DE102021211405A1 - Method and control device for securing measurement data from at least one sensor of a vehicle - Google Patents

Abstract

The present invention relates to a method for securing measurement data (108) from at least one sensor (104) of a vehicle (100), using a measurement location (112) at which a measurement (110) of the measurement data (108) is measured , a stored location-based error probability (120) of the sensor (104) at the measurement location (112) is read from a location-based database (122) and the measurement (110) is processed taking into account the error probability (120).

Description

field of invention

The invention relates to a method for securing measurement data from at least one sensor of a vehicle, a corresponding control unit and a corresponding computer program product.

State of the art

A sensor can detect a detection area and provide measurements on detected objects within the detection area as measurement data. The measurement data consists of individual measured values of the measurements. For example, a reading can represent a distance between the sensor and a detected point on an object. Likewise, the measured value can depict a direction from the sensor to the detected point.

However, the measurements can also be incorrect. Erroneous measurements can be false-positive or false-negative. In a false positive measurement, a point is mapped to an object in the measurement data even though no object exists at that point. In the case of a false-negative measurement, an existing object is not shown in the measurement data.

Incorrect measurements can influence the behavior of a vehicle's assistance systems. For example, an object detected as a false positive can cause the assistance system to take evasive action and/or brake, which is particularly unpredictable for other road users.

Therefore, an attempt is made to distinguish correct measurements from incorrect measurements before further processing of the measurement data.

For example, describes the DE 10 2017 209 667 A1 a storage of speed information for the prediction of the future speed trajectory.

Disclosure of Invention

Against this background, with the approach presented here, a method for securing measurement data from at least one sensor of a vehicle, a corresponding control unit and a corresponding computer program product are presented according to the independent claims. Advantageous developments and improvements of the approach presented here result from the description and are described in the dependent claims.

Advantages of the Invention

In the approach presented here, location-based empirical values are used to check measurements for plausibility. The empirical values show whether incorrect measurements have already been recorded at a location in the past. The empirical values are stored georeferenced in a database. For example, the database can represent a map with location-based empirical values.

The empirical values are stored for a specific sensor or a specific sensor category and represent a probability with which the sensor or a sensor of the same category measures an incorrect measurement at the location.

Measurements from locations with an increased probability of erroneous measurements are classified as less reliable or weighted less during further processing.

A method for securing measurement data from at least one sensor of a vehicle is proposed, using a measurement location at which a measurement of the measurement data is measured, a stored location-based error probability of the sensor at the measurement location being read from a location-based database and the measurement being taken into account the error probability is processed.

Ideas for embodiments of the present invention can be regarded as being based, among other things, on the ideas and findings described below.

A sensor can detect a detection area and map objects in the detection area into measurement data. The sensor can be an active sensor such as a radar sensor or a lidar sensor. The sensor can also be a passive sensor such as a camera. The sensor can be part of a sensor system of a vehicle.

The measurement data consists of individual measurements. A measurement represents a detected point on an object within the detection area. With an active sensor, the measurement represents, for example, a reflection on the object. With a passive sensor, the measurement represents, for example, a pixel of an image of the object. The object can be represented by several measurements.

A measurement may include measurement values such as a distance value and/or a direction value from the measurement location to the measurement.

A measurement location can represent a position of the sensor during a measurement. The measurement is thus carried out at the measurement location. The measurement location can also represent a position of the vehicle during the measurement.

A faulty measurement can represent a supposedly recorded point on a non-existent object. The non-existent object can be referred to as a ghost object. The erroneous measurement cannot depict an actually existing object either.

Ghost objects or incorrect measurements can be caused by reflections on flat objects, for example. The reflection thereby forms a ghost object at a position where the reflected ghost object is not arranged. The position of the ghost object is dependent on an angle of the flat object to the measurement location. Both active sensors and passive sensors can be susceptible to reflections.

In the case of passive sensors such as cameras, for example, reflective objects such as panes of glass can lead to ghost objects. If the object is highly reflective, the angle of the object to the measurement location is almost unimportant. The ghost object can then also be your own vehicle, for example.

With active sensors, ghost objects can occur particularly on objects whose surfaces are hit at a flat angle of incidence. For example, ghost objects can occur on tunnel walls, crash barriers or other surfaces that are approximately parallel to the road.

An error probability can be a numerical value. The error probability can show how likely it is that the sensor will measure incorrectly at a measurement location. The error probability can be different for different sensor categories.

