Method for pre-controlling a driver assistance system
The method leverages vehicle fleet data to identify and pre-control driver assistance systems, addressing lighting and marking issues, enhancing system reliability and avoiding failures.
Patent Information
- Application Number
- DE102023004450
- Authority / Receiving Office
- DE · DE
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2023-11-04
- Publication Date
- 2024-11-07
- Estimated Expiration
- 2043-11-04
AI Technical Summary
Existing driver assistance systems are hindered by temporary lighting fluctuations and missing or imperceptible road markings, leading to system failures and malfunctions, which existing methods fail to adequately address.
A method that utilizes data from a fleet of vehicles to identify 'hotspots' where such issues occur, allowing for proactive pre-control of driver assistance systems by adjusting sensors to anticipate and mitigate these conditions, without requiring additional sensors.
Enhances the functional availability of driver assistance systems by preventing undesirable interruptions and ensuring smooth operation in changing lighting conditions or missing markings, leveraging data from other vehicles to optimize sensor adjustments.
Smart Images

Figure 00000000_0000_ABST
Abstract
Description
[0001] The invention relates to a method for pre-controlling a driver assistance system.
[0002] Driver assistance systems are primarily used to enable autonomous or at least semi-automated vehicle control. The vehicle's sensors can be used to implement features such as lane keeping assist. To guide the vehicle autonomously within its lane, the sensors must detect the surrounding lane markings. Furthermore, other driver assistance systems are conceivable that, for example, automatically adjust the speed limit. In this case, the vehicle's sensors must automatically detect speed limits displayed on the road or on signage.
[0003] Worn road surfaces, damaged or missing signage, and varying lighting conditions can negatively affect the functionality of driver assistance systems. For example, certain fluctuations in lighting, such as those caused by shadows in avenues, urban canyons, under bridges, or when entering or exiting a tunnel, can lead to a failure of the driver assistance system.
[0004] From EP 3 291 134 A1, a method for controlling a vehicle's visual sensor is known, wherein the visual sensor is adjusted at a predictable point along a host vehicle's route where a lighting fluctuation greater than or equal to a threshold value is expected to occur. Images captured by the visual sensor, map data from the vehicle, and GPS coordinates are used to determine this point along the route. The location of the lighting fluctuation is derived from an underlying map. A disadvantage of this method is that only lighting fluctuations derived from known buildings can be considered. Temporarily occurring shadows, such as those caused by changes in vegetation or varying sun positions, cannot be taken into account.
[0005] German Patent DE 10 2020 108 531 A1 describes a method for identifying potential hazards in road traffic. An event indicating a potential hazard is recorded and transmitted with its geolocation to a central processing unit. If multiple vehicles detect numerous similar events with this geolocation, the area is entered as a hotspot on a map. The hotspots can be visualized on the map and reported to an authority for review. However, a disadvantage is the lack of notification to road users, meaning the hazard remains for other vehicles involved in traffic.
[0006] Furthermore, the publications DE 10 2007 041 121 A1, DE 10 2016 213 913 A1 and DE 10 2019 131 118 A1 describe various procedures in vehicles in the event of malfunctions and / or errors in optical environment detection systems.
[0007] The object of the present invention is now to provide a method which overcomes the aforementioned disadvantages.
[0008] According to the invention, this problem is solved by a method for pre-controlling a driver assistance system with the features in claim 1, and in particular in the characterizing part of claim 1. Advantageous embodiments and further developments are described in the dependent claims.
[0009] At the core of the inventive method, during successive deactivations and activations of the driver assistance system within a given period, at least geolocation data of vehicles in a fleet is transmitted to a central computing unit. This data is then used to derive information and make it available to other vehicles in the fleet. When such events occur frequently at the same geographical location, a hotspot is identified, and the reason for the deactivation of the driver assistance system is evaluated. This allows all vehicles in a fleet to benefit from information gathered from different vehicles within the fleet. For example, information from the vehicle's backend can be used to determine the geographical location at which assistance systems need to intervene to compensate for an impending outage.The information from the backend is provided by the other vehicles in the fleet.
[0010] The method is designed to proactively control a driver assistance system to prevent malfunctions and / or errors caused by missing and / or visually imperceptible markings. Missing markings can result, for example, from wear or fading due to environmental influences. Visually imperceptible markings can result, for example, from shadows or suddenly changing light conditions when entering or exiting an avenue, a narrow street, a bridge underpass, or a tunnel. This method prevents unwanted interruptions in the functionality of the driver assistance systems. Overall, it provides increased operational availability for the vehicle driver. In particular, it eliminates the need for additional sensors specifically designed for this type of proactive control.
[0011] In particular, an intelligent decision can be derived from the mass of individual data points in the backend, preferably in the central processing unit. Therefore, it is preferable not to transmit the raw data, but rather the information derived from it to other vehicles. In other words, an intelligent decision, especially for predictive control, can be derived from the mass of data points.
