Method for operating a driver assistance system, driver assistance system and motor vehicle with a driver assistance system
The driver assistance system reduces incorrect steering interventions by using sensor data and historical vehicle behavior to validate lane boundaries, ensuring accurate lane recognition and reducing unwanted steering corrections.
Patent Information
- Application Number
- DE102021208245
- Authority / Receiving Office
- DE · DE
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2021-07-29
- Publication Date
- 2025-03-13
- Estimated Expiration
- 2041-07-29
AI Technical Summary
Existing driver assistance systems often incorrectly interpret road features, leading to unwanted steering interventions due to misidentification of lane boundaries.
A procedure for operating a driver assistance system that recognizes possible lane track limits using vehicle sensors, performs a plausibility check based on historical data from external vehicles, and only initiates steering interventions when the vehicle is about to cross the lane limit.
This approach reduces the frequency of incorrect steering interventions by leveraging historical data to validate lane boundaries, providing a more robust interpretation of lane limits than vehicle sensors alone.
Smart Images

Figure 00000008_0000 
Figure 00000008_0001 
Figure 00000009_0000
Abstract
Description
[0001] The present invention relates to a method for operating a driver assistance system for a motor vehicle, in particular a method for operating a lane keeping assistant for the motor vehicle. The invention also relates to a driver assistance system and a motor vehicle with such a driver assistance system.
[0002] Driver assistance systems that support lane keeping, or so-called lane keeping assistants, are well known. Typically, such a driver assistance system detects a lane marking using a camera or other optical sensor, for example, and prevents the driver from crossing the lane markings by corrective steering interventions. It is known that the camera of such a driver assistance system can interpret a structure, such as a tar joint or an old and incompletely removed lane marking, as the currently valid lane marking of the lane in which a vehicle equipped with the driver assistance system is traveling. If the driver assistance system uses such an incorrectly interpreted lane marking as its guide, this can lead to unintentional steering intervention by the driver assistance system.
[0003] Various methods are known for improving a driver assistance system of the type described above. For example, DE 10 2016 201 249 A1 describes a method for determining a road model. If any provided first sensor data is consistent with the provided second sensor data, a model of a road is calculated using the first sensor data. If no first sensor data was provided or this data is inconsistent with the second sensor data, the model is calculated using the second sensor data.
[0004] From DE 10 2010 053 964 A1, a driver assistance system for supporting a driver in the lateral guidance of a vehicle and a corresponding method are known.
[0005] A start time and / or end time and / or a strength of the support action is controlled depending on local lane information.
[0006] DE 10 2017 219 297 A1 describes a method for operating a vehicle. In this method, a future path for the vehicle is determined based on sensor data. The vehicle is provided with a behavior model that is representative for a statistical evaluation of the path selection by a number of other vehicles. A plausibility check of the path is performed based on the behavior model.
[0007] DE 10 2009 007 342 A1 describes a method for determining a valid lane marking using a vehicle's camera system. A first and a second lane marking are detected, and the positions of a preceding vehicle relative to them are determined. Depending on the determined positions, one of the detected lane markings is selected as the valid lane marking.
[0008] DE 10 2016 216 335 A1 deals with the analysis of driving trajectories for a route section. Data is received that represents a first trajectory of a vehicle controlled by a human driver, environmental conditions, and vehicle-related characteristics. Several such first trajectories are clustered based on the associated received data. A request to transmit a second trajectory is then received, and the requested second trajectory is transmitted.
[0009] Furthermore, DE 10 2017 212 908 A1, for example, describes a method for improving the lateral and / or longitudinal guidance of a vehicle. This method detects when a steering intervention performed by the driver assistance system is corrected by a driver. The driver assistance system is trained based on the detected corrections.
[0010] DE 10 2019 112 413 A1 describes a method for estimating the course of a road in the environment of a vehicle. Input variables for estimating the course of the road are environmental measurement data. This environmental measurement data can be collected, for example, by a vehicle camera. Using known environmental data from a map, errors in the environmental measurement data can be detected. For example, it can be detected if lane markings that are not stored in the map have been changed due to construction work. This information can then be used to decide whether autonomous driving should be deactivated in the area of these faulty lane markings.
