Validation of an object recognition device of a rail vehicle

Abstract

The invention relates to a validation system (10) for checking the function of an object recognition device (ORU) installed on a rail vehicle (RV). The system comprises a reference object (20) which is positioned at an object position (P20) in the region of a railway track (RT) and has at least one target object dimension (M-ref) and a target object type (T-ref) which are stored, together with the object position (P20), in a database (30). The system also comprises a distance determination system (40) for determining a current target object distance (D-ref) between the rail vehicle (RV) and the reference object (20), said distance determination system having a vehicle position determination module (41) with a receiving unit (42) for navigation satellite signals for determining the current vehicle position (PRV) and a computing unit (43) which communicates with the database (30) and is intended to calculate the current target object distance (D-ref) from the current vehicle position (PRV) and the object position (P20). A test unit (50) is designed to calculate deviations (ΔD, ΔM, ΔT) between the determined actual object distance (D) and the calculated target object distance (D-ref) and between the determined actual object data (M, T) and the target object data (M-ref, T-ref). The test unit is also designed to determine whether the deviations (ΔD, ΔM, ΔT) are within predetermined validity ranges (VAL-D, VAL-M,VAL-T).

Description

[0001] Description

[0002] Validation of an object recognition device of a rail vehicle

[0003] The invention relates to a validation system and a validation method for functional testing of an object recognition device installed on a rail vehicle.

[0004] For the partially or fully automated operation of a rail vehicle, such as a tram, but also a regional or high-speed train, object detection systems are increasingly being used. These systems have a sensor array arranged at the front of the vehicle. Such a sensor array has a detection area—also called the field of view or image area—that is oriented forward in the direction of travel of the rail vehicle. The sensor array comprises one or more environmental sensors for recording sensor data representing objects located within the detection area.Objects in the environment can include tracks used by the rail vehicle and tracks adjacent to it, buildings and vegetation in the area of ​​the tracks, operating equipment of an overhead line system or signaling systems, and in particular stationary or moving objects that could pose an obstacle with a risk of collision for the rail vehicle. Potential obstacles that should be detected include, in particular, other rail vehicles, trucks and cars, motorcyclists and cyclists, pedestrians, but also track ends such as buffer stops. Monocular video cameras, stereo cameras, time-of-flight cameras, lidar sensors, laser scanners, radar sensors, ultrasonic sensors, thermal imaging cameras, and the like, the sensor data of which represent an object in the detection range.The recorded sensor data are fed to an evaluation unit of the object recognition device and processed by it, wherein in particular sensor data from different environmental sensors are fused in order to determine current actual values ​​of an object. The actual values ​​here should be the object position determined relative to the sensor arrangement of the object recognition device, which is given by a longitudinal object distance between the object and the current vehicle position of the rail vehicle and by a lateral object distance between the object and the track being traveled on. Furthermore, the actual values ​​of the object include further object data characterizing it, in particular at least one object dimension and one object type. As at least one object dimension, an object height, an object width, an object diagonal, an object area, an object outline or several of the above-mentioned object dimensions can be determined.Object types can be a rail vehicle, a motor vehicle, a motorcycle or bicycle, a pedestrian, or a buffer stop. Depending on the detected object type and several consecutively detected object positions, the object detection device can determine whether an object represents an obstacle with a risk of collision for the rail vehicle in order to, if necessary, trigger an alarm to the driver and / or road users in the vehicle's surroundings or to automatically trigger a braking intervention or emergency braking by the vehicle control system of the rail vehicle.

[0005] The reliability of the actual values ​​determined by the object recognition device is therefore of crucial importance for the safety of partially or fully automated operation of a rail vehicle.

[0006] The invention is therefore based on the object of proposing measures for checking the function of an object recognition device installed on a rail vehicle.

[0007] The object is achieved, on the one hand, by a validation system according to the invention for functional testing of an object recognition device installed on a rail vehicle, which has a sensor arrangement having a detection range for recording sensor data representing an object located in the detection range and an evaluation unit processing the recorded sensor data to determine a current actual object distance between the rail vehicle and the object as well as actual object data characterizing the object, wherein an actual object type and at least one actual object dimension are determined as actual object data. In addition to the longitudinal actual object distance between the rail vehicle and the object, a lateral actual object distance between the object and the track center of the track traveled by the rail vehicle can also be determined.The actual object dimensions that can be determined are the actual object height, the actual object width, the actual object diagonal, the actual object surface area, and / or the actual object boundary, i.e., the shape of the outer boundary of an object cross-section. The actual object types that can be determined and differentiated from one another are other rail vehicles traveling on the track being used or an adjacent track, a truck, a passenger car, a motorcyclist, a cyclist, a pedestrian, and a buffer stop on the track being used by the rail vehicle.

