Procedure for determining the probability of an accident for a motor vehicle

Abstract

The invention relates to a method for determining an accident probability (P) for a motor vehicle, comprising the steps of acquiring (100) data for locating a potential collision object (K) within the environment of the motor vehicle, determining (200) a critical driving situation based on the acquired data, wherein a critical driving situation is determined when a potential collision object (K) is located within the environment of the motor vehicle, determining (300) a plurality of possible trajectories for the motor vehicle based on current vehicle dynamics data (F1) and vehicle control variables (F2), and performing (400) a collision evaluation for the determined trajectories, wherein for each of the determined trajectories,whether the trajectory results in a collision (C) with the potential collision object (K) and performing (500) a context-related weighting of the results of the collision evaluation based on a statistical driving behavior of drivers of a motor vehicle in relation to the driving control variables (F2) of the motor vehicle.

Description

[0001] The present invention relates to a method and a system for determining the probability of an accident for a motor vehicle. Furthermore, the present invention relates to a motor vehicle comprising such a system.

[0002] In the field of collision risk assessment in road traffic, predominantly time-to-collision (TTC)-based methods are used today. These methods calculate the time until a collision assuming constant speeds and directions of movement of the objects involved. Although TTC approaches allow for a basic assessment of collision risk, they have several disadvantages. For example, they do not take into account sudden driving maneuvers, varying reaction times, or complex traffic situations involving multiple objects. Therefore, the course or outcome of critical driving situations cannot be predicted with sufficient accuracy using known TTC-based methods to determine the probability of an accident for a motor vehicle. Consequently, proactive control of passive safety systems before a collision is not possible.

[0003] DE 10 2012 009 297 A1 discloses a method for assisting a driver in operating a vehicle; DE 10 2008 062 916 A1 discloses a method for determining the probability of a collision between a vehicle and a living being; DE 10 2022 206 099 A1 discloses a method for safety-checking a target trajectory for an ego-vehicle and a method for controlling an ego-vehicle; and DE 10 2010 033 776 A1 discloses a method for detecting and predicting the actions of at least two moving objects.

[0004] It is therefore the object of the present invention to at least partially eliminate the disadvantages described above; in particular, it is the object of the present invention to provide a method and a system that enable a precise and reliable prediction of the probability of an accident for a motor vehicle in a critical driving situation in a simple and cost-effective manner, thereby significantly improving the safety of a motor vehicle driver.

[0005] The foregoing problem is solved by a method with the features of claim 1, a system according to claim 10, and a motor vehicle according to claim 11. Further features and details are set forth in the dependent claims, the description, and the drawings. Technical features disclosed with respect to the method according to the invention also apply in connection with the system and the motor vehicle according to the invention, and vice versa, so that the disclosure of the individual aspects of the invention is always, or can always be, mutually interdependent. Advantageous embodiments of the invention are set forth in the dependent claims.

[0006] According to the invention, a method for determining the probability of an accident for a motor vehicle is provided. The method according to the invention comprises the steps of acquiring data to locate a potential collision object within the vicinity of the motor vehicle, determining a critical driving situation based on the acquired data (a critical driving situation is determined when a potential collision object is located within the vicinity of the motor vehicle), determining a plurality of possible trajectories for the motor vehicle based on current vehicle dynamics data and vehicle control parameters, and performing a collision evaluation for the determined trajectories (for each of the determined trajectories, the probability of an accident is determined).whether the trajectory results in a collision with the potential collision object, as well as carrying out a context-related weighting of the results of the collision evaluation based on statistical driving behavior of drivers of a motor vehicle in relation to the driving control parameters of the motor vehicle.

[0007] According to the invention, a method is provided which, by determining possible trajectories for a motor vehicle based on current vehicle dynamics data and driving control parameters, a collision evaluation for the determined trajectories, and a context-related weighting of the results of the collision evaluation, guarantees the most realistic possible prediction of the course or outcome of critical driving situations and thus enables a particularly proactive control of passive safety systems before a collision.

[0008] The method in question can preferably be used in a motor vehicle, such as a car or truck. However, its use in other vehicles, such as commercial vehicles or the like, is also conceivable.

[0009] Within the scope of the invention, the "surroundings of a motor vehicle" can be understood to mean, in particular, a closer area detectable by the sensors of a motor vehicle. This radius can range from 1 to 500 m and extend in an angle of up to 360° around the motor vehicle. Alternatively, only an angle of up to 270° or up to 180° in the direction of travel can be considered.

