DE102020214203A1 - Method for determining a lane change intention of another motor vehicle located in the immediate vicinity of a motor vehicle, and assistance system - Google Patents

Abstract

The invention relates to a method for determining a lane change intention (3) of another motor vehicle (4) using an assistance system (1) of the motor vehicle (2), in which and on the basis of a detected environment (7) an environment model (8) is generated, and in which at least one situational feature (9, 10, 11, 12, 13) characterizing a lane change is determined as a function of the environmental model (8) generated, with the intention to change lanes (3) being determined by means of a Bayesian network model (14) as a function of the characterizing situation feature (9, 10, 11, 12, 13) is determined, wherein a lane change motif (9) for the other motor vehicle (4) is determined as the situation feature (9, 10, 11, 12, 13) characterizing the lane change, the lane change motif (9) depending on a relative distance (Srel) of the further motor vehicle (4) to an object (15), depending on a relative speed (vrel) of the further motor vehicle (4) to the object (15) and is determined as a function of an object type. The invention also relates to an assistance system (1).

Description

The invention relates to a method for determining a lane change intention of another motor vehicle located in the immediate vicinity of a motor vehicle using an assistance system of the motor vehicle, in which at least one environment detection device of the assistance system is used to detect the environment of the motor vehicle and based on the detected environment an environment model for the Environment is generated by means of an electronic computing device of the assistance system, and in which at least one situational feature characterizing a lane change of the other motor vehicle is determined by means of the electronic computing device as a function of the generated environment model, with the intention to change lanes being dependent on a Bayesian network model of the electronic computing device is determined by the at least one characterizing situational feature. Furthermore, the invention relates to an assistance system.

In particular, for the autonomous operation of a motor vehicle, a situation interpretation that is as generic as possible is necessary in order to be able to plan corresponding driving maneuvers autonomously. This also includes the lane change detection of other road users, which in particular can be predicted early and reliably. This in turn can then be used in order to be able to carry out a corresponding evaluation for subsequent driving functions. The recognition of driving maneuvers with object-oriented Bayesian networks in motorway scenarios, for example, is already known for these problems from the prior art. Furthermore, the detection of an intention to change lanes for adaptive cruise control is already known.

the DE 10 2014 003 343 A1 relates to a method for determining a lane change requirement of a system vehicle or a vehicle driving ahead of it, a number of feature parameters and a time trend development of the feature parameters and a probability of a number of hypotheses created from the feature parameters being determined using an environmental representation of a current vehicle environment. It is provided that the probability of an actual value of the respective feature parameter is determined using a continuous Gaussian distribution function and the temporal trend development of the respective feature parameter and the respective probabilities are determined by hypotheses using a sigmoid function, with at least one of the feature parameters being a feature parameter of a relative dynamic between the system vehicle or a vehicle driving ahead of this and another vehicle or object located in the vicinity of the vehicle.

The object of the present invention is to create a method and an assistance system by means of which an improved prediction of a lane change intention of another motor vehicle can be implemented.

This object is achieved by a method and by an assistance system according to the independent patent claims. Advantageous embodiments are specified in the dependent claims.

One aspect of the invention relates to a method for determining a lane change intention of another motor vehicle located in the immediate vicinity of a motor vehicle using an assistance system of the motor vehicle, in which the surroundings of the motor vehicle are detected using at least one environment detection device of the assistance system and an environment model is based on the detected environment for the environment is generated by means of an electronic computing device of the assistance system, and in which, depending on the generated environment model, at least one situation feature characterizing a lane change of the other motor vehicle is determined by means of the electronic computing device, with the intention to change lanes being determined by means of a Bayesian network model of the electronic computing device is determined as a function of the at least one characterizing situational feature.

Provision is made for the at least one situational feature characterizing the lane change of the additional motor vehicle to be determined as a reason for changing lanes for the additional motor vehicle, with the reason for changing lanes depending on a relative distance of the additional motor vehicle from an object in the area, depending on a relative speed of the motor vehicle to the object and is determined as a function of an object type.

A generic model can thus be created, on the basis of which the intention of the other motor vehicle to change lanes can be reliably determined.

In particular, in the present case a geometric approach or different geometric approaches are selected for assessing the situation, as a result of which they are more powerful than the model-based or numerical methods according to the prior art. It is therefore the task to recognize lane-changing maneuvers of the other motor vehicle in the area at an early stage. In a first step, at least one situational feature, in particular a large number of situational features, is determined, in particular on the basis of a feature extraction, which serves as quantified indications of a possible lane change of the other motor vehicle. The prerequisite for this is the existing description of the situation in the environment model. For this purpose, information obtained from different sensor data from different environment detection devices is processed and made available in a uniform environment model. This environment model can contain, for example, track information, object information, boundaries or other information. In a second step, a Bayes network model is used, which is modulated based on different events, for example lane markings are crossed, lane changes are possible or other lane change maneuvers. The events are related as random variables by conditional probabilities. This makes it possible to make a statement or forecast about unmeasurable or unobservable events.

In a last step, the situation characteristics determined in advance are fed into the Bayes network model as “evidence”. Evidence is an event that can be observed. Integrating evidence is a prerequisite for being able to calculate forecasts/probabilities for unobservable events. Consequently, this step makes it possible to make a statement about the probability of the event “lane change” occurring for the respective additional motor vehicle. The event corresponds to the intention to change lanes.

