Control device, method and sensor device for self-monitoring ground positioning

Abstract

A control device for a vehicle is disclosed, the control device being used to analyze a positioning system, wherein the control device can be connected in a data-transmitting manner to at least two positioning systems that can be operated independently of one another, the at least two positioning systems being used to determine a system-specific position, wherein the control device is configured to evaluate the position information determined by the positioning systems in a plausibility-checked manner. In addition, a method and a sensor device are disclosed.

Description

Technical Field

[0001] The invention relates to a control device for a vehicle, a method for positioning a sensor device and a sensor device. Background Art

[0002] As demands in the automotive sector increase and the degree of automation increases, more and more sensors are used. These requirements are particularly imposed on automated driving functions, such as highly automated driving or driverless driving. Since failures or functional errors of the sensor system and the subsequent processing systems cannot be completely ruled out, continuous or frequent self-monitoring is required, especially in the case of driverless controlled vehicles.

[0003] Driver assistance systems and automated driving functions rely on sensors that have been calibrated or initially set since leaving the factory. Due to the increasing number of sensors, errors can occur during the factory process of calibrating the sensors. Furthermore, during operation or after delivery of the sensor system or vehicle, the sensors can deviate from the original factory specifications. Such deviations can be caused by external influences or by wear phenomena. In this context, sensor errors or sensor misalignments that cannot be compensated or detected autonomously during operation are particularly problematic, since this results in the use of erroneous sensor data for safety-related functions without being noticed. Summary of the invention

[0004] The object on which the present invention is based can be seen as providing a control device and a method which allow a reliable operation of a plurality of positioning systems.

[0005] This object is achieved by means of the corresponding subject matter of the method according to the invention. Advantageous embodiments of the invention are the subject matter of the individual embodiments.

[0006] According to one aspect of the present invention, a control device for a vehicle is provided, the control device being used to analyze a positioning system, wherein the control device can be connected in a data transmission manner to at least two positioning systems that can be operated independently of each other, the at least two positioning systems being used to determine a system-specific position. The control device is configured to analyze and process position information determined by the positioning system in a plausibility-checking manner.

[0007] According to another aspect of the invention, a method for localizing a sensor device, in particular a sensor device of a vehicle, by means of a control device is provided. In one step, a system-specific position is determined by using sensors of at least two localization systems. The system-specific positions of the different localization systems are compared with one another. In the event of a lack of a predetermined consistency of the system-specific positions, a signal is generated. According to an advantageous embodiment, after the signal has been generated, an error analysis can be performed by the control device.

[0008] According to another aspect of the invention, a sensor device, in particular a sensor device for a vehicle, is provided. The sensor device has at least two positioning systems that can be operated independently of one another and are used to determine the system-specific position of the sensor device. Each positioning system accesses at least one sensor and corresponding sensor data, wherein the sensor device has a control device coupled to the positioning system in a data-conducting manner. The control device is configured to evaluate the position information determined by the positioning system in a plausibility-checked manner.

[0009] The control device, the sensor arrangement and / or the method can ensure a self-monitoring of an existing positioning system, which can reveal faulty calibration of the sensors or errors in the use of digital maps.

[0010] Although sensors can be calibrated from the factory or already in-process, this procedure does not permanently guarantee successful calibration and functional capability of the sensors. Later environmental influences can in particular change the position of the sensors relative to one another, a so-called external calibration. In addition, the sensors used can subsequently change their intrinsic properties. Such changes can be caused, for example, by forces acting on the sensors.

[0011] In addition to GPS sensors, lidar sensors, radar sensors, camera sensors, ultrasonic sensors, etc. can be used as sensors. Features in the surroundings of the sensor device that are identified by the sensor can be compared and identified with a digital map. Based on this comparison, the position of the sensor device in the map can be obtained. The different positioning systems used can, for example, access one type of sensor and preferably operate independently of each other. Therefore, a positioning system can extract features of the surroundings based on the sensor data of a camera sensor. Another positioning system can obtain surrounding features based on the sensor data of a radar sensor. Such a positioning system can use sensor data of the same or different sensors to calculate a position specific to the corresponding positioning system.

