METHOD FOR DETERMINING A SENSOR SETTING

Description

[0001] The invention relates to a method and a system for determining a setting of a sensor.

[0002] Determining a suitable sensor setting for a sensor, such as a laser scanner or the like, is a complex process and is typically performed by a corresponding expert. Based on their knowledge and experience, the expert manually selects, for example, an optimal mounting position and / or orientation as well as an optimal sensor configuration of the sensor-specific sensor parameters. For example, the sensor setting of a real sensor is iteratively adjusted, and the respective sensor data is evaluated by the expert until a desired sensor setting is achieved that meets the expert's subjective requirements.

[0003] However, the existing approach to determining a suitable sensor setting has the disadvantage that manual iteration to determine the sensor setting is very time-consuming, especially on a real sensor. Furthermore, the sensor setting process requires an expert to be present on-site during sensor assembly and configuration. Furthermore, dynamic adjustment of the sensor setting during ongoing sensor operation, for example, adaptation to a changing sensor environment, is not possible.

[0004] The document US 11 663 377 B2 discloses a method for determining a setting of a sensor according to the preamble of claim 1 and a system for determining a setting of a sensor according to the preamble of claim 14. The documents US 2023 / 169684 A1 and US 9 740 944 B2 disclose similar methods and systems.

[0005] An object underlying the invention is therefore to provide a method and a system for determining a setting of a sensor which eliminates the above problems.

[0006] This problem is solved by the subject matter of the independent claims.

[0007] The invention is described in the appended set of claims.

[0008] The invention relates to a method for determining a setting of a sensor, the method being carried out by means of a computing device and comprising: (1) a simulation model is defined which comprises a virtual test environment and at least one virtual sensor; (2) in the virtual test environment the virtual sensor is initialized with a sensor setting, wherein the sensor setting comprises a sensor position, a sensor orientation and / or at least one sensor parameter; (3) virtual sensor data generated by the virtual sensor which are associated with the sensor setting are evaluated based on at least one evaluation metric, wherein a detection range of the virtual sensor is included in the respective evaluation metric, wherein the evaluation of the sensor data comprises: (4) determining whether the evaluation metric satisfies at least one predetermined condition; (5) in the event of a negative determination, the sensor setting of the virtual sensor is adjusted and the sensor data associated with the adjusted sensor setting are evaluated according to (3);(6) if the determination is positive, the sensor setting is output as the preferred sensor setting by the computing device; and (7) a real sensor is configured in a real application based on the preferred sensor setting.

[0009] A sensor is therefore simulated in a virtual, in particular two- or three-dimensional, world in order to generate virtual sensor data that arise from the virtual sensor capturing the virtual world. According to the invention, a sensor setting of the virtual sensor is adjusted manually or automatically in the virtual test environment as often as desired until a preferred sensor setting is found, which is then used to configure a real sensor in a real application. The criterion for deciding whether a sensor setting represents a preferred sensor setting is defined by an evaluation metric, whereby the evaluation metric must fulfill a predetermined condition for a sensor setting to be output as a preferred sensor setting.If the evaluation metric does not fulfill the specified condition, the sensor setting of the virtual sensor is adjusted, in particular using an optimization procedure, and the associated sensor data is evaluated again.

[0010] With the present method, a preferred sensor setting of a virtual sensor is determined in a virtual test environment, and the determined preferred sensor setting is then used to configure a real sensor in a real application. The determination of the preferred sensor setting can thus be carried out purely virtually, without the need for a real sensor to generate real sensor data. In particular, all steps for determining the preferred sensor setting can be carried out purely virtually. For example, the simulation model can be implemented on a computing device that performs all the calculations required to simulate the virtual test environment and the virtual sensor.

[0011] The virtual test environment can, for example, correspond to a 1:1 replica of a real environment, e.g., a real environment in which the real sensor is to be used. Thus, the virtual sensor data generated by the virtual sensor can also at least substantially correspond to that of a real sensor. Because the virtual test environment and the real environment, as well as the generated virtual sensor data and the real sensor data, correspond to one another or are at least comparable, the sensor setting used for the virtual sensor can also be used, at least substantially unchanged, as the sensor setting of the real sensor.Known differences between the virtual test environment and the real environment and / or the virtual sensor data and the real sensor data can be taken into account when configuring the real sensor, for example by presetting individual parameters or by adjusting them based on a preset adaptation rule.

[0012] As already described above, the sensor setting comprises a sensor position, a sensor orientation, and / or at least one sensor parameter. In particular, the sensor setting comprises intrinsic sensor data, such as a scan pattern of the sensor, a measurement frequency, and properties of the measurement signal, e.g., the shape, frequency swing, and / or the slope of the frequency modulation of the transmitted signal (e.g., in the case of FMCW). The preferred sensor setting depends in particular on the specific application. Accordingly, different value ranges can be specified for the respective sensor parameters, which cannot or must not be exceeded depending on the respective sensor specifications.

