METHOD FOR DETERMINING AN UPDATED TRAJECTORY FOR A VEHICLE

Description

[0001] The invention relates to a method and a corresponding device for determining a trajectory for controlling and / or regulating the lateral and / or longitudinal guidance of a vehicle during a driving maneuver.

[0002] Vehicles (especially road vehicles) increasingly feature partially or highly automated driving functions. A key aspect of a partially or highly automated driving function is the planning of the most optimal vehicle trajectory possible, thereby avoiding collisions with other road users and enabling the highest possible level of driving comfort.

[0003] For example, DE 10 2014 223 000 A1 describes a method for determining a trajectory for the longitudinal and / or lateral guidance of a vehicle. The method involves determining the values ​​of one or more trajectory parameters. These one or more trajectory parameters indicate the dynamics and / or accuracy of the trajectory to be determined.

[0004] DE 10 2016 009 760 A1 describes a control system that, starting from the instantaneous location of a motor vehicle, determines a set of trajectories with a predetermined number, differing in length and / or course, for possible paths of the motor vehicle, wherein the course of adjacent trajectories differs by a predetermined difference of possible different steering angles of the motor vehicle, and varies the predetermined number of trajectories, the length and / or course of the trajectories depending on a driving situation of the motor vehicle, and generates at least one signal that assists a driver of the motor vehicle in controlling the motor vehicle to guide the motor vehicle along a selected of these trajectories, or at least generates an associated control command that causes the motor vehicle toto follow one of these selected trajectories.

[0005] German patent DE 10 2014 215 244 A1 describes a method and a corresponding device for determining a target trajectory for controlling the lateral / longitudinal guidance of a vehicle. The method includes detecting one or more objects in the vehicle's environment based on environmental data. Furthermore, the method includes performing global planning to determine a multitude of potentially possible driving maneuvers based on the detected objects, as well as selecting one driving maneuver from this multitude. Finally, the method includes performing local planning to determine a target trajectory for the longitudinal and / or lateral guidance of the vehicle for the selected driving maneuver.

[0006] The repeated determination of a trajectory (e.g., an evasive trajectory) during the execution of a driving maneuver involves a high computational effort, which can typically only be provided to a limited extent by control units in a vehicle.

[0007] This document addresses the technical problem of enabling the repeated determination of an optimal trajectory for a vehicle during a driving maneuver with reduced computational effort.

[0008] The problem is solved by the independent claims in each case. Advantageous embodiments are described, among other things, in the dependent claims.

[0009] According to one aspect, a method for determining an updated trajectory at a specific time (e.g., at time (n + 1) of a sequence of (possibly equidistant) time points) is described. The trajectory can be used for the (automated) longitudinal and / or lateral guidance of a vehicle, particularly a road vehicle. Specifically, an updated trajectory can be used to determine a steering input for the vehicle's power steering system, a deceleration input for the vehicle's braking system, and / or an acceleration input for the vehicle's propulsion system. The longitudinal and / or lateral guidance of a vehicle can thus be dependent on a (continuously updated) trajectory.

[0010] A trajectory shows a temporal progression or sequence of values ​​of one or more state variables of the vehicle, starting from an initial state and ending at a final state. Example state variables include position. x 1 ( k ) of the vehicle, a speed e 1 ( k ) of the vehicle, an acceleration ẍ 1 ( k ) of the vehicle and / or a jerk x 1 3 k of the vehicle.

[0011] The procedure involves identifying nearby trajectory candidates in the immediate vicinity of a trajectory planned at a previous time. The procedure can be repeated for a sequence of time points, particularly a sequence of equidistant time points. The previous time point n can be the one immediately preceding the determined time point. nThe next point in time in the sequence of time points is + 1. At each of these time points, a trajectory determined at a previous time point can be updated. The vehicle's state can change along the sequence of time points. This change in the vehicle's state can be caused, at least in part, by the repeatedly determined and updated trajectory.

[0012] The short-range trajectory candidates can be determined with a relatively fine resolution of the values ​​of one or more state variables. Thus, an updated version of the trajectory can be searched for with relatively high accuracy in the immediate vicinity of the trajectory determined at the previous time step (where the updated trajectory starts from an updated initial state).