[0011] The object of the invention is to reduce faulty steering interventions of a driver assistance system.
[0012] This object is achieved by the subject matter of the independent patent claims. Further possible embodiments of the invention are disclosed in the subclaims, the description, and the figures.
[0013] The invention provides a method for operating a driver assistance system for a motor vehicle. The driver assistance system is a lane keeping assistance system or lane assist. The method according to the invention provides that along a route to be traveled, a possible lane marking is detected and its course is recorded by at least one on-board sensor of the motor vehicle. Such a sensor can be implemented, for example, as a camera sensor. In other words, with the aid of the sensor, the motor vehicle can observe its surroundings or a route ahead in such a way that the possible lane marking can be both detected and its course recorded. In addition to an actual lane marking, a tar seam or an old, invalid construction site marking can also be considered as a possible lane marking.
[0014] The procedure provides for a plausibility check of the possible lane markings, and for a steering intervention to be performed or suppressed depending on the result of the plausibility check. In other words, the driver assistance system does not simply react to the detected and recorded possible lane markings, but rather first subjects them to a plausibility check. The steering intervention is not exclusively dependent on the result of the plausibility check. Therefore, a plausibility-checked lane marking does not automatically lead to the steering intervention being performed. Rather, the steering intervention occurs when the host vehicle threatens to cross the plausibility-checked lane markings.
[0015] The plausibility check is based on historical driving history data from a large number of other vehicles, with the historical driving history data describing the respective driving behavior of each other vehicle within the range of the possible lane markings. In other words, the driver assistance system can access data provided by any number of other vehicles. This can be achieved, for example, by the driver assistance system establishing a communication connection to an external server device in order to receive the historical driving history data from the other vehicles. Alternatively or additionally, it is possible for the other vehicles to exchange the data in question directly with the motor vehicle via car-to-car communication. The data from the other vehicles can also be provided to the driver assistance system as swarm data or REM data (REM - Road Experience Management).
[0016] Historical journey data refers to data that describes the past driving behavior of other vehicles. Such historical journey data can also be transmitted by a third-party vehicle currently driving directly in front of the owner vehicle, whereby the historical journey data can include information about the last distance traveled by the third-party vehicle.
[0017] By utilizing the historical driving history data of the multitude of other vehicles or by using the swarm data, the driver assistance system can be supported in its interpretation of the detected possible lane markings without having to perform a computationally intensive training process, as is known from the state of the art. The described method also addresses situations in which all of the vehicle's sensors exhibit weaknesses and in which even redundant sensors cannot solve the problem. The multitude of historical driving history data from other vehicles, which describe their behavior in the area of a questionable lane marking and which can be used to verify the plausibility of the lane marking in the ego vehicle, can provide a more robust interpretation of the possible or questionable lane markings than sensors in the vehicle without knowledge of other vehicles and / or drivers.
[0018] A variant of the invention provides that the historical driving history data describe historical driving trajectories of the other vehicles in the area of the possible lane boundary, and that the plausibility check is based not only on historical driving history data, but also on current driving history data of at least one other vehicle currently driving ahead, which describes a current driving trajectory of the other vehicle currently driving ahead in the area of the possible lane boundary. Steering intervention is suppressed if both the historical driving trajectories and the current driving trajectory cross the course of the possible lane boundary. In other words, the historical driving history data can describe the route or lane along which a respective other vehicle traveled in the area of the possible lane boundary.The individual driving trajectories of individual other vehicles can also be combined and calculated to form an averaged swarm trajectory. If it turns out that all or most of the historical driving trajectories do not treat the possible lane marking as such, or even cross it, the driver assistance system can reject or suppress the steering intervention. In other words, the driver assistance system can use the historical driving trajectories in the area of the possible lane marking to determine whether the possible lane marking is actually a lane marking. If the majority of the historical driving trajectories cross or at least touch the possible lane marking, this suggests that the possible lane marking is not actually a lane marking.In this case, the steering intervention, which would have been based on the possible lane boundary, can be suppressed.