[0010] For the simple, cost-effective, and precise localization of a potential collision object within the vicinity of a motor vehicle, it is advantageous to acquire the data for locating the potential collision object within the vehicle's environment using the vehicle's environmental sensors, preferably LiDAR and / or radar sensors. In addition to the aforementioned advantages, LiDAR and radar sensors also exhibit high weather resistance and are less susceptible to adverse weather conditions such as fog, rain, or glaring sunlight.

[0011] Furthermore, for an accurate and reliable determination of the probability of an accident, it can be advantageous if the potential collision object is static. It is understood that determining the probability of an accident is much more difficult and less precise when considering two dynamic objects. Therefore, determining the probability of an accident in a collision between a motor vehicle and a static object, such as a wall, a guardrail, or a stationary vehicle, promises lower uncertainties in the motion prediction and more stable algorithms for driver assistance systems.

[0012] In the context of a particularly precise and meaningful determination of possible trajectories for a motor vehicle, it may advantageously be provided that the vehicle dynamics data include the current speed of the motor vehicle and / or the current friction coefficients of the motor vehicle and / or the current acceleration of the motor vehicle and / or the current yaw rate of the motor vehicle.

[0013] In the context of a particularly precise and meaningful determination of possible trajectories for a motor vehicle, it may also be advantageously provided that the vehicle's driving control variables include a plurality of possible brake pedal positions and / or a plurality of possible accelerator pedal positions and / or a plurality of possible steering wheel angle positions.

[0014] In order to have as large a number of meaningfully usable data as possible, it can advantageously be provided according to the invention that the driving control variables of the motor vehicle include steering wheel angle positions between -180° and +180°, preferably at least 20 different steering wheel angle positions between -180° and +180°, in particular at least 35 different steering wheel angle positions between -180° and +180°.

[0015] Similarly, with a view to obtaining the largest possible number of meaningfully usable data, it may be provided that the vehicle control variables include accelerator pedal positions and / or brake pedal positions between 0% and 100%, preferably at least three different accelerator pedal positions and / or brake pedal positions, in particular at least five different accelerator pedal positions and / or brake pedal positions.

[0016] For a particularly precise and comprehensive determination of possible trajectories of a motor vehicle with regard to a potential collision with an object, it may be advantageous to determine at least 100 possible trajectories, preferably at least 150 possible trajectories, and in particular at least 180 possible trajectories.

[0017] With a view to qualitatively improving the determination of the probability of an accident between a motor vehicle and a potential collision object, it is intended that the statistical driving behavior, for the context-related weighting of the collision analysis results, takes into account how the driver of a motor vehicle reacts in a critical driving situation with regard to the vehicle's driving controls, preferably with regard to the driving controls of a brake pedal position and a steering wheel angle. Advantageously, a naturalistic driving study, in particular the naturalistic driving study AMP, can be used for the context-related weighting, whereby, for example, the maximum steering wheel angle and the maximum brake pedal position of a pre-accident phase can be extracted for each relevant case.

[0018] It is understood that individual, several, or all mandatory and / or optional steps of the method according to the invention can be carried out in the proposed sequence, but also in a different sequence. In particular, individual, several, or all mandatory and / or optional steps of the method according to the invention can be carried out repeatedly, e.g., cyclically. It is further understood that individual, several, or all of the mandatory and optional steps of the method according to the invention can also be carried out automatically, in particular by means of a computer.

[0019] The invention also relates to a system for determining the probability of an accident for a motor vehicle, preferably for carrying out a method according to one of the preceding claims, comprising a sensor unit for acquiring data for locating a potential collision object within a driving path of the motor vehicle, a processing unit for determining a critical driving situation based on the acquired data, a computing unit for determining a plurality of possible trajectories for the motor vehicle based on current driving dynamics data and driving control parameters of the motor vehicle and for carrying out a collision evaluation for the determined trajectories and for carrying out a context-related weighting of the results of the collision evaluation based on statistical driving behavior in relation to the driving control parameters of the motor vehicle.The system according to the invention thus exhibits the same advantages as those already described in detail with regard to the method according to the invention. Furthermore, the system in question can advantageously be designed in the form of a self-learning system (machine-learning system).

[0020] The invention also relates to a motor vehicle comprising a system as described above. The motor vehicle according to the invention thus has the same advantages as those already described in detail with regard to the method and system according to the invention.

[0021] The invention also relates to a computer program product comprising commands that cause the system according to the invention to execute the process steps of a previously described method.

[0022] Furthermore, the invention relates to a computer-readable, in particular non-volatile, storage medium on which such a computer program product is stored. Thus, the computer program product and the storage medium according to the invention also offer the advantages described above.