[0013] The immediate surroundings or near-surround typically represent a limited subspace of the overall solution space for possible trajectory candidates. In particular, the updated trajectory can be part of an overall solution space of possible trajectory candidates. The immediate surroundings of the trajectory planned at the previous time point typically represent a sub-solution space of near-surround trajectory candidates from the overall solution space of possible trajectory candidates.

[0014] The immediate vicinity of the trajectory planned at the previous time point may be such that the sub-solution space of nearby trajectory candidates comprises 20%, 10%, or less of the total solution space of trajectory candidates. This allows for an efficient, spatially limited, detailed search for an updated trajectory in the immediate vicinity of the previously determined trajectory. This covers the relatively likely case that the updated trajectory is located in the immediate vicinity of the previously determined trajectory.

[0015] Furthermore, the method includes identifying at least one (typically several) distant trajectory candidates outside the immediate vicinity of the trajectory planned at the previous time point. The one or more distant trajectory candidates can thus be selected from the portion of the overall solution space that is not part of the sub-solution space of nearby trajectory candidates.

[0016] At least one long-range trajectory candidate can be determined with a relatively coarse resolution of one or more state variables. In particular, the resolution for determining the long-range trajectory candidate can be 2, 3, 4, 5, 10 times or more coarser than the resolution for determining the short-range trajectory candidates. Thus, one or more long-range trajectory candidates for the updated trajectory to be determined can be efficiently, albeit with relatively little accuracy, determined. This allows for consideration of the relatively unlikely case that the updated trajectory is not in the immediate vicinity of the previously determined trajectory (e.g., due to a short-term change in the driving situation around the vehicle).

[0017] The procedure further includes determining the updated trajectory at the specified time based on the identified short-range trajectory candidates and on the basis of one or more identified long-range trajectory candidates. For this purpose, a trajectory candidate can be selected as the updated trajectory from the identified short-range trajectory candidates and the identified one or more long-range trajectory candidates. The selection can be made such that one or more constraints regarding at least one obstacle in the vehicle's vicinity are fulfilled (e.g., to determine a collision-free updated trajectory). Alternatively or additionally, the selection can be made such that the value of a quality functional is improved, in particular optimized. The value of the quality functional for a trajectory candidate can, for example, be...depend on the driving comfort of the trajectory candidate during its implementation by the vehicle.

[0018] This method thus enables efficient and precise updating of a trajectory for the (automated) lateral and / or longitudinal guidance of a vehicle. The update can be performed at a specific frequency (e.g., 10 Hz, 20 Hz, 50 Hz, or more), for example, during the execution of an automated driving maneuver. The update process utilizes the fact that an updated trajectory is relatively likely to lie in the immediate vicinity of the previously determined trajectory. Therefore, precise trajectory candidates are identified in the immediate vicinity of the previously determined trajectory.

[0019] To cover the relatively unlikely case that the updated trajectory is not in the immediate vicinity of the previously determined trajectory, one or more trajectory candidates are determined outside the immediate vicinity with relatively low accuracy. If one of these distant trajectory candidates is selected, a precise determination of the trajectory (within the immediate vicinity of the determined trajectory) is performed directly during the next trajectory update (i.e., directly at the following time step).

[0020] The value resolution of a state variable describes, for example, how finely the possible values ​​of a state variable are rasterized when determining trajectory candidates. The value resolution can be the (uniform) grid spacing between different grid points of a rasterization of the possible values ​​of a state variable.

[0021] The procedure can involve determining a distance value or distance measure between a final state for the updated trajectory and a final state for the (previously determined) trajectory. The final state can be defined, for example, by a driving function and / or a driver assistance system of the vehicle. Typically, this does not result in a significant change to the desired final state. However, a significant change to the desired final state may occur (e.g., if a driving maneuver is aborted).

[0022] The determination of short-range trajectory candidates and / or the determination of at least one long-range trajectory candidate can be carried out depending on the distance value.

[0023] In particular, the identification of short-range trajectory candidates and the identification of at least one long-range trajectory candidate with varying levels of resolution cannot be performed if the distance value is greater than a distance threshold, and / or can only be performed if the distance value is less than the distance threshold. This allows, for example, an efficient switch to a new final state for a newly planned trajectory if a driving maneuver is aborted.