[0013] The virtual sensor data generated by the virtual sensor includes, in particular, spatial data of the virtual test environment, spatial data of the virtual objects, and / or position and / or orientation data of the virtual sensor. The virtual sensor may, for example, include a simulated LiDAR sensor, FMCW-based sensor, radar sensor, a simulated camera, and / or other suitable sensors.

[0014] As already described above, the evaluation metric takes into account a detection range of the virtual sensor, in particular a spatial one. This means that the evaluation metric, on the basis of which a decision is made as to whether a sensor setting is selected as the preferred sensor setting, is based on the detection range of the virtual sensor. For example, the evaluation metric can indicate the extent to which a specified area, such as a specified room or a specified area, in the virtual test environment is covered by the detection range of the virtual sensor. In particular, it can be determined whether the evaluation metric fulfills at least one specified condition, e.g. whether the detection range of the virtual sensor covers at least 70% of the specified area.In such a case, the sensor setting of the virtual sensor can be adjusted until the above condition is met. Only when the above condition is met, i.e., when the detection range of the virtual sensor covers at least 70% of the specified range, is the corresponding sensor setting output as the preferred sensor setting and can be used to configure the real sensor in the real application. The evaluation metric can vary depending on the application and may consider additional or different metrics than those mentioned above.

[0015] Once a preferred sensor setting has been determined, this and / or at least part of the virtual sensor data can be saved and exported in a format readable by a real sensor in order to configure the real sensor with the preferred sensor setting. The real sensor can, for example, comprise a LiDAR sensor, an FMCW-based sensor, a radar sensor, a camera, and / or other suitable sensors.

[0016] Further embodiments of the invention can be found in the description, the subclaims and the drawings.

[0017] According to a first embodiment, the evaluation of the sensor data takes place automatically. For example, the method is implemented using an algorithm, wherein the above method steps are automatically executed by the algorithm. In particular, the evaluation metric can be designed such that a computer can check whether the evaluation metric fulfills the specified condition. The evaluation metric can, for example, be designed such that it can be represented by a numerical value. As described above, the numerical value can, for example, indicate what percentage of a specified spatial area is covered by the detection range of the virtual sensor.

[0018] The advantage of an automatic implementation of the method is that a preferred sensor setting can be determined quickly and without the presence of a user. Rather, the user knowledge can, for example, be implemented in the algorithm and in particular in the configuration of the evaluation metric. For example, based on their knowledge, a user can set the evaluation metric such that the execution of the algorithm leads to a preferred or optimal sensor setting, which would be achieved by manual determination of the sensor setting by a user. The evaluation metric can, for example, comprise an optimization function whose arguments comprise sensor setting parameters and which is designed such that an extreme value, i.e. a minimum or maximum, of the optimization function corresponds to a preferred sensor setting or preferred sensor setting parameters.By determining the extreme value of the optimization function, the preferred sensor setting can be determined. The extreme value of the optimization function can also be determined using known optimization algorithms, such as the gradient method, the interval halving method, and the like.

[0019] Alternatively, the sensor data can also be evaluated manually, for example by a user evaluating the evaluation metric or checking whether the evaluation metric fulfills the specified condition. To do this, the user can use the virtual sensor data and / or a visualization of the virtual sensor data, for example. For example, the user can iteratively evaluate the virtual sensor data and / or the visualization of the virtual sensor data until they define a sensor setting as the preferred sensor setting based on the evaluation metric. The advantage of manually evaluating the sensor data is that special cases, such as exceptional cases that an algorithm cannot handle optimally, i.e. which are unsuitable for automatic evaluation, can be treated individually.

[0020] It is also conceivable that the evaluation of the sensor data takes place automatically in a first step and that in a second step the sensor setting found in the first step is manually adjusted in order to obtain an improved sensor setting or to check a sensor setting found in the first step.

[0021] According to a further embodiment, the evaluation metric is based on a comparison between a detection range of the virtual sensor and a predetermined range in the virtual test environment. For example, the detection range of the virtual sensor is compared with a predetermined range of the virtual test environment to determine which portion of the range is detected or covered by the detection range of the virtual sensor. As already described above, the evaluation metric can indicate the proportion of coverage of the predetermined range by the detection range of the virtual sensor. The detection range of the virtual sensor can be determined, for example, by the scan pattern of the virtual sensor and / or by the measurement distance range, i.e., the range, of the virtual sensor.

[0022] According to a further embodiment, a predefined measurement distance range of the virtual sensor is included in the evaluation metric, wherein the predefined measurement distance range defines a range within which the virtual sensor generates plausible measurement values. Plausible means, for example, that the virtual sensor does not generate any error values ​​or any values ​​that lie outside a meaningful value range for the respective application. The virtual sensor can, for example, be designed for a specific measurement distance range.

