WO2014111408A1 - Prévision d'une manœuvre de conduite d'un véhicule - Google Patents

Prévision d'une manœuvre de conduite d'un véhicule Download PDF

Info

Publication number
WO2014111408A1
WO2014111408A1 PCT/EP2014/050673 EP2014050673W WO2014111408A1 WO 2014111408 A1 WO2014111408 A1 WO 2014111408A1 EP 2014050673 W EP2014050673 W EP 2014050673W WO 2014111408 A1 WO2014111408 A1 WO 2014111408A1
Authority
WO
WIPO (PCT)
Prior art keywords
vehicle
distance
driving maneuver
reference point
speed
Prior art date
Application number
PCT/EP2014/050673
Other languages
German (de)
English (en)
Inventor
Felix Klanner
Horst KLÖDEN
Martin LIEBNER
Original Assignee
Bayerische Motoren Werke Aktiengesellschaft
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Bayerische Motoren Werke Aktiengesellschaft filed Critical Bayerische Motoren Werke Aktiengesellschaft
Priority to EP14700989.8A priority Critical patent/EP2945825A1/fr
Publication of WO2014111408A1 publication Critical patent/WO2014111408A1/fr

Links

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W40/00—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models
    • B60W40/10—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models related to vehicle motion
    • B60W40/105—Speed
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W40/00—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models
    • B60W40/10—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models related to vehicle motion
    • B60W40/107—Longitudinal acceleration
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W50/00—Details of control systems for road vehicle drive control not related to the control of a particular sub-unit, e.g. process diagnostic or vehicle driver interfaces
    • B60W50/0097—Predicting future conditions
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/16—Anti-collision systems
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/16—Anti-collision systems
    • G08G1/166—Anti-collision systems for active traffic, e.g. moving vehicles, pedestrians, bikes
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/16—Anti-collision systems
    • G08G1/167—Driving aids for lane monitoring, lane changing, e.g. blind spot detection
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B60—VEHICLES IN GENERAL
    • B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
    • B60W2552/00—Input parameters relating to infrastructure
    • B60W2552/30—Road curve radius