[0024] Alternatively or additionally, the immediate surroundings of the trajectory planned at the previous time point, in particular the size of the immediate surroundings or the proportion of possible trajectory candidates within the immediate surroundings, can depend on the distance value. The surroundings, especially their size and / or the proportion of possible trajectory candidates within the immediate surroundings, can be enlarged with an increasing distance value. This allows trajectory planning to be adapted reliably and flexibly to changing objectives.

[0025] The method can include determining a tolerance range around a given final state of the updated trajectory. In other words, tolerances can be allowed in the planning of the final values ​​of one or more state variables. This allows the total solution space of possible trajectory candidates to be expanded, which typically increases the quality of a determined trajectory (with respect to the quality functional underlying the trajectory planning). The consideration of tolerance ranges becomes possible due to the reduced computational effort of the described method.

[0026] The short-range trajectory candidates and / or the one or more long-range trajectory candidates can then be determined, taking into account the tolerance range around the specified final state. This allows the quality of an updated trajectory to be improved.

[0027] Determining a trajectory candidate can include determining the initial state of the updated trajectory, where the initial state comprises initial values ​​for one or more state variables of the vehicle. The initial state can be determined, for example, based on sensor data from one or more sensors of the vehicle.

[0028] Furthermore, the (desired or predefined) final state of the updated trajectory can be determined, where the final state comprises final values ​​for one or more of the vehicle's state variables. A temporal sequence of values ​​for these one or more state variables can then be determined as a trajectory candidate, with this temporal sequence transforming the initial state into the final state. This temporal sequence can be determined based on a state model of the vehicle. In this way, trajectory candidates can be identified that can be reliably implemented or driven by a vehicle.

[0029] According to another aspect, a procedure for determining an updated trajectory at a specific point in time (e.g., the time) is needed. n+ 1) for the longitudinal and / or lateral guidance of a vehicle. The aspects described in this document are also applicable to this method. A trajectory shows a temporal progression of one or more state variables of the vehicle from an initial state to a final state.

[0030] The procedure involves identifying trajectory candidates in a specific environment of a previous time (e.g., at time ). n The method involves determining the trajectory. The resolution of the one or more state variables is reduced (e.g., continuously) when determining trajectory candidates, based on the distance from the trajectory determined at the previous time. Furthermore, the method includes determining the updated trajectory at the specified time based on the determined trajectory candidates.

[0031] A trajectory can encompass the values ​​of one or more state variables at a sequence of time points (starting from an initial time point, representing the initial state, to a final time point, representing the final state). These time points can be spaced 100 ms, 50 ms, 20 ms, or less apart. To determine the distance between two trajectories, a distance value (e.g., a squared or absolute distance) can be calculated for each state variable and for each time point in the trajectories. Thus, a sequence of distance values ​​can be determined for each state variable, and based on this sequence, an average distance value for that state variable can be calculated. The distance between the trajectories can then be determined as the sum or (optionally weighted) average of the (average) distance values ​​of the state variables in the trajectories.An increasing distance between two trajectories thus indicates an increasing distance between the values ​​of the state variables of the two trajectories.

[0032] According to another aspect, a device is described which is set up to perform one of the procedures described in this document.

[0033] According to another aspect, a road motor vehicle (in particular a passenger car or a truck or a bus) is described that includes the device described in this document.

[0034] Another aspect described is a software (SW) program. The SW program can be configured to run on a processor (e.g., on a vehicle's control unit) and thereby execute one of the procedures described in this document.

[0035] Another aspect describes a storage medium. This storage medium can contain a software program configured to run on a processor and thereby execute one of the procedures described in this document.

[0036] The invention will now be described in more detail using exemplary embodiments. Figure 1 an exemplary traffic situation; Figure 2 a flowchart of an exemplary procedure for determining a trajectory or a trajectory candidate; Figure 3 an exemplary transformation into a coordinate system relative to a reference line of a vehicle; Figure 4a an exemplary close-range area of ​​a planned trajectory; Figure 4b an exemplary tolerance range for a final state of a trajectory; and Figure 5 a flowchart of an exemplary procedure for determining an updated trajectory.