[0019] The current journey data of at least one other vehicle currently driving ahead, which is used as a basis for the plausibility check in addition to the historical journey data, can be determined, for example, from live observation of the other vehicle currently driving ahead using the vehicle's own camera. The current journey data describe the driving behavior of the other vehicle currently driving ahead in the area of the possible lane markings. This is the current driving trajectory of at least one other vehicle currently driving ahead in the area of the possible lane markings. Live monitoring of a other vehicle currently driving ahead offers the advantage that a detected lane marking can be further verified for plausibility.
[0020] Additionally or alternatively, the invention provides that the historical journey data describe a plurality of steering interventions by the other vehicles. Each steering intervention is assigned a respective geographical location and a respective course of a lane marking on the basis of which the respective steering intervention was carried out. In the event that the historical journey data show that a frequency of steering interventions exceeding a predetermined level is registered at the respective geographical location and on the basis of the respective course of the lane marking, the lane marking located at this location is recognized as faulty. A lane marking recognized as faulty is marked, for example, on a digital road map, and can therefore receive a flag. The other vehicles or vehicles in the swarm can preferably report each steering intervention by their respective lane keeping assistant to an external server device.This message can contain the location of the steering intervention and the lane boundary path that was intervened in. If steering interventions on this lane boundary path occur frequently at a particular geographical location, this indicates that the lane boundary was incorrectly detected. If a predetermined percentage, for example 40%, of the vehicles in the swarm or of other vehicles that were driving with the lane keeping assistant ready to intervene received a steering intervention at a particular location and in response to a particular lane boundary path, this location and / or path is marked as a potentially incorrectly detected lane boundary. Freely applicable limit values can be provided in a backend or the external server device for the predetermined percentage. These limit values can specify the frequency at which an incorrect steering intervention is defined, for example per 100 journeys.Optionally, several such limit values can be sent and / or generated, which indicate, for example, how likely an incorrect steering intervention is.
[0021] The invention also includes embodiments which provide additional advantages.
[0022] A further embodiment provides that the steering intervention is suppressed if the geographical location and the course of the possible lane marking detected by the ego vehicle essentially coincide with the geographical location and the course of a lane marking detected as faulty as described above. In other words, it can be provided that the driver assistance system detects a possible lane marking and records its course, checks a digital road map to determine whether a lane marking marked as faulty exists at its current location on the map, and, if such a lane marking marked as faulty exists, carries out, changes, or suppresses the steering intervention.
[0023] A further embodiment provides that the historical driving history data include information on a generation or development stage of a respective sensor of a respective other vehicle, by means of which a respective lane boundary was detected. This allows the quality of a respective lane boundary detection to be estimated.
[0024] A further embodiment provides that a respective steering intervention is stored in a digital map of the surroundings with its geographical location and / or the course of the lane marking on the basis of which it was carried out and / or with the information on the generation or development stage of the respective sensor of the respective other vehicle by means of which the respective lane marking was detected.
[0025] The invention also includes a driver assistance system configured to implement the method according to the invention. Furthermore, the invention includes a motor vehicle with such a driver assistance system. Advantageous developments of the driver assistance system and / or the motor vehicle may have features as already mentioned in connection with the described method.
[0026] Further features of the invention can be derived from the following description of the figures and from the drawings. The features and combinations of features mentioned above in the description, as well as the features and combinations of features shown below in the description of the figures and / or in the figures alone, can be used not only in the respective combinations specified, but also in other combinations or on their own, without departing from the scope of the invention.
[0027] The drawing shows: Fig. 1 a first exemplary driving situation; Fig. 2 a second exemplary driving situation; Fig. 3 a schematic representation of a method for operating a driver assistance system in an exemplary driving situation; Fig. 4 a possible communication between a motor vehicle, a third-party vehicle and an external server device.