Definitions

  • the invention relates to a method for predicting a driving maneuver of a vehicle and to a corresponding device,
  • a driving maneuver may be, for example, the passage of an intersection, a stop at a stop line, a turn and a turn with stopping at an overpass.
  • a driving maneuver may be, for example, the passage of an intersection, a stop at a stop line, a turn and a turn with stopping at an overpass.
  • other road users may become relevant and be considered for warnings.
  • Driver Intent Interface at Urban Intersections using the Intelligent Driver Model IEEE Intelligent Vehicle Symposium, Jun. 2012, pp.
  • a method for predicting driving maneuvers is presented, according to which the trajectories and speed curves of test drives are displayed
  • desired speeds and their progressions are calculated for individual driving maneuvers, on the basis of which desired speeds can be determined on subsequent cruises at the considered intersection using then measured vehicle speeds
  • the method proposed in the abovementioned publication is based on the records of test runs carried out and limits the application of the method in this respect to such sections of track for which there are recorded trajectories with velocities. It is therefore the task of a person skilled in the art to make the method available also for sections and in particular intersections for which no speed profiles have been recorded.
  • a method of predicting a driving maneuver of a vehicle comprises: detecting the speed of the vehicle at first and second distances of the vehicle along a path from a reference point; Wherein the vehicle is at a first time at the first distance from the reference point and at a second time at the second distance; Establishing at least a first and a second estimate of the speed of the vehicle for the second time and the distance of the vehicle from the reference point for the second time; Wherein the first and second estimates are respectively based on the speed of the vehicle at the first distance, the first distance from the reference point, and the second time; The first estimate also being on a first in advance certain course of the desired speed of the vehicle based on the assumption of a first driving maneuver as a function of the distance of the vehicle from the reference point based; Where the second estimate is based on an assumption of a second driving maneuver; Determining a measure of the probability of occurrence of an estimate for each estimate taking into account the detected speed of the vehicle at the second distance from the reference point and the second distance itself;
  • the determination of the first course of the desired speed is based on the geometry of the path. A record of test track speed characteristics is not used for this. In this way, runs of desired speeds can also be created for routes or intersections for which there are no records of speed profiles. However, the geometry of the first path can still be obtained by recording and statistical evaluation of test drives. Likewise, however, it is also conceivable to obtain the geometries of the paths for the first driving maneuver from geometries of route sections or intersections existing in maps for determining the first and second course of the desired speeds. If necessary, highly accurate maps must be used. However, no centimeter-accurate description of the complete topology of the intersection or curvature profiles obtained from the path is needed for the path. The suggested method is thus robust to the quality of the vehicle localization.
  • the first path may be the trajectories typically traveled by a vehicle at an intersection and possibly a few meters, 10 meters, 20 meters to hundreds of meters describe before and after. It can also be provided that for a whole
  • Map the exact gradients of paths are given, including the first path. Then this source can be used for the first path, which may represent a section of a specified path.
  • the prediction of the driving maneuver based on the determined measures may consist in the selection of the driving maneuver having the highest measure.
  • the method may include outputting the measure determined in each case for the first and the second driving maneuvers.
  • This output of the considered driving maneuvers and the determined measures can be used by a driver assistance system to provide warnings. It may occur in the process that although the measure of a driving maneuver is higher than the other or the other measure, but the highest and another or more metrics close together. In such a case, the issuing of warnings by a driver assistance system may be expedient for the closely spaced driving maneuvers.
  • the determined measures can be used to prioritize warnings of possible conflict situations with regard to the probability of occurrence of the associated maneuver.
  • the highest dimensioned warning is given the highest priority and thus output to the driver.
  • it may happen that a maneuver with a smaller measure than the highest measure is prioritized higher than the maneuver with the highest measure. In particular, this is the case if the associated conflict situation is to be assessed as particularly critical.
  • the measure of the likelihood of an estimate can be a decision-relevant share (term) or the relevant share (or term) of a complete calculation of an absolute probability of occurrence; that is, the part of the calculation by which the estimates can be ranked, at least approximately or with sufficient accuracy. Under certain assumptions, the measure may even be the probability of occurrence itself. The measure may be a so-called score or an abnormal probability of occurrence.
  • a method for predicting a probability of occurrence can be a decision-relevant share (term) or the relevant share (or term) of a complete calculation of an absolute probability of occurrence; that is, the part of the calculation by which the estimates can be ranked, at least approximately or with sufficient accuracy. Under certain assumptions, the measure may even be the probability of occurrence itself. The measure may be a so-called score or an abnormal probability of occurrence.
