Method for operating a motor vehicle

Abstract

The invention relates to a method for operating a motor vehicle (2) having a sensor (4) for detecting the surroundings thereof and / or a further vehicle (6), information (I1 to I11), in particular parameters, relating to a roadway (10) used by the further vehicle (6) and / or relating to the further vehicle (6) being determined on the basis of measurement data (D) of the sensor (4), a probability distribution (P) of nodes, in particular parent nodes, of the static Bayesian network is determined on the basis of the information, and wherein the probability of a turning process of the further vehicle (6) is determined by means of the static Bayesian network (12).

Description

Technical Field

[0001] The invention relates to a method for operating a motor vehicle having sensors for detecting its surroundings and / or other vehicles. Background Art

[0002] A method for operating a vehicle is known, in which it is determined whether another vehicle performs a turning maneuver. Thus, for example, a system and a method are known from DE 10 2017 100 029 A1, which are arranged to predict whether the driver of a carrying vehicle or a distant vehicle intends to turn left or right at an intersection or drive straight through the intersection before the carrying vehicle or the distant vehicle arrives at the intersection, and the method is based on a probabilistic model using a dynamic Bayesian network. The method comprises: obtaining a plurality of surrounding factors representing external parameters at or around the intersection, wherein the surrounding factors include the position and speed of the distant vehicle; and obtaining a plurality of carrying vehicle factors that define the operation of the carrying vehicle. Then, the method predicts the turning intention of the carrying vehicle and / or the distant vehicle at the intersection based on the external factors and the vehicle factors by means of the model using the model. The model can use information learned about the driver's previous turning operation at the intersection.

[0003] In addition, DE 10 2019 132 006 A1 discloses a method for identifying a turning object, which includes: detecting a vehicle traveling ahead by means of a surrounding sensor system of the own vehicle; determining a turning probability of the vehicle traveling ahead; and marking the vehicle traveling ahead as a turning object when the turning probability is equal to or greater than a predetermined threshold value.

[0004] If the vehicle's turn is not classified or is classified only relatively unreliably for a driver assistance system, such as an adaptive cruise control (ACC), or if such information is not provided to the driver assistance system, there is the risk that the host vehicle will not react to such a maneuver at all or in a manner that is uncomfortable for the driver, based on the control of the driver assistance system, in particular the speed control.

[0005] Thus, for example, in an adaptive cruise control system (ACC), the own vehicle is braked until the vehicle ahead has completely completed the turn. This results in a relatively long braking process, which is followed by a correspondingly long and / or strong acceleration to the target speed. Summary of the invention

[0006] The object of the present invention is to detect the turning of another vehicle relatively reliably and / or to predict such a turning process relatively easily.

[0007] According to the invention, this object is achieved by the method according to the invention. Furthermore, according to the invention, this object is achieved by a motor vehicle according to the invention, by a computer program according to the invention and by a computer-readable medium according to the invention. In this case, the embodiments described in connection with the method also apply expediently to the motor vehicle, the computer program and the computer-readable medium, and vice versa.

[0008] The method is used for operating a motor vehicle which has sensors which are arranged and configured to detect the surroundings of the motor vehicle and / or other vehicles, in particular vehicles travelling ahead.

[0009] According to the method, the environment and / or other traffic participants, in particular other vehicles, are first detected by means of sensors, and corresponding measurement data are provided as a result of the detection. Information about the other vehicles and / or about the lane used by the other vehicles is determined based on the measurement data of the sensors. For example, the information forms parameters that represent properties of the lane and / or the other vehicles. Preferably, the information is a numerical value or is presented by means of such a numerical value.

[0010] For example, for this purpose the measurement data are evaluated directly, in particular by means of an evaluation unit of the motor vehicle, wherein the information is a result of the evaluation.

[0011] Furthermore, for example, a (virtual) environment model is first generated based on the measurement data, which are also referred to as sensor data. The environment model represents the surroundings of the motor vehicle and the detected traffic participants. For example, within the scope of forming the environment model, a coordinate system is determined in which the motor vehicle, its surroundings, in particular the detected lanes, the detected traffic participants and / or the detected objects (for example obstacles or traffic signs) are arranged. This information is then derived from the environment model or calculated based on the environment model.

[0012] Subsequently, a probability distribution, in particular an unconditional probability distribution, of the nodes of the static Bayesian network is determined based on this information.