The lane change motive, which can also be referred to as a motivation hypothesis, makes a statement about whether the other motor vehicle wants to change lanes for reasons such as cooperative behavior, jostling or speeding, stationary objects or active direction indicators, for example a turn signal device. For example, with a corresponding algorithm, the current scene can be evaluated for pushers/speeders and stationary objects. Thus, for example, the hypothesis is calculated from the relative speed and the route (distance) to the object driving ahead, which can also be a vehicle, for example. The time until the possible meeting of both objects is determined from these two variables. In addition, the information about, for example, the object type or vehicle type of the vehicle driving ahead or the other motor vehicle is also taken into account. This can enable a better statement of the motivation hypothesis, especially in traffic scenes with high traffic density. Because in this case, the motor vehicles have shorter distances and lower relative speeds among themselves.

To determine the reason for changing lanes, in a first step the object or the vehicle is checked to see whether it is a vehicle driving ahead in the same lane as the motor vehicle. If this is the case, the lane change motive can be determined. As a second step, the distance or distance S _rel between the two motor vehicles is determined. The distance is defined by, for example, the front axle of the motor vehicle behind to the rear axle of the motor vehicle in front. The center line of the lane can be used to determine the distance between the two objects so that a correct distance measurement can also be determined for large curve radii. As a result, the road geometry is taken into account when determining the distance. A third feature for the lane change motif is the consideration of the speed v _rel between the objects. Here, the absolute speeds of the motor vehicles from the environment model are used and a relative speed of the objects to one another is determined from this.

The method is in particular a computer-implemented method. A computer program product is thus also disclosed, which has program instructions which cause an electronic computing device to carry out the method. A computer-readable storage medium is also disclosed, which has the computer program product and is embodied in particular on the electronic computing device.

According to an advantageous embodiment, depending on the relative speed and the relative distance, a collision time between the other motor vehicle and the object is predicted by the electronic computing device and the collision time is taken into account in the lane change motive. In particular, the so-called time-to-collision (TTC), which corresponds to the time of the collision, is calculated on the basis of the relative speed and the distance to the vehicle driving in front. definitely. Based on this, the intention to change lanes can now be reliably determined. If, for example, the time-to-collision is very low, the probability that a lane change will take place is increased. If the time-to-collision is high, the likelihood of a lane change is lower.

Furthermore, it has proven to be advantageous if a vehicle type of the other motor vehicle is taken into account in the lane change motive by means of the electronic computing device. In particular, the vehicle type of the additional motor vehicle is thus considered. Furthermore, the object type, which can also correspond to the vehicle type, of the vehicle can also be considered. The vehicle type characteristic can have a value between 0 and 1. A value of 1 means that it is very likely that this type of vehicle will motivate a lane change. This applies, for example, to trucks on a freeway. If a car drives in front of the motor vehicle under consideration, the feature is set to a value of 0.5, for example. The corresponding hypotheses for the characterizing situational feature can thus be refined, so that the intention to change lanes can be determined in an improved manner.

It has also proven to be advantageous if a second situational feature characterizing the lane change of the other motor vehicle is carried out by means of the electronic computing device, with context information for a potential lane change in the area being taken into account. For example, provision can be made for the environment to be evaluated with regard to this context information. In particular, the context information indicates whether a lane change is possible at all. If this is the case, for example, in the traffic situation at hand, then this situation feature is also taken into account, so that an improved determination of the intention to change lanes can be implemented.

In a further advantageous embodiment, traffic sign monitoring and/or line monitoring of a roadway on which the motor vehicle is located are used as context information. In particular, the context evaluation can thus also be viewed as a traffic sign hypothesis. The traffic sign hypothesis makes a statement as to whether a motor vehicle is theoretically allowed to change lanes due to traffic signs, for example no overtaking, or not. The value of the hypothesis is either 0 if there is no overtaking ban, for example, or 1 if there is no overtaking ban, based on the current lane on which the other motor vehicle is located. Using the traffic sign hypothesis, it is possible to add an offset to the overall lane change probabilities, which in some cases leads to a sharp increase or decrease in a shorter time and can thus have a positive influence on the detection horizon. In order to enable the traffic sign hypothesis, statements about the currently valid traffic signs, for example no overtaking, are necessary.

It is also advantageous if an error assessment is carried out by means of the electronic computing device as a function of the environment model and the error assessment is taken into account when determining the intention to change lanes. In particular, error handling can thus be taken into account, which, for example, treats the quality of the lane markings, for example on the basis of a fitting error, from corresponding fusion data and thus downregulates the credibility of a path hypothesis, for example. This error handling can be applied holistically to differently characterizing situational features. In this way, potential errors can also be taken into account when determining the intention to change lanes.

Furthermore, it has proven to be advantageous if the error assessment takes into account a curvature of a roadway on which the other motor vehicle is located and/or noise during the detection of the surroundings and/or at least one current environmental parameter and/or at least one parameter characterizing the detection of the surroundings will. In particular, curve geometries can thus be taken into account when determining the characterizing situational features. Furthermore, the effect of noise for object and lane information on the determination of the situation features can be taken into account. Furthermore, other vehicles without lane information can also be taken into account. In particular, all features work reliably for this, even if only the lane information of the motor vehicle is available. This is particularly important if, for example, neighboring lanes cannot be modeled in the environment model due to bad weather, in particular the current environmental parameter, and/or due to being covered by a truck, for example, which in particular describes the parameter characterizing the environment detection. In the case of several lanes, one reference lane in particular is then sufficient to determine the lane change probabilities of all motor vehicles.

