DE102007044761A1 - Driving tube determining method for vehicle, involves determining driving tube edges based on vehicle parameter and lane information determined based on lane data measured by sensors, where edges are described in form of frequency polygon - Google Patents

Abstract

The method involves limiting a driving tube (FS) to both side of a moving path of a vehicle (2) by driving tube edges (FSR1, FSR2). Lane information (FI) i.e. lane marker (FSM1-FSM3) representing a lane, and the driving tube edges are determined depending on determination of adapted driving tube edge (aFSR1, aFSR2). The driving tube edges are determined based on a vehicle parameter and the lane information (FI) determined based on lane data measured by sensors, where the driving tube edges are described in form of a frequency polygon (PZ1.1-PZ1.n, PZ2.1-PZ2.n). An independent claim is also included for a device for determining a driving tube that is extended along a moving path of a vehicle.

Description

The The invention relates to a method for determining a path along a trajectory of a vehicle extending driving tube according to the preamble of claim 1 and an apparatus for performing the method according to the preamble of claim 3.

modern Vehicles are equipped with so-called driver assistance devices, which the trajectory along which the vehicle is moving will, predict. The probability of getting the vehicle Moving along this trajectory is essentially depends on, whether the current driving condition of the vehicle remains constant or Not. To determine the trajectory of a vehicle it is known, for example, the speed, the steering angle and / or to grasp the yaw rate of the vehicle in order to deduce the momentary, static curvature of the path to determine.

Furthermore It is known not only to determine the trajectory, but to determine a driving route based on the trajectory within that, in all likelihood the vehicle will move. The edges of the Driving hose through a fixed distance to the predetermined trajectory established. Objects captured during the environment capture must be located within the driving tube to be driven by a driving assistance device, such as for example, an adaptive distance and speed device (also referred to as Adaptive Cruise Control or ACC for short), as to be considered relevant.

The Goodness of Warning or control intervention in the known driver assistance devices (also called driver assistance systems (= FAS)) depends on the described relevance consideration of example by means of detected objects of a vehicle sensor. An important component This relevance consideration includes the analysis of whether a recognized Object inside or outside certain driving hoses (FSe) is or due to a movement forecast in future this will move in or out of these.

was standing the technology is the one for the evaluation of the object relevance underlying the driving tube from the current driving condition of the own vehicle (also system vehicle called), z. B. from its driving speed, yaw rate, Steering wheel angular velocity and other input variables such as z. B. stationary Radar targets on the roadside, to determine. A derived from this data is driving lane hereinafter referred to as state or standard drive hose. The boundary lines of the state drive hose are often through circular arcs described.

In the DE 10 2005 039 525 A1 A method is described which improves the state drive tube such that in dynamic traffic situations (eg when entering narrow bends, exits from tight bends, S-curves or roundabouts) a boundary line is widened in the direction of the yawing motion. This is done by describing the corresponding circular arc by a smaller radius. As a result, an improved evaluation of the object relevance is achieved in dynamic driving situations, in particular with regard to comfort systems.

Of the Invention is based on the object, an improved method to specify the determination of the driving tube, with an improvement the evaluation of inside and / or outside the driving tube lying objects allows should be. About that In addition, a suitable device for carrying out the method should be specified.

Regarding of the method, the object is achieved by the features of the claim 1. With regard to the device, the object is achieved by the features of claim 3.

At the inventive method for determining a movement along a trajectory of a vehicle extending driving hose within which the vehicle with high probability will move on a lane, wherein the travel tube on both sides of the trajectory of one each Will be limited to at least one of the driving lane edges and additionally at least one lane information, in particular one the lane and the lane mark representing the respective raceway edge determined on the basis of which an adapted driving tube, in particular an adapted Fahrfahrrand edge is determined. By consideration of lane information in addition to the determined on the basis of vehicle condition data driving tube (also called state or standard tube) becomes the prediction the future Trajectory and the relevance of in the field of this trajectory improved objects improved.

For the purposes of this invention, the term driving lane is understood to mean, in particular, a geometric subarea of the potential driving area (eg the street level) defined by a left and a right driving lane edge (= delimiting line), which divides the entire driving lane into two areas (inner half or outside of the driving tube) shares. The driving-tube edges are advantageously described by polygons. However, other geometric description forms are possible, for example by means of Bezier curves or so-called splines.