[0008] The validation system comprises at least one reference object, which is positioned at an object position in the area of ​​a railway line such that it lies within the detection range of the sensor arrangement of the rail vehicle approaching the reference object on the railway line. The reference object has a target object type and at least one target object dimension as target object data, which are also referred to as target values. Along a railway line, which can be designed, for example, as a branch line or circular route for a regional train, several validation points can be set up permanently or temporarily, which the rail vehicle passes one after the other on its journey. At one validation point, several reference objects can also be positioned, which are positioned at different object positions and can have different target object types and / or different target object data.The object positions, target object types and target object data of a reference object are known with a predefined accuracy. For example, the reference objects can be measured at their object positions using high-precision position measuring technology. A reference object can have the target object dimensions characteristic of a particular target object type. A reference object can be designed as a dummy representing a person, a wild animal, a car, a stroller, a lamppost, an overhead line mast, and the like. When the rail vehicle approaches a reference object or a group of reference objects, their target values ​​serve as a comparison basis for the corresponding actual values ​​determined by the object recognition device.During operation, the rail vehicle repeatedly approaches reference objects after certain sections of the route in order to be able to carry out or repeat a comparison between the actual values ​​currently determined by the object recognition device and known target values.

[0009] The validation system also includes a database in which the object position and the target object data of the positioned reference object are stored. If there are multiple reference objects, these target values ​​are stored in the database for each reference object. Depending on the design of the validation system, the database can be located on the rail vehicle or on the reference object(s), or it can be provided centrally by a rail vehicle operator or via the cloud.

[0010] The validation system also includes a distance determination system for determining a current target object distance between the rail vehicle and the reference object. For this purpose, the distance determination system has a vehicle-side positioning module with a receiving unit for navigation satellite signals to determine the current position of the rail vehicle. In addition, the distance determination system has a computing unit that communicates with the database and calculates the current target object distance from the current vehicle position and the object position. To determine the current vehicle position, signals from so-called global satellite systems for positioning and navigation (Global Navigation Satellite System, GNSS for short), such as NAVSTAR GPS, GLANOSS, Galileo, or Beidou, are received by the receiving unit.The computing unit of the distance determination module continuously calculates at a high frequency, for example 10 Hz or 100 Hz, the target object distance between the object position of the reference object detected by the object recognition device and the currently determined vehicle position of the rail vehicle.

[0011] The validation system further comprises a test unit configured to calculate deviations between the determined actual object distances and the calculated target object distances, as well as between the determined actual object data and the target object data. The test unit is further configured to determine whether the deviations lie within predetermined validity ranges. The validity range is understood here as a tolerance range for actual object values ​​around a target object value; actual object values ​​determined by the object recognition device that lie within the tolerance range are then qualified as valid.If the actual object values ​​determined for a reference object are within the tolerance ranges, the functional test of the object recognition device has been positive and it can be assumed that the object recognition device recognizes the target object type and the target object distance of the reference object correctly and accurately enough.

[0012] The validation system can be used during ongoing rail operations at any time of day or night and under appropriate environmental conditions to monitor the proper functioning of a rail vehicle's object detection system. However, it can also be used to comprehensively test the functionality of an object detection system for rail vehicles that is still under development. Should the deviations between actual and target object values, for any reason, lie outside the assigned validity ranges, immediate action can be taken. The validation system according to the invention thus significantly increases the operational reliability of a partially or fully automated rail vehicle with an object detection system for obstacle detection.

[0013] In an advantageous embodiment, the validation system according to the invention comprises a digital map system for a rail network having the railway line, in which map the object position of the positioned reference object is recorded. The distance determination system is designed to determine when a predetermined triggering distance of the rail vehicle has been reached before the object position. Reaching the triggering distance then triggers that, as the rail vehicle approaches the reference object, sensor data is continuously recorded by the sensor arrangement, current actual object distances and actual object data are determined by the evaluation unit, current target object distances are determined by the distance determination system, and deviations are calculated and in each case it is determined whether these lie within the validity ranges.The digital map system can be stored directly on the rail vehicle; at least, it can be accessed from the rail vehicle if it is stored elsewhere. The validation system's test unit does not need to be operated continuously, but only upon approaching a reference object. The sensor array and the evaluation unit of the object recognition device operate continuously during normal operation; whereas, during an evaluation phase of an object recognition device under development, this can also only be started with the validation system upon approaching a reference object.In a further advantageous embodiment, the validation system according to the invention further comprises a communication system with a vehicle-side communication module and a reference object-side communication module, which are configured to exchange data wirelessly between the rail vehicle and the reference object, and optionally additionally with an operations center and a data storage device in a computer cloud. For wireless communication between the rail vehicle and the reference object, common radio systems in the 433 MHz or 868 MHz frequency range can be used. For communication with an operations center of the rail vehicle and with the data storage device in the computer cloud, wireless local area networks (WLAN) or mobile radio systems according to the LTE or 5G standard can be used.The communication modules can also be configured as terminal devices of a satellite network for internet access, such as the Starlink satellite network. The communication system allows data generated by the components of the validation system to be transferred to other components of the validation system, regardless of whether they were generated or stored on the rail vehicle, on the reference object, in an operations center, or in the cloud.