A location-based database can in particular contain data from a road map. Different error probabilities can be assigned to different route sections. If a measurement location is in the area of a road section, the error probability stored for this road section can be read out. Different error probabilities can be stored in the database for different sensor classes.

In the case of data fusion with other measurement data, the measurement can be discarded if the error probability is greater than a threshold value. The threshold value can depend on a use of the measurement data. For example, the threshold value for data fusion can be between 10 percent and 50 percent with an error probability. In particular, measured values with an error probability of more than 50 percent can be discarded. The threshold value can be lower, the more safety-relevant the use of the measurement data is.

The location-related database can be read in by a higher-level data processing system. The database can be stored centrally and many vehicles can access the database. The database can thus be maintained centrally. For example, construction sites and/or changes in the course of the road can be entered centrally and many participants can benefit from this.

Using the measurement, the location-based error probability of the sensor at the measurement location can be determined and stored in the location-based database. The error probability already stored can be verified. If the determined error probability differs from the stored error probability, the stored error probability can be changed. The stored error probability can be changed incrementally. Changes in the error probability can be mapped quickly and easily. The error probability determined can also be stored on the superordinate data processing system. For example, trained methods of machine learning can be used to determine the error probability.

The error probability can be determined with a time delay using further measurements recorded before and/or after the measurement location. The error probability for a measurement location can be determined after the vehicle has already driven past it. By retrospectively determining the error probability, the error probability for a future passing of the measurement location can be learned. The error probability can be determined, for example, by a retrospective plausibility check of objects. A longer observation period can be considered here than with a time-synchronous determination. For example, an object can be retrospectively recognized as implausible and an increased probability of error be determined at a measuring location where the object suddenly appeared, although it should actually have been detectable for a longer period of time. Likewise, an object can be retrospectively recognized as implausible and the increased error probability for past measurement locations can be determined if the object suddenly disappears although it should still be detectable. Furthermore, an object can be retrospectively recognized as implausible and the increased error probability can be determined if the object is physically impossible to move, ie, for example, moves too quickly, moves with too great an acceleration and/or moves jerkily. Likewise, the object can be retrospectively recognized as implausible and the increased error probability can be determined if the object appears to be moving through one's own vehicle.

The error probability can be determined using at least one further measurement, measured at the measuring location, of at least one further sensor. The additional sensor can have a detection range that at least partially overlaps with the sensor. The additional sensor can be an additional sensor of the vehicle. The sensor can also be an external sensor that at least partially detects the detection area. The additional sensor can have a different measuring principle. The additional sensor can be arranged within an area around the measurement location, that is to say have a different sensor perspective, or be arranged at an adjacent measurement location. A cross-comparison of the measurements can be carried out by taking several measurements from the same or similar measurement locations.

A further error probability of the further sensor at the measurement location can be determined and stored in the database. Due to different error probabilities from different sensors, the measurements of the sensors with low error probabilities can be used.

The error probability can be determined using environmental information at the measurement location. Information about the surroundings can be read from a stored map, for example. The environmental information can also be provided by the vehicle's sensor system. Disturbing objects and/or causes can be mapped in the environmental information. For example, tunnels and crash barriers can be depicted in the environmental information. Furthermore, possible traffic areas can be mapped in the environmental information. Also, a location where a ghost object is imaged may be imaged as impossible in the environmental information. For example, the ghost object can be arranged outside the traffic areas and can thus be recognized as an impossible ghost object. The ghost object can also float in the air and thus be recognized as an impossible ghost object.

Error probabilities determined at different points in time at the measurement location can be aggregated in the database. For example, a counter can be assigned to the measurement location in the database. The counter can be increased for each false detection. Another counter can be increased each time the vehicle drives past the measurement location. A ratio of the two counters directly represents the error probability. The probability of error can be determined more and more precisely through multiple trips. Error probabilities can be determined for many measurement locations along the commuter route, particularly on a commuter route that is driven almost every day.

Certain error probabilities can be aggregated in the database from different vehicles at the measurement location. Different vehicles may have different sensors. However, the operating principles of the sensors can match. Information collected centrally can be used to determine many error probabilities for a measurement location. If there is a change in the error probability of a measurement location, a vehicle driving past the measurement location earlier in time can store the error probability to be expected for a vehicle driving past the measurement location later in time.

The method can be implemented, for example, in software or hardware or in a mixed form of software and hardware, for example in a control unit.

The approach presented here also creates a control device which is designed to carry out, control or implement the steps of a variant of the method presented here in corresponding devices.