[0012] The positioning system can be implemented in the form of a positioning algorithm on a control unit or on a special control unit and connected to the sensor directly or via a drive in a data-conducting manner. In addition, the positioning system can have one or more machine-readable memories, in which the positioning system is stored and can be implemented. The memory can also be used as a temporary or permanent memory for sensor data.

[0013] The positioning system can be composed of one or more positioning filters (such as Kalman filters or particle filters), one or more positioning optimizers (such as graphic optimizers) or any combination thereof. The positioning system can determine the position with high certainty based on features in the surrounding environment and thus the position determination is convergent.

[0014] If the sensors have low noise, the maps used are precise and still up-to-date, the association of surrounding features with map features has been correctly carried out and if the optimization of the features with respect to one another or of the features for the position of the device relative to the features was successful, the position can be determined with high certainty by the positioning system.

[0015] If one of the listed points is implemented incorrectly or with insufficient accuracy, the position is determined with low certainty and is therefore non-convergent.

[0016] The control device can receive the position data determined by the positioning system and analyze them in a plausibility-checked manner. In the case of the analysis by the control device in a plausibility-checked manner, the corresponding position data can be evaluated and / or compared with each other. Thus, only the position data determined by the positioning system are used for further applications, which have been determined logically and / or correctly by the positioning system. Thus, different unrelated position data can be optimally associated or processed with each other. Illogical position data can be excluded from the application by the control device.

[0017] For example, a solution which best describes the positioning of the position data can be calculated by the control device. Such a solution can, for example, have an area which has been determined from the known position data by the method of least squares. All position data outside this area can be ignored. In particular, outliers can be identified by evaluation in the form of a plausibility check and ignored in the case of further evaluation by the control device.

[0018] According to one specific embodiment, the control device is configured to evaluate the convergence behavior of the locating systems in order to ascertain the respective system-specific positions, wherein differences in the convergence behavior of the respective locating systems can be ascertained by the control device.

[0019] In particular, the position information ascertained by the positioning system can be evaluated in a plausibility-checked manner by evaluating the convergence behavior of the positioning system.

[0020] The specific positions determined by the positioning systems can form an intersection or set by spatial superposition, and this intersection or set is regarded as the convergence of the corresponding positioning systems.

[0021] Therefore, the convergence position determined by the plurality of positioning systems is preferably used as the absolute position of the device. In the case of a plurality of convergence positions, the position determined based on the majority of the positioning systems can further be used.

[0022] Such uncertainties in determining the position are caused by sensor errors and sensor noise, for example due to interference effects and weather influences. In addition, incorrect individual associations and outdated maps with missing surrounding features can lead to uncertain positioning. Such uncertainties in positioning can be checked and identified by analyzing the convergence characteristics of the positioning system by the control device. Thus, errors in the mapping process, such as features that are not correctly aligned with each other or so-called "alignment" errors, can also be determined.

[0023] The method can be used in particular for vehicles that can be operated automatically and have corresponding automated driver assistance systems. With the help of the method, self-monitoring positioning, map verification, and sensor checking and recalibration can be performed. Here, positioning systems, such as positioning systems of highly automated vehicles, are operated independently of each other so that the cause of the lack of convergence of the positioning can be inferred in the case of possible differences in the convergence characteristics. As a result, the independent operation of multiple positioning systems can be used to achieve redundancy in positioning. In addition, multiple system-specific positions obtained by different positioning systems can be compared with each other in order to determine the positioning system with errors. If each positioning system is independent of other positioning systems, for example, by the characteristics and / or sensors used and their corresponding combinations, the cause of the lack of convergence can be determined by the differences in the convergence characteristics. This can be detected, for example, in the case of multiple positioning systems based on the deviations of the majority of positioning.