[0037] As stated at the outset, this document addresses the technical task of efficiently determining or updating a safe longitudinal and / or lateral trajectory for a vehicle (also referred to as an ego-vehicle). In this context, it shows Fig. 1 An exemplary traffic situation. The ego vehicle 100 is traveling on a multi-lane roadway 105. A vehicle 101 in the same lane of roadway 101 ahead of the ego vehicle 100 (i.e., a vehicle 101 in the ego lane) may be traveling at a relatively low speed. The ego vehicle 100 then has the option, for example, of performing an overtaking maneuver and changing lanes along a trajectory 112 to overtake vehicle 101. A collision with other vehicles 102, 103 must be avoided.

[0038] To carry out the in Fig. 1During the maneuver depicted, a control unit or device of the ego vehicle 100 can determine and periodically update a trajectory 112 that satisfies one or more boundary or secondary conditions. Vehicle dynamics aspects can be considered when determining a trajectory 112. In particular, a trajectory 112 that can be realistically driven by the vehicle 100 can be determined based on one or more vehicle parameters and / or the current driving situation. A curvature that the vehicle 100 can handle can be taken into account. Further examples of vehicle parameters that can be considered are lateral and / or longitudinal acceleration or deceleration of the vehicle 100 (achievable in the current driving situation).

[0039] Furthermore, a trajectory 112 is typically determined in such a way that a collision with the detected objects or obstacles 101, 102, 103 in the vicinity of the ego vehicle 100 can be avoided. The trajectory 112 determined in this way can then be transferred to one or more controllers for the lateral and / or longitudinal guidance of the vehicle 100 and used by these controllers.

[0040] The determination of a trajectory 112 is preferably carried out in a curved coordinate system, relative to a road alignment. The procedure described in this document for determining or updating a trajectory can therefore include the step of transforming state data or values ​​of state variables of the vehicle 100 (such as the position of the vehicle 100, a yaw angle of the vehicle 100, and / or a steering angle of the vehicle 100) from a Cartesian coordinate system into a (curved) Frenet coordinate system.

[0041] The straightening of the road surface (using a curved coordinate system) is exemplified in Fig. 3The following is shown. For the straightening process, measurement signals regarding the state of vehicle 100 are transformed into a lane coordinate system. The trajectory planning itself therefore does not take place in a Cartesian coordinate system 301, but in a Frenet coordinate system. The Frenet coordinate system is described with respect to a reference curve 300 (e.g., the center of a lane). The vehicle position is thus defined by the variables s ( t ) 303 in the longitudinal direction and d ( t ) 302 described in the transverse direction. ṡ ( t ) and ḋ ( t ) describe the longitudinal and lateral speeds and s̈ ( t ) and d̈ ( t ) describe the longitudinal and lateral acceleration.

[0042] Both the vehicle's own motion and the road users or objects 101, 102, 103 to be considered can be taken into account in the Frenet coordinate system. Intuitively, this transformation corresponds to the straightening of the coordinate system 301 and thus allows the separate optimization of the longitudinal and lateral motion of the vehicle 100. After determining a trajectory 112 (i.e., after determining a temporal sequence of values ​​of the state variables of the vehicle 100), the determined values ​​of the state variables can be transformed back into the Cartesian coordinate system 301 before being used to control the longitudinal and lateral guidance of the vehicle 100.

[0043] The lateral and longitudinal motion of a vehicle 100 can be described as an optimal control problem with output s ( t ) = x 1 ( t ) (in the case of longitudinal planning) or d ( t ) = x 1 ( t) (in the case of cross-planning) of an integrator system (i.e., a model of the dynamics of a vehicle 100). Here, x 1 ( t ) a first state variable of vehicle 100, which describes the position of vehicle 100 (in the longitudinal or lateral direction). The jerk can be used as input to the integrator system. x 1 3 t (i.e., the 3rd derivative of the state variable) x 1 ( t )) and / or the derivation of the jerk x 1 4 t (i.e., the 4th derivative of the state variable) x 1 ( t )) be defined.

