METHOD FOR CHECKING AT LEAST ONE DRIVING ENVIRONMENT SENSOR OF A VEHICLE

Abstract

A method for checking at least one driving environment sensor of a vehicle is provided. The vehicle is located on a digital map, features of stored stationary objects in an environment of the vehicle, which are expected to be recognized by the driving environment sensor, are identified in the digital map and the environment of the vehicle is sensed using the driving environment sensor. A degradation of the driving environment sensor is deduced if the features to be recognized according to expectations are not recognized by the driving environment sensor or if features actually recognized by the driving environment sensor deviate strongly from the features to be recognized according to expectations.

Claims

1-8. (canceled)

9. A method for checking at least one environment detection sensor on a vehicle, comprising: locating the vehicle on a digital map, identifying features of stored stationary objects in an environment around the vehicle on the digital map, and the environment detection sensor is adapted recognize the objects based on those features, scanning the environment around the vehicle by the environment detection sensor, wherein it is deduced that the environment detection sensor has degraded if the features whose recognition is expected are not recognized by the environment detection sensor or if features actually recognized by the environment detection sensor deviate from the features whose recognition is expected.

10. The method of claim 9, wherein a deviation occurs when the deviation exceeds a preset tolerance range.

11. The method of claim 9, wherein degradation of the environment detection sensor is deduced if an expected distance between the vehicle and one of the objects stored in the map varies from an actual detected distance between the vehicle and the object by more than a preset distance threshold value, wherein the expected distance taken from the map is determined as the feature of the object that is expected to be recognized and the actual distance is determined by the environment detection sensor.

12. The method of claim 9, wherein based on sensor data from the environment detection sensor, the detected objects are classified to determine recognized classification types of the objects as actual recognized features of the objects, wherein the recognized classification types are checked for agreement with expected classification types that are obtained from the map as features expected to be recognized, and wherein degradation of the environment detection sensor is deduced if a determined classification rate is lower than a preset classification threshold.

13. The method of claim 9, wherein degradation of the environment detection sensor is not deduced if it is determined that the features whose recognition is expected are not recognized by the environment detection sensor due to being concealed by at least one dynamic object, or the features actually recognized by the environment detection sensor deviate significantly from the features whose recognition is expected.

14. The method of claim 9, wherein features of stored stationary objects in the environment around the vehicle, whose recognition by the environment detection sensor is expected, are identified in the digital map by sensor-specific detection information stored in the digital map.

15. The method of claim 9, wherein features of stored stationary objects in the environment around the vehicle, whose recognition by the environment detection sensor is expected, are identified in the digital map by calculation of the expected recognizability, based on a three-dimensional model of the digital map.

16. The method of claim 15, wherein the expected recognizability is calculated using a ray tracing method.

Description

[0017] Examples of the invention are explained in more detail below, with reference to figures.

[0018] The figures show:

[0019] FIG. 1 a schematic example of environment detection by means of an environment detection sensor on a vehicle, and

[0020] FIG. 2 another schematic example of environment detection by means of an environment detection sensor on a vehicle.

[0021] The same items are marked with the same references in all figures.

[0022] Based on FIGS. 1 and 2, which show examples of environment detection by means of an environment detection sensor 1 on a vehicle 2, a method for checking the at least one environment detection sensor 1 on the vehicle 2 is described below. In this method, the vehicle 2 is located on a digital map, in particular wherein a position of the vehicle 2 is determined and a corresponding position is determined on the digital map. Features of stationary objects 3 in the environment around the vehicle 2 stored in the digital map are identified on that map, which are expected to be recognized by the environment detection sensor 1 if it is working properly.

[0023] The environment around the vehicle 2 is scanned by the environment detection sensor 1. It is deduced that the environment detection sensor 1 has degraded if the features whose recognition is expected are not recognized by the environment detection sensor 1 or if the features actually recognized by the environment detection sensor 1 deviate significantly from the features whose recognition is expected.

[0024] In other words, if the environment detection sensor 1 recognizes features, the features whose recognition is expected are compared to the features actually recognized by the environment detection sensor 1. If the features whose recognition is expected agree with the features actually recognized by the environment detection sensor 1, or at least if there is not too much deviation between the features whose recognition is expected and the features actually recognized by the environment detection sensor 1, no degradation of the environment detection sensor 1 is deduced. However, if the environment detection sensor 1 does not recognize any features or if the features actually recognized by the environment detection sensor 1 in this comparison do not agree with the features whose recognition is expected, with strong, and in particular too strong, deviation, degradation of the environment detection sensor 1 is deduced.