[0013] For example, for this purpose, a value between 0 and 1 (event value) is associated with the information in a suitable manner according to a mathematical function, according to a table or according to a characteristic curve as the probability of an event occurring, in particular the probability of a turning process of another vehicle. In addition, the information is associated with a probability that the event does not occur. Accordingly, the probability is determined as 1 minus the event value. Thus, a normalized probability of the event occurring and a normalized probability of the event not occurring are determined as a result of such an association based on the information.

[0014] It is particularly expedient to use this probability distribution as the probability distribution of a parent node, in particular a parent node which itself has no parent node. The probability distribution determined according to the method thus forms an unconditional probability distribution.

[0015] Subsequently, the probability of a turning process of the further vehicle is determined with the aid of a static Bayesian network. For example, this includes determining the probability of whether the further vehicle is turning left, and / or determining the probability of whether the further vehicle is turning right. Alternatively or additionally, this includes determining the probability of whether the vehicle is turning (whether turning left or right).

[0016] In summary, therefore, according to the method, a turning process of a further vehicle is predicted. By using a static Bayesian network, a relatively reliable prediction of the turning process based on the determined information and based on the measured data is achieved. In contrast to a dynamic Bayesian network, in which the nodes of adjacent time steps are interconnected, a static Bayesian network only includes the probability distribution of a (only) time step. Therefore, the prediction is advantageously relatively simple to carry out.

[0017] The determined probability of a turning process is preferably provided to an assistance function, such as an adaptive cruise control system. The assistance function then preferably adjusts the vehicle speed and / or the distance to other vehicles as a function of the determined probability of a turning process. For example, the assistance function compares the determined probability of a turning process with a predefined threshold value and outputs a theoretical variable, in particular a theoretical variable for the vehicle speed and / or for the distance to other vehicles, as a function of the result of the comparison.

[0018] According to a particularly suitable embodiment of the method, the probability distribution determined from the information is determined with the aid of a Sigmoid function with a value range between 0 and 1. In particular, the information, in particular configured as a parameter, forms the independent variable (input variable) of the Sigmoid function. Particularly preferably, the Sigmoid function comprises preset and / or presettable parameters (function parameters, shape variables). Suitably, the shape variables are selected and preset according to the detected environment or another vehicle. For example, the shape variables are derived from an environment model or calculated according to the environment model. In this way, the Sigmoid function can be adapted to the environment so that the turning probability of another vehicle is relatively reliable.

[0019] According to a suitable embodiment, the lateral position of the other vehicle with respect to the center line of the roadway, also called the center line, and / or the time derivative (change) of the lateral position is used as information. Thus, the position of the other vehicle in the transverse direction of the vehicle with respect to the roadway used by the vehicle is determined and used as information. In particular, the distance between the center line of the roadway and the rear position of the other vehicle (i.e. the intersection of the center line of the vehicle and its end on the rear side) is used as the lateral position.

[0020] The lateral position thus describes the lateral arrangement of the other vehicle relative to the center line of the roadway, also referred to as the lane. In this case, a relatively large lateral position and / or a positive value of the time derivative of the lateral position is an indication of a turning process.

[0021] According to a suitable design, a rectangle that is fixed in rotation with respect to the center line of the roadway and that encloses the vehicle is determined. Such a rectangle is also called a "fixed rotation bounding box". Therefore, this is a rectangle with a minimum area that is fixed in rotation with respect to the center line (i.e., cannot be rotated) and completely surrounds the vehicle in a top view. Subsequently, the area of ​​the rectangle and / or the time variation of the area are determined. The area and / or the time variation of the area are used as information.

[0022] If this area is greater than the area of ​​a rectangular (non-rotationally fixed) envelope (i.e., a non-rotationally fixed bounding box) for the other vehicle, this indicates a turning maneuver. This is the case in particular when the vehicle longitudinal direction of the other vehicle is not parallel to the longitudinal direction of the roadway, i.e., when the yaw angle is not equal to 0°. Compared to using the yaw angle itself as information, a relatively reliable (i.e., less error-prone) prediction of the turning maneuver of the other vehicle is achieved, since both the roadway geometry and the vehicle geometry are taken into account.

[0023] According to a suitable design, the turning point at the intersection is determined based on the environment model. For example, the intersection of the center line of the lane currently used by another vehicle (current lane) and the center line of the (turning) lane into which another vehicle may turn and / or intersects with the current lane is used as the turning point. In addition, for example, the point on the center line of the current lane where the turning lane begins with respect to the longitudinal direction of the lane is used as the turning point.