  • Driving maneuvers of a vehicle Detecting the speed of the vehicle at a first distance and the acceleration of the vehicle at a second distance of the vehicle along a path from a reference point, wherein the first and second distances are in particular equal; Wherein the vehicle is at a first time at the first distance from the reference point and at a second time at the second distance, wherein the first time and the second time are in particular equal; Creating at least a first and a second estimate of the acceleration of the vehicle for the second time; Wherein the first and second estimates are respectively based on the speed of the vehicle at the first distance, the first distance from the reference point, and the second time; Wherein the first estimate is further based on a first predetermined course of the desired speed of the vehicle for the assumption of a first driving maneuver in dependence on the distance of the vehicle from the reference point; Where the second estimate is based on an assumption of a second driving maneuver; Determining a measure of the probability of occurrence of an estimate for each estimate taking into account the detected acceleration of the vehicle at the second distance from the reference
  • the speed and the distance are not calculated as an estimate, but the acceleration. Based on the estimation results, the measure is then calculated and the prediction made.
  • the estimation can also be done in a "point" by simultaneously measuring speed, distance and acceleration, at which point the estimates are determined and the prediction made.
  • the measure for each estimate may also be calculated from the mean of the measures in the event that the driver actually performs the assumed maneuver over a preferably selected period of time.
  • a plurality of estimates can be carried out over a period of time and in each case the measures for each estimate can be determined. These measures can then be averaged for all estimates for an assumption of a maneuver. The prediction can then be based on this averaging. The averages can be output together with the underlying assumptions on the driving maneuver.
  • the prediction of the driving maneuver based on the determined measures may consist in the selection of the driving maneuver having the highest measure. Furthermore, the method may include outputting the measure determined in each case for the first and the second driving maneuvers. This output of the considered driving maneuvers and the determined measures can be used by a driver assistance system to provide warnings. It may happen in the process that, although the measure for a maneuver is higher than the others or the other measure, but the highest and a further or further measures close to each other. In such a case, issuing warnings by a driver assistance system may be useful for the closely spaced driving maneuvers.
  • the determined measures can be used to phorise warnings of possible conflict situations with regard to the probability of their associated maneuver.
  • the highest dimensioned warning is given the highest priority and thus output to the driver.
  • it may happen that a maneuver with a smaller measure than the highest measure is prioritized higher than the maneuver with the highest measure. In particular, this is the case if the associated conflict situation is to be assessed as particularly critical.
  • the measure of an occurrence probability of an estimate can be a decision-relevant part (term) or the decision-relevant part (or term) of a complete calculation of an absolute probability of occurrence, ie the part of the calculation by which the estimates can be ranked, at least approximately or with sufficient accuracy. Under certain assumptions, the measurement number even the probability of occurrence itself.
  • the measure may be a so-called score or an abnormal probability of occurrence.
  • the geometry of the first and the second path is the two-dimensional course of the respective path.
  • the first and the second predetermined course of the desired speed was determined based on the curvature of the first or second path.
  • the curvature can be specified in sections using so-called smooth circular arc splines, as described in the above publication "Driver Intent Interface at Urban Intersections using the Intelligent Driver or de.
  • the first and the second predefined course of the desired speed were determined based on the geometry of a section of the preceding first or second path that would still have to be traveled by a vehicle.
  • the desired speed thus depends on the course of the path that is still to drive for a vehicle that is on this path.
  • the first and the second course of the desired speed was determined based on a first driver model.
  • the driver model specifies one or more of the following: the maximum amount of lateral acceleration, the maximum speed, the minimum derivative of the desired speed after the distance. These specifications decisively influence the course of the desired speed.
  • the specification of the maximum lateral acceleration limits the speed with which a curve is traversed. For example, a sporty driver will accept higher lateral acceleration than a comfort-oriented driver. Likewise, different drivers may prefer other maximum speeds.
  • the course of the desired speed in the approach of the curve that is to say before a turning operation, is essentially influenced by the minimum amount of the derivative of the desired speed after the distance, that is to say the distance traveled by the vehicle.
  • a smaller minimum amount represents a rather late but strong braking before a turn, while a higher minimum amount of Derivative represents a rather gentle braking or negative acceleration or low negative accelerations.
  • the property "minimal" of values herein denotes the negative number ( ⁇ 0) whose magnitude is greatest, for example, given the values -2 and -5, the -5 is the minimum value
  • the further or each estimate of the group is further based on a third or each further predetermined course of the desired speed of the vehicle as a function of the distance of the vehicle from the reference point for the first assumption Driving maneuver based, wherein the third or each further course of the desired speed is determined based on the geometry of the first path to be traveled by the vehicle from the reference point for the first assumption of the driving maneuver, wherein the third or each further course of the desired speed was determined based