[0014] In a further advantageous embodiment of the validation system according to the invention, the reference object has a position determination module on the reference object side with a receiving unit for navigation satellite signals, which is designed to determine the object position of the reference object after it has been positioned in the area of ​​the railway line. Instead of manually measuring the object position when positioning a reference object, the same position determination technology as in the rail vehicle can be used for the reference object, so that the object position is automatically determined from the received GNSS signals after a reference object has been positioned. The determined object position can be entered into the database with the other target values ​​of the reference object.It is transmitted via the communication system to the rail vehicle, where the computing unit of the distance determination system calculates an exact distance between the rail vehicle and the object, the so-called current target object distance, from the currently determined object and vehicle positions at high frequency.

[0015] In a further advantageous embodiment of the validation system according to the invention, the vehicle-side distance determination system additionally has an inertial measuring unit for measuring a vehicle position value, wherein the current vehicle position is determined from the received navigation satellite signals and the measured vehicle position value. A known inertial measuring unit combines acceleration and yaw rate sensors in order to obtain - with reference to a reference point - among other things, its spatial position by twice integrating the determined acceleration values. This determined spatial position represents the measured vehicle position value. By additionally taking this vehicle position value into account, the accuracy is increased by a factor of 5 to 10 compared to determining the vehicle position based only on GNSS signals. This in turn increases the accuracy of the calculated current target object distances.An inertial measuring unit can also be provided in the reference object, which also improves the accuracy of the determined object position.

[0016] In a further advantageous embodiment, the validation system according to the invention further comprises a time synchronization system with a vehicle-side synchronization module and with a reference object-side synchronization module, which are designed so that the recording of the sensor data, the determination of the actual object distances and the actual object data, the determination of the target object distances, and the calculation of the deviations are carried out in a time-synchronized manner. Furthermore, it has a data memory for recording the time-synchronized sensor data, actual object distances, actual object data, target object distances, and deviations.The synchronization modules have their own internal clocks, which are synchronized using known methods (for example the Precision Time Protocol according to IEEE 1588 or the GPS time signal). This ensures that the synchronization modules and the components connected to the synchronization modules in the rail vehicle or reference object are coordinated with one another and carry out the right actions at the right time. This includes sensor data recording, actual object distance and actual object data determination, target object distance determination, deviation calculation and data recording. This synchronicity makes it possible to relate data that belong to one another in time. The data storage can be on the rail vehicle and / or the reference object and / or in the cloud.The data storage may include a management system that sends a notification when the amount of data stored in the data storage exceeds a threshold, so that measures can be taken to ensure that sufficient storage space is always available in the data storage.

[0017] In a further advantageous embodiment, the validation system according to the invention further comprises a weather system with an environmental sensor for recording environmental data of the reference object, which can influence the recording of sensor data by the sensor arrangement, wherein the environmental data is recorded in a time-synchronized manner and recorded in the data memory in a time-synchronized manner. Depending on the environmental sensors used, the recording of sensor data by the sensor arrangement for object recognition can be influenced by the weather prevailing in the area of ​​the reference object, i.e. by the air temperature and / or the air humidity and / or the air pressure and / or the precipitation and / or the wind speed and / or the solar radiation and the like.To record this environmental data, one or more environmental sensors are arranged in the area of ​​the reference object. These sensors can be designed as thermometers and / or hygrometers and / or barometers and / or ombrometers and / or anemometers and / or pyranometers. The environmental sensors can be combined to form a weather station arranged on the reference object. However, environmental sensors from a weather service provider can also be used, from which only the environmental data relating to the environment around the object position is received. The environmental data can also be recorded in a time-synchronized manner, so that after data has been transmitted, for example to the rail vehicle, it can be stored in the storage unit in time-synchronized fashion with the other data.If deviations between actual and target values ​​lie outside the validity ranges, this may be due to the object recognition device's evaluation unit not having complete sensor data available for processing, or to faulty sensor data due to weather-related influences. Instead of a negative result for the functional test, the test unit may, for example, reject the test result as invalid if adverse weather conditions are present.