[0024] Next, an advantageously circular turning area around the turning point is determined, in particular defined. The turning point is arranged in the turning area, advantageously centrally in the turning area. For example, an area arranged within a circle with a predetermined radius around the turning point as a center point is used as the turning area. This area is also referred to as a "region of interest".

[0025] Next, the expected (i.e., predicted) time required for the other vehicle to reach the turning point is determined. Additionally or alternatively, the distance of the other vehicle from the turning area is determined. Furthermore, additionally or alternatively, the expected time and / or the time variation of the distance is determined. The time, the time variation of the time, the distance and / or the time variation of the distance are then used as information.

[0026] Preferably, the turning area, in particular the radius for determining the turning area, can be preset and / or adapted. Thus, the turning area can be adapted for each intersection. For example, the radius is selected in proportion to the lane width of the current lane and / or in proportion to the speed of the vehicle.

[0027] According to a suitable design, in particular the type and / or presence of road markings (e.g. directional arrows, solid or dashed lines) as lane restrictions, in particular as information, is used. Additionally or alternatively to this, in particular the type and / or presence of traffic signs and / or the position of traffic lights (e.g. "red", "green") are used as information. Suitably, the type and / or presence of road markings, traffic signs and / or the state of traffic lights are determined based on an environmental model or directly from measurement data. In addition, the type and / or presence of road markings, traffic signs and / or the state of traffic lights are associated with a probability distribution of nodes of a Bayesian network. In particular, a table or the like is provided for this purpose in which the type or presence of one of these features is associated with a probability distribution.

[0028] According to a suitable design, the expected (i.e. predicted) (turning) duration required or possibly required for the other vehicle to completely leave the current lane during the turn is determined as information. For example, for this purpose, the trajectory for the turning process is predicted for the right end on the rear side and the left end on the rear side of the other vehicle respectively. For this purpose, in particular, the circle equation is used. Based on the current acceleration, current speed and / or current position of the other vehicle, the duration can be calculated for each of the two trajectories, wherein the larger of the two durations is used as the turning duration.

[0029] According to one expedient embodiment, in particular the current speed of the other vehicle, its acceleration, in particular the longitudinal acceleration, the yaw angle and / or the time variation of the yaw angle are used as information.

[0030] According to a suitable embodiment, the turn signal state of the turn signal of the other vehicle is detected (in particular “left turn signal on”, “right turn signal on”) and used as information.

[0031] To determine the probability distribution of a (first) node of the nodes of the Bayesian network, the detected turn signal state is assigned a predefined value, in particular 0 or 1. For example, for each possibility of the turn signal state, an associated probability distribution is stored in a memory.

[0032] Additionally or alternatively to this, the distance of the motor vehicle to the other vehicle is determined as (further, second) information, and the probability distribution of the other (second) node in the nodes of the Bayesian network is determined based on this information. In this case, the first and second nodes form the parent node for the common (third) node. In summary, the distance to the other vehicle is taken into account for determining the probability of the turning process based on the turn signal state. Advantageously, this corresponds to a stabilization of the recognition of the turn signal state (i.e., an improvement in reliability).

[0033] Another aspect of the invention relates to a motor vehicle. The motor vehicle can be operated and / or operates according to the method in one of the variants shown above. For this purpose, the vehicle includes sensors set up to detect the environment of the vehicle and / or other vehicles, in particular vehicles traveling in front. In addition, the vehicle includes a device for executing the method steps of the method in one of the variants shown above. Preferably, the device is or includes a computing device, such as a computer and / or a processor. Preferably, the motor vehicle also includes a computer-readable medium on which a Bayesian network or a calculation rule corresponding to the Bayesian network is stored.

[0034] Another aspect of the invention relates to a computer program. The computer program includes commands that cause a motor vehicle to implement the method steps of the method in one of the variants described above. In particular, the computer program includes commands that cause information about lanes used by other vehicles and / or about the other vehicles themselves to be determined, the probability distribution of nodes of a static Bayesian network is determined based on this information, and the probability of a turning process of the other vehicles is determined with the help of the static Bayesian network. Preferably, the computer program also includes commands that cause conditional and / or unconditional probability distributions of nodes, especially all nodes, of the Bayesian network to be determined. Alternatively to this, these probability distributions are stored in a memory and loaded during the method.