  • the method further comprises: creating another or a group of further estimates of the acceleration of the vehicle for the second time; And the one more or each estimate of the group is based on the speed of the vehicle at the first distance, the first distance from the reference point, and the second time; Wherein the one or more estimates of the group are further based on a third or each further predetermined course of the desired speed of the vehicle as a function of the distance of the vehicle from the reference point for the first assumption to the driving maneuver and in each case on a model assumption for the maximum longitudinal acceleration based, where multiple estimates can be based on the same model assumption; wherein the third or each further course of the desired speed is determined based on the geometry of the first path to be traveled by the vehicle from the reference point for the first assumption on the driving maneuver, wherein the third or each further course of the desired speed is based on a second or respectively further Driver model, each of which specifies one or more of the following, the maximum amount of lateral acceleration, the maximum speed
  • a device comprises programmable computer means, in particular a computer, wherein the apparatus is configured to execute one of the methods presented above.
  • the apparatus may also include interfaces adapted to receive the speed, distance and / or acceleration of the vehicle.
  • this reception is understood as the detection of the speed, the distance or the acceleration.
  • interfaces to driver assistance systems can be provided.
  • a computer program includes instructions for performing one of the methods presented above. The computer program may cause a computer to execute the presented procedures while executing the instructions.
  • FIG. 1a shows the view of an exemplary intersection.
  • FIG. 1b shows the geometry of multiple trajectories at the exemplary intersection according to one embodiment.
  • FIG. 1c shows by way of example two determined paths for the passage through the intersection and for the right turn of the vehicle.
  • FIGS. 2 a to 2 c show, by way of example, steps in the development of the course of the desired speed for a path according to one exemplary embodiment.
  • FIG. 2d shows a course of a desired speed according to the calculation according to FIGS. 2a to 2c according to an exemplary embodiment.
  • FIG. 3 shows in a flow chart the prediction of a driving maneuver according to an exemplary embodiment.
  • FIG. 5 shows in a flowchart the prediction of a driving maneuver according to an exemplary embodiment.
  • Fig. 1a shows a view of an exemplary intersection.
  • a vehicle shown at the front of the picture
  • makes a first maneuver by holding the right-turn (indicated by the arrow) on a cycle path to let cyclists pass.
  • Another driving maneuver would be if the vehicle turns right without stopping at the bicycle crossing, crosses the intersection without turning off, or stops at the stop line.
  • FIG. 1b shows the geometry of multiple trajectories at the exemplary intersection according to one embodiment.
  • Each trajectory corresponds to the track a test drive of a vehicle for which the positions of the vehicle have been determined by means of GPS.
  • an averaging can be formed by circular arc splines in order to obtain a path for this driving maneuver.
  • Fig. 1c shows two determined paths, one for the passage through the intersection and one for the right turn of the vehicle. Both paths are initially superimposed.
  • sections of the path are considered as a circular arc.
  • the sections are identified in FIG. 1c by the long lines perpendicular to the path.
  • For the path of the right turn some of the circle centers are drawn, whose arcs form the averaging.
  • Thicker markers perpendicular to the path indicate the beginning and end of the paths.
  • FIG. 2 a shows the course of the curvature k 1 (s) obtained from the circular arc splines for a turning maneuver.
  • the curvature for a distance s is the reciprocal of the radius of the respective circular arc.
  • the vehicle executes cornering in the turn-off maneuver, especially at a distance of between 60 and 90 m. The measurement of the distance begins about 60 m before the actual curve.
  • the curvature is then first smoothed, k 2 (s), as shown in Fig. 2b.
  • a moving average of the curvature ki can be formed.
  • the actual course of the curvature can be approximated, since roads typically have a continuous curvature (without cracks).
  • a smoothing of the curvature can take into account the cutting of curves.
  • Fig. 2c shows the velocity V j (s) calculated from the smoothed curve k 2 taking into account three different driver models (represented by the solid line, the dashed line and the dotted line in Fig. 2c).
  • the exemplary values used here for the parameters v maxJ and a max. I at j are given in Table 1.
  • the velocity Vj (s) is then calculated on the basis of the parameters according to the following formula:
  • Fig. 2d shows the course of the desired speed.
  • the desired speed takes account of different degrees of deceleration or braking before the curve.
  • the derivative of the velocity Vj (s) after the distance s, g min; j is limited downwards, to different values depending on the driver model. The values used are given in Table 1.
  • the derivation of the direction vector of a path according to the path or the distance could also be used as the basis for calculating the course of desired speeds or the amount thereof.
  • the curvature can be considered as the second derivative of the transverse coordinate after the longitudinal coordinate (in the direction of a path).
  • the velocities Vj (s) and desired velocities Uj (s) with respect to Figs. 2c and 2d are calculated for a turn off maneuver without stopping at a pedestrian or cycle crosswalk. When it is desired to provide progressions of desired speeds for a turn maneuver with a pedestrian or bicycle crossing, this is taken into account in determining an estimate, as described below.
  • FIG 3 shows in a flow chart the prediction of a driving maneuver according to an exemplary embodiment.
  • the distance s () and the speed v (t t ) of the vehicle is determined from a reference point.
  • the position of the vehicle ascertained for example, via satellite navigation is mapped to a position on a path and the distance of the beginning of the path, that is the reference point, to the position shown is determined.