[0023] The computer program product can be implemented as machine-readable instruction code in any suitable programming language and / or machine language, such as Java, C++, C#, and / or Python. The computer program product can be stored on a machine-readable storage medium such as a data disk, removable drive, volatile or non-volatile memory, or onboard memory / processor. The instruction code can program a computer or other programmable devices, such as a control unit (which may be part of the control system), to perform the desired functions. Furthermore, the computer program product can be made available on a network, such as the internet, from which it can be downloaded by a user as needed.The computer program product can be implemented using software, one or more special electronic circuits (i.e., in hardware), or in any hybrid form (i.e., using both software and hardware components). The computer program product according to the invention thus offers the same advantages as those already described in detail with regard to the inventive method and system.

[0024] Further advantages, features, and details of the invention will become apparent from the following description, in which exemplary embodiments of the invention are described in detail with reference to the drawings. The features mentioned in the claims and in the description can each be essential to the invention individually or in any combination.

[0025] They show schematically: Fig. 1 the individual steps of a method according to the invention for determining an accident probability for a motor vehicle, Fig. 2 a set of trajectories with a multitude of possible trajectories for a motor vehicle based on current vehicle dynamics data and vehicle control parameters, Fig. 3. Illustrative representation of the individual steps of the inventive method for determining the probability of an accident for a motor vehicle, Fig. 4 Graphical representation of values ​​for carrying out the step of a context-related weighting of the results of the collision evaluation according to a first embodiment.

[0026] Fig. Figure 1 shows the individual steps of a method according to the invention for determining an accident probability P for a motor vehicle.

[0027] As per Fig. As can be seen from Figure 1, the method according to the invention comprises the steps of acquiring 100 data to locate a potential collision object K within the environment of the motor vehicle, determining 200 a critical driving situation based on the acquired data, wherein a critical driving situation is determined when a potential collision object K is located within the environment of the motor vehicle, determining 300 a plurality of possible trajectories for the motor vehicle based on current vehicle dynamics data F1 and vehicle control variables F2, and performing 400 a collision evaluation for the determined trajectories, wherein for each of the determined trajectories,whether the trajectory results in a collision C with the potential collision object K, as well as carrying out a context-related weighting of the results of the collision evaluation based on statistical driving behavior of drivers of a motor vehicle in relation to the driving control variables F2 of the motor vehicle.

[0028] The data for locating the potential collision object K within the environment of the motor vehicle can advantageously be acquired using the environmental sensors of the motor vehicle, preferably using LIDAR sensors and / or radar sensors.

[0029] The potential collision object K can be, for example, a static object such as a wall, a guardrail, or a tree.

[0030] The vehicle dynamics data F1 may include the current speed of the vehicle and / or the current friction coefficients of the vehicle and / or the current acceleration of the vehicle and / or the current yaw rate of the vehicle.

[0031] In contrast, the vehicle control variables F2 can include a plurality of possible brake pedal positions and / or a plurality of possible accelerator pedal positions and / or a plurality of possible steering wheel angle positions.

[0032] The vehicle control parameters F2 can include steering wheel angle positions between -180° and +180°, preferably at least 20 different steering wheel angle positions between -180° and +180°, in particular at least 35 different steering wheel angle positions between -180° and +180°.

[0033] Similarly, the vehicle control parameters F2 can include accelerator pedal positions and / or brake pedal positions between 0% and 100%, preferably at least three different accelerator pedal positions and / or brake pedal positions, in particular at least five different accelerator pedal positions and / or brake pedal positions.

[0034] In total, at least 100 possible trajectories can be determined according to the described method, preferably at least 150 possible trajectories, and in particular at least 180 possible trajectories.

[0035] The statistical driving behavior for the context-related weighting of the results of the collision evaluation can also advantageously take into account how drivers of a motor vehicle react in a critical driving situation with regard to the driving control variables F2 of the motor vehicle, preferably with regard to the driving control variables F2 of a brake pedal position and a steering wheel angle position.

[0036] Fig. Figure 2 shows a family of trajectories with a multitude of possible trajectories (aligned along an x ​​and y positioning) for a motor vehicle based on current vehicle dynamics data F1 and vehicle control variables F2 together with a collision object K.

[0037] Based on the current vehicle dynamics (e.g., speed and friction coefficients), a "family of trajectories" consisting of, for example, 185 individual trajectories can be determined, with, for example, 37 steering wheel angles (-180°, -170°, ..., 170°, 180°) and five brake pedal positions (0%, 25%, ..., 100%). Subsequent collision analysis can then be performed to determine which of the trajectories intersect the collision object K.