[0025] Strong deviation, and in particular too strong deviation, occurs in particular when the deviation exceeds a preset tolerance range. In other words, if the features actually recognized by the environment detection sensor 1 do not agree with the features whose recognition is expected, but the deviation between the features actually recognized by the environment detection sensor 1 and the features whose recognition is expected is not too strong, and in particular lies within the preset tolerance range, no degradation of the environment detection sensor 1 is recognized. The threshold value for a tolerance measurement can be based on safety requirements, for example, so that the distance measurement value cannot exceed a preset error, for example, or the classification rate of a sensor in the sensor array cannot lie below a preset threshold.

[0026] The method makes it possible for a self-driving vehicle 2, such as a shuttle or robo-taxi, to determine whether an environment detection sensing system, including the at least one environment detection sensor 1 or multiple similar or different environment detection sensors 1, is experiencing a performance reduction, i.e., a reduction in its efficiency. In the event of a recognized performance reduction, an expanded system reaction by the vehicle 2 can be initiated to prevent hazardous situations.

[0027] The core of the method consists of the fact that, for the sensor technologies available in the vehicle 2, i.e., for the respective environment detection sensor 1, which should be checked using the method, at each point in time it is known which infrastructure objects, i.e., stationary objects 3, for example buildings G and/or vegetation in the form of bushes B, for example, as shown in FIGS. 1 and 2, should be recognized in which locations. If one or more of the expected stationary objects 3, and in particular their features, cannot be determined by the respective environment detection sensor 1, and if it cannot be determined that this was caused by view concealment, in particular due to other dynamic objects, this is an indication of performance degradation, i.e., degradation of the respective environment detection sensor 1.

[0028] Such degradation can be caused, for example, by atmospheric factors, such as fog and/or rain and/or snow, and/or by mechanical factors. FIG. 2 shows an example of an atmospheric effect, i.e., an atmospheric impediment 5 of the environment detection sensor 1.

[0029] The method allows for referential identification of respective causes. Therefore, atmospheric factors systematically predominate in all affected environment detection sensors 1 of the technologies recommended for such atmospheric factors, while mechanical effects on an environment detection sensor 1 or several environment detection sensors 1 remain limited. In other words, a mechanical factor can be ruled out in particular if one or more environment detection sensors 1 on the vehicle 2 are experiencing degradation and one or more other environment detection sensors 1 on the vehicle 2, which are configured in the same way as the environment detection sensor 1 that is experiencing degradation and/or should also have been affected by atmospherically caused degradation, have no degradation.

[0030] Such a determination, whether the respective existing degradation is atmospherically or mechanically caused, can be advantageous because atmospherically caused degradation changes as the atmospheric conditions change, so that once atmospheric conditions improve there is no longer any degradation of the environment detection sensor 1. Mechanical degradation, caused by damage to the environment detection sensor 1 and/or to an area of the vehicle 2 on which it is installed, for example, does not improve by itself but rather requires repair or replacement or adjustment and/or calibration of the environment detection sensor 1.

[0031] In one possible embodiment of the method, features of stored stationary objects 3 in the environment around the vehicle 2, whose recognition by the environment detection sensor 1 is expected, are identified in the digital map by means of sensor-specific detection information stored in the digital map. The stationary objects 3 and their features are therefore sensor-specifically coded into the digital map, so that from the digital map it is possible to read directly which stationary objects 3 and their corresponding features the respective environment detection sensor 1 must recognize.

[0032] In another embodiment of the method, this does not happen, i.e., the digital map is not coded to indicate which environment detection sensor 1 should recognize the stationary object 3 and its features or which of the stationary objects 3 and their features recorded in the digital map should be recognized by the respective environment detection sensor 1 to be checked, but instead an expected visibility of a respective stationary object 3 and its features are actively calculated at each point in time for the respective sensor technology, i.e., for the respective environment detection sensor 1 to be checked, for example by one or more ray tracing methods based on a three-dimensional model of the digital map. Thus, in this embodiment of the method, features of stored stationary objects 3 in the environment around the vehicle 2, whose recognition by the environment detection sensor 1 is expected, are identified in the digital map by means of a calculation of the expected recognizability, in particular through at least one ray tracing method based on a three-dimensional model of the digital map.