[0018] In a further advantageous embodiment, the validation system according to the invention further comprises a reporting unit connected to the testing unit, which is designed to send an electronic message to an output unit in a driver's cab of the rail vehicle and / or in an operations center of the rail vehicle if the check has shown that deviations lie outside the respective validity ranges.If one or more deviations between target and actual values ​​lie outside the relevant validity ranges, a message can be sent by email or another messaging service to a rail vehicle driver's cab or to a rail vehicle operations center. There, the message can be presented to the train driver in the case of semi-automated vehicle operation and / or to a control center employee in the case of fully automated vehicle operation via an output unit, e.g. a display or monitor. The message can contain the values ​​of the deviations in relation to the validity ranges, but also the magnitude of the deviations and recommendations for action as to whether, for example, imminent maintenance or immediate decommissioning of the object detection device is recommended.

[0019] In a further advantageous embodiment of the validation system according to the invention, the reference object has a rechargeable energy source for the autonomous supply of its electrical consumers with electrical energy. This means that reference objects are free from power grid connections in their energy supply and can be positioned flexibly along the railway line. A management system for an energy source designed as a rechargeable battery detects when the charge level falls below a threshold, whereupon a corresponding message can be sent to recharge the rechargeable battery. Alternatively, the reference objects can also be supplied with electrical energy via power supplies.

[0020] The object is also achieved by a validation method according to the invention for testing the functionality of an object recognition device installed on a rail vehicle. Accordingly, a reference object which has at least one target object dimension and one target object type as target object data is positioned at an object position in the area of ​​a railway track to be traveled by the rail vehicle. Sensor data characterizing the reference object are recorded by a sensor arrangement of the object recognition device. From the recorded sensor data, an evaluation unit of the object recognition device determines a current actual object distance between the rail vehicle and the reference object, as well as at least one actual object dimension and one actual object type as actual object data.A current target object distance between the rail vehicle and the reference object is determined by determining the current vehicle position of the rail vehicle using received navigation satellite signals and calculating the current target object distance from the current vehicle position and the object position. Furthermore, deviations between the determined actual object distance and the calculated target object distance as well as between the determined actual object data and the target object data are calculated. Finally, it is determined whether the deviations lie within predetermined validity ranges. The validation system according to the invention described at the outset is particularly suitable for carrying out this method.

[0021] Further features and advantages of the invention will become apparent from the following description of a specific embodiment with reference to the drawing, in the sole

[0022] FIG Components and data streams of the validation system according to the invention are illustrated schematically.

[0023] According to the figure, a partially depicted rail vehicle RV with a driver's cab DRC arranged at the front is traveling to the right on a rail track RT. The rail vehicle RV can be, for example, a regional train, but also a tram or a high-speed train. For partially or fully automated operation, an object detection device ORU is installed on the rail vehicle RV in order to detect potential obstacles on the rail track RT ahead. The object detection device ORU has for this purpose a sensor arrangement SEN arranged at the front, which is made up of one or more environmental sensors not shown in detail.Depending on the detection task, one or more lidar, camera, radar and / or infrared sensors can be used as environmental sensors. Their fields of vision at least partially overlap in terms of range and aperture angle and form a detection area FOV of the sensor arrangement SEN. Depending on the maximum travel speed of the rail vehicle RV, the detection area FOV can cover a section of track in front of the rail vehicle RV of at least 5 m up to 500 m as well as the surrounding area to the side of this section of track. The sensor arrangement SEN is used to record sensor data SD that represent an object located in the detection area FOV. There can also be several objects in the detection area FOV. The recorded sensor data SD is processed in an evaluation unit EVU of the object detection device ORU in order to determine the current actual values ​​of the object(s).From the sensor data SD of individual environmental sensors or the fused sensor data of several environmental sensors, a current actual object distance D between the rail vehicle RV and the object as well as actual object data M, T characterizing the object can be determined as actual values ​​of the object(s), namely at least one actual object dimension M and an actual object type T of an object. As actual object dimension M, an actual object height and / or an actual object width and / or an actual object diagonal and / or an actual object area and / or an actual object boundary, i.e. a shape of the outer boundary of an object cross-section, can be determined.From this, the actual object type T of an object can be identified, i.e. whether the detected object is another rail vehicle traveling on the same track or an adjacent track, a truck, a passenger car, a motorcyclist, a cyclist, a pedestrian (e.g., an adult or a child), a stroller, an animal, particularly a wild animal, a lamppost, an overhead line mast, a buffer stop, or other objects in the track-side infrastructure. Reliable, automated recording of the actual values ​​D, M, T of objects ahead is of key importance for a rail vehicle (RV) in automated operation if the observation of the environment ahead of the rail vehicle (RV) no longer falls solely within the responsibilities of a train driver in the DRC driver's cab, or if the deployment of a train driver is to be eliminated entirely.

