Computer-implemented method for providing a control signal for a vehicle

Abstract

A computer-implemented method (100) is described for providing a control signal for a vehicle (10) having an environmental sensor (14) for providing a sensor signal indicating data about an environment of the vehicle, and a position sensor (15) for providing a position signal indicating a current position of the vehicle, comprising the following steps: receiving (102) a sensor signal from the environmental sensor and a position signal from the position sensor and determining three-dimensional map data (32) around the current position of the vehicle; creating (104) a three-dimensional virtual image (38) in real time comprising a virtual surface section (42) of an environment around the current position of the vehicle based on the three-dimensional map data and the data about the environment of the vehicle;Deriving (106) an effective field of view (60) for the one environmental sensor in real time using a virtual reference obstacle (46) arranged at a virtual position (48) on the virtual surface section (42) and a virtual ray (50-58) emanating from the position of the vehicle in the virtual image and bounded by the surface section; and providing (108) a control signal for the vehicle based on the effective field of view. The method provides an improvement in the performance of the driving function in real vehicles.

Description

[0001] The invention relates to a computer-implemented method for providing a control signal for a vehicle.

[0002] Semi-autonomous and autonomous vehicles rely on various sensors for environmental detection, such as camera sensors, radar sensors, lidar sensors, ultrasonic sensors, etc., and for navigation. The actual sensor performance depends heavily on the environment in which the vehicle is located. Influences from the road layout, e.g., inclines, curves, etc.; obscuration by objects, e.g., buildings along the road, other road users, or vegetation; and weather conditions, e.g., rain, snow, hail, etc., can impair the sensor field of view. Accordingly, the driving function is also affected in terms of its ability to assess situations in a timely manner and react accordingly.

[0003] DE 11 2019 001 605 T5 discloses a method for virtual testing of a vehicle, in particular for training an AI. For this purpose, a simulated environment is created based on real data, in which a simulated vehicle sensor is observed. For example, virtual beams from a virtual LIDAR sensor are directed at an obstacle in the simulated environment. The obstacle is detected depending on the number of beams that hit the obstacle.

[0004] Furthermore, DE 10 2019 125 075 A1 discloses a method for the computer-implemented simulation of a LIDAR sensor in a virtual environment, comprising the steps of: reading in a VR data set representative of the virtual environment, reading in a parameter data set for parameterizing a LIDAR sensor simulation module, generating a laser beam data set representative of a laser beam emitted by the LIDAR sensor using the LIDAR sensor simulation module and the parameter data set, determining a range data set representative of a range of the laser beam by evaluating the VR data set and the laser beam data set by ray tracing, and generating image data representative of the virtual environment by combining the VR data set with the laser beam data set and the range data set.

[0005] From DE 10 2017 213 214 A1 a method for modeling a motor vehicle sensor in a virtual test environment by defining and using a sensor support, a raycast distribution shape, a group of raycast properties, a raycast reflection factor and a raycast echo is also known.

[0006] The object of the invention is to provide a method that provides an improvement in the performance of the driving function in real vehicles.

[0007] The problem is solved by the features of the independent claims. Advantageous further developments are the subject of the dependent claims and the following description.

[0008] According to a first aspect, a computer-implemented method for providing a control signal for a vehicle, in particular an autonomous vehicle, is described, wherein the vehicle has at least one environmental sensor for providing a sensor signal indicating data about the vehicle's surroundings, and at least one position sensor for providing a position signal indicating a current position of the vehicle. The method comprises at least the following steps: receiving at least one sensor signal from the environmental sensor and at least one position signal from the position sensor and determining three-dimensional map data around the current position of the vehicle;Creating at least one three-dimensional virtual image in real time, comprising at least one virtual surface section of an environment around the current position of the vehicle based on the three-dimensional map data and the data about the environment of the vehicle; Deriving at least one effective field of view for the at least one environmental sensor in real time using at least one virtual reference obstacle arranged at at least one virtual position on the virtual surface section and at least one virtual ray emanating from the position of the vehicle in the virtual image and bounded at least by the surface section; and Providing a control signal for the vehicle based on the effective field of view.