[0038] Fig. Figure 3 shows a clear illustration of the individual steps of the inventive method for determining the probability of an accident P of a motor vehicle.

[0039] Based on vehicle dynamics data F1 and the vehicle's driving control variables F2, in the form of multiple possible brake pedal positions and multiple possible steering wheel angle positions, trajectories are determined. It is then ascertained whether these trajectories collide with a detected potential collision object K, where a checkmark symbolizes a collision C and an x ​​represents no collision C. In a second step, the results are weighted with corresponding probabilities of statistical driving behavior by vehicle drivers to obtain a meaningful prediction of the accident probability P in a critical driving situation.

[0040] Fig. Figure 4 shows a graphical representation of values ​​from a naturalistic driving study for carrying out the step of a context-related weighting of the results of the collision evaluation according to a first embodiment.

[0041] In this process, a density distribution D is fitted from the frequency F of registered events (collisions) as a function of the maximum steering wheel angle (left) or the maximum brake pedal position (right), which can then be used to weight the results of the collision evaluation.

Claims

[1] Method for determining the probability of an accident (P) for a motor vehicle, comprising the steps: - Acquiring (100) data to locate a potential collision object (K) within the vicinity of the motor vehicle, - Determining (200) a critical driving situation based on the recorded data, wherein a critical driving situation is determined when a potential collision object (K) is located within the vicinity of the motor vehicle, - Determine (300) a plurality of possible trajectories for the motor vehicle based on current vehicle dynamics data (F1) and vehicle control variables (F2), - Performing (400) a collision evaluation for the determined trajectories, determining for each of the determined trajectories whether the trajectory results in a collision (C) with the potential collision object (K), - Performing (500) a context-related weighting of the results of the collision evaluation based on a statistical driving behavior of drivers of a motor vehicle in relation to the driving control variables (F2) of the motor vehicle, wherein the statistical driving behavior for the context-related weighting of the results of the collision evaluation takes into account how drivers of a motor vehicle react in a critical driving situation in relation to the driving control variables (F2) of the motor vehicle. [2] Method according to claim 1, characterized by that the data for locating the potential collision object (K) within the environment of the motor vehicle are acquired by means of the motor vehicle's environmental sensors, preferably by means of LIDAR sensors and / or radar sensors. [3] Method according to claim 1 or 2, characterized by , that the potential collision object (K) is a static object. [4] Method according to any of the preceding claims, characterized by , that the vehicle dynamics data (F1) include the current speed of the vehicle and / or the current friction coefficients of the vehicle and / or the current acceleration of the vehicle and / or the current yaw rate of the vehicle. [5] Method according to any of the preceding claims, characterized by , that the vehicle control variables (F2) include a plurality of possible brake pedal positions and / or a plurality of possible accelerator pedal positions and / or a plurality of possible steering wheel angle positions. [6] Method according to claim 5, characterized by , that the vehicle control parameters (F2) include steering wheel angle positions between -180° and +180°, preferably at least 20 different steering wheel angle positions between -180° and +180°, in particular at least 35 different steering wheel angle positions between -180° and +180°. [7] Method according to any one of claims 5 to 6, characterized by, that the vehicle control variables (F2) comprise accelerator pedal positions and / or brake pedal positions between 0% and 100%, preferably at least three different accelerator pedal positions and / or brake pedal positions, in particular at least five different accelerator pedal positions and / or brake pedal positions. [8] Method according to any of the preceding claims, characterized by that at least 100 possible trajectories are determined, preferably at least 150 possible trajectories, in particular at least 180 possible trajectories. [9] Method according to any of the preceding claims, characterized by , that the statistical driving behavior takes into account for the context-related weighting of the results of the collision evaluation how drivers of a motor vehicle react in a critical driving situation with regard to the driving control variables (F2) of the motor vehicle with regard to the driving control variables (F2) of a brake pedal position and a steering wheel angle position. [10] System for determining an accident probability (P) for a motor vehicle, preferably for carrying out a method according to one of the preceding claims, comprising: - a sensor unit for acquiring (100) data for locating a potential collision object (K) within a driving path of the motor vehicle, - a processing unit for determining (200) a critical driving situation based on the recorded data, - a computing unit for determining (300) a plurality of possible trajectories for the motor vehicle based on current vehicle dynamics data (F1) and vehicle control parameters (F2), for performing (400) a collision evaluation for the determined trajectories and for performing (500) a context-related weighting of the results of the collision evaluation based on statistical driving behavior in relation to the vehicle control parameters (F2). [11] Motor vehicle comprising a system according to claim 10.

Images