[0024] According to the invention, a validation system 10 is therefore proposed for testing the functionality of the object recognition device ORU installed on a rail vehicle RV. The validation system 10 comprises one or more reference objects 20 which are positioned along the rail route RT at one or more validation points. One or more reference objects 20 can be positioned at a validation point. The validation points are arranged at a distance from one another along the rail route RT, which is designed, for example, as a branch line or circular route, in such a way that the rail vehicle RV repeatedly passes one of the validation points during its operation, at which points the functionality of the object recognition device ORU can be checked using known reference objects 20.

[0025] Each reference object 20 is positioned at an object position P20 in the area of ​​the rail track RT such that it lies in the detection range FOV of the sensor arrangement SEN of the rail vehicle RV approaching the reference object 20 on the rail track RT. The respective object position P20 is measured with high accuracy when a reference object 20 is set up. In the illustrated embodiment, the reference object 20 has a reference object-side positioning module 21 with a receiving unit 22 for navigation satellite signals, with which the object position P20 of the reference object 20 is automatically determined with the high accuracy of the GNSS signals after its positioning.

[0026] Each reference object 20 is designed as an artificial test dummy, also called a dummy, with specific target values. Thus, a reference object 20 has a specific target object type T-ref, such as a person, a wild animal, a motor vehicle, a stroller, a lamppost, an overhead line mast, and the like, which is defined by one or more target object dimensions M-ref. As a target object dimension M-ref, for example, a target object height and / or a target object width and / or a target object diagonal and / or a target object area and / or a target object boundary, i.e. a shape of the outer boundary of an object cross-section can be considered. In this way, reference objects of 20 different types, sizes and poses can be manufactured as test dummies and positioned at a validation point.When the rail vehicle RV approaches a reference object 20 or a group of reference objects 20, their target values ​​serve as a comparison basis for the corresponding actual values ​​determined by the object recognition device ORU.

[0027] The validation system 10 also comprises a database 30 in which the object position P20 and the target object data M-ref, T-ref of a positioned reference object 20 are stored. If there are several reference objects 20, the target object dimensions M-ref and the target object type T-ref for each reference object 20 are stored in the database 30. Depending on the design of the validation system 10, the database 30 is arranged on the reference objects 20 or distributed across them, but can also be arranged on the rail vehicle RV or, for example, be made available centrally by an operator of the rail vehicle RV or via a computer cloud or cloud OLD.

[0028] The validation system 10 further comprises a distance determination system 40 for determining a current target object distance D-ref between the rail vehicle RV and the reference object 20. This highly precisely determined target object distance D-ref serves as a reference value or comparison basis for the actual object distance D detected by the object recognition device ORU. For this purpose, the distance determination system 40 has a vehicle-side position determination module 41 with a receiving unit 42 for navigation satellite signals for determining a current vehicle position PRV of the rail vehicle RV, the accuracy of which is predetermined by the GNNS signals, for example from GPS or GALILEO. In the illustrated embodiment, the vehicle-side distance determination system 40 additionally has an inertial measuring unit 44 with acceleration and yaw rate sensors in order to obtain a further vehicle position value.The current vehicle position PRV is then determined with greater accuracy from the received navigation satellite signals and the measured vehicle position value. The distance determination system 40 further comprises a computing unit 43 for calculating the current target object distance D-ref, which is calculated from the determined current vehicle position PRV and the object position P20 of the reference object 20. For this purpose, the computing unit 43 communicates with the database 30 in order to obtain the object position P20 of the reference object 20 stored there.

[0029] If the object position P20 is located on a database 30 in the reference object 20 and the computing unit 43 is located on the rail vehicle RV, the object position P20 is transmitted, for example, by radio from the reference object 20 to the rail vehicle RV. For this purpose, the validation system 10 further comprises a communication system 60 with a vehicle-side communication module 61 and a reference object-side communication module 62. These are configured to exchange data wirelessly between the rail vehicle RV and the reference object 20, for example by forming a radio system operating in the 433 MHz or 868 MHz frequency range.

[0030] The validation system 10 further comprises a test unit 50 which is designed to calculate deviations AD, AM, AT between the determined actual object distances D and the calculated target object distances D-ref as well as between the determined actual object data M, T and the target object data M-ref, T-ref:

[0031] AD = D - D-ref

[0032] AM = M - M-ref AT = T - T-ref

[0033] The test unit 50 then determines whether the calculated deviations AD, AM, AT lie within predetermined validity ranges VAL-D, VAL-M, VAL-T. The validity range VAL-D, VAL-M, VAL-T is understood here as a tolerance range for actual object values ​​D, M, T around a target object value D-ref, M-ref, T-ref; actual object values ​​D, M, T determined by the object recognition device ORU that lie within the tolerance range are then qualified as valid. If the actual object values ​​D, M, T determined for a reference object 20 are within the tolerance ranges, the functional test of the object recognition device ORU has been positive and it can be assumed that the object recognition device ORU recognizes the target object type T-ref with its target object dimensions M-ref and the target object distance D-ref of the reference object 20 correctly and accurately enough.