The control device can be an electrical device with at least one computing unit for processing signals or data, at least one memory unit for storing signals or data, and at least one interface and/or one communication interface for reading in or outputting data that are embedded in a communication protocol, be. The computing unit can be, for example, a signal processor, a so-called system ASIC, or a microcontroller for processing sensor signals and outputting data signals as a function of the sensor signals be. The storage unit can be, for example, a flash memory, an EPROM or a magnetic storage unit. The interface can be designed as a sensor interface for reading in the sensor signals from a sensor and/or as an actuator interface for outputting the data signals and/or control signals to an actuator. The communication interface can be designed to read in or output the data in a wireless and/or wired manner. The interfaces can also be software modules that are present, for example, on a microcontroller alongside other software modules.

A computer program product or computer program with program code, which can be stored on a machine-readable carrier or storage medium such as a semiconductor memory, a hard disk memory or an optical memory and for carrying out, implementing and/or controlling the steps of the method according to one of the embodiments described above, is also advantageous used, especially when the program product or program is run on a computer or device.

It is noted that some of the possible features and advantages of the invention are described herein with reference to different embodiments. A person skilled in the art recognizes that the features of the control device and of the method can be combined, adapted or exchanged in a suitable manner in order to arrive at further embodiments of the invention.

character list

• 1 zeigt eine Darstellung eines Fahrzeugs mit einem Steuergerät gemäß einem Ausführungsbeispiel; und
• 2 zeigt eine Darstellung einer Datenbank gemäß einem Ausführungsbeispiel.

Embodiments of the invention are described below with reference to the accompanying drawings, neither the drawings nor the description being to be construed as limiting the invention.

• 1 shows an illustration of a vehicle with a control device according to an embodiment; and
• 2 shows a representation of a database according to an embodiment.

The figures are merely schematic and not true to scale. In the figures, the same reference symbols denote the same features or features that have the same effect.

Embodiments of the invention

1 10 shows a depiction of a vehicle 100 on a road. The vehicle drives into a left-hand bend in the road. A crash barrier 102 is arranged on the outside of the left turn. The vehicle 100 includes at least one forward facing sensor 104 . Sensor 104 is a radar sensor here and sends radar waves into a detection range in front of vehicle 100. The radar waves therefore hit the road in front of vehicle 100, the crash barrier 102 and another vehicle 106 driving in front of vehicle 100 on the road. Sensor 104 provides measurement data 108 depicting measurements 110 in the coverage area. A measurement 110 represents radar waves returning to a measurement location 112 of the sensor 104 . Measurement location 112 thus essentially corresponds to a position of vehicle 100 at the time of measurement 110. Measurement 110 contains at least directional information about a direction from which the radar waves arrive at measurement location 112 and distance information about a propagation time of the radar waves between a transmission time and a Time of receipt.

The other vehicle 106 has already made the left turn further than vehicle 100 . The radar waves of the radar sensor are reflected on the other vehicle 106 and thrown back to sensor 104 . The other vehicle 106 is depicted in the measurement data 108 by at least one direct measurement 110 as a moving object at an actual distance from the vehicle 100 and in an actual direction relative to the measurement location 112 or to the vehicle 100 . The direct measurement 110 thus depicts the other vehicle 106 in its actual position on the left in front of the vehicle 100 .

However, the other vehicle 106 is also represented here by at least one incorrect measurement 116 as a moving ghost object 118 at an apparent distance and an apparent direction in relation to the measurement location 112 or the vehicle 100 in the measurement data 108 . In this case, the ghost object 118 is imaged in the measurement data 108 essentially just in front of the vehicle 100 . In the case of incorrect measurement 116, the other vehicle 106 is therefore depicted in the measurement data 108 at a point where it is not actually located.

The incorrect measurement 116 results here from a reflection of the radar waves on the crash barrier 102. The radar waves hit the crash barrier 102 at a flat angle and are deflected in the direction of the other vehicle 106, i.e. they hit the other vehicle 106 from the direction of the crash barrier 102. The The radar waves reflected on the other vehicle 106 likewise impinge on the crash barrier 102 at a flat angle and are deflected in the direction of the sensor 104 by the reflection. These radar waves therefore hit sensor 104 from the direction of crash barrier 102. The moving ghost object 118 is thus mapped behind the crash barrier 102 in the measurement data 108.

The stationary crash barrier 102 is also shown in the measurement data 108 . However, the crash barrier 102 can also have been filtered out of the measurement data 108 as an infrastructure object.