[0024] In this case, the positioning system determines the position of the sensor device individually and system-specifically, which positions can agree with one another or deviate from one another in the best case. The determined absolute position of the device can thus be considered as an average of a plurality of system-specific positions. In this case, the system-specific positions can deviate slightly from one another while maintaining tolerances in order to still be classified as agreeing.

[0025] The use of various possible combinations of convergence characteristics of sensor-independent positioning systems, but overlapping with respect to the map features used, can achieve the following advantageous functions:

[0026] Detection sensor failure;

[0027] Detect miscalibration and misadjustment;

[0028] Detect map changes.

[0029] In this case, despite faulty sensors or evaluations, localization of the sensor device or of the vehicle is still possible, since, if a discrepancy with one or more localization systems is detected, the control device can continue or allow localization of the device by weighting and by checking and monitoring the localization system as a whole.

[0030] According to one embodiment, the system-specific position of at least one positioning system is determined based on sensor data of at least one sensor via a map stored in the control device or extracted from the control device. Preferably, the positioning system uses one or more a priori independent map features and one or more types of sensors that are also a priori independent of one another. Based on the sensors and map features, the system-specific position in the mapped world can be calculated relatively by detecting, linking and optimizing surrounding features and map features. This makes it possible to perform positioning independent of the GPS system.

[0031] According to another embodiment, the control device can determine a sensor error or miscalibration based on the determined differences in the convergence characteristics of the positioning system. The corresponding positioning systems can converge to a specific position respectively. However, one or more specific positions may deviate from other specific positions. In this case, the sensor data of the system that has determined the deviated position may be erroneous and at least one of the sensors used may therefore be erroneous. In this case, an online calibration or calibration can be performed during operation. Alternatively or additionally, a recalibration of the corresponding sensors can be performed in the workshop. As long as the majority of the positioning systems of different sensors converge to a common position, the position can still be used as an available position of the sensor device or an available position of a vehicle with a sensor device.

[0032] According to another embodiment, the control device can determine map errors based on the determined differences in the convergence characteristics of the positioning systems. As a result, certain map features, such as lane markings, signs, etc., can be used for positioning. In the event of errors in the map, a positioning system using certain map features may converge to a different position than the remaining positioning systems using different map features. This characteristic indicates a change in the actual surroundings relative to the map or indicates map features that were incorrectly aligned with each other when the map was created. For example, new lane markings that are slightly shifted can be drawn. Alignment errors in the map may be caused, for example, by algorithmic errors or calibration errors when the map was created.

[0033] According to another embodiment, the measurement data of different sensors that can be operated independently of one another are run. As a result, the positioning system can be operated redundantly, wherein errors are isolated and limited to a single positioning system.

[0034] According to another embodiment, at least one system-specific position is determined by comparing and identifying features of a digital map.Therefore, a positioning can be provided which is independent of GPS sensors and which can therefore operate reliably and precisely even in densely built-up areas.

[0035] According to another embodiment, the system-specific positions of different positioning systems are compared with one another while taking into account tolerances. As a result, the corresponding positions determined by the positioning systems can have slight deviations in order to compensate for system deviations and basic noise of the sensors.

[0036] According to another embodiment, a check is made according to the error analysis whether one or more sensors of the positioning system are out of adjustment or have a sensor error. Thus, a self-monitoring of the sensor arrangement can be provided. If an inconsistency is determined by the control device, corresponding measures can be carried out, such as an online calibration or a transition to a safe state. Furthermore, the positioning of the sensor arrangement can be continued based on the result of the positioning system, when the control device verifies the correctness of the result.

[0037] According to another embodiment, a check is performed based on an error analysis to determine whether a map error exists. High-precision maps are a basic prerequisite for safe automated driving and are part of highly developed assistance systems. Here, too, these functions require the position of the vehicle relative to the map. The more features of the real world are stored in the map, the easier it is for the map to become outdated, since each of the features may change over time. If assistance systems and systems for automated driving use map information, the map's up-to-dateness must first be ensured. This method can be used to detect errors that are caused by maps that are no longer up-to-date. Such errors may occur when the positioning system is based on a certain feature type, such as lane markings, and the correspondingly calculated position deviates from other positioning systems.