  • three estimates a, b, and c are computed, which are the velocities Vestim, a (t 2 ), v esflm , b (t 2 ), v esBm, c (t 2 ), and distances s estimi a (t 2 ) , s estim. b (t 2 ), Sesüm, c (t2) has the vehicle at a second time.
  • the first estimate a is based on the assumption of a first driving maneuver, in this embodiment, a right turn without stopping at a bicycle transition (hereinafter also called 13).
  • the second estimate b is based on the assumption of a second driving maneuver, in this embodiment a straight passage through the intersection (also referred to below as 11).
  • the third estimate c assumes the assumption of a third driving maneuver versed, in this embodiment, a right turn with holding at a bicycle transition (hereinafter also called 14).
  • the first estimate a is further based on a predetermined course of the desired speed u (s) for the first driving maneuver.
  • the so-called Intelligent Driver Model (IDM) described in the publication "Driver Intent Interface at Urban Intersections using the Intelligent Driver Model "is also used and explained after this, the derivative of the velocity v determined estirn (t) after the time according to the following formula.:
  • the variable d denotes the distance of the vehicle to an obstacle, for example a preceding vehicle.
  • the maximum longitudinal acceleration ai ong . k is determined by an applicable model assumption. In In a simple variant of the invention, only one model is used and, for example, the value 4 m / s 2 is used for ai ong, k . In a further development, different values could also be used for the other parameters in Table 1, depending on the model assumption. Based on the speed and the distance at the first point in time ti, the speed of the vehicle and the distance of the vehicle at time t 2 can then be calculated by means of the IDM and the known course of the desired speed, v deslim, a (t 2 ) and s estim , a (t 2 ).
  • the second estimate b is determined, this time for an assumption of a second driving maneuver.
  • a predetermined course of the desired speed can also exist. This can be, for example, the course as it would result in a transit (straight ahead) of the vehicle.
  • This second predetermined course of the desired speed can also be determined based on the geometry of a path for the second driving maneuver (passage). The determination of the speed and distance at the second time can then be done in the same way as for the estimate a, namely using the IDM.
  • the third estimate c is determined in a modified way.
  • the third estimate c is generally based on the course of the desired speed for the assumption that the vehicle is making a right-turn without stopping at the bicycle crossing, and the IDM.
  • a virtual standing obstacle is placed such that the vehicle stops at the bicycle crossing or generally at a breakpoint.
  • the obstacle is "taken away” as soon as the vehicle reaches the obstacle.
  • the obstacle is placed 1 to 2 meters after this point, ie around d 0 behind the obstacle.
  • This obstacle is taken into account by the IDM in calculating the speed to decelerate the vehicle toward the obstacle and stand on the obstacle comes. This is taken into account by the variables and expressions d and d * of the IDM.
  • the third estimate c could also be referred to as the second estimate b and the second estimate b as the third estimate c.
  • the measure regarding the probability of occurrence for an estimate takes into account in the present example only one decision relevant term of a calculation of an absolute probability of occurrence.
  • the measure in the present example is the probability density of a normal distribution for the estimation:
  • index i stands for one of the estimates a, b or c.
  • a s and ⁇ ⁇ empirical values can be determined. Typically, these are in the range of 0.6 m / s to 1, 6 m / s, preferably 1, 2 m / s.
  • the estimate with the highest measure is selected and the corresponding assumption (turn without or with hold, passage) is output as a prediction of the driving maneuver.
  • the driving maneuver for whose estimation the highest measure was determined is identified as driving maneuver and thus predicted.
  • Estimates a, b, and c are output together with the measures determined for the respective estimates.
  • the prior knowledge can be expressed by a probability.
  • the inflow of the prior knowledge can also be done by multiplying the measure with the probability of the prior knowledge.
  • the measures and / or their multiplication with the probability for the prior knowledge can be set in relation to each other, either by calculating an absolute probability of occurrence or, for example, by dividing each measure by the sum of all measures.
  • the result for each estimate is thus normalized. This can be of importance for driver assistance systems whose function (for example the output of warnings) depends on absolute threshold values.
  • the results for each estimate and the associated driving maneuvers can be output.
  • variants of the estimation can be calculated for every assumption of a driving maneuver. These variants represent different driver models according to Table 1 and model assumptions for the maximum ionic acceleration k .
  • the case is considered that for each of the maneuvers right-turn without holding, right-turn with hold and passage variants of the estimate are calculated in accordance with the following Table II.
  • FIGS. 1 b and 1 c Furthermore, reference is made in this development to the intersection according to FIGS. 1 b and 1 c.
  • Estimation results for the further training are represented in FIG. 4.
  • Each estimate is represented by its own line.
  • For each estimate its measure of probability of occurrence is calculated, in the same way as shown in relation to FIG.
  • Estimates that apply to the same assumption for the maneuver have the same linetype.
  • the solid lines represent the estimates for accepting a transit.
  • the dashed lines represent the estimates for assuming a right-turn without holds, and the dotted lines represent the estimates for right-turn with holds.
  • the measures of the estimates for a Considering the assumption for driving maneuver together and calculating a measure of the probability of occurrence of this maneuver, for example by forming an average of the measures for a driving maneuver.
  • the driving maneuver, for the estimation of which the highest measure was determined is identified as driving maneuver and thus predicted.