[0009] The computer-implemented method is used to first determine an effective sensor range for the vehicle's environmental sensors in real time. This can then be used to provide control signals for the vehicle that are tailored to the effective sensor range. For this purpose, sensor signals from at least one environmental sensor of the vehicle are used in combination with three-dimensional map data of the vehicle's surroundings. The three-dimensional map data of the vehicle's surroundings can be determined using the vehicle's position signal. The sensor signals and the three-dimensional map data are combined in real time to create a three-dimensional virtual image of the environment. The three-dimensional virtual image has at least one virtual surface section. The at least one virtual surface section can, for example, represent the topography of the environment and / or an object in the vehicle's surroundings.Preferably, all surface sections that can be determined from the three-dimensional map data and the sensor signals are provided in the virtual image. The position of the vehicle in the virtual image is known. Starting from the position of the vehicle in the virtual image, a virtual beam can be generated. The virtual beam indicates a line of sight within the virtual image that is clear for the environmental sensor. This means that when the virtual beam hits the surface section, the virtual beam ends at the surface section. Furthermore, a virtual reference obstacle can be arranged on the virtual surface section. The virtual beam can also end at the virtual reference obstacle when it hits the virtual reference obstacle. In this case, the virtual beam can be shorter than when it hits a virtual surface section.If the virtual beam does not encounter a surface section or a virtual reference obstacle, the virtual beam can end after a distance that corresponds to the nominal sensor range of the environmental sensor. The virtual beam, in combination with the reference obstacle, can be used to determine an effective field of view for the at least one environmental sensor in real time. A separate effective field of view can be determined for each environmental sensor present in the vehicle. The effective field of view is used to provide a control signal for the vehicle. The control signal can, for example, indicate a reduction in speed if the effective field of view is smaller than the nominal field of view of the environmental sensor. This can ensure that the vehicle can perform a braking maneuver despite a possible obstacle that may be outside the effective field of view.This can improve the performance of the driving function in a real vehicle while driving.

[0010] According to some embodiments, it is conceivable that the step of deriving an effective field of view may comprise at least the following sub-steps: positioning the at least one virtual reference obstacle at at least one virtual position on the virtual surface section and checking whether the at least one virtual ray intersects the at least one virtual reference obstacle.

[0011] The virtual reference obstacle can then be positioned at a virtual position in the virtual image such that it is located on the virtual surface section. If the virtual beam intersects the virtual reference obstacle, this means that the virtual beam terminates there. From this information, the effective field of view of the environmental sensor can be derived in this embodiment.

[0012] Alternatively or additionally, the virtual beam may be designed as a scanning beam which, starting from the position of the vehicle in the virtual image, scans the surroundings at several positions, preferably by a continuous movement of the virtual beam, in order to find and detect the reference obstacle.

[0013] According to some embodiments, it is conceivable that the step of positioning the at least one virtual reference obstacle can be carried out for different virtual positions, which can preferably be arranged in a grid around the position of the vehicle in the virtual image.

[0014] The virtual surface section can thus be provided with positions in a grid-like manner, at each of which a reference obstacle is placed one after the other. After each placement of a reference obstacle, a check is carried out to determine whether at least one virtual beam hits the respective reference obstacle. Thus, by selecting the appropriate grid, a suitable resolution of the sensor range for generating the control command can be determined.

[0015] According to some embodiments, it is conceivable that in the step of deriving an effective field of view for the environmental sensor, a plurality of virtual rays can be generated which can extend in different virtual directions in the virtual image.

[0016] The plurality of virtual rays can, for example, extend within the nominal field of view of the relevant environmental sensor in the virtual image. Each of the virtual rays has a different angle from the position of the vehicle in the virtual image than the other virtual rays. In this way, a large area can be scanned with virtual rays. In some embodiments, a raycasting method can be used for this purpose.

[0017] According to some embodiments, it is conceivable that in the step of deriving an effective field of view for the environmental sensor, a number of virtual rays can be counted that intersect the virtual reference obstacle and from the number a detection value for the virtual position of the reference obstacle can be determined to derive the effective visibility.

[0018] The more virtual rays intersect the virtual reference obstacle, the more likely the reference obstacle will be detected as an obstacle by the corresponding environmental sensor. For example, the detection value can increase the more virtual rays intersect the virtual reference obstacle. This can improve the derivation of the effective field of view of the environmental sensor.

