DE102014215244A1 - Collision-free transverse / longitudinal guidance of a vehicle - Google Patents

Abstract

The invention relates to a method and a corresponding device for determining a desired trajectory for the control and / or regulation of the transverse / longitudinal guidance of a vehicle. In particular, a method is described, which includes the detection of one or more objects in an environment of the vehicle, based on environmental data. Furthermore, the method comprises performing a global planning to determine a plurality of potentially possible driving maneuvers in dependence on the one or more detected objects, as well as selecting a driving maneuver from the multiplicity of potentially possible driving maneuvers. In addition, the method comprises performing a local planning in order to determine a desired trajectory for the longitudinal and / or transverse guidance of the vehicle for the selected driving maneuver.

Description

The invention relates to a method and a corresponding device for determining a desired trajectory for the control and / or regulation of the transverse / longitudinal guidance of a vehicle.

The realization of a partially or highly automated driving operation of a vehicle (in particular a road motor vehicle) is driven forward both in research and increasingly in the automotive industry. A central aspect of the partially or highly automated driving operation is the planning of an optimal trajectory of the vehicle by which collisions with other road users are avoided. The determination of such a trajectory (eg an avoidance trajectory) is associated with a high level of computation, which typically can not be provided by control devices in a vehicle, or only to a limited extent.

The present document deals with the technical task to determine the most optimal trajectory for a vehicle with a reduced computational effort. In particular, the computational effort is to be reduced in such a way that an implementation of the trajectory planning on a control device of the vehicle is made possible.

The object is solved by the independent claims. Advantageous embodiments are u. a. in the dependent claims.

According to one aspect, a method for determining a trajectory for the longitudinal and / or transverse guidance of a vehicle (in particular a road motor vehicle) is described. The method can be listed, for example, on a control unit of the vehicle. The determined trajectory, in particular a determined target trajectory, can be used to provide automatic assistance to a driver of the vehicle with regard to the longitudinal and / or transverse guidance of the vehicle. In particular, depending on the determined desired trajectory, a steering input for an auxiliary power steering (Electronic Power Steering) of the vehicle and / or a deceleration specification for a brake system of the vehicle and / or a specification for a drive of the vehicle can be determined and possibly regulated.

The method may include determining environmental data from one or more environmental sensors of the vehicle. The environment data can be provided by an environment model (UFM) of the vehicle. For example, the vehicle may include one or more cameras, one or more radar sensors, and / or one or more lidar sensors as environment sensors. The environment data may include information regarding the environment in front of, beside and / or behind the vehicle. On the basis of the environmental data, one or more objects (and their position) can be detected in the surroundings of the vehicle. For example, an obstacle on a road in front of the vehicle (which is also referred to as an ego vehicle) can be detected. In such a case, the method can determine a trajectory for the vehicle, by means of which a collision with the obstacle can be avoided. Furthermore, z. B. further vehicles in the vicinity of the vehicle (eg., On other lanes) are detected. The trajectory can be determined in such a way that, when the trajectory is implemented, no collision occurs with the other vehicles.

The method further comprises performing global planning to determine a plurality of potentially possible driving maneuvers in dependence on the one or more detected objects. The potentially possible driving maneuvers can be determined in such a way that it is ensured that the potentially possible driving maneuvers can be driven through the vehicle (eg taking into account a maximum possible acceleration of the vehicle and the freedom from collision).

In the context of global planning can z. B. a plurality of different local target areas for the vehicle, on the basis of the environment data and taking into account (at least a part) of the one or more detected objects, are determined. The local target areas may be disjoint. For example, the local target areas may represent different areas to which the ego vehicle may travel to avoid collision with a detected obstacle. For each target area a potentially possible driving maneuver can be determined.

The variety of potentially possible driving maneuvers can be z. B. a lane change of the vehicle to a left lane one include multi-lane road. The corresponding target area can lie on the left lane of the ego vehicle. Alternatively or additionally, the plurality of potentially possible driving maneuvers may include a lane change of the vehicle to a right lane of the multi-lane road. The corresponding target area can lie on the right-hand lane. Alternatively or additionally, the plurality of potentially possible driving maneuvers may include a deceleration or acceleration on a currently traveled road. The corresponding target area can lie on a lane of the ego vehicle in front of a current position of the ego vehicle.

The method further includes selecting a driving maneuver from the plurality of potentially possible driving maneuvers. For example, a value of a global selection measure for each of the plurality of potentially possible driving maneuvers can be determined for this purpose. The global selection measure can, for. B. depending on a collision probability of the vehicle with the one or more detected objects. The global selection size can thus z. B. can be used to reduce the risk of collision of the ego vehicle. Alternatively or additionally, a global selection measure can be defined, which is aimed at increasing comfort for the occupants of the vehicle or sportiness of the driving style of the vehicle. The driving maneuver can be selected on the basis of the values of the global selection measure for the multiplicity of potentially possible driving maneuvers. In particular, the driving maneuver with the highest or lowest value of the global selection measure can be selected.