In the approach presented here, an error probability 120 of the sensor 104 at the measurement location 112 is stored in a location-based database 122 . The error probability 120 indicates for this measurement location 112, for example, that moving objects imaged in certain directions are ghost objects 118 with a high probability, while objects imaged in other directions are ghost objects 118 with a low probability. A control unit 124 of the vehicle reads error probability 120 from database 122 and assigns measurements 110 or incorrect measurements 116 to their respective error probability 120 at this measurement location 112 . A low error probability 120 is therefore assigned to the direct measurement 110 of the vehicle 106 . The erroneous measurement 116 of the vehicle 106 over the crash barrier 102 is therefore assigned a high error probability 120 . Due to the different error probabilities, the measurement 110 and the incorrect measurement 116 can be weighted differently in the further processing 126 .

In one exemplary embodiment, a measurement 110 or incorrect measurement 116 is deleted from the sensor data 108 before further processing 126 if the error probability 120 is greater than a threshold value 128 . The deletion of the measurement 110 or faulty measurement 116 can be used in particular when safety-relevant functions are controlled on the basis of the sensor data 108 during the further processing 126 .

In one exemplary embodiment, database 122 is stored on a higher-level data processing system 130 and is read in wirelessly by control unit 124 . The higher-level data processing system 130 can be a cloud server, for example. In this way, many vehicles can access the same database 122 .

In one exemplary embodiment, the measurement data 108 is evaluated in the control unit 124 in order to determine a current error probability 120 of the measurements 110 and erroneous measurements 116 . In this way, changes in the area surrounding the measurement location 112 can be recorded and mapped in the database 122 .

In one embodiment, the current error probability 120 is not determined synchronously. Rather, the error probability is determined with a time delay and retrospectively. A course of the measurement data 108 before and/or after the measurement location 112 can be recorded and viewed later. The measurements 110 can thus be plausibly checked and differentiated from incorrect measurements 116 in an improved manner. In this way, the error probability 120 can be determined with a high level of accuracy and stored in the database 122 for the next trip. Determining error probability 120 can be carried out, for example, when computing capacity is available in control unit 124, ie, for example, no other vehicle 106 is being detected and tracked.

For example, the ghost object 118 suddenly appears in the example shown at the start of the crash barrier 102 and just as suddenly disappears at the end of the crash barrier 102. The sudden appearance and disappearance can be recognized very well from the course of the measurement data 108. Suddenly appearing and/or disappearing objects are very implausible.

Furthermore, the course of movement of the ghost object 118 is far more dynamic due to the reflection on the crash barrier 102 than a course of movement of the vehicle 106 measured directly real vehicle 106 is usually impossible due to its inertia. Highly dynamic objects are also very implausible.

Since it is very unlikely that another vehicle will drive behind the crash barrier 102 , since there is a probability bordering on certainty that there cannot be a vehicle there, the faulty measurement 116 can also be distinguished from the correct measurement 110 . Measurements 110 at unlikely locations are also highly implausible. The improbable positions can be extracted from a map, for example. If the ghost object 118 is traveling at the speed of the other vehicle 106 outside of a road area depicted in the map, this is also very implausible.

In one exemplary embodiment, measurement data from at least one further sensor are evaluated in order to check measurement data 108 from sensor 104 for plausibility. The additional sensor captures the detection range of the sensor 104 at least in part. In particular, measurement data from a sensor with a different measurement principle can be used for plausibility checks. For example, can Images from a camera are evaluated. It can be clearly recognized in the images from the camera that there is no vehicle driving behind the crash barrier 102 .

The additional sensor can be part of a sensor system of vehicle 100 . Likewise, the additional sensor can be an infrastructure sensor. In particular, if the additional sensor is part of the sensor system, an error probability 120 at the measurement location 112 can also be determined for the additional sensor and stored in the database 122 .

In one embodiment, the vehicle 100 drives the same route through the left turn more often. A new error probability 120 is determined each time and combined with the previously determined error probabilities 120 in the database 124 . Slight differences from trip to trip result in slightly different error probabilities 120 each time.

In one exemplary embodiment, a number of ghost objects 116 recognized is added up over several journeys along the route for the measurement location 112 . A ratio of the number of ghost objects 116 to a number of trips results in the error probability 120 for this measurement location 112.