According to a further advantageous embodiment, a path feature of the further motor vehicle and/or a position feature of the further motor vehicle and/or a free space feature for the further motor vehicle is generated as a further situation feature characterizing the lane change of the further motor vehicle and taken into account when determining the intention to change lanes. The path feature or trajectory feature can in particular be a so-called trajectory hypothesis. The trajectory hypothesis is modeled in particular by using a sigmoid function in the lane coordinate system. In particular, a pre-alignment, a so-called heading, of the other motor vehicle is determined to determine an intersection with a lane line on the right or left. The parameters T _LCR (crossing of a lane marking) and Φ _lane (course angle relative to the course of the lane) required for feature extraction can already be extracted from this simple straight line. This is a very simple form of lane change prediction, which is why, after an intersection with the lane line has been found, another intersection point on the central line of the respective neighboring lane is searched for. Depending on the course angle and the time it takes to cross a lane marking, a future trajectory can then be predicted on the basis of geometric approximation.

The position feature can also be referred to as a lateral hypothesis. The lateral hypothesis results from the two variables lateral distance and lateral speed. Both values provide information about the position of the motor vehicle within the lane in relation to the lane markings on the left and right. Based on this, an additional hypothesis or estimate for an impending lane change can be determined. The lateral distance is determined by determining the shortest distance between the vehicle pose and an adjacent polyline segment of the lane marking. This results in the lateral distance between the vehicle pose and a point on the polyline segment and thus the position of the motor vehicle in relation to the outer lane markings. The lateral speed is calculated from the change in distance to the lane marking over a fixed period of time.

The free space feature can also be referred to as the free space hypothesis. A so-called occupancy grid is created for each object in the area. This is marked out using four segments, which run orthogonally to the respective vehicle heading and at defined distances from the vehicle pose. The edges of the grid cells are determined using these segment lines and the marking lines of the corresponding neighboring tracks. If there is no left lane for the other motor vehicle, the grid is created exclusively for the right lane and vice versa. If, for example, no track is detected, the free space feature is nevertheless determined. The algorithm iterates over all other vehicles in the vicinity, except for the own vehicle, and determines the closest point on the centerline of the lane in which the vehicle is located. This point is used to split the central line into a “heading line” to the front and a “heading line” to the back. These two heading lines are used to determine the point of intersection with occupancy time grids on the respective lane. This approach has the advantage that the track geometry is included in the distance and time calculations. In particular, a number of objects 15 can therefore also have an influence on the probability of movement in the area surrounding the vehicle. In order to take this fact into account, the individual probabilities are assumed to be stochastically independent and can be summarized as in the formula shown above. As a result, all vehicles or objects in the area are taken into account in the open space analysis.

In a further advantageous embodiment, the lane change probability is output by means of the electronic computing device as probability values for a right lane change and for a left lane change and for a following drive. In particular, the respective probability values have a value between 0 and 1. In particular, the sum of all probabilities results in the value 1. On the basis of these probability values, the intention to change lanes can then in turn be reliably determined by means of the electronic computing device. In particular, preference is given to the value which is the highest. In other words, the one of the three probability values that has the highest value is determined as the lane change intention.

A further aspect of the invention relates to an assistance system for a motor vehicle for determining a lane change intention of another motor vehicle located in the immediate vicinity of the motor vehicle, with at least one surroundings detection device and with an electronic computing device, the assistance system being designed to carry out a method according to the preceding aspect . In particular, the method is carried out using the assistance system.

Yet another aspect of the invention relates to a motor vehicle with an assistance system according to the preceding aspect. The motor vehicle is in particular designed to be at least partially autonomous.

The invention also includes developments of the assistance system according to the invention and of the motor vehicle according to the invention, which have features as have already been described in connection with the developments of the method according to the invention. For this reason, the corresponding developments of the assistance system according to the invention and the motor vehicle according to the invention are not described again here.

The invention also includes the combinations of features of the described embodiments.

• 1 ein schematisches Blockschaltbild einer Ausführungsform eines Kraftfahrzeugs mit einer Ausführungsform eines Assistenzsystems;
• 2 eine schematische Draufsicht auf eine Verkehrssituation zur Bestimmung eines Pfadmerkmals;
• 3 eine weitere schematische Draufsicht auf eine Verkehrssituation zur Bestimmung eines Positionsmerkmals;
• 4 eine weitere schematische Draufsicht auf eine Verkehrssituation zur Bestimmung eines Freiraummerkmalms;
• 5 eine weitere schematische Draufsicht auf eine Verkehrssituation zur Bestimmung eines Spurwechselmotivs; und
• 6 ein schematisches Blockschaltbild einer Ausführungsform eines Bayes-Netz-Modells.

Exemplary embodiments of the invention are described below. For this shows:

• 1 a schematic block diagram of an embodiment of a motor vehicle with an embodiment of an assistance system;
• 2 a schematic plan view of a traffic situation for determining a path feature;
• 3 a further schematic plan view of a traffic situation for determining a position feature;
• 4 a further schematic plan view of a traffic situation for determining a free space feature;
• 5 a further schematic plan view of a traffic situation for determining a lane change motive; and
• 6 FIG. 12 is a schematic block diagram of an embodiment of a Bayesian network model.

The exemplary embodiments explained below are preferred exemplary embodiments of the invention. In the exemplary embodiments, the components described each represent individual features of the invention to be considered independently of one another, which also develop the invention independently of one another and are therefore also to be regarded as part of the invention individually or in a combination other than that shown. Furthermore, the exemplary embodiments described can also be supplemented by further features of the invention already described.