[0038] The system-specific positions can be relative positions or absolute positions. These positions can have, for example, a distance from an object or feature that is adjacent or arranged in the scanning area.

[0039] According to a further embodiment, the position of the sensor device is determined based on a plurality of system-specific positions that are substantially consistent with one another. Thus, despite the detection of an error in the positioning system, the function can still be continued, since the source of the error can be extracted by the error analysis of the control device. For example, in the case of most positioning systems that use different features of a map, a position can be determined when the determined positions (without taking into account an erroneous positioning system) are consistent and plausible.

[0040] According to another embodiment, at least one positioning system is taken into account to eliminate ambiguities or errors of another positioning system. This allows a redundant functioning of the positioning systems, which ensures positioning even in the event of errors, ambiguities or misalignments.

[0041] According to another embodiment, the weighting is influenced by the control device when determining the system-specific position of the positioning system. Depending on the configuration of the positioning system, the weighting or error can be adjusted, for example, in the case of a positioning optimizer according to the individual edges of the graphic optimizer. In addition, in the case of positioning filters, individual filter terms in the case of a Gaussian mixture filter or individual particles in the case of a particle filter can be evaluated and analyzed. By evaluating different convergence characteristics, the control device can decide which positioning filters or which parts of positioning filters are actually considered, evaluated and / or combined for positioning. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Preferred exemplary embodiments of the invention are explained in more detail below based on a greatly simplified schematic diagram.

[0043] Here,

[0044] Figure 1 A schematic top view of a sensor device in a vehicle according to one specific embodiment of the present invention is shown;

[0045] Figure 2-7 A schematic diagram of a scenario of a sensor arrangement is shown for illustrating the method according to the present invention. DETAILED DESCRIPTION

[0046] Figure 1 A top view of a sensor device 1 installed in a vehicle 2 is shown.

[0047] The sensor device 1 has a first positioning system 4, a second positioning system 8 and a third positioning system 12, wherein the first positioning system uses two camera sensors 6 for system-specific positioning, the second positioning system is based on a lidar sensor 10, and the third positioning system uses two radar sensors 14. In this case, the camera sensor 6 and the radar sensor 14 are arranged at the front of the vehicle and at the rear of the vehicle, respectively. The lidar sensor 10 is positioned on the top of the vehicle.

[0048] According to this exemplary embodiment, the locating system 4, 8, 12 is implemented as a component of the control device 16. However, the locating system 4, 8, 12 can also be designed as a separate or independent software module or hardware module.

[0049] The respective positioning system 4, 8, 12 can determine the relative or absolute system-specific position of the vehicle 2 by evaluating the sensors 6, 10, 14. The positioning systems 4, 8, 12 operate independently of one another and can access sensors 6, 10, 14 of different types.

[0050] The control device 16 is configured to evaluate the positioning systems 4, 8, 12 and in particular to compare the positions of the vehicle 2 or the device 1 determined by the positioning systems 4, 8, 12. In addition, the control device 16 is configured to merge the positioning systems 4, 8, 12 or the positions determined by the positioning systems 4, 8, 12 with each other.

[0051] exist Figure 2-7 , a schematic diagram of a scenario of a sensor device 1 on a four-lane road 18 is shown for illustrating the method according to the invention. The sensor-specific positions P1, P2, P3 determined by the positioning systems 4, 8, 12 are represented as planar regions describing the uncertainty of the respective positioning. In addition, one or more intersection regions X, X1, X2 are shown, which are formed by the intersection of the positions P1, P2, P3 and thus form the positions to which the respective positioning systems 4, 8, 12 converge. The respective positions P1, P2, P3 are determined, for example, by the positioning systems 4, 8, 12 relative to the roadway boundary 20 or the roadway marking 22.

[0052] exist Figure 2 In the embodiment of FIG. 1 , all positioning systems 4 , 8 , 12 converge to a common position X. The sensor device 1 has an unlimited functionality, so that the position X can be used, for example, for automated driving.