[0034] The validation system 10 shown further comprises a reporting unit 110 which is connected to the testing unit 50 and is designed to send an electronic message to an output unit DIS in the driver's cab DRC of the rail vehicle RV and / or to the operations center OCL of the rail vehicle RV. If the check by the testing unit 50 has shown that one or more deviations AD, AM, AT lie outside the respective validity ranges VAL-D, VAL-M, VAL-T, a message can be sent by email or via another messaging service and output to the train driver in the case of semi-automated vehicle operation and / or to a control center employee in the case of fully automated vehicle operation by means of an output unit DIS designed, for example, as a display or monitor.The notification may contain the values ​​of the deviations AD, AM, AT in relation to the validity ranges VAL-D, VAL-M, VAL-T, but may also contain the size of the deviations AD, AM, AT values ​​and recommendations for action as to whether, for example, early maintenance or immediate decommissioning of the object recognition device ORU is recommended.

[0035] The validation system 10 can further comprise a weather system 100 with one or more environmental sensors 101 for recording environmental data U of the reference object 20. Weather-related environmental conditions can influence the recording of sensor data SD by the sensor arrangement SEN. For example, the air temperature and / or the air humidity and / or the air pressure and / or the precipitation and / or the wind speed and / or the solar radiation and the like can lead to sensor data SD being incomplete or incorrect and object recognition therefore being unreliable or not possible at all. To record this environmental data U, one or more environmental sensors 101 are arranged in the area of ​​the reference object 20 and can be designed as thermometers and / or hygrometers and / or barometers and / or ombrometers and / or anemometers and / or pyranometers.The environmental sensors 101 can be combined to form a weather station arranged on the reference object 20. If deviations AD, AM, AT between actual and target values ​​lie outside the validity ranges VAL-D, VAL-M, VAL-T, a possible explanation can be found based on the environmental data U, without the object recognition device ORU having necessarily operated incorrectly. Instead of a negative result for the functional test, the test unit 50 can, for example, reject the test result as invalid if adverse weather conditions are present and issue a corresponding message via the reporting unit 110.

[0036] The validation system 10 also includes a digital map system 90 for a rail network comprising the rail route RT. The digital map system 90 can be stored directly on the rail vehicle RV, but is at least accessible from the rail vehicle RV if it is stored at another location. All positioned reference objects 20 are recorded in the map system 90. The distance determination system 40 is designed to determine when a predetermined trigger distance of the rail vehicle RV has been reached before the object position P20.Reaching the trigger distance triggers that as the rail vehicle RV approaches a reference object 20, sensor data SD are continuously recorded by the sensor arrangement SEN, current actual object distances D and actual object data M, T are determined by the evaluation unit EVU, current target object distances D-ref are determined by the distance determination system 40, deviations AD, AM, AT are calculated, it is determined in each case whether these lie within the validity ranges VAL-D, VAL-M, VAL-T, and environmental data U are recorded by the weather system 100.

[0037] The validation system 10 further comprises a time synchronization system 70 with a vehicle-side synchronization module 71 and with a reference object-side synchronization module 72. These are designed so that the recording of the sensor data SD, the determination of the actual object distances D and the actual object data M, T, the determination of the target object distances D-ref, the calculation of the deviations AD, AM, AT, the determinations as to whether these lie within the validity ranges VAL-D, VAL-M, VAL-T, and the recording of the environmental data U take place in a time-synchronized manner with one another.The synchronization modules 71, 72 have their own internal clocks, which are synchronized using methods known per se (for example the Precision Time Protocol according to IEEE 1588 or the GPS time signal), which ensures that the synchronization modules 71, 72 and the components connected to the synchronization modules 71, 72 in the rail vehicle RV or reference object 20 carry out the correct actions at the correct time and in a coordinated manner.

[0038] The validation system 10 also includes a data memory 80 for the time-synchronized recording of the sensor data SD, the actual object distances D, the actual object data M, T, the target object distances D-ref, the deviations AD, AM, AT, the environmental data U, and the determination results as to whether the deviations AD, AM, AT lie within the validity ranges VAL-D, VAL-M, VAL-T. The time-synchronized procedure makes it possible to relate data that belong to one another in time. The data memory 80 can be located on the rail vehicle RV and / or the reference object 20 and / or in the cloud OLD. Wireless local area networks (WLAN) or mobile radio systems according to the LTE or 5G standard can be used to transmit or upload data using the communication system 60. Likewise, the communication modules 61, 62 can be designed as terminal devices of the satellite network.The data storage 80 may have a management system which sends a corresponding notification when the amount of data stored in the data storage 80 exceeds a threshold value, so that measures can be taken to ensure that sufficient storage space is always available in the data storage 80.