[0035] Another aspect of the invention relates to a computer-readable medium, which is in particular designed as a non-volatile memory, on which a computer program is stored. The computer-readable medium is, for example, a flash memory, a hard disk, a CD, a DVD, etc. Preferably, the motor vehicle has the computer-readable medium. Preferably, the computer-readable medium is connected and / or connectable to a device for carrying out the method in terms of signal and / or data transmission technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] The embodiments of the present invention are explained in detail below with reference to the accompanying drawings.

[0037] Figure 1The flowchart shows a method sequence for operating a motor vehicle, wherein a turning probability of a vehicle preceding the motor vehicle is determined using a static Bayesian network based on measurement data of sensors of the motor vehicle.

[0038] Figure 2 shows a static Bayesian network,

[0039] Figure 3 The vehicle and its lateral position in the region of the intersection are schematically shown in a top view,

[0040] Figure 4 A schematic top view of a vehicle preceding a motor vehicle approaching an intersection is shown, wherein a turning area around a turning point has been determined.

[0041] Figure 5 A schematic top view of a vehicle traveling ahead of a motor vehicle in the region of an intersection is shown, wherein a rectangle has been defined which envelops the vehicle traveling ahead and is rotationally fixed with respect to the roadway, and

[0042] Figure 6 A vehicle traveling ahead of a motor vehicle in the area of ​​an intersection is schematically shown in a top view, wherein two trajectories intended for the vehicle to turn have been determined.

[0043] In all figures, corresponding parts and dimensions are always provided with the same reference symbols. DETAILED DESCRIPTION

[0044] exist Figure 1 , a flow chart representing a method for operating a motor vehicle 2 is shown. The motor vehicle comprises at least one sensor 4, which is arranged and configured to detect the surroundings of the motor vehicle 2 and / or another vehicle 6 (here, for example, a vehicle 6 traveling in front of the motor vehicle 2). For example, the sensor 4 is or comprises a camera, a lidar or a radar.

[0045] In the method, in a first step I, the environment and / or other traffic participants (here, a vehicle 6 traveling in front of the motor vehicle 2) are detected by means of a sensor 4 and corresponding measurement data D are provided. For example, the other vehicle 6 uses the same (current) lane 10 as the motor vehicle 2, as an alternative to which the vehicle 6 uses an adjacent lane. For example, according to the method shown here, measurement data are provided for all detected vehicles. In a second step II, a virtual environment model 8 is calculated based on the measurement data D, which represents the surroundings of the motor vehicle 2 and the detected traffic participants (here, the other vehicles 6). For example, within the scope of forming the environment model 8, a coordinate system is determined in which the motor vehicle 2, its surroundings, in particular the detected lane, the detected traffic participants and / or the detected objects (for example obstacles or traffic signs) are arranged and / or evaluated.

[0046] Based on the environment model 8 , one or more information items I1 to I2 about the other vehicle 6 and / or about the lane 10 used by the other vehicle 6 can be determined. 11 .

[0047] According to the example shown here, in a third step III, the information I1 to I2 is determined from the environment model 8 and / or directly from the measurement data D. 11 The following will also be combined with Figures 3 to 6 The corresponding information I1 to I2 are shown in relative detail. 11 of determination.

[0048] In the following fourth step IV, the information I1 to I 11 , respectively, or determine the probability distribution P n , or determine the probability distribution P n,L and the probability distribution P n,R . Let the determined probability distribution P n ,P nL and P nR Respectively and in Figure 2 The nodes K1 to K2 of the static Bayesian network 12 are shown in FIG. 19 Here, nodes K1 to K 19 is a parent node that itself has no parent node. Therefore, the probability distribution determined in step IV is an unconditional probability distribution.

[0049] The corresponding probability distribution P n It also includes two probabilities, namely, P t and P f , where P t represents the probability of vehicle 6 turning left or right, and where P f represents the probability that vehicle 6 turns. In a similar way, the probability distribution P n,LIt also includes two probabilities, namely, P t and P f , where P t represents the probability that vehicle 6 turns left, and where P f represents the probability that vehicle 6 does not turn left. Accordingly, the probability distribution P n,R Including two probabilities P t and P f , where P t represents the probability that vehicle 6 turns right, and where P f Represents the probability that vehicle 6 does not turn right.

[0050] The following apply here:

[0051] P f =1-P t

[0052] In short, the probability P t and P f The corresponding probability distribution P is formed at the corresponding associated node n or P n,L or P n,R .