[0053] Figure 3 Another scenario is shown, in which a positioning system 12 using a radar sensor 14 converges to a deviating position X1. The specific position P3 determined by the third positioning system 12 does not correspond to the specific positions P1, P2 of the first and second positioning systems 4, 8, which converge to a common position X. Such a scenario occurs in particular when the radar sensor 14 is misaligned with respect to the other sensors 6, 10. In this case, the control device 16 can perform an online calibration or request a workshop search. In the workshop, the radar sensor 14 can be recalibrated.

[0054] As soon as the majority of the localization systems 4 , 8 of the different sensors 6 , 10 converge on a common position X, this position can still be output as a usable position of the vehicle 2 .

[0055] exist Figure 4, in which the position P2 determined based on the radar sensor 14 does not converge to one position. In this case, the control device 16 can determine a sensor error. In this case, the sensor units 6, 10 can also determine an unambiguous and plausible position that corresponds to Figure 3 The position X in is consistent.

[0056] exist Figure 5 , positions P1, P2 of positioning systems 4, 8 are shown, which use lane markings 22 as features. However, these positions P11, P21 converge to a position X1 that is different from the other ascertained positions P1, P2, wherein different map features are used in this case.

[0057] This indicates either a change in actual surroundings 18 relative to the map or an error that was already present when the map was generated.

[0058] In the case of this scenario, most system-specific positions P1, P2 are configured so that they converge in position X and thus enable effective positioning. The corresponding specific positions P1, P11 and P2, P21 have been created based on the sensor data of the camera sensor 6 and the lidar sensor 10, wherein the sensor data of the same sensor can be used for feature analysis of different features, thereby determining a specific position for each considered feature. Thus, one or more positions P2, P21 can be determined by means of the lidar sensor 10, for example, by the second positioning system 8.

[0059] Figure 6 Another scenario is shown in which the positioning system 4, 8 uses map features that lead to non-convergent positions P11, P21. In this case, either the surroundings 18 have changed since the map was created so that the map can no longer describe the surroundings 18, or errors have occurred during the map creation, which has the result that the map can no longer reliably depict the surroundings. Here, the feature diversity can also always lead to a valid position X.

[0060] Figure 7A situation in which convergence cannot be obtained is shown. The positions P1, P2, P3 that have been obtained by the positioning systems 4, 8, 12 according to different characteristics diverge and / or cannot be unified to one position. In this case, there is a sensor error or a map error. In this case, it is impossible to distinguish between the two types of errors. If there is still a (mixed in this case) majority of positioning systems that are unified to one position despite the divergence and inconsistency, the positioning system can be used for positioning. According to this embodiment, the position X is not provided for further use, so that the control device can output an error notification and put the vehicle 2 in a safe state as long as there is no driver controlling the vehicle 2. In the case of a manually controllable vehicle 2, the control device 16 can transfer vehicle control to the driver.

[0061] In the case of periodic signals, such as lane markings 22, a positioning system can converge on more than one position and thus have ambiguity with respect to a specific position. In this case, the ambiguity can be resolved by superimposing a correspondingly suitable other positioning system. For this purpose, signs in the direction of travel can mostly be locally unambiguous, and a corresponding positioning system that uses the signs as a feature for positioning can resolve the ambiguity of a positioning system that uses lane markings as a feature.

Claims

1. A control device (16) for a vehicle, the control device being used to analyze a positioning system (4, 8, 12), wherein: The control device (16) can be connected to a plurality of positioning systems (4, 8, 12) that can operate independently of each other in a data transmission manner, wherein the plurality of positioning systems are respectively used to determine a system-specific position (P1, P2, P3) of the vehicle, wherein the system-specific position is represented as a planar area describing the uncertainty of the positioning and can form an intersection or set by spatial superposition, and the intersection or set is regarded as the convergence of the corresponding positioning system, and is characterized in that the control device (16) is configured to analyze the position information obtained by each of the plurality of positioning systems (4, 8, 12) in a credibility check manner by analyzing the convergence characteristics of the positioning systems (4, 8, 12), wherein when most of the system-specific positions in the plurality of positioning systems converge to a common position, the common position can be used as an available position of the vehicle.