Elements with the same function are each provided with the same reference symbols in the figures.

1 1 shows a schematic block diagram of an embodiment of an assistance system 1 of a motor vehicle 2. The motor vehicle 2 can be embodied as a motor vehicle 2 that is operated at least partially autonomously. The assistance system 1 is used to determine a lane change intention 3 for another motor vehicle 4 ( 2 ) educated. For this purpose, the assistance system 1 has at least one environment detection device 5 and an electronic computing device 6 . The environment detection device 5 can be, for example, an ultrasonic sensor device and/or a radar sensor device and/or a lidar sensor device and/or a camera.

In the method for determining the intention to change lanes 3 of the person in the immediate vicinity 7 ( 2 ) located motor vehicle 4, the environment 7 of the motor vehicle 2 is detected by means of at least the environment detection device 5 of the assistance system 1 and based on the detected environment 7, an environment model 8 for the environment 7 is generated by the electronic computing device 6. Depending on the generated environment model 8, a situational feature 9, 10, 11, 12, 13 characterizing a lane change of the other motor vehicle 4 is determined by means of the electronic computing device 6, with a Bayesian network model 14 of the electronic computing device 6 indicating the intention to change lanes 3 is determined as a function of the at least one characterizing situational feature 9, 10, 11, 12, 13.

Provision is made for a lane change motif 9 to be determined for the further motor vehicle 4 as the at least one situational feature 9, 10, 11, 12, 13 characterizing the lane change of the further motor vehicle 4, with the lane change motif 9 depending on a relative distance S _rel ( 5 ) of the other motor vehicle 4 to an object 15 ( 5 ) in the vicinity of 7, depending on a relati ven speed v _rel ( 5 ) of the other motor vehicle 4 to the object 15 and is determined as a function of an object type.

the 1 also shows that another situation feature 9, 10, 11, 12, 13 characterizing the lane change of the other motor vehicle 4 is a path feature 10 of the other motor vehicle 4 and/or a position feature 11 of the other motor vehicle 4 and/or a free space feature 12 for the other motor vehicle 4 are generated and taken into account when determining the intention to change lanes 3 .

Furthermore, in particular in 1 shown that as a further situation feature 9, 10, 11, 12, 13 characterizing the lane change of the other motor vehicle 4, a context evaluation 13 is carried out by means of the electronic computing device 6, with context information for a potential lane change in the area 7 being taken into account for this purpose. In particular, traffic sign monitoring and/or line monitoring of a roadway on which the other motor vehicle 4 is located can be used as context information, for example. In particular, information relevant to lane changes can be taken into account in the context evaluation. For example, no-overtaking signs and solid lines on the roadway can be taken into account. In this way, contextual knowledge from the environment model 8 can be taken into account.

The characterizing situational features 9 , 10 , 11 , 12 , 13 can be generated from the environment model 8 in particular on the basis of a feature extraction 16 .

Provision can also be made in particular for an error assessment to be carried out by means of the electronic computing device 6 as a function of the environment model 8 and for the error assessment to be taken into account when determining the intention to change lanes 3 . For example, the curvature of a roadway on which the other motor vehicle 4 is located and/or noise during the detection of the surroundings and/or at least one current environmental parameter and/or at least one parameter characterizing the detection of the surroundings can be taken into account in the error assessment. Error handling can thus be provided. In particular, curve geometries can thus be taken into account when determining the situational features 9 , 10 , 11 , 12 , 13 . Furthermore, the effects of noisy object and track information can be taken into account. In particular, it can thus be realized that the assistance system 1 can work reliably, even if only lane information of the motor vehicle 1 is available. This is particularly important if neighboring lanes cannot be modeled in the environment model 8 due to bad weather or being covered by trucks, for example. If there are several lanes, one reference lane is sufficient to determine the lane change probabilities of all vehicles.

In particular, provision can also be made for the intention to change lanes 3 to be output by means of the electronic computing device 6 as probability values for a right lane change and for a left lane change and for a following drive.

To use the characterizing situation features 9, 10, 11, 12, 13, in particular the lane change motif 9, the path feature 10, the position feature 11 and the free space feature 12, a sigmoid function 20 can be used.

2 shows a schematic top view of a traffic situation for determining the path feature 10. In particular, a geometric interpolation method is used for determining the features for the trajectory. The path feature 10 can also be referred to as a trajectory hypothesis (TR). Here, in particular, an intersection is determined between so-called heading segments 17 of the other motor vehicle 4 with a central line 18 of an adjacent lane 19. The resulting segment is halved and a difference angle α is halved, both of which can be parameterized in particular. These steps are repeated until the difference angle α is below a limit. A time to cross the lane marking T _LCR can then be determined. Furthermore, an angle can be determined at the point of intersection between the trajectory or path and the outer track marking, which can be referred to as the Φ track. $$P\left(T_{LCR}\right)=T_{LCR}\to f_{SIC}\left(x\right);P\left({\varphi }_{Sp\mathrm{and}\mathrm{right}}\right)={\varphi }_{Sp\mathrm{and}\mathrm{right}}\to f_{SIC}\left(x\right)$$