  • the considered driving maneuvers of the estimates a, b and c are output together with the respective estimated values for the corresponding estimates.
  • Fig. 4 further shows the probabilities P that the assumption of the driving maneuver is correct, taking into account the detected speed and distance of the vehicle.
  • the probabilities P are calculated by normalizing the summation of the probability densities.
  • the assumption for driving maneuver 13 is the most probable and is therefore used as a precursor. say. Since the stop line has already been run over at a distance of 16 m, the probability of 12 is absent.
  • the distance s (ft) and the speed v (t t ) of the vehicle are determined from a reference point.
  • the position of the vehicle determined for example, via satellite navigation is mapped to a position on a path and the distance of the beginning of the path, that is the reference point, to the position shown.
  • the speed v (t 2 ) of the vehicle that is used in the calculation of the measurement numbers of estimates, but the acceleration at time t 2 .
  • the first estimate a is based on the assumption of a first driving maneuver, in this embodiment, a right-turn without stopping at a bicycle crossing.
  • the second estimate b is based on the assumption of a second driving maneuver, in this embodiment a straight passage through the intersection.
  • the third estimate c is based on the assumption of a third driving maneuver, in this embodiment, a right turn with stopping at a bicycle crossing.
  • the first estimate a is based on a predetermined course of the desired speed u (s) for the first driving maneuver.
  • the acceleration aestim. a (t 2 ) uses the so-called Intelligent Driver Model (IDM), which is also used and explained in the above publication.
  • IDM Intelligent Driver Model
  • the acceleration of the vehicle at time t 2 can then be calculated using the IDM and the known curve of the desired speed, a estim, a (t 2 ). This can be done during the time interval from ti to t 2 in consideration of the gradient of v ⁇ v, d). For this purpose, recourse may be had to known methods. For example, it is possible to calculate for discrete temporal intermediate steps between to t 2 v (v, d) and thus to detect the travel of the vehicle and its respective speed. The estimate can then be arrived at by summation. This estimate a then holds for the assumption that the vehicle is making the first driving maneuver, for example a right turn without stopping on a bicycle over road.
  • the second estimate b is determined, this time for an assumption of a second driving maneuver.
  • a predetermined course of the desired speed can also exist. This can be, for example, the course as it would result in a transit (straight ahead) of the vehicle. This second predetermined course of the desired speed can also be determined based on the geometry of a path for the second driving maneuver (passage). The determination of the acceleration at the second time can then be done in the same way as for the estimate a, namely using the IDM.
  • the third estimate c is determined in a modified way.
  • the third estimate c is generally based on the course of the desired speed for the assumption that the vehicle is a right-turn without stopping at the bicycle crossing executes, and the IDM.
  • a virtual standing obstacle is placed so that the vehicle stops at the bicycle overpass or generally at a breakpoint.
  • the obstacle is "taken away” as soon as the vehicle reaches the obstacle
  • the obstacle is placed 1 to 2 meters after this point, ie around d 0 behind the obstacle, this obstacle being taken into account by the IDM in calculating the acceleration that the vehicle decelerates towards the obstacle and comes to stand on the obstacle, this is taken into account by the variables and expressions d and d * of the IDM
  • the third estimate c is also the second estimate b and the second estimate b could be referred to as third estimate c.
  • the measures for the estimates a, b and c are determined.
  • the actual acceleration a (t 2 ) is compared with the estimated accelerations at time t 2 .
  • the metric for an estimate is calculated using the probability density formula:
  • the index i stands for one of the estimates a, b or c.
  • Empirical values are determinable for ⁇ ⁇ , a typical value being from 0.8 m / s 2 to 1.4 m / s 2 , advantageously 1.2 m / s 7 .
  • the estimate with the highest measure is selected and the corresponding assumption (turn without or with hold, passage) is output as a prediction of the driving maneuver.
  • the driving maneuver for whose Estimate the highest measure was determined is identified as driving maneuvers and thus predicted.
  • the considered driving maneuvers of the estimates a, b and c are output together with the respective estimated values for the corresponding estimates.
  • each t m is from the time interval [t ⁇ , t 2 j and M is the number of times over which to averag.
  • the prior knowledge can be expressed by a probability.
  • the inflow of the prior knowledge can also be done by multiplying the measure with the probability of the prior knowledge.
  • the measures and / or their multiplication with the probability of the prior knowledge can be set in relation to each other, either by calculating an actual probability of occurrence or, for example, by dividing each measure by the sum of all measures.
  • the result for each estimate is thus normalized. This may be important for driver assistance systems whose function (eg, issuing warnings) depends on absolute thresholds.
  • the results for each estimate and the associated driving maneuvers can be output.
  • variants of the estimation can be calculated for each assumption of a driving maneuver. These variants represent different driver models according to Table 1 and model assumptions for the maximum longitudinal acceleration a iong , k . In this development, the case is considered that for each of the maneuvers right- ⁇ b bend without holding, right-turn with hold and transit variants of the estimate are calculated in accordance with the following Table II.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)
  • Transportation (AREA)
  • Mechanical Engineering (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Human Computer Interaction (AREA)
  • Control Of Driving Devices And Active Controlling Of Vehicle (AREA)
  • Traffic Control Systems (AREA)