[0023] The measuring distance range of the virtual sensor can also correspond to a measuring distance range of the associated real sensor. Since the evaluation metric also takes into account the specified measuring distance range of the virtual sensor, the preferred sensor setting is thus also determined depending on the specified measuring distance range of the sensor. For example, a sensor can have different sensor parameter configurations through which different measuring distance ranges can be set. Different preferred sensor settings can thus be determined and saved for the different measuring distance ranges of the virtual sensor. For example, in such a case, several preferred sensor settings, each assigned to an associated measuring distance range, can be output and saved.The real sensor can then be configured depending on the desired measuring distance range based on the multiple preferred sensor settings. If the requirements for the measuring distance range, i.e., the desired measuring distance range, change during operation of the real sensor, the real sensor can be adjusted, especially automatically, based on the multiple preferred sensor settings. The sensor settings of the real sensor are thus adaptively adjusted to the desired measuring distance range.

[0024] The evaluation metric comprises an evaluation function value, wherein the optimization function value is calculated using an optimization function and based on the virtual sensor data, wherein the predetermined condition comprises that the optimization function value corresponds to a substantially minimum value or a substantially maximum value of the optimization function. The optimization function can, for example, be designed such that by solving the corresponding optimization problem, i.e., by determining the corresponding minimum function value or the maximum function value of the optimization function, a sensor setting is determined that satisfies the predefined requirements. As previously described, one requirement can, for example, be that a predetermined portion of a spatial region of the virtual test environment is covered by the detection range of the virtual sensor, in particular for a predetermined measurement distance range.The optimization function can further be configured such that additional or different requirements are taken into account when determining the sensor setting. The optimization function value is, in particular, a numerical value, which is calculated, for example, as follows: . Optimierungsfunktionswert = f Sensoreinstellungsparameter , virtuelle Sensordaten where f represents the mathematical function that maps the sensor setting parameters and the virtual sensor data to the optimization function value. If the evaluation metric includes an optimization function value, the specified condition can be met if the difference between the

[0025] optimization function value and the optimization function value of a previous iteration is less than a specified threshold. Alternatively, as is known for optimization problems, a derivative of the optimization function can be formed, whereby the specified condition can be met if the value of the derivative of the optimization function is less than a specified threshold. In In both of the above cases, fulfilling the condition may mean that the minimum function value or the maximum function value of the optimization function has been found, i.e. that the optimization problem has been solved.

[0026] For example, the optimization function f can correspond to a function that specifies the area or the proportion of the area of ​​the specified region that is not covered by the detection range of the virtual sensor or, if applicable, by multiple virtual sensors. The minimum of this optimization function thus corresponds to the sensor setting parameters for which the optimization function value, i.e., the proportion of the undetectable area of ​​the specified region, is minimal. Using an optimization method, such as the gradient method, the minimum of the optimization function and thus the preferred sensor setting can be determined iteratively. In particular, sensor settings that approach the preferred sensor setting can be successively determined.

[0027] According to a further embodiment, the sensor data comprise information, in particular data, on landmarks, wherein the landmarks are assigned to at least one fixed location in the virtual test environment, wherein the sensor settings of the real sensor in the real application are configured based on the landmarks and the sensor data of the real sensor. The landmarks correspond, for example, to reference points in the virtual test environment that can also be found in a corresponding manner in the real environment. The landmarks can, for example, be assigned to fixed coordinates in the virtual test environment. Based on the landmarks, a spatial relationship, in particular a unique one, can be established between the virtual sensor and the landmark or the coordinates of the landmark.Depending on the sensor setting, in particular depending on the position and / or orientation of the virtual sensor, the spatial relationship between the virtual sensor and the orientation point can change, so that a sensor setting, in particular the position and / or orientation of the virtual sensor, can be assigned to a unique spatial relationship. Accordingly, the coordinate system of the real sensor, i.e., the position and / or orientation of the real sensor, can be adjusted based on the orientation points and the sensor data of the real sensor in order to obtain the preferred sensor setting corresponding to the virtual sensor.

[0028] According to a further embodiment, the position and / or orientation of the real sensor is determined based on the orientation points. For example, the sensor data of the real sensor, e.g., the acquired environmental data, is analyzed or processed using the orientation points. The analysis of the sensor data of the real sensor can, for example, comprise detecting and / or classifying the spatial points of the acquired environment corresponding to the orientation points in order to determine a spatial relationship between the real sensor and the spatial points of the real environment assigned to the orientation points. Thus, a current sensor setting of the real sensor, in particular the position and / or orientation of the real sensor in relation to the spatial points assigned to the orientation points, can be determined. Subsequently, the sensor setting of the real sensor, ieThe position and / or orientation of the real sensor can be adjusted such that the spatial relationship between the real sensor and the spatial data of the real environment corresponding to the landmarks corresponds to the spatial relationship between the virtual sensor and the respective landmarks. The position and / or orientation of the real sensor can be determined, for example, based on at least one landmark. The more landmarks used to determine the position and / or orientation of the real sensor, the greater the accuracy with which the positioning and / or orientation of the real sensor can be adjusted to the positioning and / or orientation of the virtual sensor.

