basis of route markings which comprise the reference measuring points (RMP).

5. Method according to one of the preceding claims, wherein the known reference position (RP) is assigned to the reference object (3) by identifying the reference object (3) and determining the known reference position (RP) on the basis of a look-up table in which a geographical position (P-RMP) is specified for each reference object (3).

6. Method according to one of the preceding claims, wherein the reference position (RP) of the reference object (3) is determined by determining a nearest reference measuring point (RMP) to the reference object (3) and by assigning the geographical position (P-RMP) of the nearest reference measuring point (RMP) to the reference object (3).

7. Method according to one of the preceding claims, wherein the determination of a reference distance (ds R) to the reference object (3) comprises the addition of the distances between adjacent reference measuring points (RMP), which comprise - the first reference measuring point (RMP1) and - the second reference measuring point (RMP2) and - all reference measuring points (RMP) which lie between the first reference measuring point (RMP1) and the second reference measuring point (RMP2).

8. Method according to one of the preceding claims, wherein the reference measuring points (RMP) and their reference positions (RP) are stored in a look-up table and in the selection of the first reference measuring point (RMP1) and the second reference measuring point (RMP2) and the determination of the reference distance (ds R ) the look-up table is accessed to determine the reference positions (RP) of the reference measurement points (RMP1, RMP2, RMP).

9. Method according to one of the preceding claims, wherein the reference object (3) is movable and can assume different reference positions (RP) and determines its current reference position (RP) by a self-localization unit arranged on the reference object (3).

10. Method according to one of the preceding claims, wherein the rail route map (SLK), which shows the course of the rail route (2), shows the reference measuring points (RMP) and their reference positions (P-RMP).

11. An arrangement (1) for dynamically validating a function of a sensor unit (7) of a rail vehicle (6) for object localization, comprising: - a rail line (2) with a plurality of reference measuring points (RMP) distributed over the rail line (2), the positions (P-RMP) of which are known, - at least one reference object (3) on or at the rail line (2), the reference position (RP) of which is also known, - a rail vehicle (6) comprising: - a sensor unit (7) configured to detect a reference object (3) by measuring a distance (d) of the reference object (3) to the sensor unit (7) and to determine a relative position (RLP) of the reference object (3) to the sensor unit (7) based on the measured distance (d) of the reference object (3) to the sensor unit (7) and based on an angle (α) from which the reference object (3) is detected by the sensor unit (7),- a self-localization unit (4) for determining an ego pose (EP) of the rail vehicle (6), - a route distance determination unit (5) for determining a route distance (ds) of a route still to be traveled to the reference object (3) along the rail line (2) on the basis of the relative position (RLP) and on the basis of a rail line map (SLK) which reproduces the course of the rail line (2), - a reference unit (8) for selecting a first reference measuring point (RMP1) from the plurality of distributed reference measuring points (RMP) which is closest to an ego position (EPS) encompassed by the ego pose (EP), and a second reference measuring point (RMP2) from the plurality of distributed reference measuring points (RMP) which is closest to the reference object (3), - a reference distance determination unit (9) for determining a reference route distance (ds, R) a distance still to be traveled to the reference object (3) on the basis of the first reference measuring point (RMP1) and on the basis of the second reference measuring point (RMP2), - a validation unit (10) for validating the function of the sensor unit (7) for object localization by determining a deviation (AW) of the determined distance (ds) from the reference distance (ds R ).

12. Arrangement according to claim 11, wherein the rail section (2) comprises a curvilinear course.

13. Arrangement according to claim 11 or 12, wherein the rail vehicle (6) has at least one of the following units: - the track distance determination unit (5), - the reference unit (8), - the reference distance determination unit (9), - the validation unit (10).

14. A computer program product comprising instructions which, when executed by a computer, cause the computer to perform the steps of determining a distance (ds), selecting a first reference measuring point (RMP1) and a second reference measuring point (RMP2), determining a reference distance (ds R ) and validating a function of the sensor unit (7) for object localization of the method according to one of claims 1 to 10.

15. A computer-readable storage medium comprising instructions which, when executed by a computer, cause the computer to perform the steps of determining a distance (ds), selecting a first reference measuring point (RMP1) and a second reference measuring point (RMP2), determining a reference distance (ds R ) and validating a function of the sensor unit (7) for object localization of the method according to one of claims 1 to 10.

Citation Information

Patent Citations

  • Method and arrangement for calibrating at least one sensor of a rail vehicle

    DE102016225595A1

  • Method for verifying the calibration of an environmental sensor mounted on a rail vehicle

    DE102020212374A1

  • Obstacle detection for a rail vehicle

    DE102021203014A1

  • Method for the computer-assisted monitoring of the operation of a vehicle travelling along a pre-determined section, especially a railborne rail vehicle

    WO2007096273A1

Cited By

  • Method for determining a relative position indication in a track

    US20250035465A1