Abstract

L'invention concerne un procédé pour prévoir une manœuvre de conduite d'un véhicule, comprenant : la détection de la vitesse du véhicule à une première et à une seconde distance du véhicule le long d'une trajectoire à partir d'un point de référence, le véhicule se trouvant à une première distance du point de référence à un premier instant et à une seconde distance à un second instant ; établissement d'une première et d'une seconde évaluation de la vitesse du véhicule au second instant et de la distance du véhicule par rapport au point de référence pour le second instant, la première et la seconde évaluation reposant respectivement sur la vitesse du véhicule à la première distance, la première distance par rapport au point de référence, et le second instant, la première évaluation reposant également sur une première allure, prédéterminée, de la vitesse souhaitée du véhicule pour la compatibilité avec une première manœuvre de conduite en fonction de la distance du véhicule par rapport au point de référence, la seconde évaluation reposant sur une compatibilité avec une seconde manœuvre de conduite ; détermination d'une valeur de la probabilité de survenue d'une évaluation pour chaque évaluation en prenant en considération la vitesse détectée du véhicule à la seconde distance par rapport au point de référence et la seconde distance elle-même ; prévision de la manœuvre de conduite en fonction de la valeur déterminée. L'invention se caractérise en ce que la première allure prédéterminée de la vitesse souhaitée pour la première manœuvre de conduite est déterminée en fonction de la géométrie d'une première trajectoire pour la première manœuvre de conduite. L'invention concerne également un procédé pour prévoir une manœuvre de conduite d'un véhicule en fonction de l'accélération du véhicule.
PCT/EP2014/050673 2013-01-18 2014-01-15 Prévision d'une manœuvre de conduite d'un véhicule WO2014111408A1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
EP14700989.8A EP2945825A1 (fr) 2013-01-18 2014-01-15 Prévision d'une manoeuvre de conduite d'un véhicule

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
DE102013200724.8A DE102013200724A1 (de) 2013-01-18 2013-01-18 Vorhersage eines Fahrmanövers eines Fahrzeugs
DE102013200724.8 2013-01-18

Publications (1)

Publication Number Publication Date
WO2014111408A1 true WO2014111408A1 (fr) 2014-07-24

Family

ID=49999904

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/EP2014/050673 WO2014111408A1 (fr) 2013-01-18 2014-01-15 Prévision d'une manœuvre de conduite d'un véhicule

Country Status (3)

Country Link
EP (1) EP2945825A1 (fr)
DE (1) DE102013200724A1 (fr)
WO (1) WO2014111408A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110954894A (zh) * 2018-09-25 2020-04-03 通用汽车环球科技运作有限责任公司 对象的速度估计

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE102013203908A1 (de) 2013-03-07 2014-09-11 Bayerische Motoren Werke Aktiengesellschaft Schätzung der zukünftigen Geschwindigkeit eines Fahrzeugs
DE102013212359A1 (de) 2013-06-27 2014-12-31 Bayerische Motoren Werke Aktiengesellschaft Vorhersage von Fahrpfaden eines Fahrzeugs
DE102013218497A1 (de) 2013-09-16 2015-03-19 Bayerische Motoren Werke Aktiengesellschaft Vorhersage von Fahrpfaden eines Fahrzeugs

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120203440A1 (en) * 2009-12-18 2012-08-09 Toyota Jidosha Kabushiki Kaisha Travel control device
EP2537727A1 (fr) * 2011-06-22 2012-12-26 Volvo Car Corporation Procédé d'estimation du profil de vitesse pour un véhicule

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE102011103096B4 (de) * 2011-06-01 2022-03-31 Audi Ag Kraftfahrzeug umfassend ein Fahrerassistenzsystem mit einer Steuerungseinrichtung zur automatischen Längsführung
DE102011113019A1 (de) * 2011-07-14 2012-05-10 Daimler Ag Verfahren zur Ermittlung und Bewertung von Gefahren einer Situation zwischen zumindest zwei Verkehrsteilnehmern in einem Straßenkreuzungsbereich und Verfahren zur Unterstützung eines Fahrers beim Führen eines Fahrzeugs