[0029] In other words, the coordinate system of the real sensor is adapted to the coordinate system of the virtual sensor. The spatial points in the real environment corresponding to the landmarks can be detected using classification algorithms, e.g., machine learning. The landmarks can include, for example, objects in the virtual test environment, prominent points in the virtual test environment, and / or any other coordinate points in the virtual test environment. The landmarks can also be determined automatically by an algorithm and / or manually by a user.

[0030] According to a further embodiment, the real sensor is connected to a sensor configuration application, wherein the sensor device of the real sensor is configured based on the preferred sensor setting by means of the sensor configuration application. The sensor configuration application can thus represent an interface via which the real sensor can be configured by the user. The sensor configuration application can, for example, comprise an app on a mobile phone, a laptop, a PC, and / or another suitable computing device. The sensor configuration application can further be connected to the real sensor wirelessly or via a cable in order to exchange data, in particular in real time.A user of the sensor configuration application can receive instructions from the sensor configuration application, for example via a display of the sensor configuration application, to change a position and / or orientation of the sensor.

[0031] For example, the user of the sensor configuration application can receive instructions until the preferred sensor setting is achieved. In particular, the instructions can be provided to the user in the form of a guide.

[0032] Furthermore, the sensor data of the real sensor can be visualized using the sensor configuration application, e.g., on the aforementioned display, so that the user of the sensor configuration application gets a sense of the effect a change in the sensor settings has on the generated sensor data. This can provide an improved user experience. For example, the detection range of the real sensor is visualized. Alternatively, the sensor configuration application can automatically configure the sensor settings of the real sensor, so that the user does not have to perform any manual steps. In particular, the sensor configuration application can be operated independently of the simulation model.

[0033] According to a further embodiment, a final sensor setting of the real sensor is determined manually or automatically based on the preferred sensor setting and the information on the landmarks. For example, after an automatic configuration of the sensor setting has taken place, the user of the sensor configuration application still has the option of manually fine-tuning the sensor setting, i.e., the preferred sensor setting is adjusted by the user. Alternatively, a final fine-tuning of the sensor setting can be performed automatically by the sensor configuration application. For example, the sensor configuration application uses the aforementioned landmarks to adjust the preferred sensor setting and obtain a final sensor setting.The landmarks can also be used by the user to determine the final sensor setting during manual fine-tuning. For example, the landmarks can be graphically visualized in the sensor configuration application to facilitate fine-tuning of the physical sensor and improve the user experience.

[0034] According to a further embodiment, suggestions for changing the sensor setting are presented to a user of the sensor configuration application based on the preferred sensor setting and the information on the landmarks. As already described above, the landmarks can be used to adjust the position and / or orientation of the real sensor. For example, as described above, the sensor configuration application can determine a current sensor setting of the real sensor based on the landmarks and compare it with the preferred sensor setting. Based on the deviation of the current sensor setting of the real sensor from the preferred sensor setting, the sensor configuration application can then present the user with suggestions for changing the sensor setting.For example, the user has the option of accepting the suggestions for changing the sensor setting by means of an acceptance command, so that the sensor configuration application implements the change to the sensor setting in response to the acceptance command in order to obtain the final sensor setting. Alternatively, the user can also implement the suggestions for changing the sensor setting manually. For example, the user can make a corresponding adjustment to the sensor setting via a suitable user interface of the sensor configuration application. Via the user interface, the user can, for example, make changes to the sensor setting, i.e. the sensor position, sensor orientation and / or the sensor parameters. In particular, the user can manipulate or change the values ​​for the aforementioned parameters via the user interface.

[0035] According to a further embodiment, further virtual sensors are provided in the virtual test environment, wherein the evaluation metric

[0036] Detection ranges of the additional virtual sensors are introduced. The detection ranges of the additional virtual sensors are thus taken into account in the evaluation metric. In other words, the preferred sensor setting is determined depending on the additional virtual sensors or the detection ranges of the additional virtual sensors. The preferred sensor setting is therefore not static, but can change dynamically depending on the detection ranges of the additional virtual sensors. This dependency can be implemented in the evaluation metric. For example, the evaluation metric can represent a measure of the extent to which a given area or room is detected with a given number of sensors. In particular, the evaluation metric can indicate whether the given area is completely covered by the detection ranges of the virtual sensors, especially in combination.Even in the case of a multi-sensor arrangement, a preferred sensor setting can be determined.