[0033] If performance degradation, i.e., degradation of the respective checked environment detection sensor 1, is recognized, the system, i.e., the vehicle 2, and in particular a system for performing automatic operation of the vehicle 2, reacts, advantageously with an adequate process. Therefore, if there is a reduction in a view width, i.e., a detection range width, of the respective checked environment detection sensor 1, advantageously a maximum speed of the vehicle 2 is reduced. Alternatively or additionally, in the event of such an error, i.e., recognized degradation of the respective checked environment detection sensor 1, the vehicle 2 can also be actively shut off, for example. In this case, for example, the vehicle 2 automatically drives to a suitable position, such as a roadside, paved shoulder, emergency pull-off area, or parking lot, and is shut off there. The respective way to proceed, i.e., whether, in what fashion, at what speed, and how far the vehicle 2, in particular automated, can still go, depends especially on the degree of the determined degradation of the respective environment detection sensor 1 and on how many and which environment detection sensors 1 of the vehicle 2 are experiencing such degradation.

[0034] Alternatively or in addition to such an especially automated reaction of the vehicle 2 to the recognized degradation of the respective checked environment detection sensor 1, it is possible, for example, that a remote operator, i.e., a person who is not in the vehicle 2 or in the immediate vicinity of the vehicle 2 but instead has remote access to the vehicle 2, in particular to steering and/or control equipment of the vehicle 2, assesses the current sensor performance, i.e., efficiency, of the respective environment detection sensor 1 for which degradation was determined when it was checked and initiates appropriate additional steps, such as reducing the maximum speed of the vehicle 2, changing the driving route of the vehicle 2, and/or shuts off or initiates the shutoff of the vehicle 2, in particular in an appropriate location for such a shutoff.

[0035] The invented method thereby makes it possible to avoid potentially hazardous situations such as failed or late detection of dynamic objects and stationary objects 3 in the environment of the vehicle 2, due to unrecognized performance reduction in environment detection, i.e., due to unrecognized efficiency reduction in one or more environment detection sensors 1 of the vehicle 2. By means of the method, an error situation caused by degradation of one or more environment detection sensors 1 of the vehicle 2 can be recognized, and supported system reactions can thereby be initiated, such as slower driving to an emergency stopping point.

[0036] The method is described again below, with reference to FIGS. 1 and 2. FIGS. 1 and 2 respectively show a view area, i.e., a detection range 4, of at least one environment detection sensor 1 in the environment detection sensing system of the vehicle 2. The environment detection sensing system includes, for example, one or more video, lidar, and/or radar systems. The environment detection sensor 1 to be checked using the method is therefore, for example, a video, lidar, or radar sensor.

[0037] Information is stored in the digital map as to which stationary objects 3 in the environment of the vehicle 2 can be detected with the respective sensing system, i.e., with the respective environment detection sensor 1 to be checked, looking outward from a specific current position of the vehicle 2. In the example provided, the detection range 4, also known as the view range or field of view, of the environment detection sensor 1, configured here as a video sensor or lidar sensor, for example, is shown. In the case shown in FIG. 1, this environment detection sensor 1 detects, as features, one side of the stationary object 3 designated as Building G and the respective contour of two stationary objects 3 designated as Bushes B, wherein their detection conforms to the corresponding expectations. Therefore, in the example according to FIG. 1, no degradation of the environment detection sensor 1 on the vehicle 2 checked by means of the method is deduced.

[0038] In the example according to FIG. 2, there is degradation of this environment detection sensor 1. Here also, the detection range 4, also known as the view range or field of view, of this environment detection sensor 1 to be checked is shown. This detection range 4 is impaired by an atmospheric disturbance 5, schematically represented as a cloud, i.e., there is atmospherically caused degradation of the environment detection sensor 1.

[0039] In this example, the environment detection sensor 1 does detect, as a feature, the sides of the stationary object 3 designated as Building G, but due to changed reflection caused by the atmospheric disturbance 5 and/or due to changed reflection conditions caused by the atmospheric disturbance 5, the stationary objects 3 stored in the digital map and designated as Bushes B, in particular their contours as their features, can no longer be perceived by the environment detection sensor 1 at a preset, specifically ideal, distance.

[0040] This can be proven In particular through comparison with the expected efficiency using historic and/or digital maps. In other words, the environment detection sensor 1 is expected to recognize the features of the stationary objects 3 stored in the digital map, and therefore the stationary objects 3, from the current position of the vehicle 2. However, this does not occur for the stationary objects 3 designated as Bushes B, so degradation of the environment detection sensor 1 is deduced.

LIST OF REFERENCE INDICATORS

[0041] 1 Environment detection sensor [0042] 2 Vehicle [0043] 3 Stationary object [0044] 4 Detection range [0045] 5 Atmospheric disturbance [0046] B Bush [0047] G Building