In one exemplary embodiment, the error probabilities 120 determined by different vehicles are combined in the higher-level data processing system 130 . As a result, the control unit 124 can already access error probabilities 120 stored in the database 122 during a first trip along the route and thus secure the measurement data 108 .

2 12 shows a representation of a database 122 according to an embodiment. The database is presented here in the form of a map. A displayed section of the map forms the left turn 1 away. Areas 200 with a high error probability are marked along the left curve. Each area 200 is assigned to a measurement location 112 and is located on the outside of the curve in the direction of the crash barrier, that is to say in the direction from which ghost objects are increasingly detected.

In one exemplary embodiment, each area 200 is assigned the error probability as a numerical value. Measurements from these areas 200 are assigned the stored error probabilities before further processing.

In other words, a detection of incorrect radar measurement data on a repeatedly driven route is presented.

Driver assistance functions and automated driving require detailed information about objects in the environment of the ego vehicle. The objects are e.g. other road users, obstacles or the course of the road.

Environment sensors such as radar, video or lidar scan the environment and provide the necessary measurement data about the objects in the vehicle's surroundings. Using a tracking and fusion system, the measurement data is temporally aggregated or filtered and supplemented with additional attributes, such as speed or acceleration. This creates a consistent environment model on which the driver assistance function or the automated driving function can be based.

Unfortunately, the measurement data from the environmental sensors are often not perfect. On the one hand, they exhibit measurement inaccuracies, which can be partially corrected by temporal filtering. However, objects are also not measured even though they actually exist (false negatives) and measurement data are supplied even though no objects actually exist (false positives).

These sensory deficiencies can ultimately lead to an incorrect system reaction, which in the worst case can result in a traffic accident. An example would be an unjustified emergency braking on an object that does not actually exist, which can lead to a rear-end collision with the following traffic.

The approach presented here addresses the problem of false positive measurement data, especially from radar sensors. In particular, an application in the field of automated driving on a repeatedly driven route (commuter route) is addressed.

Automotive radar sensors that are currently available can have a comparatively high number of false-positive measurement data. This can be in the single-digit percentage range. For example, 1% of all measured values supplied could be incorrect. There are many reasons for these incorrect measurement data.

Modern tracking methods can be used to filter out measurement data that is not temporally correlated by means of a temporal plausibility check.

Methods are also known from the literature which attempt to distinguish and filter out correct measurement data from incorrect measurement data with the aid of classifiers or machine learning.

If there are other environment sensors in the vehicle that have an overlapping field of view, correct and incorrect measurement data can be distinguished by cross-comparison. The prerequisite for this is that different sensors do not deliver the same measurement error at the same time, which is unfortunately not always the case.

Particularly problematic are measurement errors that are correlated in terms of time and are repeated over a number of measurement cycles. An example of this would be an ambiguity in the angle measurement in the radar sensor, which can be caused by the measuring principle.

A real object that is at an angle of 45°, for example, can be incorrectly measured at an angle of 0°. If this error is repeated several times in a row, the error can no longer be detected simply by checking the plausibility over time. The system then assumes an object where none actually exists. These objects are also known as ghost objects and are a well-known problem for a driver assistance function or automated driving.

The method proposed here can be used for a driver assistance function or automatic driving function in which one or more vehicles drive repeatedly on the same route. If there is only one vehicle, the route can be, for example, a repeatedly driven commuter route.

The basic idea is to reduce the technical challenges of automated driving through precise knowledge of the application area and its challenges. The knowledge of the route is used to specifically compensate for the above-mentioned deficits of radar sensors.

Known methods for detecting incorrect measured values in radar sensors are only effective to a limited extent. Incorrect measured values will always remain, which can then potentially lead to undesired system reactions.

Using the approach presented here, when driving through the same route several times, areas can be marked on a map in which incorrect measured values often occur (approaches for recognition see below). This knowledge can be used on the next trip to assume an increased error probability for the measured values that occur there. This can be done, for example, in the form of an additional attribute that is appended to the measured value. This means that incorrect measured values cannot be taken into account in the system, or at least to a lesser extent. In this way, false detections can be eliminated at an early stage and the system behavior can be improved.

With every trip on the commuter route, potentially incorrect measured values, e.g. from a radar sensor, are marked with common methods and the location of their occurrence is marked on a map. An example of an incorrect measured value could be a crash barrier reflection, which is detected by a radar sensor and can hardly be distinguished from a direct reflection.