[0037] According to a further embodiment, the virtual test environment, the virtual sensor and / or the further virtual sensors are visualized on a display, wherein the detection range of the virtual sensor and / or the detection ranges of the further virtual sensors and / or an overlap of the detection range of the virtual sensor and / or the detection ranges of the further virtual sensors and / or non-detected areas are visually highlighted on the display, in particular in color. For example, the respective detection areas can be displayed in different colors. Overlaps of the respective detection areas and / or non-detected areas can also be displayed in a uniform, unique color so that the user can recognize and identify these areas directly on the display.By visualizing the respective areas on the display, the user can adjust the sensor settings of the virtual sensor such that the individual detection ranges of the different virtual sensors, i.e. the virtual sensor and the other virtual sensors, are coordinated with one another. For example, the user can adjust the sensor settings of the virtual sensor such that a predetermined area is covered by the detection ranges of all existing virtual sensors, in particular completely or at least to the greatest possible extent. By visualizing the respective detection ranges on the display, it is easier for the user to determine the preferred sensor setting, which in particular fulfills the above condition. In this respect, the display can be designed as part of the sensor configuration application.

[0038] According to a further embodiment, a plurality of, in particular different, sensor settings of the virtual sensor are output as preferred sensor settings, wherein a respective preferred sensor setting is assigned to an associated application area and / or an associated weather condition. The requirements for a sensor can, for example, differ for different application areas and / or weather conditions. For example, a steeper modulation ramp for modulating the laser beam may be necessary for a short-range application than for a long-range application. In particular, the desired sensor setting can vary depending on the location, purpose, specified measuring distance range, weather, for example visibility conditions, and / or other factors. Accordingly, different optimal or preferred sensor settings can be provided for the different circumstances.

[0039] For example, in an aircraft docking application where an aircraft is to be brought into a predefined position after landing, different preferred sensor settings can be provided depending on the distance of the aircraft from the sensor or the aircraft stop position. Using the simulation model, the different preferred sensor settings can be determined and output as described above. The real sensor can then switch between the respective preferred sensor settings during operation, depending on the distance of the aircraft from the sensor. To detect aircraft at a great distance, for example, a sensor setting may be required in which the field of view (FoV) is reduced, so that the measurement time per scan point and / or the number of scan points is optimized to provide a higher detection probability and measurement performance.Such a setting can be particularly advantageous in situations where environmental influences are present. For example, if the aircraft is closer to the sensor, a larger FoV may be required to observe and capture the aircraft's surroundings, thus simplifying and accelerating aircraft handling and the correct positioning of the aircraft at a designated position. A lower detection probability and expected poorer measurement performance can be accepted in this case.

[0040] As described in a previous embodiment, additional virtual sensors may also be provided. In such a case, the preferred sensor setting may depend on the detection ranges of the virtual sensors used and / or on the number of virtual sensors. Accordingly, in a real application, the sensor setting of the real sensor can be flexibly adapted to changing circumstances, i.e., application areas and / or weather conditions.

[0041] According to a further embodiment, the sensor setting of the real sensor is automatically adjusted using the plurality of preferred sensor settings and depending on the application area and / or the weather conditions. For example, the sensor setting of the real sensor is automatically adjusted based on the acquired sensor data if the sensor data indicate that a different preferred sensor setting is provided for the current circumstances. For example, the sensor data can include different parameters and associated parameter values. The sensor data can include, for example, a measured distance to an object, a signal strength of the received, reflected signal, and / or other suitable parameters. A value range for a parameter and / or a combination of value ranges of the different parameters can, for example, be assigned to a respective preferred sensor setting.If the sensor data indicates that one or more parameters are within predefined value ranges, a corresponding preferred sensor setting can be automatically selected for the actual sensor. For example, a measured distance to an object can be assigned to a predefined measurement distance range, which is associated with one of the plurality of preferred sensor settings.

[0042] A further aspect of the invention relates to a system for determining a setting of a sensor, comprising: a real sensor and a simulation device which is designed to: (1) to define a simulation model comprising a virtual test environment and at least one virtual sensor; (2) to initialize the virtual sensor in the virtual test environment with a sensor setting, wherein the sensor setting comprises a sensor position, a sensor orientation and / or at least one sensor parameter; (3) to evaluate virtual sensor data generated by the virtual sensor and associated with the sensor setting based on at least one evaluation metric, wherein a detection range of the virtual sensor is included in the respective evaluation metric, wherein the evaluation of the sensor data comprises: (4) determining whether the evaluation metric fulfills at least one predetermined condition; (5) in the event of a negative determination, the sensor setting of the virtual sensor is adjusted and the virtual sensor data associated with the adjusted sensor setting is evaluated according to (3);(6) upon a positive determination, the sensor setting is output as the preferred sensor setting; and (7) transmit the preferred sensor setting to the real sensor, wherein the system is configured to configure the real sensor based on the preferred sensor setting.

[0043] The statements regarding the method according to the invention apply accordingly to the system. This applies in particular with regard to advantages and embodiments.