Furthermore can determined depending on the underlying vehicle assistance system different driving hoses and taken into account become. For Security systems (eg so-called CMS) are available for evaluation the relevance of the detected objects, eg. B. other road users, Obstacles, vehicles in front, such as a driving tube Ideally, that of the driver of the own vehicle or of the system vehicle in the future indeed worn area corresponds. For comfort systems (eg the ACC system) For example, offers a driving hose, with the by the on the street applied lane markings defined lane corresponds (at least as long as the driver of his own vehicle or the system vehicle maintains the currently traveled lane and performs no lane change). Especially offer driving hoses different width or width profiles. Because the driving behavior of road users in general by existing lane markings is influenced leaves also the prognosis to future Support object movements by appropriate driving hoses. Object-related play here Sizes, like z. B. length of stay in the route, probability of occurrence in and exit probability from the driving tube, which predicted Time until entering or exiting the driving tube a role.

By the consideration of lane information, such as actual lane width, location own vehicle within the traffic lane, lane edges, or double-sided and / or central lane markings, and their inclusion in the determination of a customized driving tube (also called fusion the evaluation of preceding objects is clear improved. In addition, one is opposite the prior art earlier and secure recognition of inputs and Ausschervorgängen of lying objects possible. to Determination of the additional Lane information can different methods, for. B. optical lane detection and digital Maps, which are used the lane or parts of the lane or determine parameters of the traffic lane.

Under circumstances can the lane information (also called additional information) incorrect or only valid up to a certain distance range. High availability this lane information can not always be ensured for example, with optical track detection switched off. As well can the description of the respective driving tubes and lanes with different Models are made. The improvement approach is based on the description the driving hoses and lanes in generic form by two polylines, the in an advantageous embodiment by polygons with a predefinable number of nodes or polygon nodes be approximated. Geometric polylines can always be model-based Descriptions are calculated and are therefore suitable as homogeneous Description with regard to the combination of the determined driving line edge and the determined lane information.

embodiments The invention will be explained in more detail with reference to a drawing. Showing:

1 schematically a block diagram of a method for determining a driving tube adapted to a state driving tube taking into account lane information,

2 a schematic representation of another example of a according to the method according to 1 determined adapted driving tube,

3 schematically a block diagram of a method for determining the adapted driving tube according to 1 considering a weighting function,

4 a diagram for an exemplary embodiment of a weighting function as a function of the pathway ahead,

5 a schematic representation with an example of a according to the method according to 3 and 4 determined adapted driving line,

6 schematically in block diagram an embodiment of a method for determining the weighting function,

7 a diagram for a further embodiment of a weighting function as a function of time,

8th schematically in block diagram an embodiment of a filter function for suppressing and disregarding sudden changes in the lane information in the determination of the adapted driving tube, and

9 a schematic representation with an example of a according to the method according to 8th determined adapted driving tube.

each other corresponding parts are in all figures with the same reference numerals Mistake.

1 shows an embodiment of a signal flow of the combination or geometric fusion of at least one parameter of a driving tube FS, z. B. a Fahrfahrrandrand FSR1 (eg left edge of the driving lane), FSR2 (eg right Fahrfahrrandrand), a curvature FSK, and at least one lane information FI, z. A lane marking FSM1 (eg left lane marking), FSM2 (eg right lane marking), FSM3 (eg central lane marking) or a lane width FSB, which are each described on the basis of an associated traverse PZ1 and PZ2 and to a fused or adapted driving tube aFS, z. B. an adapted Fahrfahrrandrand aFSR1, aFSR2, in the form of an associated adapted polygonal course aPZ3.1 to aPZ3.n by means of a data processing unit 1 are processed. In this case, the parameters of the driving tube FS and the lane information FI of the data processing unit 1 supplied as input variables.

The Parameters of the driving tube FS are, for example, based on Vehicle condition data such as vehicle speed, steering angle, yaw rate, path curvature and / or the like predetermined instantaneous vehicle states determined. Therefore, this is determined based on vehicle condition data driving tube FS also as state or standard Fahrfahr be distinguished. The lane information FI will be over On-board sensors, eg. B. optical recording units (also Additional sensors called) determined.