[0044] An exemplary integrator system or state model of a vehicle 100 can be defined as follows: x ˙ = 0 1 0 0 0 0 1 0 0 0 0 1 0 0 0 0 x + 0 0 0 1 u where the input variable u the derivation of the jerk x 1 4 t This corresponds to the state of a vehicle 100 at a specific time t. The state of a vehicle 100 at a specific time t can be described by the state vector. x T< = [ x 1 , x 2 , x 3 , x4 ] are described, whereby x 2 ( t ) = e 1 ( t ) , x 3 ( t ) = e 2 ( t ) and x 4 ( t ) = Yes 3 ( t ) is.

[0045] It can now be a trajectory, i.e., a temporal sequence of states. x ( t ) or in the discrete-time domain x ( k ), with k = 1, ... . , N lon , where N lon The planning horizon must be determined. The planning horizon can be, for example, 5 seconds, 10 seconds, or more. The individual time points can be spaced 100ms, 50ms, 20ms, or less apart. Within the context of planning a trajectory, the sequence of states can be determined. x ( kThe cost function is determined to reduce, in particular minimize, or optimize a cost or performance function. The cost function can be specified as requiring vehicle 100 to reach a specific target position at the end of the planning horizon. Alternatively or additionally, it can be specified that the sequence of states fulfills one or more comfort criteria (e.g., regarding jerk).

[0046] To calculate a transverse trajectory, a target area can be used as the desired endpoint of a trajectory. d ziel a specified range, for example, indicating an area on an adjacent lane (as in Fig. 1 (shown). This target area can, for example, be the desired final position of the final state. x ( tf ) or x ( N lon ) be determined, possibly with x 1 ( tf ) = x 1 ( N lon ) = d zielThe planning horizon for planning a transverse trajectory and the planning horizon for planning a longitudinal trajectory can differ. Typically, the planning horizon for a transverse trajectory is shorter than for a longitudinal trajectory (e.g., 6s vs. 10s).

[0047] The following function can be used (especially reduced or minimized) as a lateral selection parameter or as a lateral quality parameter for determining a trajectory for the lateral guidance of the vehicle 100: J quer = 1 2 ∫ 0 t f d 4 t 2 dt + k q 1 d ziel − d t f 2 + k q 2 t f

[0048] The first expression evaluates the development of the jerk derivative along trajectory 112 (and thus the comfort). The second expression evaluates the deviation of the final position. d ( tf ) from the target position d ziel Furthermore, the third expression evaluates the time length of trajectory 112. Regarding the weighting factors k q 1 and k q2. The shape of trajectory 112 can be influenced.

[0049] Longitudinal planning can be carried out in a similar manner. For longitudinal planning, the following longitudinal selection parameter or longitudinal quality function can be used (in particular, reduced or minimized). J l ä ngs = 1 2 ∫ 0 t f s 4 t 2 dt + k l 1 s ziel + s t f 2 + k l 2 t f , especially if a specific target position s ziel to be achieved. Alternatively, the following longitudinal selection dimension or longitudinal quality function can be used. J l ä ngs = 1 2 ∫ 0 t f s 4 t 2 dt + k l 1 s ˙ ziel + s ˙ t f 2 + k l 2 t f , especially when a certain target speed is required ṡ ziel to be achieved.

[0050] Thus, a (longitudinal and / or transverse) trajectory can be determined, where a trajectory defines the state x of vehicle 100, in particular the position s ( k ) of vehicle 100, at a large number of sampling times k, with k = 1, .... , N lon , displays, whereby N lonThe planning horizon is defined. It is then necessary to check which of the multitude of trajectories fulfills one or more constraints regarding obstacles, particularly regarding other vehicles 101, 102, 103. Specifically, a multitude of identified trajectory candidates can be sorted according to an increasing value of the respective performance functional. The trajectory candidate from the multitude of candidates can be selected as trajectory 112, which fulfills the one or more constraints regarding obstacles 101, 102, 103 and exhibits the lowest possible, or optimal, value of the performance functional.

[0051] Therefore, an optimal trajectory can be determined at a specific time n. x opt ( k ) , with k = 1, ... . , N lon ,The optimal trajectory can be determined. At the specified time, it can be used for automated longitudinal and / or lateral guidance of the vehicle 100.