2. The control device according to claim 1, wherein: The control device (16) is configured to evaluate the convergence characteristics of the positioning systems (4, 8, 12) in order to determine corresponding system-specific positions (P1, P2, P3), wherein differences in the convergence characteristics of the corresponding positioning systems (4, 8, 12) can be determined by the control device (16).

3. The control device according to claim 2, wherein: A system-specific position (P1, P2, P3) of at least one positioning system (4, 8, 12) can be determined based on sensor data of at least one sensor (6, 10, 14) by comparison with features of a map stored in the control device (16) or features of a map extracted from the control device (16).

4. The control device according to claim 2, wherein: The control device (16) is capable of determining a sensor error or a miscalibration based on the ascertained difference in the convergence behavior of the positioning system (4, 8, 12).

5. The control device according to claim 3, wherein: The control device (16) is capable of determining an error in the map based on the ascertained difference in the convergence behavior of the positioning systems (4, 8, 12).

6. A method for positioning a sensor device (1) by means of a control device (16) according to any one of claims 1 to 5, wherein: Determining system-specific positions (P1, P2, P3) respectively by analyzing and processing measurement data of sensors (6, 10, 14) of at least two positioning systems (4, 8, 12); comparing the system-specific positions (P1, P2, P3) of the positioning systems (4, 8, 12) with one another; In the event of a lack of a predefined match of the system-specific positions ( P1 , P2 , P3 ), a signal is generated.

7. The method according to claim 6, wherein: An error analysis is performed after generating the signal.

8. The method according to claim 6 or 7, wherein: The measurement data originate from different sensors (6, 10, 14) which can operate independently of each other.

9. The method according to claim 6 or 7, wherein: At least one system-specific location (P1, P2, P3) is determined by comparing and identifying features of the digital map.

10. The method according to claim 6 or 7, wherein: The system-specific positions (P1, P2, P3) of different positioning systems (4, 8, 12) are compared with one another taking into account tolerances.

11. The method according to claim 7, wherein: According to the error analysis, it is checked whether one or more sensors (6, 10, 14) of a positioning system (4, 8, 12) are out of adjustment or have a sensor error.

12. The method according to claim 7, wherein: According to the error analysis, it is checked whether there is a map error.

13. The method according to claim 6 or 7, wherein: The position (X) of the sensor device (1) is determined based on a plurality of system-specific positions (P1, P2, P3) that substantially coincide with one another.

14. The method according to claim 6 or 7, wherein: Measurement data of at least one positioning system (4, 8, 12) are taken into account in order to eliminate ambiguities or errors of other positioning systems (4, 8, 12) and / or in order to influence weighting when determining a system-specific position (P1, P2, P3) of the positioning system (4, 8, 12).

15. The method according to claim 6, wherein: The sensor device (1) is a sensor device of a vehicle (2).

16. A sensor device (1) for a vehicle (2), the sensor device having a plurality of positioning systems (4, 8, 12) which can be operated independently of one another, the plurality of positioning systems being respectively used to determine a system-specific position (P1, P2, P3) of the sensor device (1), the system-specific position being represented as a planar area describing the uncertainty of the positioning and being able to form an intersection or set by spatial superposition, the intersection or set being regarded as a convergence of the corresponding positioning systems, wherein: Each positioning system (4, 8, 12) has at least one sensor (6, 10, 14), characterized in that the sensor device (1) has a control device (16) which is coupled to the positioning system (4, 8, 12) in a data-conducting manner, wherein the control device (16) is configured to analyze the position information determined by each of the plurality of positioning systems (4, 8, 12) in a plausibility-checking manner by analyzing the convergence characteristics of the positioning systems (4, 8, 12), wherein when a majority of the system-specific positions of the plurality of positioning systems converge to a common position, the common position can be used as a usable position of the vehicle.

Images