The data processing unit 1 For example, integral part, eg. B. in the form of an extension of a function for determining driving hoses and evaluation of objects ahead, a driving assistance system, eg. As an ACC system.

in the Below are the driving hoses FS on which the invention is based and their driving-route edges FSR1, FSR2 or the traffic lane information FI, in particular their Lane markings FSM1 to FSM3 in each case on the basis of the associated polygons PZ1.1 to PZ1.n or PZ2.1 to PZ2.n. described. One polyline each PZ1.1 to PZ1.n describes a left or right driving tube edge FSR1, FSR2 of a travel tube FS. If necessary, however, too only one polygon PZ1.1 or PZ2.1 exists, z. In that case, by only one driving rim FSR1 is used or the optical Track detection recognizes only one line of lane markings.

All Processing steps relate equally to both polygons PZ1.1 to PZ1.n and PZ2.1 to PZ2.n. It is assumed that a polygon PZ1.1 to PZ1.n, PZ2.1 to PZ2.n in the vehicle longitudinal direction (x-direction) is monotonic and excludes ambiguities in the x-direction are. A polyline PZ1.1 to PZ1.n must therefore start at a yaw angle stopped by a maximum of 90 ° become. This requirement does not exist in real traffic situations restriction with regard to the function of the considered driver assistance system Based on the traverse FS representing polygon PZ1, and the determined from at least one lane information FI Polygonzuges PZ2 is by means of geometric fusion of the polygons PZ1, PZ2 the fused and adapted polygon aPZ for a adapted driving hose aFS) determined by the consideration the lane information FI opposite the standard travel tube FS is improved and expanded.

In 2 is an example of one according to the method according to 1 determined adapted driving hose aFS shown. In this case, the determined parameters of the driving tube FS are the left and right driving-tube edge FSR1 or FSR2 and the course of curvature FSK of the driving tube FS. The lane information FI is the left lane marking FSM1 or the distance a between the own vehicle 2 and the left lane marking FSM1 has been detected. The parameters of the driving tube FS and the determined lane information FI are described in more detail on the basis of associated polygons PZ1.1, PZ1.2 or PZ2.1. Depending on the type and structure of the data processing unit 1 and the implemented function may, as exemplified in 5 shown, based on one of the polygons PZ1.1 to PZ1.n of the standard travel tube FS (in the 5 , z. B. PZ1.1) and one of the polygons PZ2.1 to PZ2.n the lane information FI (in the 5 , z. B. PZ2.1) of the adapted driving tube aFS and its associated traverse aPZ3.1 to aPZ3.n (in the 5 , z. B. aPZ3.1) can be determined.

at the fusion or processing of the parameter (s) of the driving tube FS and the lane information FI should also be considered be that there are no sudden changes in the adjusted polygon aPZ of one time to the next comes to misinterpretation at downstream object assessment to avoid.

In 3 is the signal flow of the geometric fusion of the standard Fahrschlauch FS associated polygon PZ1 and the lane information FI associated polygonal course PZ2 to the fused or adapted driving tube aFS and its representing matched Po Traverse aPZ shown. In addition, a time-dependent weighting factor G (t) (see, for example, FIG. 6 ), a distance-dependent and / or curvature-dependent weighting function f (x) (see, for example, FIG. 4 and 5 ) be used.

The distance- and / or curvature-dependent weighting function f (x), G (t) and g (x, t) can, as in 4 and 5 represented as a function of the curvature, in particular of the distance a in the y-direction and as a function of the distance or distance x in the x-direction. The value range of the distance-dependent and / or curvature-dependent weighting function g (x, t) for the lane information FI representing the polygon PZ2.1, for example, lies between 0 (no weight results curvature course polygon PZ1.3, "FSK", see 2 , or polygon PZ2.1, "FSM1", see 5 correspondingly) and a weighting value of the weighting function g (x, t) (weighting value yields the polygon PZ2.1, FSM1, see 2 or Traverse aPZ3.1, "aFS", see 5 accordingly), wherein the weighting factor G (t) is between 0 (no weight) and 1 (maximum weight) and may additionally change with time. On the basis of this weighting function g (x, t), lane changes or transitions from one to the other lane can be taken into account. If the course of the lane edges is not known, but only the distance of at least one lane edge of the vehicle longitudinal axis, a defined based on driving dynamics data standard driving hose is used to determine the further course of the optical lane markings. This is preferably done by orthogonal ablation, z. For example, a trajectory with curvature in the form of a tangent vector with a lateral offset via an orthogonal, e.g. B. distance a of the own vehicle to the lane boundary FSR1 or FSR2, half lane width FSB / 2, or via a vector which is orthogonal to the lane FS shown. Close to the sensor vehicle, the quality of the additional sensors is very good and can be highly weighted. The greater the distance x from the sensor vehicle (= own vehicle 2 ), the less trustworthy the additional information is and the weighting must be reduced accordingly. The beginning and end of the transition range from full to low weight may vary.