[0012] Predictive control means that the vehicle can react proactively to a future situation. Specifically, sensors such as cameras, radar sensors, or lidar can anticipate future situations. For example, the aperture of the vehicle's cameras can be adjusted to prevent glare or to capture enough light and brighten the image.
[0013] The period t1 can be a particularly short period, especially one that is atypical compared to normal vehicle behavior. Vehicle assistance systems are not usually deactivated, activated, deactivated, and activated in quick succession. For example, the period t1 could be a few seconds, such as 0.2 to 2 seconds, or 0.5 to 1 second.
[0014] A hotspot can therefore be understood, for example, as a spatial cluster of events of a specific type. This type might be characterized, for instance, by a particular kind of change in lighting or a specific way in which a marker disappears.
[0015] Preferably, the vehicle assistance systems of vehicles that will pass the geographical location in the future can be pre-controlled in such a way that they can react automatically to missing or visually imperceptible markings. For example, the aperture of at least one vehicle camera can be adjusted to obtain a good or sufficient camera image for analysis even under changing lighting conditions. Similarly, an active steering function can be continued automatically.
[0016] According to a highly advantageous further development of this idea, it can be provided that vehicles in the fleet that have received a pre-control signal send feedback to the central processing unit, whereby pre-control parameters can be continuously optimized in a location- and / or time-specific manner. For example, specific opening angles of the camera apertures, particularly as a percentage, as well as an indexing start time or the duration of the aperture opening, can be stored. The vehicles can therefore both contain information and send feedback back to the central processing unit, especially regarding any bridging success.
[0017] In an advantageous embodiment, the missing markings may be lane markings. These may be incomplete, for example, due to weather or other environmental factors. If such an insufficient lane or road marking occurs geostationarily, the method can, so to speak, "bridge" this gap.
[0018] Another advantageous embodiment can provide for the markings to be generated by abrupt changes in lighting, even if they are not visually perceptible. This can occur, for example, when entering or exiting covered areas such as a tunnel, underpass, or tree-lined avenue. Such an abrupt change in lighting means that the sensors typically cannot react as abruptly. However, this method effectively "prepares" the sensors for the future situation.
[0019] According to a highly advantageous further development of the idea, it can be assumed that the abrupt changes in lighting occur at a tunnel exit or entrance. The method can therefore be specifically designed to address functional limitations at tunnel exits or entrances. Strong changes in lighting result at tunnel exits or entrances, so the cameras should be adjusted to these changes early on, i.e., especially shortly before the car enters or exits the tunnel.
[0020] In an advantageous embodiment, the pre-control can be achieved by partially closing the aperture of at least one vehicle camera, so that the marking, which is not optically perceptible with the aperture fully open, can be detected. This can, for example, prevent abrupt changes in lighting conditions from causing misinterpretation or deactivation of the driver assistance systems.
[0021] Another advantageous embodiment can provide for the pre-control to be achieved by continuing the current steering function if the road layout is known and / or the missing and / or visually imperceptible markings exist only for a spatially limited area. This spatially limited area is particularly small in relation to the distance traveled per unit of time. Specifically, it must be ensured in the backend that the unavailability of the driver assistance system at the aforementioned geolocation typically only lasts for short periods and that no excessively strong steering interventions are necessary, since the road layout is more or less straight and / or could be reconstructed by extrapolation within the vehicle. This, in particular, prevents the active steering function from "toggling."
[0022] According to a highly advantageous refinement of this idea, the pre-control can be achieved by providing trajectories from the backend. This is particularly beneficial when the road layout necessitates at least limited steering. For relatively short distances, typical, promising trajectories for the relevant location can therefore be provided from the backend.
[0023] In an advantageous embodiment, the system can be configured to deactivate the driver assistance system if the marker remains missing after the geographically located hotspot is crossed. For example, information about the expected period of unavailability can be provided to the driver assistance system from the backend. Preferably, the driver assistance system can employ bridging solutions for this duration, or slightly longer, i.e., for the duration plus a delta. The duration should only be extended by a delta that is very small compared to the duration of unavailability. Preferably, the driver assistance system ceases bridging as quickly as possible and becomes available again.
[0024] Further advantageous embodiments of the method according to the invention also result from the exemplary embodiment, which is described in more detail below with reference to the figures.
[0025] This shows: Fig. 1 a vehicle on a road with partially missing markings; Fig. 2 a schematic sectional view of the activation and deactivation of a driver assistance system according to Fig. 1; and Fig. 3 An exemplary embodiment of the method when a vehicle exits a tunnel.
[0026] In the presentation of the Fig. Figure 1 shows a vehicle 3a, 3b on a road with partially missing markings 2. The vehicle 3a, 3b is part of a fleet of vehicles, which are connected to a central computing unit, in particular via a wireless network connection. If the vehicle 3a, 3b is at the position shown on road 7, there is an area directly in front of the vehicle in which markings 2 are either missing or not visible to the vehicle's sensors. Marking 2 is, in particular, a lane marking. It could also be a sign indicating the maximum permissible speed, which is, for example, painted on the road surface or displayed as signs at the roadside. The vehicle assistance system must be able to bypass this area 2, as it cannot operate within it.For example, the vehicle assistance system is automatically deactivated if it cannot detect any markings. Furthermore, nonsensical misinterpretations can occur, indicating the absence of markings or markings that are not perceptible.