The magnitude for feeding the Bayesian network model 14 is then determined by: $$P\left(TR\right)=P{T}_{LCR}\ast P\left({\varphi }_{Sp\mathrm{and}\mathrm{right}}\right).$$

$$P\left(v_{LAT}\right)=v_{LAT}\to f_{SIG}\left(x\right)$$

3 shows a further schematic plan view of a further traffic situation. In the present exemplary embodiment, the position feature 11 is determined in particular. The position feature 11 is in particular a so-called lateral hypothesis (LE). In order to now determine the lateral evidence, the position of the other motor vehicle 4 in the lane is determined in relation to the outer lane marking. A distance O _LAT between a pose 22 and a lane marking 23, in particular to a point P ₁ , is determined. A lateral speed v _LAT can be derived from this. The value ranges of O _LAT and v _LAT are normalized using the sigmoid function 20 in the value range between 0 and 1 as probability: $$P\left(O_{LAT}\right)=O_{LAT}\to f_{SIG}\left(x\right)$$

$$P\left(v_{LAT}\right)=v_{LAT}\to f_{SIG}\left(x\right)$$

The following quantities are now fed into the Bayes network model 14: $$P\left(LE\right)=P\left(O_{LAT}\right)\ast P\left(v_{LAT}\right)$$

$$P\left(S_{TD}\right)=S_{TD}\to f_{SIG}\left(x\right);P\left(T_{TD}\right)=T_{TB}\to f_{SIG}\left(x\right)$$

4 shows a further schematic plan view of a further traffic situation. In the 4 In particular, the free space feature 12 can be determined. The free space feature 12 may also be referred to as a free space hypothesis (FR). For this purpose, an occupancy grid 24 is set for the other motor vehicle 4 if an adjacent lane is present. The initial value for each cell Z1 through Z4 is 100 percent, which means there is some space. When the vehicle approaches the cells Z1, Z2, Z3, Z4, which is represented here by the object 15, the distance up to the entry and exit into the cells Z1, Z2, Z3, Z4 is calculated. The distance S _TE shown describes the distance up to entry into a respective cell Z1, Z2, Z3, Z4. The distance STD describes the distance to exit from the cell Z1, Z2, Z3, Z4. The time it takes to enter and exit cells Z1, Z2, Z3, Z4 is determined using the equation of motion for uniform motion. T _TE describes the time until entry into cells Z1, Z2, Z3, Z4 and T _TD describes the time until exit from cells Z1, Z2, Z3, Z4. The probabilities are then determined using the sigmoid function 20 . $$P\left(S_{TE}\right)=S_{TE}\to f_{SIG}\left(x\right);P\left(T_{TE}\right)=T_{TE}\to f_{SIG}\left(x\right)$$

$$P\left(S_{TD}\right)=S_{TD}\to f_{SIG}\left(x\right);P\left(T_{TD}\right)=T_{TB}\to f_{SIG}\left(x\right)$$

$$P\left(TD\right)=P\left(S_{TD}\right)\ast P\left(T_{TE}\right)$$

The probabilities result from the summary of the characteristics: $$P\left(TE\right)=P\left(S_{TE}\right)\ast P\left(T_{TE}\right)$$

$$P\left(TD\right)=P\left(S_{TD}\right)\ast P\left(T_{TE}\right)$$

The variables for feeding into the Bayesian network model 14 are the movement probabilities: $$P\left(fR\left(I\right)\right)=P\left(TE\right)+P\left(TD\right)$$

Insbesondere können somit auch mehrere Objekte 15 einen Einfluss auf die Bewegungswahrscheinlichkeit im Fahrzeugumfeld haben. Um diesen Umstand zu berücksichtigen, werden die Einzelwahrscheinlichkeiten als stochastisch unabhängig angenommen und können wie in der oben gezeigten Formel zusammengefasst werden. If several objects 15 have an influence on the cells Z1, Z2, Z3, Z4, then the movement probabilities are determined by the formula: $$P_{GES}\left(f{R}_{Zelle\left(i\right)}\right)=P_{OBJ1}\left(f{R}_{Zelle\left(i\right)}\right)\ast P_{OBJ2}\left(f{R}_{Zelle\left(i\right)}\right)$$

In particular, a number of objects 15 can therefore also have an influence on the probability of movement in the area surrounding the vehicle. In order to take this fact into account, the individual probabilities are assumed to be stochastically independent and can be summarized as in the formula shown above.

As a result, all vehicles or objects in the area are taken into account in the open space analysis.

The vehicle environment is not discretized by grids, but by lines. The calculation of the features is carried out taking into account the geometry of the road. The discretization lines follow the orientation of the object 15. In particular, only four discretization lines are necessary for this.

5 shows a further schematic plan view of a further traffic situation. In particular, the lane change motif 9 is shown here. The lane change motif 9 can also be referred to as a motivational hypothesis. The motivation for the other vehicle 4 to change lanes is estimated. The basis for this are relative sizes to the object 15 of the lane change candidate. In order to set up this hypothesis, the following variables are determined between the lane-changing candidate and his vehicle in front, that is to say in the present case the other motor vehicle 14 and the object 15 . The relative distance S _rel , the relative speed v _rel and the predicted time until the two vehicles meet, in particular a so-called collision time TTC, are determined. Furthermore, the vehicle types of vehicles 4 and objects 15 can also be taken into account.