[0019] According to some embodiments, it is conceivable that in the step of deriving an effective field of view for the environmental sensor, virtual rays can be generated such that their arrangement can substantially correspond to a resolution of the environmental sensor.

[0020] This allows the environment sensor to be simulated as realistically as possible using the virtual rays in the virtual image. This prevents the detection of a reference obstacle due to an excessive number of virtual rays emanating from the vehicle's position in the virtual image, which would otherwise not occur in reality. The accuracy of the derived effective field of view can thus be increased.

[0021] According to some embodiments, it is conceivable that the control signal may indicate a command to adjust a speed of the vehicle.

[0022] If the effective visibility of all environmental sensors is reduced compared to the nominal visibility, for example, reducing the vehicle's speed can prevent an obstacle that may be located outside the effective visibility of the environmental sensors from causing a collision with the vehicle or causing the vehicle to perform a sudden evasive maneuver. This can improve the performance of the vehicle's driving function.

[0023] According to a second aspect, a computer program product is described, comprising instructions which, when the program is executed by a computer, cause the computer to carry out the steps of the method according to the preceding description.

[0024] Advantages and effects as well as further developments of the computer program product arise from the advantages and effects as well as further developments of the method described above. In this regard, reference is therefore made to the preceding description. A computer program product can be understood, for example, as a data carrier on which a computer program element is stored that has instructions executable by a computer. Alternatively or additionally, a computer program product can also be understood, for example, as a permanent or volatile data storage device, such as flash memory or RAM, that has the computer program element. However, this does not exclude other types of data storage devices that have the computer program element.

[0025] According to a third aspect, a vehicle is described, comprising at least one control unit for the vehicle, at least one environmental sensor and at least one position sensor, wherein the control unit is connected to the environmental sensor and the position sensor via at least one signal connection and is designed to carry out the method according to the description.

[0026] Advantages and effects, as well as further developments of the vehicle, arise from the advantages and effects, as well as further developments of the method described above. To avoid repetition, reference is made to the previous description in this regard.

[0027] According to some embodiments, it is conceivable that the control unit can be designed to control the vehicle at least partially autonomously.

[0028] When the control unit performs the procedure described above, sudden evasive maneuvers or collisions with obstacles beyond the effective line of sight can be further avoided, thus improving the performance of the vehicle's driving function.

[0029] The invention is described below using an exemplary embodiment with the aid of the accompanying drawings. They show: Fig. 1 a flowchart of the process; Fig. 2 a schematic representation of the vehicle; Fig. 3 a schematic representation of an environment of the vehicle; Fig. 4 a schematic representation of the creation of the virtual image of the environment; and Fig. 5 a schematic representation of the derivation of the effective visibility.

[0030] In Fig. 1, the computer-implemented method in its entirety is designated by the reference numeral 100.

[0031] The computer-implemented method 100 is configured to provide a control signal for a vehicle 10.

[0032] The vehicle 10 has, according to Fig. 2 includes at least one environmental sensor 14 and a position sensor 15. The environmental sensor 14 provides a sensor signal that indicates data about the surroundings of the vehicle 10. The environmental sensor can be embodied, for example, as a camera sensor, a radar sensor, a LiDAR sensor, or an ultrasonic sensor. Other sensor types are also conceivable.

[0033] The position sensor 15 provides a position signal indicating the current position of the vehicle 10. The position sensor 15 can be connected, for example, to a global satellite navigation network via a wireless signal connection.

[0034] The vehicle 10 can be controlled semi-autonomously or autonomously. For this purpose, the vehicle 10 can have a control unit 12, which can be connected to the environmental sensor 14 and the position sensor 15 via at least one signal connection 18. The signal connection 18 can be wireless or wired.

[0035] The method 100 can be performed in real time, i.e., when the vehicle 10 is driving or is being controlled semi-autonomously or autonomously. The control unit 12 can perform the method 100.

[0036] Fig. 3 shows, by way of example, a hill 24 over which the vehicle 10 is traveling. Furthermore, an obstacle 22 is shown, which is arranged on the side of the hill 24 not visible to the vehicle 10. At least one environmental sensor 14 of the vehicle 10 has a field of view 26. The field of view 26 has a nominal sensor range 28. However, the hill 24 prevents the obstacle 22 from being detected, which, in terms of distance along the surface of the hill 24, would be within the nominal sensor range 28 on a flat surface. Due to the hill 24, the field of view 26 therefore has an effective sensor range 30. The effective sensor range 30 is smaller than the nominal sensor range 28.