[0027] Fig. Figure 2 shows a schematic cross-sectional view of the activation and deactivation of a driver assistance system according to Fig. 1. The diagram shown depicts a trajectory where axis B represents a driving distance and axis A represents the activated or deactivated state of the vehicle assistance system. The vehicle assistance system is activated in the area marked 10. Area 2 corresponds to area 2 in Fig. 1. As can be seen from the lowered curve, a temporary function shutdown occurs in this area, particularly due to glare affecting the sensors or the absence of markings. Therefore, in this area 2, the functions are bypassed by a feedforward control 9.
[0028] In Fig.Figure 3 shows an exemplary embodiment of method 1 as a vehicle 3a, 3b exits a tunnel. Different vehicles 3a, 3b in a fleet can send data, such as the geolocation of the deactivation, the vehicle's speed, and / or the lighting condition, to a central processing unit each time a relevant driver assistance system is activated or deactivated. For example, activations of the driver assistance system that closely follow deactivations can be aggregated as reactivation events in the central processing unit. This allows clusters to be identified and stored as so-called hotspots. In the illustration, hotspot 5 is located at the tunnel exit 6, where there is a significant change in lighting.The processing unit allows for the automated evaluation of the underlying shutdown cycle of the vehicle assistance system, enabling the calculation of specific characteristics of hotspot 5. For example, it can determine the expected duration of the vehicle assistance system's unavailability.
[0029] Data exchange 8 can take place between each of the vehicles 3a, 3b and the central processing unit 4. This allows a fault report from vehicle 3a, 3b to be sent to the processing unit 4, and a feedforward trigger with a location and / or time reference can be sent from the processing unit 4 to vehicle 3a, 3b. In particular, a feedback loop can be implemented, whereby the feedforward trigger can be continuously optimized in a location-specific and / or time-specific manner. For example, the ego vehicle can receive location-based and / or event-based feedforward control of its sensors or ADAS function.
Claims
[1] Method (1) for the pre-control of a driver assistance system to avoid malfunctions and / or malfunctions due to missing and / or visually imperceptible markings (2), characterized by that in the case of successive deactivations and activations of the driver assistance system in a period (t1), at least data for the geolocation of vehicles (3a, 3b) of a fleet are transmitted to a central processing unit (4) in order to derive information from the data and to make it available to other vehicles (3a, 3b) of the fleet, wherein in the case of an accumulation of such events at the same geographical location, a hotspot (5) is identified and a reason for the deactivation of the driver assistance system is evaluated. [2] Method (1) according to claim 1, characterized bythat a pre-control of the vehicle assistance systems of the vehicles (3a, 3b) which drive past the geographical location in the future is carried out in such a way that the missing or optically imperceptible markings (2) can be reacted to automatically. [3] Method (1) according to claim 1 or 2, characterized by that vehicles (3a, 3b) of the fleet which have received a feedforward control send feedback to the central processing unit (4), whereby parameters of the feedforward control can be continuously optimized in a location- and / or time-specific manner. [4] Method (1) according to one of claims 1 to 3, characterized by that the missing markings (2) are missing lane markings. [5] Method (1) according to one of claims 1 to 4, characterized by that the optically imperceptible markings (2) result from abrupt changes in lighting. [6] Method (1) according to claim 5, characterized bythat the abrupt lighting changes occur at a tunnel exit (6) or tunnel entrance. [7] Method (1) according to claim 5 or 6, characterized by that the pre-control is carried out by partially closing the aperture of at least one vehicle camera so that the marking (2), which is not visually perceptible when the aperture is fully open, can be detected. [8] Method (1) according to one of claims 1 to 7, characterized by that the pre-control is carried out by a continuation of the current steering function if the course of the road is known and / or the missing and / or optically imperceptible markings (2) are only present for a spatially limited area. [9] Method (1) according to one of claims 1 to 8, characterized by that the feedforward control is carried out by providing trajectories from the backend. [10] Method (1) according to one of claims 1 to 9, characterized bythat if the marking (2) is still missing after the geographically localized hotspot (5) is exceeded, the driver assistance system is switched off.
Citation Information
Patent Citations
Sensor data processing method for driver assistance system e.g. modern motor vehicle, involves adjusting error value belonging to measured value of one sensor by another measured value of another sensor
DE102007041121A1
Sensor device for a vehicle, vehicle, method for operating a sensor device
DE102016213913A1
SYSTEM AND METHOD FOR EVALUATING THE OPERATION OF VEHICLE ENVIRONMENTAL SENSING SYSTEMS
DE102019131118A1
Procedures for identifying potential hazards
DE102020108531A1
Method and apparatus for controlling vision sensor for autonomous vehicle
EP3291134A1
Cited By
Lane determination for a driver assistance system of a motor vehicle
DE102025101842A1