6 shows a schematic block diagram of an embodiment of the Bayesian network model 14. In the present case, in particular, a multiplicity of different nodes K1 to K17 are shown. With the Bayesian network model 14 conclusions can be drawn taking into account uncertainties and missing a priori information. In the present case, it is in particular a discrete Bayesian network model 14, which means that an event can only assume discrete values. In our case, for example, it's just true and false. The events are random variables. The parameterization takes place via expert knowledge, in particular via findings from measurement runs and tests. The calculated situation features 9, 10, 11, 12, 13 are fed into the Bayes network model 14 as observable events, in particular as so-called a priori knowledge. The Bayesian network model 14 is used to infer a possible lane change.

All determined situation features 9, 10, 11, 12, 13 must be normalized to a value between 0 and 1 for processing in the Bayesian network model 14. The sigmoid function 20 is used for this, which maps a function value between 0 and 1 for each input value with a firmly defined value range. This function value is then also fed into the Bayesian network model 14 as observable variables (parent nodes).

The value range of the input variables is determined from measured values. The extreme values of the value range always represent the safe state, i.e. whether a lane change has taken place or not.

From this point on, situational characteristics 9, 1, 11, 12, 13 are interpreted as events that will or will not occur. The function values of the sigmoid function 20 indicate the measure, ie a probability, for both events.

The general form of the sigmoid function 20 for determining the probability of event X is described as follows: $$\text{P}\left(\text{X}=\text{Yes}|\text{E}\right)=\mathrm{n}\cdot 1\text{a}+\text{ex}\left(\text{b}\cdot \text{E}\right)$$

E: Eingangswerte aus den Situationsmerkmalen, Normierungsfaktor a, b Gradient der Kurve von 0 auf 1The non-occurrence of the event is calculated as follows. $$\text{P}\left(\text{X}=\text{no}|\text{E}\right)=1-\text{P}\left(\text{X}=\text{Yes}|\text{E}\right)$$

E: Input values from the situation characteristics, normalization factor a, b gradient of the curve from 0 to 1

Value ranges and parameters for normalizing the situation characteristics 9, 10, 11, 12, 13

$$\text{P}\left(\text{LE}=\text{Ja}|\text{o\_lat}\right):\text{a\_o\_lat=60}\text{, b\_o\_lat=3}\text{.5}\text{, \_o\_lat=60 o\_lat}\left\{0.0\dots 2.0\right\}$$

Value ranges and parameters for normalizing the hypothesis lateral evidence (LE): $$\text{P}\left(\text{LE}=\text{Yes}|\text{v\_lat}\right):\text{a\_v\_lat=0}\text{.04}\text{, b\_v\_lat=9}\text{, \_v\_lat=0}\text{.04 v\_lat}\left\{-1.5...0.0\right\}$$

$$\text{P}\left(\text{LE}=\text{Yes}|\text{o\_lat}\right):\text{a\_o\_lat=60}\text{, b\_o\_lat=3}\text{.5}\text{, \_o\_lat=60 o\_lat}\left\{0.0...2.0\right\}$$

Calculation of the probability of the lateral evidence event. Serves as an input variable for the node LE in the Bayes network model 14. $$\text{P}\left(\text{LE}=\text{Yes}|\text{v\_lat o\_lat}\right)=\text{P}\left(\text{LE}=\text{Yes}|\text{v\_lat}\right)\text{P}\left(\text{LE}=\text{Yes}|\text{o\_lat}\right)$$

For further consideration in the Bayes network model 14 we use the abbreviated notation for the probability: $$\text{P}\left(\text{LE}=1\right)=\text{P}\left(\text{LE}=\text{Yes}|\text{v\_lat}\right)\text{P}\left(\text{LE}=\text{Yes}|\text{o\_lat}\right)$$

$$\text{P}\left(\text{TR}=\text{Ja}|\text{\_spur}\right):\text{a\_spur=0}\text{.01}\text{, b\_spur=200}\text{, \_spur=0}\text{.14 \_spur}\left\{-0.05\dots 0.0\right\}$$

Value ranges and parameters for normalizing the hypothesis trajectory (TR): $$\text{P}\left(\text{TR}=\text{Yes}|\text{T\_lcr}\right):\text{a\_T\_lcr=150}\text{, b\_T\_lcr=2}\text{.5}\text{, \_T\_lcr=151 T\_lcr}\left\{-1.0...4.0\right\}$$

$$\text{P}\left(\text{TR}=\text{Yes}|\text{\_track}\right):\text{a\_track=0}\text{.01}\text{, b\_track=200}\text{, \_track=0}\text{.14 \_track}\left\{-0.05...0.0\right\}$$

Calculation of the probability for the trajectory event. Serves as an input node for node TR in Bayesian network model 14. $$\text{P}\left(\text{TR}=\text{Yes}|\text{T\_lcr},\_\text{track}\right)=\text{P}\left(\text{TR}=\text{Yes}|\text{T\_lcr}\right)\text{P}\left(\text{TR}=\text{Yes}|\text{\_track}\right)$$

For further consideration in the Bayes network model 14 we use the abbreviated notation for the probability: $$\text{P}\left(\text{TR}=1\right)\text{P}\left(\text{TR}=\text{Yes}|\text{T\_lcr}\right)\text{P}\left(\text{TR}=\text{Yes}|\text{\_track}\right)$$

$$\text{P}\left(\text{Eintritt}=\text{Ja}|\text{S\_TE}\right):\text{a\_S\_TE=0}\text{.4}\text{, b\_S\_TE=-0}\text{.4}\text{, \_S\_TE=0}\text{.41 S\_TE}\left\{-10.0\dots 10.0\right\}$$

$$\text{P}\left(\text{Austritt}=\text{Ja}|\text{T\_TD}\right):\text{a\_T\_TD=0}\text{.2}\text{, b\_T\_TD=-3}\text{, \_T\_TD=0}\text{.24 T\_TD}\left\{-1.0\dots 1.0\right\}$$

$$\text{P}\left(\text{Austritt}=\text{Ja}|\text{S\_TD}\right):\text{a\_S\_TD=-0}\text{.4}\text{, b\_S\_TD=-0}\text{.4}\text{, \_S\_TD=0}\text{.41 S\_TD}\left\{-10.0\dots 10.0\right\}$$