[0037] As a consequence, vehicle 10 would have to slow down in order to be able to decelerate in time before approaching obstacle 22 or to avoid having to perform a sudden evasive maneuver. However, this cannot be done because the effective sensor range 30 is not yet known.

[0038] According to the method 100, in a first step 102, at least one sensor signal can be received from the at least one environmental sensor 14. This is exemplified in Fig. 4 shown.

[0039] Furthermore, at least one position signal can be received from the at least one position sensor 15 in order to determine the current position of the vehicle.

[0040] Based on the current position of the vehicle 10, three-dimensional map data 32 can be determined, which is also Fig. 4. The three-dimensional map data 32 can be transmitted to the vehicle 10, for example, via a wireless signal connection. Alternatively or additionally, the three-dimensional map data 32 can be located in a memory (not shown) of the vehicle 10 and then retrieved based on the current position.

[0041] In Fig. 4, the three-dimensional map data 32 shows a roadway 34 and contour lines 36 representing various elevations of the terrain in the vicinity of the roadway 34. This provides a topography of the surroundings of the vehicle 10.

[0042] By combining the three-dimensional map data 32 and the sensor data of the at least one environmental sensor 14, a virtual image 38 of the surroundings of the vehicle 10 is created according to step 104. The virtual image 38 can, for example, represent the roadway and a position 40 of the vehicle 10. Furthermore, the topographical data can, for example, have surface sections 42, which in this exemplary embodiment are represented as hills next to the roadway. An obstacle 22 displayed in the sensor data can also be displayed in the virtual Fig. be displayed.

[0043] The entire topography of the environment in combination with all obstacles in the virtual image 38 can form a single surface section 42.

[0044] In a further step 106, at least one effective field of view 60 for the at least one environmental sensor 14 can be derived.

[0045] For this purpose, according to the optional sub-step 110 of step 106, a virtual reference obstacle 46 can be arranged on the surface section 42, as for example in Fig. 5. The arrangement of the reference obstacle 46 can be done at a virtual position 48.

[0046] Starting from the position 40 of the vehicle in the virtual image 38, a plurality of virtual beams 50-58 extending in different directions can be generated. The virtual beams 50-58 can be arranged such that they substantially correspond to a resolution of the environmental sensor 14. This means that if the environmental sensor 14 has a resolution of, for example, 1 mm at a distance of, for example, 5 m, the virtual beams can be arranged such that they also have a resolution of 1 mm in the virtual image at the distance of 5 m. The distance and resolution mentioned here are merely examples and are not intended to limit the scope of protection. Any desired resolutions can be used for the environmental sensors 14 and, accordingly, for the virtual beams 50-58.

[0047] When a virtual beam 50-58 encounters the surface portion 42, the virtual beam 50-58 is limited by the surface portion 42. This also occurs when the virtual beam 50-58 encounters an obstacle 22.

[0048] According to a further optional sub-step 112, it is then checked whether the at least one virtual ray 50-58 intersects the at least one virtual reference obstacle 46. This means whether the virtual ray 50-58 has an intersection point with the reference obstacle 46, so that the virtual ray 50-58 hits the virtual reference obstacle 46.

[0049] When the virtual beam 50-58 hits the virtual reference obstacle 46, the effective visibility for the effective field of view 60 can be derived for this position.

[0050] Furthermore, the surface section 42 or the virtual image 38 can have a grid of a plurality of virtual positions 48. In some embodiments, the reference obstacle 46 can be arranged sequentially at all virtual positions 48. Then, for each virtual position 48, it can be checked whether and which of the virtual rays 50-58 intersect the virtual reference obstacle 46.

[0051] Furthermore, it can optionally be provided that the number of virtual rays 50-58 that intersect the virtual reference obstacle 46 is detected. The more virtual rays 50-58 intersect the virtual reference obstacle 46, the higher the detection value that can be used to determine the effective field of view 60. For example, it can be assumed that only when the detection value exceeds a certain threshold value would an environmental sensor in the real environment detect the virtual reference obstacle as an obstacle.