[0053] The information I1 to I9 are each a value x or are represented by means of a value x. Preferably, the information I1 to I9 are represented by means of a sigmoid function f S (x) (Equation (1)) to determine the probability P t :

[0054]

[0055] Here, the Sigmoid function f S (x) is defined, for example, as follows:

[0056]

[0057] a,x min ,x max are parameters (shape variables). Preferably, these parameters can be preset and / or adapted. In this way, the Sigmoid function can be adapted to the detected environment. For example, the corresponding shape variables are derived from the environment model 8 or calculated based on the environment model. Alternatively, the parameters a, x min ,x max Fixedly (ie, in an unchangeable manner) predefined. In this case, the parameters are determined on the basis of a priori knowledge. For example, they are determined on the basis of test data or on the basis of simulations or, as an alternative, estimated.

[0058] Here, the parameters a,x min ,x maxIt is expedient to select accordingly for each of the information items I1 to I9 so that a distinction can be made in each case for two of the information items.

[0059] In summary, the corresponding associated probability distribution P is determined according to the corresponding information I1 to I9. n ,P n,R ,P n,L .

[0060] In step V, the probability P of the turning process of the further vehicle 6 is determined with the aid of the static Bayesian network 12 . Abbiegen , the probability of turning left L and / or the probability P of a right turn R In addition, for example, the probability P that the vehicle 6 is not turning right or left (ie, is traveling straight) is determined. G .

[0061] exist Figure 2 The static Bayesian network 12 is shown in FIG. The static Bayesian network includes a network with an unconditional probability distribution P n ,P n,R ,P n,L The parent nodes K1 to K 19 In other words, the probability distribution P determined in step IV n ,P n,R ,P n,L One of them is respectively connected to nodes K1 to K 19 is associated with one of the .

[0062] In addition, the Bayesian network 12 includes nodes K 21 and K 22 Here, nodes K1, K3, K4, K6, K8, K 10 ,K 12 ,K 13 ,K 18 and node K 20 (See Figure 6 ) is the node K 21 The parent node of . And nodes K2, K3, K5, K7, K9, K 11 ,K 12 ,K 12 ,K 19 and K 20 is node K 22 The parent node of .

[0063] In summary, node K 21 There is a probability distribution P associated with n,L Or associated with a probability distribution P n In addition, node K 21 With node K 20Therefore, node K 21 Not associated with probability distribution P n,R Node connection. Accordingly, node K 22 There is a probability distribution P associated with n,R Or associated with a probability distribution P n In addition, node K 22 With node K 20 Therefore, node K 22 Not associated with probability distribution P n,L Node connection.

[0064] In addition, the Bayesian network 12 includes nodes K 23 , whose parent node is node K 21 and K 22 .

[0065] At node K 20 ,K 21 ,K 22 and K 23 The corresponding probability distribution P v is predefined and stored in a suitable manner in a memory of the motor vehicle 2. For example, the corresponding probability distribution P is determined based on prior knowledge. v The probability distribution is determined, or alternatively estimated, for example, by means of experiments or by means of simulations.

[0066] For example, using the probability distribution P shown in the following table for node 23 v :

[0067]

[0068] exist Figure 3 , another vehicle 6 is shown in the region of the intersection of a first (current) lane 10, in which the vehicle 6 and / or the motor vehicle 2 (own vehicle) is traveling, and a second lane 14, into which the vehicle 6 can turn. The center line of the current lane 10 is provided with the reference numeral M. In addition, in Figure 2 The lateral position p of the vehicle 6 with respect to the center line M is shown in lat , wherein the rear position p of the vehicle 6 is used here H The lateral position p is calculated, for example, from the environment model 8 or directly from the measurement data D. lat Here, the lateral position p is used lat As the (first) information I1. Suitably, according to equation (1), according to the lateral position p lat To determine the probability distribution P for the left turn process at node K1 n,Land the probability distribution P for the right turn at node K2 n,R . For this purpose the following apply:

[0069] x=p lat

[0070] According to an alternative, not shown in detail, additionally or alternatively to this, the lateral position p lat The time derivative of is used as information and the corresponding probability distributions of the nodes of the Bayesian network 12 are determined therefrom.