Value ranges and parameters for normalizing the free space (FR) hypothesis: $$\text{P}\left(\text{entry}=\text{Yes}|\text{T\_TE}\right):\text{a\_T\_TE=0}\text{.2}\text{, b\_T\_TE=-3}\text{, \_T\_TE=0}\text{.24 T\_TE}\left\{-1.0...1.0\right\}$$

$$\text{P}\left(\text{entry}=\text{Yes}|\text{S\_TE}\right):\text{a\_S\_TE=0}\text{.4}\text{, b\_S\_TE=-0}\text{.4}\text{, \_S\_TE=0}\text{.41 S\_TE}\left\{-10.0...10.0\right\}$$

$$\text{P}\left(\text{exit}=\text{Yes}|\text{T\_TD}\right):\text{a\_T\_TD=0}\text{.2}\text{, b\_T\_TD=-3}\text{, \_T\_TD=0}\text{.24 T\_TD}\left\{-1.0...1.0\right\}$$

$$\text{P}\left(\text{exit}=\text{Yes}|\text{HOURS}\right):\text{a\_S\_TD=-0}\text{.4}\text{, b\_S\_TD=-0}\text{.4}\text{, \_S\_TD=0}\text{.41 S\_TD}\left\{-10.0...10.0\right\}$$

Calculation of the probability of the event that a vehicle enters the cell: $$\text{P}\left(\text{entry}=\text{Yes}|\text{T\_TE}\text{, S\_TE}\right)=\text{P}\left(\text{entry}=\text{Yes}|\text{T\_TE}\right)\text{P}\left(\text{entry}=\text{Yes}|\text{S\_TE}\right)$$

Calculation of the probability of the event that a vehicle exits the cell: $$\text{P}\left(\text{exit}=\text{Yes}|\text{T\_TD}\text{, HOURS}\right)=\text{P}\left(\text{exit}=\text{Yes}|\text{T\_TD}\right)\text{P}\left(\text{exit}=\text{Yes}|\text{HOURS}\right)$$

Calculation of the free space probability:

P_ze: Freiraumwahrscheinlichkeit eines Fahrzeugs für eine ZelleThe probabilities for the events “entry” and “exit” of the other motor vehicle 4 with regard to a cell Z1, Z2, Z3, Z4 can be combined via an addition due to their stochastic dependency. The free space probability that a further motor vehicle 4 exerts on a cell Z1, Z2, Z3, Z4 thus results as follows: $$\text{P\_ze}\left(\text{entry}=\text{Yes exit}=\text{Yes}\right)=\text{P}\left(\text{entry}=\text{Yes}\right)+\text{P}\left(\text{exit}=\text{Yes}\right)$$

P_ze: Free space probability of a vehicle for a cell

P_z: Gesamt Freiraumwahrscheinlichkeit für eine Zelle bei zwei FahrzeugenIf there is another motor vehicle in front of a cell Z1, Z2, Z3, Z4 and another vehicle behind a cell Z1, Z2, Z3, Z4, both vehicles can have the same influence on the free space probability of a cell Z1, Z2, Z3, Z4. This means that the free space probability of the vehicle in front of the cell can be just as great as the free space probability of the vehicle behind the cell Z1, Z2, Z3, Z4. This means that both events are stochastically independent and the free space probability, taking both vehicles into account, can be calculated by multiplying the individual probabilities {P_{ze}}. $$\text{P\_z}\left(\text{<figure-callout id="1" label="P\_ze" filenames="DE102020214203A1\_0030.png" state="{{state}}">P\_ze</figure-callout>}1\text{P\_ze2}\right)\text{<figure-callout id="1" label="P\_ze" filenames="DE102020214203A1\_0030.png" state="{{state}}">P\_ze</figure-callout>}1\text{P\_ze2}$$

P_z: Total free space probability for a cell with two vehicles

• P_z(P_z_i P_z_i+1 P_n)= _i=1^nP_ze_i ,n>0 n : Anzahl Fahrzeuge welche auf eine Zelle Einfluss haben

General formula for calculating the free space probability of a cell Z1, Z2, Z3, Z4:

• P_z(P_z_i P_z_i+1 P_n)= _i=1^nP_ze_i ,n>0 n : Number of vehicles that affect a cell

The Bayesian network model 14 for detecting a lane change is carried out via the determination of the intermediate nodes “cross lane marking left” K11 and “cross lane marking right” K2. Both partial results flow into the “lane change” result node K6. The “lane change” hypothesis provides us with a probability for the events “lane change to the left”, “lane change to the right” and “following” per vehicle.

Input nodes of the situation characteristics 9, 10, 11, 12, 13 are "free space" K7, K10, "trajectory" K1, K9 and "lateral evidence" K0, K8. The calculated situation features 9, 10, 11, 12, 13 from the feature extraction 16 are modeled in the Bayes network model 14 as parent nodes. Input variables of these nodes are the calculated probabilities from the normalization of the situation features 9, 10, 11, 12, 13 for the Bayes network model 14.