[0044] It should be noted that any combination of the above and below embodiments is possible unless explicitly excluded.

[0045] The invention is described below purely by way of example with reference to the drawings. They show: Fig. 1 schematically shows a system for determining a setting of a sensor; Figs. 2A and 2B show a two-dimensional representation of a virtual sensor in a virtual test environment with a first and a second sensor setting; Fig. 3A show a two-dimensional representation of a virtual sensor with a first sensor setting and a further virtual sensor; Fig. 3B show a two-dimensional representation of the virtual sensor with a second sensor setting and a further virtual sensor; and Figs. 4A, 4B and 4C show visualizations of the virtual sensor data and the virtual test environment on a display of the simulation device.

[0046] Fig. 1 shows a system 10 for determining a sensor setting, which comprises a simulation device 12, for example a computing device, and a real sensor 14. The simulation device 12 is designed to define a simulation model that comprises a virtual test environment 18 and at least one virtual sensor 16. The virtual sensor 16 is initialized in the virtual test environment 18 with a sensor setting, wherein the sensor setting comprises a sensor position, a sensor orientation and / or at least one sensor parameter. Virtual sensor data generated by the virtual sensor 16 and associated with the sensor setting are evaluated based on at least one evaluation metric, wherein a detection range of the virtual sensor 16 is included in the evaluation metric.The evaluation of the sensor data comprises determining whether the evaluation metric fulfills at least one predetermined condition; if the determination is negative, the sensor setting of the virtual sensor 16 is adjusted, for example using an optimization method, and the virtual sensor data associated with the adjusted sensor setting is evaluated; and if the determination is positive, the sensor setting is output as the preferred sensor setting by the simulation device 12 and transmitted to the real sensor 14. The system 10 is designed to configure the real sensor 14 based on the preferred sensor setting, i.e., to overwrite an existing sensor setting of the real sensor 14 with the preferred sensor setting. This changes, for example, the position, orientation, and / or at least one sensor parameter of the real sensor 14.

[0047] In Fig. 2A und 2B an adjustment of the sensor setting of the virtual sensor 16 in the virtual test environment 18 is visualized in 2D. Fig. 2A shows the virtual sensor 16, which has a first sensor setting, i.e. a first sensor position, a first sensor orientation and a first FoV 20, i.e. field of view, of the virtual sensor 16. The virtual sensor 16 is initialized, for example, in the virtual test environment 18 with the first sensor setting. In principle, the sensor setting can include additional parameters than those mentioned here. Furthermore, a predetermined area 22 can be seen, which is partially detected or covered by the detection range of the virtual sensor 16. The predetermined area 22 represents, for example, an area to be detected, which is to be covered by the detection range of the virtual sensor 16.The virtual sensor data generated by the virtual sensor 16, which were generated based on the first sensor setting, can then be evaluated to determine whether the specified area 22 is covered by the detection range of the virtual sensor 16. The evaluation of the sensor data takes place automatically by using an evaluation metric which, for example, indicates what percentage of the specified area 22 is not covered by the detection range of the virtual sensor 16. The sensor setting of the virtual sensor 16 is thus adjusted until 100% of the specified area 22 is covered by the detection range of the virtual sensor 16. Alternatively, the evaluation of the virtual sensor data can also be carried out manually, for example by using one of the . Fig. 2A similar representation on a display for a user, so that the user can adjust the sensor setting in such a way that the specified area 22 is completely covered by the detection range of the virtual sensor 16. A corresponding illustration is shown in the Fig. 4A bis 4C shown.

[0048] Fig. 2B shows the virtual sensor 16 with a second, adjusted sensor setting, ie with a second sensor position, a second sensor orientation and a second FoV 24. The first sensor setting was adjusted in such a way that the specified condition is met. As can be seen from Fig. 2B As can be seen, the orientation of the virtual sensor 16 and the FoV have changed. However, a change in the position of the virtual sensor 16 cannot be detected, since a change in position is not necessary in order to cover the predetermined area 22 with the detection range of the virtual sensor 16. The adjusted, second sensor setting can be determined iteratively, in particular, by continuously changing the sensor setting starting from the initial, first sensor setting until a sensor setting is found in which the predetermined area 22 is completely covered by the detection range of the virtual sensor 16.

[0049] The determined second sensor setting is then transmitted to the real sensor 14 as a preferred sensor setting and used to configure the real sensor 14, ie to initialize or adapt the sensor setting of the real sensor 14 with the preferred sensor setting.

[0050] Fig. 3A und 3B show one of the Fig. 2A und 2B similar representation. The Fig. 3A und 3B However, they differ from Fig. 2A und 2B in that an additional virtual sensor 26 is provided in the virtual test environment 18. The specified area 22 is thus not only detected or covered by one virtual sensor, but by two virtual sensors. In such a case, the sensor setting of the virtual sensor 16 can be adjusted such that the specified area 22 is covered by the, in particular combined, detection areas of the virtual sensor 16 and the additional virtual sensor 26.