The drawing shows the ego vehicle, a vehicle in front and a crash barrier at the edge of the road. A radar sensor installed in the ego vehicle detects the vehicle ahead directly (left star) and indirectly via a reflection on the crash barrier (right star). The measured value from the guardrail reflection potentially leads to an unwanted ghost object because the reflection is stable over time and is not filtered out by a classic tracking algorithm. Depending on the curvature of the curve and the speed ratio, the ghost object could become relevant for the driving function.

A common procedure attempts to detect the incorrect measured value. This could be done, for example, by using information from a map that states that there is no drivable area at the location beyond the crash barrier and that therefore no real, moving road users can be there. Likewise, the incorrect measured value can be detected by a cross comparison with other sensors (e.g. video) that may not detect an object there. Furthermore, the measurement data can be classified using a classifier that has been trained using machine learning to distinguish correct from incorrect measurement values. A plausibility check can also detect incorrect measured values if objects are moving in a physically impossible way (too fast, accelerated too quickly, too high a jerk, through the ego vehicle, through another, particularly highly plausible object, collision with another moving the object without any discernible change in movement).

If an incorrect measured value, such as the crash barrier reflection, is detected, the location is marked on a map (circle). The marker can also receive a counter that will be increased with each trip on the commuter route if incorrect readings occur there again.

In addition to the location, other object properties can be used to distinguish between a repeated incorrect measurement and a real object.

For example, a velocity vector, an acceleration vector, and/or an object class can be used.

If the number of trips is also logged, this counter gives an indication of the probability of incorrect readings at this location.

During the next trip, all radar readings are compared with the map and receive the error probability estimated there as an additional attribute.

If the error probability is above a certain threshold, the measured value can be discarded directly.

The probability of error can also be used to weight the measured value less when forming an object (object tracking), in order to give a tracked object a low probability of existence or to prevent a new object track from being set up at this point.

areas are marked in the example map in which faulty radar data was detected during previous journeys.

In the approach proposed here, the occurrence of incorrect object tracks on a repeatedly driven route is evaluated. As a result, fewer and fewer false object tracks with a highly estimated probability of existence can occur over the course of the repeated trips.

The proposed method can be used for driver assistance functions or automated driving if these are based on measurement data from surroundings sensors, in particular radar sensors, and the same route is traveled repeatedly.

Finally, it should be noted that terms such as "comprising," "comprising," etc. do not exclude other elements or steps, and terms such as "a" or "an" do not exclude a plurality. Any reference signs in the claims should not be construed as limiting.

QUOTES INCLUDED IN DESCRIPTION

This list of documents cited by the applicant was generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent Literature Cited

• DE 102017209667 A1 [0006]

Claims (13)

Method for securing measurement data (108) of at least one sensor (104) of a vehicle (100), using a measurement location (112) at which a measurement (110) of the measurement data (108) is measured, a stored location-based error probability (120 ) of the sensor (104) at the measurement location (112) is read from a location-based database (122) and the measurement (110) is processed taking into account the error probability (120). procedure according to claim 1 , in which the measurement (110) is discarded in a data fusion with other measurement data (108) if the error probability (120) is greater than a threshold value (128). Method according to one of the preceding claims, in which the location-based database (122) is read in by a higher-level data processing system (130). Method according to one of the preceding claims, in which the location-related error probability (120) of the sensor (104) at the measurement location (112) is determined using the measurement (110) and is stored in the location-related database (122). procedure according to claim 4 , in which the error probability (120) is determined with a time delay using further measurements (110) recorded before and/or after the measuring location (112). Method according to one of Claims 4 until 5 , in which the error probability (120) is determined using at least one further measurement (110) measured at the measuring location (112) of at least one further sensor (104). procedure according to claim 6 , in which a further error probability (120) of the further sensor (104) at the measurement location (112) is determined and stored in the database (122). Method according to one of Claims 4 until 7 , in which the error probability (120) is determined using environmental information at the measurement location (112). Method according to one of Claims 4 until 8th , in which certain error probabilities (120) are aggregated in the database (122) at different points in time at the measurement location (112). Method according to one of Claims 4 until 9 , in which certain error probabilities (120) from different vehicles (100, 106) at the measurement location (112) are aggregated in the database (122). Control unit (124), wherein the control unit (124() is designed to execute, implement and/or control the method according to one of the preceding claims in corresponding devices. Computer program product that is set up to guide a processor when executing the computer program product to perform the method according to one of Claims 1 until 10 to execute, implement and/or control. Machine-readable storage medium on which the computer program product according to claim 12 is saved.

Images