[0071] The vehicle 6 is traveling at a current speed v and has a current acceleration. The magnitude of the current speed v is used as the (second) information I2. According to an alternative, not shown in detail, in addition or as an alternative, the current acceleration of the vehicle 6, in particular its magnitude, is used as the information I2 or as further information and the corresponding probability distribution of the Bayesian network 12 is determined therefrom. Preferably, according to equation (1), the probability distribution P of the node K3 of the static Bayesian network 12 is determined as a function of the magnitude of the current speed. n . For this purpose the following apply:

[0072] x=|v|

[0073] The time variation of the yaw angle α (yaw-Winkel) (i.e., the time variation of the angle between the longitudinal direction of the vehicle and the longitudinal direction of the center line M) is used as the (third) information I3. Preferably, the probability distribution P for the left turn process of the node K4 is determined according to the time variation of the yaw angle α according to equation (1). n,L and the probability distribution P for the right turn process at node K5 n,R . For this purpose the following apply:

[0074]

[0075] Furthermore, the yaw angle α itself is determined as information I4 and the probability distribution P for the left turn process of the node K6 is determined according to equation (1) from this yaw angle. n,L and the probability distribution P for the right turn process at node K7 n,R . For this purpose, it applies that: x = α

[0076] exist Figure 4 Schematically shows another vehicle 6 approaching the intersection relative to the motor vehicle 2. In addition, Figure 4 The turning point p is schematically shown in FIG. A Here, the turning point is defined, for example, as the point on the center line M of the roadway 10 at which the intersecting roadway 14 begins with respect to the longitudinal direction of the roadway. For example, the turning point p is determined within the scope of the environment model 8.A In addition, Figure 4 The turning area ROI is shown in FIG. 1 , which is designed around the turning point p as the center point. A For example, the radius of the circle is twice the width of the roadway 10. For example, the radius is selected in proportion to the current speed of the vehicle 6.

[0077] Here, the distance d between the vehicle 6 and the turning area ROI is used. ROI In addition, the expected time t required for vehicle 6 to reach the curve is used as information I5. e As information I6 , for example, the equations of motion and the known position, current speed v and current acceleration of vehicle 6 are used to determine the expected duration.

[0078] In summary, in this method, the turning point p is determined A , determine the turning area ROI, and use the expected time t until the vehicle 6 reaches the turning area ROI e or the distance d between the other vehicle 6 and the turning area ROI F As information I5 and I6.

[0079] Suitably, according to equation (1), according to the information I5 (ie, according to the distance d ROI ) to determine the probability distribution P for the left turn process of node K8 n,L and the probability distribution P for the right turn process at node K9 n,R . For this purpose the following apply:

[0080] x=d ROI

[0081] Suitably, according to equation (1), according to the expected duration t e To determine the node K 10 The probability distribution of the left turn process is P n,L and for node K 11 The probability distribution of the right turn process P n,R . For this purpose the following apply:

[0082] x=t e

[0083] exist Figure 5 In FIG. 6 , a further vehicle 6 is schematically shown in the area of ​​the intersection. Figure 5The so-called “fixed rotation bounding box” of the vehicle is shown in FIG. 8 . For example, the fixed rotation bounding box is calculated within the scope of the environment model 8 . The fixed rotation bounding box is thus a rectangle B with a minimum area which completely encloses (i.e. surrounds) the vehicle 6 , wherein the sides of the rectangle are parallel or perpendicular to the center line M. Furthermore, the area A of the rectangle is determined B .

[0084] The area is given by:

[0085] A B =(sinα·L+cosα·W)·(cosα·L+sinα·W)

[0086] Wherein W is the width of the vehicle 6, L is the length of the vehicle 6, and α is the yaw angle. Here, the length L and width B of the vehicle 6 are determined, for example, based on the measurement data D, or estimated, for example, based on the identification of the type or model of the vehicle 6.

[0087] In summary, therefore, in this method, a rectangle B enveloping the other vehicle 6 is determined, and its area A is determined. B As information I7. Suitably, according to equation (1), according to the area A B To determine the node K of the static Bayesian network 12 12 The probability distribution P n . For this purpose the following apply:

[0088] x=A B

[0089] According to an alternative, not shown in detail, additionally or alternatively, the area A is determined B The time profile of is used as information I7 or as further information and the corresponding probability distribution of the Bayesian network 12 is determined therefrom.

[0090] exist Figure 6 FIG. 6 shows a vehicle 6 in the area of ​​an intersection. In addition, two trajectories T1 and T2 are shown. These trajectories show the left end p of the rear side of the vehicle 6. L or the right end p of the tail side RThe expected path to the lateral end of the current lane 10 during the turn at the intersection. For modeling, for example, a circle equation is used for the corresponding trajectory T1 or T2. In this case, the radius is selected in particular as a function of the speed v of the vehicle 6. Subsequently, based on the equation of motion and the current acceleration and speed of the vehicle, the time until the left end of the rear side of the vehicle passes through trajectory T1 and the time until the right end of the rear side of the vehicle 6 passes through trajectory T2 are determined for each of trajectories T1 and T2 based on the equation of motion based on the current acceleration a and speed v of the vehicle. The larger of the two time durations thus corresponds to the expected time duration required for the vehicle 6 to completely leave the current lane 10. Using this time duration t ab As information I8.