[0039] For the autonomous supply of its electrical consumers with electrical energy, the reference object 20 can have a rechargeable energy source 21. This makes the reference objects 20 at least temporarily free of power grid connections in their energy supply and allows them to be positioned flexibly along the rail route RT. A management system for an energy source 21 designed as a rechargeable battery detects when the charge level falls below a threshold value, whereupon a corresponding message can be sent to recharge the rechargeable battery.

[0040] The validation system 10 can be used during ongoing rail operations at any time of day or night and under corresponding ambient conditions to monitor the proper functioning of the object recognition device ORU of a rail vehicle RV. However, it can also be used for the comprehensive functional testing of an object recognition device ORU that is still under development for rail vehicles RV. If the deviations AD, AM, AT between the actual and target object values ​​​​deviate from the assigned validity ranges VAL-D, VAL-M, VAL-T for any reason, measures can be taken immediately. The validation system 10 according to the invention thereby considerably increases the operational reliability of a partially or fully automated rail vehicle RV with an object recognition device ORU for obstacle detection.

[0041] The reference symbols used in the description of the figures are clearly shown in the following list of reference symbols

[0042] RV rail vehicle

[0043] DRC driver's cab

[0044] RT railway line

[0045] ORU object recognition device

[0046] SEN sensor arrangement

[0047] FOV detection area

[0048] SD sensor data

[0049] D Actual value: Actual object distance

[0050] M Actual value: Actual object date: Actual object dimensions

[0051] T Actual value: Actual object date: Actual object type

[0052] 10 Validation system

[0053] 20 reference objects

[0054] 21 Positioning module, reference object side

[0055] 22 Receiving unit

[0056] P20 Object position

[0057] T-ref Target value: Target object date: Target object type

[0058] M-ref Target value: Target object date: Target object dimension

[0059] 30 Database

[0060] OLD computer cloud

[0061] 40 Distance determination system

[0062] D-ref target object distance

[0063] 41 Positioning module, vehicle side

[0064] 42 Receiving unit

[0065] PRV vehicle position

[0066] 43 computing unit

[0067] 44 measuring unit, inertial

[0068] 60 Communication system

[0069] 61 Communication module, vehicle side

[0070] 62 Communication module, reference object side

[0071] OCL Operations Center

[0072] DIS output unit

[0073] EVU evaluation unit

[0074] 23 Energy source

[0075] 50 test units

[0076] AD Deviation between D and D-ref AT Deviation between T and T-ref

[0077] AM deviation between M and M-ref

[0078] VAL-D Validity range for object distance

[0079] VAL-T Validity range for object type VAL-M Validity range for object dimensions

[0080] 110 reporting unit

[0081] 100 weather system

[0082] 101 Environmental sensor

[0083] U Environmental data 70 Time synchronization system

[0084] 71 Synchronization module, vehicle side

[0085] 72 Synchronization module, reference object side

[0086] 80 data storage

[0087] 90 card system

Claims

Patent claims 1. Validation system (10) for functional testing of an object recognition unit (ORU) installed on a rail vehicle (RV), which - a sensor arrangement (SEN) having a detection area (FOV) for recording sensor data (SD) representing an object located in the detection area (FOV), and - an evaluation unit (EVU) processing the recorded sensor data (SD) for determining a current actual object distance (D) between the rail vehicle (RV) and the object as well as actual object data (M, T) characterising the object, wherein at least one actual object dimension (M) and one actual object type (T) are determined as actual object data (M, T), comprising - a reference object (20) which - is positioned at an object position (P20) in the area of a rail track (RT) such that it lies in the detection range (FOV) of the sensor arrangement (SEN) of the rail vehicle (RV) approaching the reference object (20) on the rail track (RT), and - has at least one target object dimension (M-ref) and one target object type (T-ref) as target object data (T-ref, M-ref), - a database (30) in which the object position (P20) and the target object data (T-ref, M-ref) of the positioned reference object (20) are stored, - a distance determination system (40) for determining a current target object distance (D-ref) between the rail vehicle (RV) and the reference object (20), which - a vehicle-side position determination module (41) with a receiving unit (42) for navigation satellite signals for determining the current vehicle position (PRV) of the rail vehicle (RV) and - a computing unit (43) communicating with the database (30) for calculating the current target object distance (D-ref) from the current vehicle position (PRV) and the object position (P20), and - a test unit (50) which is designed to calculate deviations (AD, AT, AM) between the determined actual object distance (D) and the calculated target object distance (D-ref) as well as between the determined actual object data (M, T) and the target object data (M-ref, T-ref), and which is further designed to determine whether the deviations (AD, AT, AM) lie within predetermined validity ranges (VAL-D, VAL-T, VAL-M).