[0052] The process of determining an optimal trajectory in each case, i.e., in particular the one in Fig. 2 The described procedure 200 can be repeated for a sequence of time points n. At each time point n, The current initial state x(0) is determined (step 201); the desired final state or target point. x ( N lon ) are specified (step 202); and an optimal trajectory is determined using the optimization procedure described above. x opt ( k ) , with k = 1, ... . , N lon , to be determined (step 203).

[0053] Fig. 4 shows an example of a planned or optimal trajectory x opt ( k ) 403 between an initial state x(0) 401 and a final state x ( N lon ) 402. The trajectory 403 is represented in a grid 400, where the grid 400 represents the possible values ​​of the different state variables as grid areas 405. x 1 , x 2 , x 3 , x 4 is displayed. The planned trajectory x opt ( k ) 403 can be done by means of procedure 200 at a specific time n have been determined.

[0054] Furthermore, it shows Fig. 4 an updated initial state x (0) 411 of vehicle 100 at a subsequent time n + 1. Vehicle 100 can, depending on the time n planned trajectory x opt ( k ) 403 at least partially automated to the updated initial state x (0) 411 have been recorded.

[0055] Furthermore, an updated final state can be x ( N lon) 412 for the following time n + 1 can be specified, for example by the driver assistance system for which trajectory 403, 413 is to be determined. For example, a lane change assistant can specify that a lane change should still be performed. The updated final state x ( N lon ) 412 for the following time n +1 can be associated with the final state x ( N lon ) 402 for the time n coincide, or are located in close proximity to them.

[0056] It can be assumed that the planned or updated trajectory 413 for the following time point n + 1 in the immediate vicinity of the already planned trajectory 403 for the time n is located, especially when the target point specifications (i.e., the final state to be achieved) x ( N lon) 412) does not change significantly. Therefore, a short range 414 of possible trajectory candidates can be found in the immediate vicinity of the planned trajectory. x opt ( k ) 403 for the time n The trajectory candidates from the immediate vicinity 414 of the last planned trajectory will be determined. x opt ( k ) 403 can be designated as short-range trajectory candidates. The short-range 414 is in Fig. 4 by two dashed curves around the last planned trajectory x opt ( k ) 403. The short-range trajectory candidates can be determined with a relatively high or fine resolution of the different state variables. x 1 , x 2 , x 3 , x 4 will be determined.

[0057] The trajectory that satisfies one or more constraints and improves, in particular optimizes, the performance functional can then be selected from the identified short-range trajectory candidates.

[0058] Even though it is relatively likely that the trajectory to be planned or updated, 413, will be available for the following time point n + 1 in the immediate vicinity of the already planned or determined trajectory 403 for the time n If the trajectory is ordered, it cannot be ruled out that (e.g. due to an abrupt change in the situation in the vicinity of vehicle 100) the planned trajectory 413 for the following time will change. n + 1 is located outside the near range 414. For this reason, one or more far ranges 415, 416 can be defined for trajectory candidates outside the near range 414. The one or more far ranges 415, 416 are in Fig. 4abounded by a dashed line on one side and a dash-dot line on the other. In the one or more distant regions 415, 416, the different state variables can be measured with a relatively low resolution. x 1 , x 2 , x 3 , x 4. One or more long-range trajectory candidates are identified.

[0059] It can then be checked whether one of the long-range trajectory candidates, which also satisfies the one or more constraints, yields a better value for the performance functional than the (optimal) short-range trajectory candidate. Depending on the value of the performance functional, either the short-range trajectory candidate or one of the long-range trajectory candidates can then be selected as the updated trajectory 413 for the subsequent time. n + 1 can be selected. The updated trajectory 413 is in Fig. 4a represented by a dotted line.

[0060] By using a relatively high or fine resolution only in the near range 414 of the previously planned trajectory 403 and by using a relatively small or coarse resolution outside the near range 414, the computational effort in determining a trajectory 413 can be significantly reduced.