[0052] In Fig. In Figure 5, the effective field of view 60 is shown two-dimensionally for clarity. However, the effective field of view 60 can also be three-dimensional. This means that for different elevation angles starting from the vehicle 10, corresponding limits for the effective field of view 60 can also be determined.

[0053] In a further step 108, a control signal for the vehicle 10 can be provided based on the effective field of view 60 determined in the previous steps. The control signal can, for example, indicate a control command for adjusting the speed of the vehicle 10. For example, a speed adjustment of the vehicle 10 can be triggered based on a smaller effective field of view 60. If the effective field of view increases again, a further speed adjustment of the vehicle 10 can be triggered with another control signal.

[0054] The example described above does not limit the invention in any way. Rather, the invention can be modified in many ways. All of the features of the invention described above can be essential to the invention alone or in combination with one another.

Claims

[1] Computer-implemented method (100) for providing a control signal for a, in particular autonomous, vehicle (10), which has at least one environmental sensor (14) for providing a sensor signal indicating data about an environment of the vehicle (10), and at least one position sensor (15) for providing a position signal indicating a current position of the vehicle (10), comprising at least the following steps: a. Receiving (102) at least one sensor signal from the environmental sensor (14) and at least one position signal from the position sensor (15) and determining three-dimensional map data (32) around the current position of the vehicle (10); b. Creating (104) at least one three-dimensional virtual image (38) in real time comprising at least one virtual surface section (42) of an environment around the current position of the vehicle (10) based on the three-dimensional map data (32) and the data about the environment of the vehicle (10); c. Deriving (106) at least one effective field of view (60) for the at least one environmental sensor (14) in real time by means of at least one virtual reference obstacle (46) arranged at at least one virtual position (48) on the virtual surface section (42) and at least one virtual ray (50-58) emanating from the position of the vehicle (10) in the virtual image (38) and limited at least by the surface section (42); and d. Providing (108) a control signal for the vehicle (10) based on the effective field of view (60). [2] Computer-implemented method (100) according to claim 1, characterized by that the step: deriving (106) an effective field of view (60) comprises at least the following sub-steps: a. Positioning (110) the at least one virtual reference obstacle (46) at at least one virtual position (48) on the virtual surface section (42); and b. Checking (112) whether the at least one virtual ray (50-58) intersects the at least one virtual reference obstacle (46). [3] Computer-implemented method (100) according to claim 2, characterized by that the step: positioning (110) of the at least one virtual reference obstacle (46) is carried out for different virtual positions (48), which are preferably arranged in a grid around the position of the vehicle (10) in the virtual image (38). [4] Computer-implemented method (100) according to one of claims 1 to 3, characterized byin that in the step of deriving (106) an effective field of view (60) for the environmental sensor (14), a plurality of virtual rays (50-58) are generated which extend in different virtual directions in the virtual image (38). [5] Computer-implemented method (100) according to claim 4, characterized by in that in the step: deriving (106) an effective field of view (60) for the environmental sensor (14), a number of virtual rays (50-58) which intersect the virtual reference obstacle (46) is counted and from the number a detection value for the virtual position of the reference obstacle (46) is determined for deriving the effective field of view (60). [6] Computer-implemented method (100) according to one of claims 1 to 5, characterized bythat in the step: deriving (106) an effective field of view (60) for the environmental sensor (14), virtual rays (50-58) are generated such that their arrangement substantially corresponds to a resolution of the environmental sensor (14). [7] Computer-implemented method (100) according to one of claims 1 to 6, characterized by that the control signal indicates a command to adjust a speed of the vehicle (10). [8] A computer program product comprising instructions which, when executed by a computer, cause the computer to carry out the steps of the method (100) according to any one of claims 1 to 7. [9] Vehicle (10) comprising at least one control unit for the vehicle (10), at least one environmental sensor (14) and at least one position sensor (15), wherein the control unit (12) is connected to the environmental sensor (14) and the position sensor (15) via at least one signal connection (18) and is designed to carry out the method (100) according to one of claims 1 to 7. [10] Vehicle (10) according to claim 9, characterized by that the control unit (12) is designed for at least partially autonomous control of the vehicle (10).

Images