A special feature here is the hypothesis free space. This has the nodes K3, K4, K5 and K12, K13, K14 which represent the individual cells Z1, Z2, Z3, Z4 of the grid with the associated free space probabilities. The free space probabilities of the individual cells in nodes K7 and K10 are summarized. In this node, the possible combinations of all individual events, i.e. whether a cell Z1, Z2, Z3, Z4 is occupied or not, and probabilities are firmly assigned (expert knowledge), from which the overall probability of the "free space" hypothesis is calculated.

In the child nodes K2 and K11, the hypothesis “lane markings crossed” is calculated for the respective lane markings on the right and left of motor vehicle 1, as seen. All possible combinations of the events from the nodes of the situation characteristics 9, 10, 11, 12, 13 free space, trajectory and lateral evidence are parameterized with probabilities, whereby a probability for the hypothesis "lane marking crossed" can be calculated.

The probabilities from the hypotheses “lane marking crossed” are entered as an event in the “lane change” node. In the "Lane change" node, the event "Lane change to the right", "Lane change to the left" or "Following drive" can be calculated.

Overall, the invention shows a lane change detection or intention detection.

Reference List

1

assistance system

2

motor vehicle

3

lane change intention

4

another motor vehicle

5

environment sensing device

6

electronic computing device

7

vicinity

8th

environment model

9

lane change motif

10

path feature

11

position feature

12

clearance feature

13

context evaluation

14

Bayesian network model

15

object

16

feature extraction

17

heading

18

central line

19

adjacent lane

20

sigmoid function

21

adjusted lane

22

vehicle pose

23

line

24

grid

a

differential angle

OLAT

Route

vLAT

lateral speed

P1

Point

HOURS

exit distance

STE

entry distance

TTE

entry time

TTD

exit time

srel

relative distance

vrel

relative speed

TTC

time of collision

K1 to K17

node

QUOTES INCLUDED IN DESCRIPTION

This list of documents cited by the applicant was generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent Literature Cited

• DE 102014003343 A1 [0003]

Claims (10)

Method for determining a lane change intention (3) of another motor vehicle (4) located in the immediate vicinity (7) of a motor vehicle (2) by means of an assistance system (1) of the motor vehicle (2), in which by means of at least one environment detection device (5) of the Assistance system (1) the environment (7) of the motor vehicle (4) is detected and based on the detected environment (7) an environment model (8) for the environment (7) by means of an electronic computing device (6) of the assistance system (1) is generated , and in which at least one situation feature (9, 10, 11, 12, 13) characterizing a lane change of the other motor vehicle (4) is determined by means of the electronic computing device (6) as a function of the environment model (8) generated, with a Bayes - Network model (14) of the electronic computing device (6) the lane change intention (3) depending on the at least one characterizing situation feature (9, 10, 11, 12, 13), characterized in that the at least one situational feature (9, 10, 11, 12, 13) characterizing the lane change of the other motor vehicle (4) is a lane change motif (9) for the further motor vehicle (4) is determined, the lane change motive (9) depending on a relative distance (S _rel ) of the further motor vehicle (4) to an object (15) in the environment (7), depending on a relative speed ( v _rel ) of the other motor vehicle (4) to the object (15) and is determined as a function of an object type. procedure after claim 1 , characterized in that depending on the relative speed (v _rel ) and the relative distance (S _rel ), a collision time (TTC) between the other motor vehicle (4) and the object (15) is predicted by means of the electronic computing device (6) and the time of collision (TTC) is taken into account when changing lanes (9). procedure after claim 1 or 2 , characterized in that a vehicle type of the other motor vehicle (4) is taken into account when changing lanes (9) by means of the electronic computing device (6). Method according to one of the preceding claims, characterized in that as a further situation feature (9, 10, 11, 12, 13) characterizing the lane change of the further motor vehicle (4) a context evaluation (13) is carried out by means of the electronic computing device (6), wherein To this end, context information for a potential lane change in the area (7) is taken into account. procedure after claim 4 , characterized in that traffic sign monitoring and/or line monitoring of a lane on which the other motor vehicle (4) is located is used as context information. Method according to one of the preceding claims, characterized in that an error assessment is carried out by means of the electronic computing device (6) as a function of the environment model (8) and the error assessment is taken into account when determining the intention to change lanes (3). procedure after claim 6 , characterized in that the error assessment takes into account a curvature of a roadway on which the other motor vehicle (4) is located and/or noise during the detection of the surroundings and/or at least one current environmental parameter and/or at least one parameter characterizing the detection of the surroundings . Method according to one of the preceding claims, characterized in that as a further situation feature (9, 10, 11, 12, 13) characterizing the lane change of the further motor vehicle (4) a path feature (10) of the further motor vehicle (4) and/or a position feature (11) of the additional motor vehicle (4) and/or a free space feature (12) for the additional motor vehicle (4) are generated and taken into account when determining the intention to change lanes (3). Method according to one of the preceding claims, characterized in that the intention to change lanes (3) is output by means of the electronic computing device (6) as probability values for a right lane change and for a left lane change and for a following drive. Assistance system (1) for a motor vehicle (2) for determining a lane change intention (3) of another motor vehicle (4) located in the immediate vicinity (7) of the motor vehicle (2), with at least one surroundings detection device (5) and with an electronic computing device (6), wherein the assistance system (1) for performing a method according to one of Claims 1 until 9 is trained.

Images