[0091] Suitably, according to equation (1), according to the time t ab To determine the node K 13 The probability distribution of the left turn process is P n,L and for node K 14 The probability distribution of the right turn process P n,R . For this purpose the following apply:

[0092] x=t ab

[0093] In addition, the turn signal state of the turn signal of the vehicle 6 is detected and used as the (tenth) information I 10 In particular, the turn signal state is one of the following possibilities, namely, “left turn signal on” and “right turn signal on”. 15 The probability distribution of the left turn process is P n,L Suitably, when the turn signal state is equal to "left turn signal on", in this P t is equal to 1, otherwise P t =0. In addition, the turn signal state is associated with node K 16 The probability distribution of the right turn process P n,R Suitably, when the turn signal state is equal to "right turn signal on", in this P t is equal to 1, otherwise P t = 0. Expediently, for each of the possible turn signal states, an associated probability distribution P is stored in a memory of the motor vehicle. n,R and the associated probability distribution P n,L .

[0094] Furthermore, the distance d between the motor vehicle 2 and the vehicle 6 is determined. F , and use it as information I9. Conveniently, according to equation (1), the node K of the static Bayesian network 12 is determined 17 The probability distribution P n. For this purpose the following apply:

[0095] x=d F

[0096] Node K associated with the turn signal state 15 and K 16 and the distance d F Associated Node K 17 Form common nodes K 20 ( Figure 2 ). Therefore, at node K 20 This corresponds to taking into account the distance d between the motor vehicle and the other vehicle 6. F The turn signal activity is identified in the case of , thereby reducing the risk of misidentifying the turn signal activity. 20 With node K 21 and K 22 connect.

[0097] According to an alternative solution not shown, for this purpose, similarly, the node K 15 and K 17 Form a parent node for the first child node, and node K 16 and K 17 Form a parent node for the second child node, wherein the first and second child nodes are nodes K 20 The parent node of .

[0098] The situation information detected by means of the sensor 4 (ie the type and / or presence of road markings, traffic signs or the position of traffic lights) is associated with a predefined 18 The probability distribution P n,L and / or for node K 19 P n,R Therefore, the situation information is used as the (eleventh) information I 11 For example, let the solid line as a road marking be associated with the probability P t The value of 0. In addition, for example, when the lane has only a direction arrow turning left, the probability distribution P n,L The probability P t Correlation value 1. For each of the possible pieces of situation information, an associated probability distribution P is advantageously stored in the memory of the motor vehicle 2 . n , the associated probability distribution P n,R and / or the associated probability distribution P n,L .

[0099] In addition, as in Figure 4As can be seen, the motor vehicle 2 includes a computer-readable medium 16, which is designed as a non-volatile memory. The computer-readable medium is connected to a computing device 18 in terms of signal and / or data transmission technology. The computing device is or includes, for example, a computer or a processor. The computing device is also connected to the sensor 4 in terms of signal and / or data transmission technology, so that the measurement data of the sensor 4 can be transmitted to the computing unit 18. A computer program 20 is stored on the computer-readable medium 16. The computer program includes commands that cause the motor vehicle 2 to execute the method steps of the method in one of the variants shown above. For example, in particular when the command is implemented by the computing device 18, the command causes the determination of the environment model Um based on the measurement data D, and the determination of the information I1 to I8 based on the environment model. 11 and its associated probability distribution P I,N , these probability distributions are used in the static Bayesian network 12, and the probability P is determined based on the static Bayesian network 12 Abbiegen , probability P L and / or probability P R .

[0100] The computing device 18 therefore forms or is a component of a device of the motor vehicle 2 for carrying out the method.

[0101] The present invention is not limited to the embodiments described above. On the contrary, other variants of the present invention may also be derived therefrom by a person skilled in the art within the scope of the claims without departing from the subject matter of the present invention. In addition, in particular, all individual features described in the embodiments and / or in the claims may also be combined with one another in other ways without departing from the subject matter of the present invention.