In 4 By way of example, a weighting function f (x) is shown which linearly drops in the mixing range between the initial distance x _a and the final distance x _{e of} the mixture. Alternatively, the weighting function f (x) can be adjusted in a manner not shown in detail in a view-width-dependent and / or speed-dependent manner. This function f (x) is then multiplied by the time-dependent weighting factor G (t), resulting in the weighting function g (x, t).

The initial distance x _a is the distance between the position of the own vehicle 2 and a forward first position and below the end distance x _e, the distance between the position of the own vehicle 2 and a preceding second position lying after the first position understood. An exemplary embodiment of the course of the adapted polygonal course aPZ3.1 as a function of the weighting function g (x, t) is shown in FIG 5 shown in more detail. Below the starting distance x _a , the adapted traverse aPZ3.1 corresponds to the traverse PZ2.1 of the traffic lane information FI (track generated from additional information such as driving lane width FSB, lane marking FSM1 to FSM3). Above the end distance x _e , the adapted traverse aPZ3.1 corresponds to the traverse PZ1.1 of the trailing edge FSR1 of the standard traction hose FS.

In other words: the 5 shows a schematic representation of an example of a method according to the 3 and 4 determined adapted driving tube aFS. From the polygons PZ1.1, PZ2.1 and the predetermined weighting function g (x, t), an adapted polygon aPZ3.1, the below a initial distance x _a the polygon PZ2.1 and above an end distance x _e the polygon PZ1.1 equivalent. In the mixing area between the starting distance x _a and the end distance x _e , a transition takes place. A separate consideration is required if the additional lane information FI is limited to the latitude and the position within a (eg visually recognized) lane FSM1. At first, no independent polygon PZ2.1 emerges from this additional information. In this case, the curvature of the polygonal course PZ1.1 can be transferred to the additional information and thus derive a polygon PZ2.1, which can be incorporated into the geometric fusion described above.

• – der Sichtweite der fahrzeugeigenen Sensorik, z. B. optische Spurerkennung,
• – der Geschwindigkeit, z. B. einer geschwindigkeitsabhängigen Überblendlänge mit L = x_e – x_a, und/oder
• – potenziellen Schnittpunkten der Polygonzüge PZ1.n und PZ2.n

vorgegeben werden und zeitlich variieren.The parameters initial distance x _a and final distance x _e and the weighting values of the distance-dependent and / or curvature-dependent weighting function f (x) can also depend on:

• - The visibility of the vehicle's sensors, z. B. optical lane detection,
• - the speed, z. As a speed-dependent cross-fading length with L = x _e - x _a , and / or
• - potential intersections of the polygons PZ1.n and PZ2.n

be specified and vary in time.

The generation of a possible polygon PZ2.1 by orthogrates queries for missing curvature information is in 2 exemplified. This case has particular then practical meaning, if a curvature can not be detected at all or only badly by the additional sensor. This is true, for example, for lane detection systems that are within sight of the immediate area of their own vehicle 2 (= System vehicle) are limited and next to the width of the lane FSB and the location of the own vehicle 2 in the traffic lane FS are unable to detect the distance-dependent curvature course FSK of the traffic lane FS.

In 6 is shown schematically in block diagram an embodiment of a method for determining a time-varying weighting factor G (t). A special consideration is necessary in case of failure of the vehicle's own sensors (also called additional sensors). In order to avoid discontinuities in the description of the fused or adapted driving hose aFS with the traverse aPZ, in case of failure of the additional sensors of the weighting factor G (t) via a delay element 3 , z. B. a low-pass, a stationary weighting factor G of 0 (with polygon PZ2 without influence) can be set stepwise or continuously. Similarly, with a valid polygon PZ2, the weighting factor G is set to a stationary value of 1. Through the low-pass filter 3 a time-dependent weighting factor G (t) is generated, which runs continuously between the weighting values 0 and 1.