[0051] As in Fig. 3A As shown, the detection range of the virtual sensor 16, which was initialized with the first sensor setting, and the detection range of the further virtual sensor 26 do not completely cover the specified area 22. Rather, there are two undetectable partial areas 28 of the specified area 22 that are not covered by the combined detection ranges of the virtual sensor 16 and the further virtual sensor 26. As already Fig. 2A und 2B As described above, the virtual sensor data from the virtual sensor 16, which are assigned to the first sensor setting, are evaluated to determine what percentage of the specified area 22 is not detected or whether undetected subareas 28 exist. In contrast to Fig. 2 When evaluating the virtual sensor data of the virtual sensor 16, the virtual sensor data of the additional virtual sensor 26 are also taken into account. Thus, it can be determined whether the combined detection ranges of the virtual sensor 16 and the additional virtual sensor 26 cover the specified range 22. As in Fig. 3A This is not the case. Consequently, the sensor setting of the virtual sensor 16 is adjusted so that a second sensor setting is found. As shown in Fig. 3B As shown, the second sensor setting includes a second FoV 24. A change in the position and orientation of the virtual sensor 16 is not necessary in this case. As shown in Fig. 3B As shown, the predetermined area 22 with the second sensor setting is completely covered by the combined detection ranges of the virtual sensor 16 and the further virtual sensor 26. It should be noted that the predetermined area 22 is not covered by the detection range of the virtual sensor 16 on its own, but only in combination with the detection range of the further virtual sensor 26. The determined second sensor setting can then be used to configure the real sensor 14 in a real application in which, for example, another real sensor is arranged in a position corresponding to the further virtual sensor.

[0052] Fig. 4A, 4B und 4C show one of the Fig. 2A similar representation of the virtual sensor 16 and the virtual test environment 18 on a display 30 of the simulation device 12. In particular, the overlaps of the detection range of the virtual sensor 16 and the predetermined range 22 are shown hatched on the display 30. Additionally or alternatively, the overlaps can be highlighted in color or by means of other suitable techniques. On the display 30 of the simulation device 12, selection options 32 are displayed to the user to adjust the sensor setting. The selection options 32 include the options for changing an orientation, a position, and a FoV, as in Fig. 4A By selecting a corresponding option, the user is taken to a view similar to that shown in Fig. 4B und 4C shown to adjust a respective sensor setting parameter. The display 30 is a touch display, allowing the user to select the appropriate option directly via the display.

[0053] In Fig. 4B The representation shown on the display 30 after selecting the "Orientation" option is shown. In this setting option, the user can adjust the orientation of the virtual sensor 16 to the left or right by pressing the respective button. In principle, the orientation can be adjusted in any conceivable direction. For example, in a three-dimensional representation of the sensor data, the orientation can be changed upwards, downwards, or in any other desired direction. As shown in Fig. 4B As can be seen, the orientation of the virtual sensor 16 has been adjusted to the left by the user, so that a larger portion of the specified area 22 is covered by the detection range of the virtual sensor 16. Despite the adjustment of the orientation, however, undetectable portions 28 of the specified area 22 still exist.

[0054] In order to also cover the undetectable sub-areas 28 of the specified area 22 with the detection range of the virtual sensor 16, the user can adjust the FoV in a final step. The FoV of the virtual sensor represents the sensor's field of view, i.e., the detection angle, which is determined, for example, by a scan pattern of a laser scanner. For example, the user can select the "FoV" selection option, which allows them to enlarge or reduce the FoV. Fig. 4C The display 30 is shown after such an adjustment. As shown in Fig. 4C As can be seen, the user has enlarged the FoV such that the specified area 22 is completely covered by the detection range of the virtual sensor 16. The FoV determined in this way can thus be defined and saved as a preferred sensor setting parameter. In the real application, the real sensor can then be configured based on this determined FoV by adjusting the scan range of the real sensor accordingly.

[0055] The step-by-step adjustment of the sensor setting described here can also be performed automatically by an algorithm. In such a case, the step-by-step adjustment of the sensor setting can also be displayed to the user on the display 30 of the simulation device 12 to visually illustrate the adjustment process for the user. Bezugszeichenliste

[0056] 10System 12Simulation setup 14Real sensor 16Virtual sensor 18Virtual test environment 20First FoV 22Specified area 24Second FoV 26Further virtual sensor 28Uncovered areas 30Display 32Selection options