[0028] Identical or functionally equivalent elements are provided with the same reference numerals in the figures.
[0029] Fig. 1 shows a first exemplary driving situation, based on which an embodiment of the described method for operating a driver assistance system is to be described. Fig. 1 shows a motor vehicle 10 and a third-party vehicle 12. The motor vehicle 10 and the third-party vehicle 12 meet along a road 14. Each of the shown motor vehicles 10, 12 moves along a respective lane 16 of the road 14. The motor vehicle 10 has an on-board sensor 18. With the aid of the sensor 18, a driver assistance system 11 (not shown) of the motor vehicle 10 detects a possible lane boundary 20. In addition, the driver assistance system 11 of the motor vehicle 10 detects a course of the possible lane boundary 20 with the aid of the sensor 18. In the Fig. In the example shown in Figure 1, the course of the possible lane boundary 20 extends from one edge of the road toward the center of the road. If, in the present situation, the driver assistance system 11 were to perform a steering intervention based on the possible lane boundary 20, the motor vehicle 10 would continue driving in lane 16, in which the oncoming vehicle 12 is traveling, as a result of the steering intervention. To prevent such an erroneous steering intervention, the driver assistance system 11 performs a plausibility check of the possible lane boundary 20.
[0030] The plausibility check is based on historical journey data, which in this case describe a historical journey trajectory 22 of a large number of third-party vehicles 12 not shown here. Fig. In the driving situation illustrated by way of example in Figure 1, the historical driving trajectory 22 crosses the detected possible lane boundary 20. This is detected by the driver assistance system 11, and the steering intervention planned based on the detected possible lane boundary 20 is suppressed. According to the embodiment shown here, the motor vehicle 10 can receive historical driving history data 24 and / or the historical driving trajectory 22 from an external server device 26. The external server device 26 can be implemented, for example, as an external cloud server.
[0031] Fig. Figure 2 shows another exemplary driving situation. This will be described below with reference to the information already provided in connection with Fig. 1 are described in more detail. As already mentioned in connection with Fig. 1, the motor vehicle 10 detects a course of a possible lane boundary 20 by means of a vehicle-specific sensor 18 and checks the plausibility of this course using a historical driving trajectory 22. In addition, the motor vehicle 10 detects a third-party vehicle 28 driving ahead of it and its current driving process data 30 in the form of a current driving trajectory 32. As in Fig. 2, both the historical driving trajectory 22 and the current driving trajectory 32 cross the possible lane boundary 20. Consequently, the driver assistance system suppresses the steering intervention originally planned due to the possible lane boundary 20.
[0032] Fig. 3 schematically shows the course of a method for operating a driver assistance system 11 with reference to the Fig. 1 and Fig. 2 Described. In a method step S1, the driver assistance system 11 detects a possible lane marking 20 using an on-board sensor 18 of the motor vehicle 10 and records its course. This can be done, for example, using a camera sensor of the motor vehicle 10. In a method step S2, the driver assistance system 11 performs a plausibility check of the possible lane marking 20, wherein the plausibility check is based on historical driving history data 24 of a plurality of other vehicles 12, 28, wherein the historical driving history data 24 describe the respective driving behavior of the other vehicles 12, 28 in the area of the possible lane marking 20. In a method step S3, the driver assistance system 11 carries out, changes, or suppresses the steering intervention depending on the result of the plausibility check.
[0033] Fig.4 schematically shows a possible communication between a motor vehicle 10, a third-party vehicle 12, 28, and an external server device 26. A first communication connection K1 exists between the motor vehicle 10 and the external server device 26. There is a second communication connection K2 between the third-party vehicle 12, 28 and the external server device 26. In addition, there is a communication connection K3 between the motor vehicle 10 and the third-party vehicle 12, 28. Via the communication connection K1, the motor vehicle 10 can, for example, receive historical trip history data 24 from the external server device 26. Alternatively or additionally, the motor vehicle 10 can transmit its own trip history data to the external server device 26 via the communication connection K1.Via the communication connection K2, the third-party vehicle 12, 28 can transmit its own journey history data to the external server device 26 and / or receive journey history data 24, 30 of other third-party vehicles 12, 28 from the external server device 26. Via the communication connection K3, the motor vehicles 10, 12, 28 can communicate directly with each other using known car-to-car communication methods and, among other things, exchange respective journey history data 24, 30.