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120203440A1 (en) * 2009-12-18 2012-08-09 Toyota Jidosha Kabushiki Kaisha Travel control device
EP2537727A1 (fr) * 2011-06-22 2012-12-26 Volvo Car Corporation Procédé d'estimation du profil de vitesse pour un véhicule

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
"Driver Intent Interface at Urban Intersections using the Intelligent Driver Model", IEEE INTELLIGENT VEHIDE SYMPOSIUM, June 2012 (2012-06-01), pages 1162 - 1167
MARTIN LIEBNER ET AL: "Driver intent inference at urban intersections using the intelligent driver model", INTELLIGENT VEHICLES SYMPOSIUM (IV), 2012 IEEE, IEEE, 3 June 2012 (2012-06-03), pages 1162 - 1167, XP032452922, ISBN: 978-1-4673-2119-8, DOI: 10.1109/IVS.2012.6232131 *
See also references of EP2945825A1 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110954894A (zh) * 2018-09-25 2020-04-03 通用汽车环球科技运作有限责任公司 对象的速度估计
CN110954894B (zh) * 2018-09-25 2024-01-02 通用汽车环球科技运作有限责任公司 对象的速度估计

Also Published As

Publication number Publication date
DE102013200724A1 (de) 2014-07-24
EP2945825A1 (fr) 2015-11-25

Similar Documents

Publication Publication Date Title
DE102015111535B4 (de) Algorithmus zur genauen Krümmungsschätzung für die Bahnplanung von autonom fahrenden Fahrzeugen
DE102015114465B4 (de) Verfahren zur Wegplanung für ein Ausweichlenkmanöver
DE102013105046B9 (de) Zielspurauswahlverfahren mittels Navigationseingabe in Straßenwechselszenarien
EP2561419B1 (fr) Procédé pour déterminer le parcours sur une voie d'un véhicule
DE19749086C1 (de) Vorrichtung zur Ermittlung fahrspurverlaufsindikativer Daten
DE112014002019B4 (de) Kurvenformmodelliereinrichtung, Fahrzeuginformationsverarbeitungssystem, Kurvenformmodellierverfahren und Kurvenformmodellierprogramm
DE102015108605A9 (de) Fahrspurwechselpfad-Planungsalgorithmus für ein autonom fahrendes Fahrzeug
EP3490862B1 (fr) Procédé et dispositif permettant de déterminer un modèle de chaussée pour un environnement de véhicule
DE102014009066A1 (de) Vorrichtung und Verfahren zur Regelung der Geschwindigkeit eines Fahrzeugs
EP1680317A1 (fr) Procede d'assistance au conducteur et dispositif fonctionnant sur la base d'informations de voies de circulation
WO2014206699A1 (fr) Procédé de traitement de données de mesure d'un véhicule pour l'établissement du commencement d'une recherche d'une place de stationnement
DE102013221662A1 (de) Systeme und Verfahren zur Geschwindigkeitsregelung von Fahrzeugen
WO2017198443A1 (fr) Procédé de détermination d'une vitesse de sécurité au niveau d'un point de passage à venir
DE102018215448B3 (de) Verfahren zur Schätzung der Geometrie eines Bewegungsweges
EP2964503B1 (fr) Estimation de la vitesse future et/ou distance d'un véhicule d'un point de référence et estimation de l'accélération future
DE102020203756A1 (de) Autobahnausfahrt-erfassung und linienspiegelung zur fahrzeugtrajektorien-bestimmung
EP2945825A1 (fr) Prévision d'une manoeuvre de conduite d'un véhicule
WO2021139974A1 (fr) Procédé de réunion de plusieurs ensembles de données pour la génération d'un modèle de voie actuel d'une voie de circulation et dispositif de traitement de données
DE112019000325T5 (de) Vorrichtung zur vorhersage des verhaltens eines sich bewegenden objekts
DE102008062796A1 (de) Fahrerassistenzsystem und Verfahren zum Betreiben eines Fahrzeugs
DE102014217654A1 (de) Verfahren zum Verarbeiten von Messdaten eines Fahrzeuges zur Bestimmungdes Beginns eines Parksuchverkehrs und Computerprogrammprodukt
EP3279049B1 (fr) Système de commande et procédé de commande pour déterminer une voie de circulation
DE102018203064A1 (de) Kollisionsrisiko-Vorhersageeinheit
DE102013217060A1 (de) Spurgenaue Positionierung eines Fahrzeugs
DE102013217061A1 (de) Spurgenaue Positionierung eines Fahrzeugs

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 14700989

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 2014700989

Country of ref document: EP

NENP Non-entry into the national phase

Ref country code: DE