Claims

1. A method for determining a setting of a sensor, wherein the method is performed by means of a computing device and comprises that: (1) a simulation model is defined that comprises a virtual test environment (18) and at least one virtual sensor (16); (2) in the virtual test environment (18), the virtual sensor (16) is initialized with a sensor setting, wherein the sensor setting comprises a sensor position, a sensor orientation and / or at least one sensor parameter; (3) virtual sensor data generated by the virtual sensor (16) and associated with the sensor setting are evaluated based on at least one evaluation metric, wherein a detection zone of the virtual sensor (16) is included the evaluation metric, wherein the evaluation of the sensor data comprises that: (4) it is determined whether the evaluation metric fulfills at least one predefined condition; (5) in the event of a negative determination, the sensor setting of the virtual sensor (16) is adjusted and the virtual sensor data associated with the adjusted sensor setting are evaluated in accordance with (3); (6) in the event of a positive determination, the sensor setting is output by the computing device as the preferred sensor setting; and (7) a real sensor (14) is configured in a real application based on the preferred sensor setting, characterized in that the evaluation metric comprises an optimization function value, with the optimization function value being calculated by means of an optimization function and based on the virtual sensor data, with the predefined condition comprising that the optimization function value corresponds to a substantially minimum value or a substantially maximum value of the optimization function.

2. A method according to claim 1, wherein the evaluation of the sensor data takes place automatically.

3. A method according to one of the preceding claims, wherein the evaluation metric is based on a comparison between a detection zone of the virtual sensor (16) and a predefined zone (22) in the virtual test environment (18).

4. A method according to any one of the preceding claims, wherein a predefined measurement distance range of the virtual sensor (16) is included in the evaluation metric, wherein the predefined measurement distance range defines a range in which the virtual sensor generates plausible measurement values.

5. A method according to any one of the preceding claims, wherein the sensor data comprise information on orientation points, wherein the orientation points are associated with at least one fixed location point in the virtual test environment, wherein a configuration of the sensor setting of the real sensor (14) in the real application takes place based on the orientation points and the sensor data of the real sensor (14).

6. A method according to any one of the preceding claims, wherein the position and / or orientation of the real sensor (14) is / are determined based on the orientation points.

7. A method according to any one of the preceding claims, wherein the real sensor (14) is connected to a sensor configuration application, wherein the sensor setting of the real sensor (14) is configured based on the preferred sensor setting by means of the sensor configuration application.

8. A method according to any one of the preceding claims, wherein a final sensor setting of the real sensor (14) is manually or automatically determined based on the preferred sensor setting and the information on the orientation points.

9. A method according to any one of the preceding claims, wherein suggestions for changing the sensor setting are made to a user of the sensor configuration application based on the preferred sensor setting and the information on the orientation points.

10. A method according to any one of the preceding claims, wherein further virtual sensors (26) are provided in the virtual test environment (18), wherein a detection zone of the further virtual sensors (26) is introduced into the evaluation metric.

11. A method according to any one of the preceding claims, wherein the virtual test environment (18), the virtual sensor (16) and / or the further virtual sensors (26) are visualized on a display (30), wherein, on the display (30), the detection zone of the virtual sensor (16) and / or the detection zones of the further virtual sensors (26) and / or an overlap of the detection zone of the virtual sensor (16) and / or of the detection zones of the further virtual sensors (26), and / or undetected zones (28) are visually highlighted, in particular in color, on the display.

12. A method according to any one of the preceding claims, wherein a plurality of, in particular different, sensor settings of the virtual sensor (16) are output as preferred sensor settings, wherein a respective preferred sensor setting is assigned to an associated application area and / or an associated weather condition.

13. A method according to claim 12, wherein the sensor setting of the real sensor (14) is automatically adjusted using the plurality of preferred sensor settings and depending on the application area and / or weather condition.

14. A system (10) for determining a setting of a sensor, comprising: a real sensor (14) and a simulation device (12) which is configured: (1) to define a simulation model that comprises a virtual test environment (18) and at least one virtual sensor (16); (2) in the virtual test environment (18), to initialize the virtual sensor (16) with a sensor setting, wherein the sensor setting comprises a sensor position, a sensor orientation and / or at least one sensor parameter; (3) to evaluate virtual sensor data, which are generated by the virtual sensor (16) and associated with the sensor setting, based on at least one evaluation metric, wherein a detection zone of the virtual sensor (16) is included the evaluation metric, wherein the evaluation of the sensor data comprises that: (4) it is determined whether the evaluation metric fulfills at least one predefined condition; (5) in the event of a negative determination, the sensor setting of the virtual sensor (16) is adjusted and the virtual sensor data associated with the adjusted sensor setting are evaluated in accordance with (3); (6) in the event of a positive determination, the sensor setting is output as the preferred sensor setting; and (7) to transmit the preferred sensor setting to the real sensor (14), wherein the system (10) is configured to configure the real sensor (14) based on the preferred sensor setting, characterized in that the evaluation metric comprises an optimization function value, with the optimization function value being calculated by means of an optimization function and based on the virtual sensor data, with the predefined condition comprising that the optimization function value corresponds to a substantially minimum value or a substantially maximum value of the optimization function.