[0034] Overall, the examples show how a method for minimizing faulty steering interventions of a lane keeping assistant can be provided using swarm data. List of reference symbols 10 motor vehicle 11 Driver assistance system 12 third-party vehicle 14 Street 16 lanes 18 Sensor 20 possible lane markings 22 historical driving trajectory 24 historical trip data 26 Server setup 28 other vehicle driving ahead 30 current trip history data 32 current driving trajectory K1 first communication connection K2 second communication connection K3 third communication connection
Claims
[1] Method for operating a driver assistance system (11) for a motor vehicle (10), wherein - along a road (14) to be travelled, a possible lane boundary (20) is detected by at least one on-board sensor (18) of the motor vehicle (10) and its course is recorded, - a plausibility check of the possible lane boundary (20) is carried out, and - the plausibility check is based on historical journey data (24) of a plurality of other vehicles (12, 28) and additionally on current journey data (30) of at least one other vehicle (28) currently traveling ahead, wherein the historical journey data (24) describe a respective driving behavior of the other vehicles (12, 28) in the area of the possible lane boundary (20) and historical travel trajectories (22) of the other vehicles (12, 28) in the area of the possible lane boundary (20), and the current journey data (30) describe a current travel trajectory (32) of the other vehicle (28) currently traveling ahead in the area of the possible lane boundary (20), characterized by , that - depending on the result of the plausibility check, a steering intervention by the driver assistance system (11) designed as a lane keeping assistance system is carried out or suppressed, whereby - the steering intervention originally planned due to the possible lane boundary (20) is suppressed if both the historical driving trajectories (22) and the current driving trajectory (32) cross the course of the possible lane boundary (20), and / or - the historical driving history data (24) describe a plurality of steering interventions by lane keeping assistants of the other vehicles (12, 28), wherein each steering intervention is assigned a respective geographical location and a respective course of a recognized possible lane boundary on the basis of which the respective steering intervention was carried out, wherein, in the event that at the respective geographical location and on the basis of the respective course of the lane boundary, a frequency of steering interventions by the other vehicles (12, 28) that exceeded a predetermined level and were driven with the lane keeping assistant ready to intervene is registered, the recognized possible lane boundary located at this location and / or this location is marked as a potentially incorrectly recognized lane boundary by an external server device (26), and the steering intervention is suppressed,if the geographical location and the course of the possible lane boundary (20) detected by the at least one on-board sensor (18) of the motor vehicle (10) coincide with the geographical location and the course of the lane boundary marked by the external server device (26). [2] Method according to one of the preceding claims, wherein the historical journey data (24) comprise information on a generation or development stage of a respective sensor of a respective other vehicle (12, 28) by means of which a respective lane boundary was detected. [3] Method according to one of the preceding claims, wherein a respective steering intervention is stored in a digital map of the surroundings with its geographical location and / or the course of the lane boundary on the basis of which it was carried out and / or with the information on the generation or development stage of the respective sensor of the respective other vehicle (2, 28) by means of which the respective lane boundary was detected. [4] Driver assistance system (11) designed to carry out a method according to one of the preceding claims. [5] Motor vehicle (10) with a driver assistance system (11) according to claim 4.
Citation Information
Patent Citations
Method and device for determining a valid lane marking
DE102009007342A1
Driver assistance system e.g. lane keeping support system, for supporting driver during transverse steering of car in e.g. location, has control unit that controls start and end times and / or strength of support action based on information
DE102010053964A1
Method for supporting a driver assistance system on a non-measurable route section
DE102015000379A1
DEVICE AND METHOD FOR DETERMINING A ROAD MODEL
DE102016201249A1
System and method for analyzing travel trajectories for a route section
DE102016216335A1