An example of this is in 7 shown, where at the times a, b and c, a change in the validity of the polygon PZ2 takes place and the weighting factor G (t) decreases or increases accordingly. At time a, the optical lane mark detection fails, and the weighting factor G (t) decreases. At time b, the optical lane marker recognition starts again and the weighting factor G (t) increases again. At time c, the optical lane marking is canceled again, the weighting factor G (t) decreases again after a short period of time.

If a jumpable polygon PZ2 is determined from the lane information FI, then, as in 8th shown by filtering these sudden changes in the polygon PZ2 (n) by means of a filter element 4 balanced and compensated. For this purpose, the individual nodes K1 to Km of the polygon PZ2 (n) by means of the designed as a low-pass filter element 4 smoothed. Setpoint or input value of the filter element 4 is, if valid, the polygon PZ2 (n). If the polyline PZ2 (n) is invalid, the previous filtered and valid polyline ptPZ2 (n-1) is kept as the setpoint or input value. If the polyline PZ2 (n) is continuously valid, the filtered polyline pt2 (n) approaches the polyline PZ2 (n) in a stationary manner. With continuously invalid polyline PZ2 (n), the filtered polyline gefPZ2 (n) remains constant, where - as described above - its weighting factor G goes to 0 and thus the influence of the polygon PZ2 (n) on the adjusted polyline aPZ3 (n) steadily decreases becomes.

9 shows an example of the low-pass filtering of the individual polygon nodes or support points K1 to Km of the filtered polygonal line gefPZ2 (n). The individual polygon nodes K1 to Km of the previous, filtered polygonal line gefPZ2 (n-1) are each drawn in the direction of the polygon nodes K1 to Km at the same x-height of the current filtered polygonal line gefPZ2 (n). The speed at which the polygon nodes approach the target node may be determined by the filter time constant Tv of the filter element 4 be determined. Analogous to this filtering, the signal must also be smoothed if there is a sudden change in the visibility information. By means of an appropriately tuned low-pass filter, the parameters initial distance x _a and final distance x _{e can be} adapted.

The described method of geometric fusion of driving hoses FS and lane information lets fi to any number of driving hoses FS to be merged and increase lane information FI. It just has to be on it ensure that the sum of all weighting functions equals 1.

Claims (6)

Method for determining a path of movement of a vehicle ( 2 ) extending travel hose (FS) within which the vehicle ( 2 ) is highly likely to travel on a lane, wherein the travel tube (FS) on both sides of the trajectory of each Fahrfahrrandrand (FSR1, FSR2) is limited, characterized in that at least one of the driving tube edges (FSR1 or FSR2) and additionally at least one lane information ( FI), in particular a lane marking (FSM1, FSM2, FSM3) representing the driving lane (FS) and the relevant driving-route edge (FSR1, FSR2), by means of which an adapted driving route (aFS), in particular an adapted driving-route edge (aFSR1, aFSR2), is determined. , is determined. Method according to claim 1, characterized in that in that the driving-route edge (FSR1, FSR2) is based on at least one vehicle parameter and the lane information (FI) based on sensors detected Lane data be true and each in the form of a polygon (PZ1.1 to PZ1.n, PZ2.1 to PZ2.n). A method according to claim 1 or 2, characterized characterized in that transitions between tracks or these descriptive polygons (PZ1.1 to PZ1.n, PZ2.1 to PZ2.n) are determined based on at least one weighting function (G (x), G (t)). Method according to claim 3, characterized that the weighting function (g (x, t)) and their weighting values visually dependent and / or depending on speed be determined. Method according to claim 3 or 4, characterized that the weighting function (G (t)) by means of a low-pass filter smoothed becomes. Device for determining a trajectory of a vehicle ( 2 ) extending travel hose (FS) within which the vehicle ( 2 ) is highly likely to travel on a lane, wherein the travel tube (FS) on both sides of the trajectory of each Fahrfahrrandrand (FSR1, FSR2) is limited, characterized in that at least one means for determining at least one of the driving tube edges (FSR1, FSR2) and at least one means for determining at least one lane information (FI), in particular a lane marking (FSM1 to FSM3) representing the traffic lane and the relevant driving-route edge (FSR1, FSR2), and at least one means for determining a driving-route edge (FSR1, FSR2) and the determined lane information (FI) adapted driving lane (aFS) are provided.