[0061] If, at a specific time n, a far-range trajectory candidate was selected as the planned trajectory 403 instead of a near-range trajectory candidate, the far-range trajectory candidate may not represent the optimal trajectory in the value range of the determined far-range trajectory candidate due to the relatively small or coarse resolution used in its determination. This is because, when the procedure is executed at the subsequent time, n+ 2 nearby trajectory candidates in the immediate vicinity of the planned trajectory 403 with a relatively high or fine resolution can be determined, but it can be ensured that at the following time n + 2 an optimal trajectory can be found. Finding an optimal trajectory is thus distributed across several time steps of the procedure. This enables the finding of an optimal trajectory over multiple computational or time steps, thereby reducing the runtime or computation time for determining the trajectory. In other words, by moving away from finding an optimal solution at every single time point. n The computational effort of trajectory planning can be significantly reduced without substantially impairing the quality of the trajectory planning (since finding an optimal trajectory is only delayed by a maximum of one time step).

[0062] As explained above, when planning a trajectory 403, 413 at a certain point in time n a specification regarding the target state x ( N lon ) 402, 412. This is done by defining a fixed target state. x ( N lon ) 402, 412 the solution space of possible trajectory candidates is restricted. If necessary, a fixed target state can be specified. x ( N lon ) 402, 412 may not be required for a specific driving function or driver assistance system. For example, it may be possible to define a specific tolerance range 422 around a desired target state. x ( N lon ) 412 to enable (see Fig. 4b ). By allowing a tolerance range of 422 around a desired target state x ( N lon) 412 the solution space 424 of trajectory candidates can be extended, making it possible to plan trajectories 413 that have an improved value of the goodness functional and thus a higher "optimality".

[0063] In other words, the number and / or accuracy of target point specifications can be reduced and at least partially replaced by soft specifications or tolerance ranges 412. This allows the solution space 424 to be expanded to improve the iterative finding of an optimal trajectory 413.

[0064] Fig. 5 shows a flowchart of an exemplary procedure 500 for determining an updated trajectory 413 at a specific time (e.g., at the time ). n+ 1) for the longitudinal and / or lateral guidance of a vehicle 100. A trajectory 403, 413 shows a temporal progression of one or more state variables of the vehicle 100 from an initial state 401, 411 to a final state 402, 412. Examples of state variables are the position, the velocity, the acceleration and / or the jerk of the vehicle 100.

[0065] The procedure 500 comprises the determination 501 of near-area trajectory candidates in a direct vicinity or in a near-area 414 of a trajectory 403 determined at a previous time (e.g., at time n). The near-area trajectory candidates are determined with a relatively fine value resolution of the one or more state variables.

[0066] Furthermore, the procedure 500 comprises determining 502 at least one far-field trajectory candidate outside the immediate vicinity 414 (i.e., outside the near-field 414) of the trajectory 403 determined at the previous time point. The at least one far-field trajectory candidate is determined with a relatively coarse resolution of the one or more state variables (e.g., with a resolution coarser by a factor of 2, 3, 4 or more).

[0067] The updated trajectory 413 at the specified time can then be determined 503 based on the identified short-range trajectory candidates and the identified long-range trajectory candidate. In particular, one of the identified trajectory candidates can be selected as the updated trajectory 413. For example, the trajectory candidate that improves or optimizes the value of a performance functional (while still satisfying one or more constraints with respect to one or more obstacles 102, 103) can be selected.

[0068] The measures described in this document can reduce the number of possible trajectory candidates and thus the required runtime for calculating a trajectory 413 on a control unit of a vehicle 100, without affecting the quality of the determined trajectory 413.

[0069] The present invention is not limited to the embodiments shown.

Claims

1. A method (500) for determining an updated trajectory (413) at a specific time for the longitudinal and / or lateral guidance of a vehicle (100); wherein a trajectory (403, 413) indicates a temporal progression of one or more state variables of the vehicle (100) from an initial state (401, 411) to a final state (402, 412); wherein the method (500) comprises: • Determining (501) near-range trajectory candidates in a direct vicinity (414) of a trajectory (403) determined at a previous time; wherein the near-range trajectory candidates are determined with a relatively fine value resolution of the one or more state variables; • Determining (502) at least one far-range trajectory candidate outside the direct vicinity (414) of the trajectory (403) determined at the previous time; wherein the far-range trajectory candidate is determined with a relatively coarse value resolution of the one or more state variables; and • Determining (503) the updated trajectory (413) at the specific time based on the determined near-range trajectory candidates and based on the determined far-range trajectory candidate.