[0102] Reference numerals list

[0103] 2 Motor Vehicles

[0104] 4 Sensors

[0105] 6 Vehicles

[0106] 8 Environmental Model

[0107] 10Current lane / First lane

[0108] 12 Bayesian Networks

[0109] 14 Second lane

[0110] 16 Memory / computer readable medium

[0111] 18 Computing Devices

[0112] 20 Computer Programs

[0113] α Yaw angle

[0114] B Envelope

[0115] D Measurement data

[0116] d F Distance between vehicles and motor vehicles

[0117] d ROI Distance between vehicle and turning area

[0118] K1 to K 23 node

[0119] L The length of the vehicle

[0120] M Center line of the driving lane

[0121] I1 to I 11 information

[0122] p A Turning point

[0123] P Abbiegen The probability of the vehicle turning

[0124] p H Tail position

[0125] p lat Horizontal position

[0126] P L The probability of the vehicle turning left

[0127] p L The left end of the rear side of the vehicle

[0128] P n Probability distribution

[0129] P n,L Probability distribution

[0130] P n,R Probability distribution

[0131] P R The probability of the vehicle turning right

[0132] p R Right end of the rear side of the vehicle

[0133] ROI Turning Area

[0134] T1 trajectory

[0135] T2 trajectory

[0136] v The speed of the vehicle

[0137] W Width of the vehicle

Claims

1. A method for operating a motor vehicle (2) having a sensor (4) for detecting its surroundings and / or another vehicle (6), in particular a vehicle traveling ahead, -in, Determining a lane (10) used by the other vehicle (6) and / or information (I1 to I2) about the other vehicle (6) based on the measurement data (D) of the sensor (4) 11 ), especially the parameters, - wherein, according to the information (I1 to I 11 ) to determine the nodes (K1 to K1) of the static Bayesian network (12) 19 ), especially the probability distribution of the parent node (P n ,P n,L ,P n,R ),and wherein the probability (P ) of the turning process of the further vehicle (6) is determined with the aid of the static Bayesian network (12) abbiegen ,P L ,P R ).

2. The method according to claim 1, It is characterized in that The probability distribution (P n ,P n,L ,P n,R ).

3. The method according to claim 1 or 2, It is characterized in that Using the lateral position (p) of the other vehicle with respect to the center line (M) of the roadway (10) lat ) and / or the lateral position (p lat ) is taken as the time derivative of (I1).

4. The method according to any one of claims 1 to 3, It is characterized in that A rectangle (B) is determined which is rotationally fixed with respect to the center line (M) of the roadway (10) and which encloses the other vehicle (6), and its area (A B ) and / or the area (A B )’s time variation as information (I6).

5. The method according to any one of claims 1 to 4, It is characterized in that -Determine the turning point (p A ), - Determine the location around the turn (p A ), especially circular turning regions (ROIs), and - using the expected time (t) until the vehicle (6) reaches the turning area (ROI) e ) or the distance (d) between the vehicle (6) and the turning area (ROI) F ) as information (I4,I5).

6. The method according to any one of claims 1 to 5, It is characterized in that The nodes (K) of the Bayesian network (12) are associated with the type and / or existence of road markings, traffic signs and / or traffic light positions 18 ,K 19 )’s probability distribution.

7. The method according to any one of claims 1 to 6, It is characterized in that Determine the expected time (t) required for the other vehicle (6) to completely leave the lane (10) during the turn ab ) as information (I8).

8. The method according to any one of claims 1 to 7, It is characterized in that The speed (v), the acceleration (a), the yaw angle (α) of the further vehicle (6) and / or the time variation of the yaw angle (α) are used as information (I2, I3).

9. The method according to any one of claims 1 to 8, It is characterized in that - using the turn signal state of the turn signal of the further vehicle (6) as information (I7), and / or - Determine the nodes (K) of the Bayesian network (12) according to the turn signal state 15 ,K 16 ) is the probability distribution (P n,L ,P n,R ), according to the distance (d F ) to determine another node (K) of the Bayesian network (12) 17 ) is the probability distribution (P n ), and the node (K 15 ,K 16 ,K 17 ) forms another one of the nodes (K) of the Bayesian network (12) 20 )'s parent node.

10. A motor vehicle (2) having a sensor (4) for detecting the environment and means (18) for carrying out the method steps of the method according to any one of claims 1 to 7.

11. A computer program (20) comprising commands which cause a motor vehicle (2) according to claim 10 to carry out the method steps of the method according to any one of claims 1 to 9.

12. A computer readable medium (16) on which a computer program according to claim 11 is stored.

Images