2. Validation system according to claim 1, further comprising - a digital map system (90) for a rail network comprising the rail route (RT), in which the object position (P20) of the positioned reference object (20) is recorded, - wherein the distance determination system (40) is designed to determine whether a predetermined triggering distance (DT) of the rail vehicle (RV) has been reached in front of the object position (P20), - whereby upon reaching the triggering distance (DT), the system is triggered so that during the approach of the rail vehicle (RV) to the reference object (20) - Sensor data (SD) are recorded by the sensor array (SEN), - current actual object distances (D) and actual object data (M, T) are determined by the evaluation unit (EVU), - current target object distances (D-ref) are determined by the distance determination system (40), and - Deviations (AD, AM, AT) are calculated and it is determined whether they lie within the validity ranges (VAL-D, VAL-M, VAL-T).

3. Validation system (10) according to one of the preceding claims, further comprising a communication system (60) with a vehicle-side communication module (61) and a re- reference object-side communication module (62) which is designed to exchange data wirelessly between the rail vehicle (RV) and the reference object (20), and optionally additionally with an operations center (OCL) and a data memory (80) in a computer cloud (CLD).

4. Validation system (10) according to one of the preceding claims, wherein the reference object (20) has a reference object-side position determination module (21) with a receiving unit (22) for navigation satellite signals, which is designed to determine the object position (P20) of the reference object (20) after its positioning in the region of the rail track (RT).

5. Validation system (10) according to one of the preceding claims, wherein the vehicle-side distance determination system (40) additionally has an inertial measuring unit (44) for measuring a vehicle position value (PRV ), wherein the current vehicle position (PRV) is determined from the received navigation satellite signals and the measured vehicle position value (PRV ).

6. Validation system (10) according to one of the preceding claims, further comprising - a time synchronization system (60) with a vehicle-side synchronization module (61) and with a reference object-side synchronization module (62), which are designed so that the recording of the sensor data (SD), the determination of the actual object distances (D) and the actual object data (T, M), the determination of the target object distances (D-ref) and the calculation of the deviations (AD, AT, AM) are time-synchronized, and - a data memory (80) for recording the time-synchronized sensor data (SD), actual object distances (D), actual object data (T, M), target object distances (D-ref) and deviations (AD, AT, AM).

7. Validation system (10) according to one of the preceding claims, further comprising a weather system (100) with an environmental sensor (101) for recording environmental data (U) of the reference object (20), which can influence the recording of sensor data (SD) by the sensor arrangement (SEN), wherein the environmental data (U) are recorded in a time-synchronized manner and recorded in the data memory (80).

8. Validation system (10) according to one of the preceding claims, further comprising a reporting unit (110) connected to the testing unit (50), which is designed to send an electronic message to an output unit (DIS) in a driver's cab of the rail vehicle (RV) and / or in an operations center (OCL) of the rail vehicle (RV) if the check has shown that deviations (AD, AT, AM) lie outside the respective validity ranges (VAL-D, VAL-T, VAL-M).

9. Validation system (10) according to one of the preceding claims, wherein the reference object (20) has a rechargeable energy source (23) for the autonomous supply of its electrical consumers with electrical energy.

10. Validation procedure for functional testing of an object detection unit (ORU) installed on a rail vehicle (RV), - wherein a reference object (20) which has at least one target object dimension (M-ref) and one target object type (T-ref) as target object data (T-ref, M-ref) is positioned at an object position (P20) in the area of a rail track (RT) to be traveled by the rail vehicle (RV), - wherein sensor data (SD) characterizing the reference object (20) are recorded by a sensor arrangement (SEN) of the object recognition device (ORU), - whereby from the recorded sensor data (SD) by an evaluation unit (EVU) of the object recognition device (ORU) a current actual object distance (D) between rail vehicle (RV) and reference object (20) and at least one Actual object dimensions (M) and an actual object type (T) are determined as actual object data (M, T), - wherein a current target object distance (D-ref) between the rail vehicle (RV) and the reference object (20) is determined by determining the current vehicle position (PRV) of the rail vehicle (RV) by means of received navigation satellite signals and calculating the current target object distance (D-ref) from the current vehicle position (PRV) and the object position (P20), - wherein deviations (AD, AT, AM) between the determined actual object distance (D) and the calculated target object distance (D-ref) as well as between the determined actual object data (M, T) and the target object data (M-ref, T-ref), and - determining whether the deviations (AD, AT, AM) lie within predetermined validity ranges (VAL-D, VAL-T, VAL-M).

Images