2. Method (500) according to claim 1, wherein the method (500) comprises: • Determining a distance value of a distance measure between a final state (412) for the updated trajectory (413) and a final state (402) for the trajectory (403) determined at the previous time; and • Determining (501) near-range trajectory candidates and / or determining (502) at least one far-range trajectory candidate depending on the distance value.

3. Method (500) according to claim 2, wherein • the direct vicinity (414), in particular a size of the direct vicinity (414), of the trajectory (403) determined at the previous time depends on the distance value, in particular is enlarged with increasing distance value; and / or • the determining (501) of near-range trajectory candidates and the determining (502) of at least one far-range trajectory candidate with different fine value resolutions does not occur if the distance value is greater than a distance threshold value, and / or only occurs if the distance value is less than the distance threshold value.

4. Method (500) according to any of the preceding claims, wherein • the updated trajectory (413) is part of a total solution space of trajectory candidates; • the direct vicinity (414) of the trajectory (403) determined at the previous time represents a partial solution space of near-range trajectory candidates; and • the direct vicinity (414) of the trajectory (403) determined at the previous time is such that the partial solution space of near-range trajectory candidates comprises 20%, 10% or less of the total solution space of trajectory candidates.

5. Method (500) according to any of the preceding claims, wherein the value resolution for determining (502) the far-range trajectory candidate is 2, 3, 4, 5, 10 times or more coarser than the value resolution for determining (501) the near-range trajectory candidates.

6. Method (500) according to any of the preceding claims, wherein • the method (500) is repeated for a sequence of times, in particular for a sequence of equidistant times; and • the previous time is a time of the sequence of times directly preceding the specific time.

7. Method (500) according to any of the preceding claims, wherein • the method (500) comprises determining a tolerance range (422) around a specified final state (412) of the updated trajectory (413); and • the near-range trajectory candidates and / or the far-range trajectory candidate are determined taking into account the tolerance range (422) around the specified final state (412).

8. Method (500) according to any of the preceding claims, wherein determining a trajectory candidate comprises: • Determining (201) the initial state (411) of the updated trajectory (413); wherein the initial state (411) comprises initial values for a plurality of state variables of the vehicle (100); wherein the plurality of state variables comprises a position (x1(k)) of the vehicle (100), a velocity (ẋ1(k)) of the vehicle (100), an acceleration (ẍ1(k)) of the vehicle (100) and / or a jerk ( x 1 3 k ) of the vehicle (100); • Determining (202) the final state (412) of the updated trajectory (413); wherein the final state (412) comprises final values for one or more of the plurality of state variables of the vehicle (100); and • Determining (203), as a trajectory candidate, a temporal sequence of values of the plurality of state variables that transitions the initial state (411) to the final state (412); wherein the temporal sequence of values of the plurality of state variables is determined depending on a state model of the vehicle (100).

9. Method (500) according to any of the preceding claims, wherein determining (503) the updated trajectory (413) comprises selecting a trajectory candidate from the determined near-range trajectory candidates and the determined at least one far-range trajectory candidate, such that • one or more constraints with respect to at least one obstacle (102, 103) in an environment of the vehicle (100) are satisfied; and / or • a value of a quality functional is improved, in particular optimized; wherein the quality functional depends in particular on a driving comfort of a trajectory candidate.

10. Method (500) according to any of the preceding claims, wherein the method (500) comprises determining a steering specification for a power steering system of the vehicle (100) and / or a deceleration specification for a braking system of the vehicle (100) and / or an acceleration specification for a drive of the vehicle (100) depending on the updated trajectory (413).

11. Device configured to carry out a method according to any of the preceding claims.

12. Road motor vehicle comprising a device according to claim 11.

13. Software program configured to be executed on a processor, and to thereby carry out a method according to any of claims 1 to 10.

14. Storage medium comprising a software program according to claim 13.