In the context of global planning, a single target area / a single driving maneuver can thus be selected from a large number of possible target areas / driving maneuvers. The determination of a desired trajectory can then be limited to the selected target area / driving maneuver. Thus, the solution space for possible trajectories can be substantially reduced, which leads to a reduction of the computational effort for the determination of the desired trajectory. At the same time, it can be ensured by a suitable selection of the target area / target trajectory (for example by a selection by means of a global selection criterion) that an (almost) optimal target trajectory can continue to be determined.

The method further comprises performing a local planning to determine a desired trajectory for the longitudinal and / or transverse guidance of the vehicle for the selected driving maneuver. In particular, the local planning can be limited to the selected driving maneuver. In other words, a desired trajectory can be calculated exclusively for the selected driving maneuver / the selected target area. Thus, the computational effort for the determination of the desired trajectory can be reduced.

Performing the global scheduling may include determining a plurality of preliminary trajectories for the corresponding plurality of potentially possible driving maneuvers. In this case, exactly one (representative) provisional trajectory can be determined for each potentially possible driving maneuver. The value of the global selection measure can be z. B. based on the provisional trajectory of a potentially possible driving maneuver can be determined. Thus, the preliminary trajectories can be used to select a particular driving maneuver. Due to the limited number of trajectories that are determined for a driving maneuver, the selection of a driving maneuver can be done efficiently.

Furthermore, a model for describing a preliminary trajectory may be used, which has a smaller number of parameters than a model for describing the desired trajectory. Alternatively or additionally, a smaller number of trajectories can be determined in the context of global planning than in determining the desired trajectory or for determining a preliminary trajectory, a smaller number of trajectories can be determined than for determining the desired trajectory. As explained in this document, in the context of determining the best possible trajectory, a large number of different trajectories can be determined for different assumptions (eg for different assumptions with respect to a final state of a trajectory). The number of determined different trajectories (and thus the number of assumptions) for a provisional trajectory (in global planning) may be lower than for a desired trajectory (in local planning). In other words, the preliminary trajectories can be determined with a lower accuracy than the desired trajectory provided by the method. In the model for describing a trajectory, it may, for. B. can be a polynomial of a certain order. The order used for preliminary trajectories may be less than the order used for the desired trajectory. By using simplified models to describe the preliminary trajectory, the computational burden of global planning can be further reduced.

The plurality of potentially possible driving maneuvers may include at least one potentially possible driving maneuver for driving the vehicle (in particular for driving forward on a road) and at least one potentially possible driving maneuver for parking the vehicle (in particular for parking). Thus, in the context of global planning, the cases "driving" and "parking" can always be considered simultaneously. When selecting a relevant driving maneuver can then always one of the two cases be excluded. Nevertheless, by considering the cases "driving" and "parking" in parallel, a smooth transition can be provided in assisting the driver between the two cases.

Detecting one or more objects in the environment of the vehicle may include detecting one or more objects that are relevant to a driving maneuver while driving the vehicle. Furthermore, detecting one or more objects may include detecting one or more objects that are relevant to a driving maneuver when parking the vehicle. More generally, different one or more objects that are relevant to the respective driving maneuver can be detected for the different potentially possible driving maneuvers. In the context of local planning, only the one or more objects that are relevant for the selected driving maneuver can then be taken into account. The other objects around the ego vehicle may be disregarded. Thus, the computational effort for the determination of a desired trajectory in the context of local planning can be reduced.

Performing the local planning may include determining an application range of the desired trajectory. The scope of application may include automatic support only for the longitudinal guidance of the vehicle (and not for the lateral guidance of the vehicle), automatic support only for the lateral guidance of the vehicle (and not for the longitudinal guidance of the vehicle), or automatic support for both the longitudinal guidance as well as for the transverse guidance of the vehicle. The application area may depend, for example, on a driver assistance function for which the desired trajectory is determined. For example, in the case of an evasion assistant, a desired trajectory can be determined only to support the transverse guidance of the vehicle or a desired trajectory only to support the longitudinal guidance of the vehicle. The desired trajectory can then be determined as a function of (that is to say taking into account) the field of application. In particular, by taking into account the respective field of application, on average the computational outlay for determining the desired trajectory can be reduced.

Performing the local planning may include determining a plurality of possible trajectories for the selected driving maneuver. The possible trajectories can be determined according to the accuracy required for the desired trajectory. By determining a multiplicity of different possible trajectories, it is possible to ensure that the best possible desired trajectory (for example, optimally with regard to collision avoidance, comfort and / or sportiness) can be determined. The desired trajectory can be selected from the multiplicity of possible trajectories.

Determining a possible trajectory for the selected driving maneuver may include determining initial values for a plurality of state variables of the vehicle at an initial instant of the possible trajectory. In this case, the plurality of state variables may include a position of the vehicle, a speed of the vehicle, an acceleration of the vehicle and / or a jerk of the vehicle. The initial values can result from the current state of the vehicle. Furthermore, end values can be determined at an end time of the possible trajectory for the plurality of state variables of the vehicle. The final values can be z. B. from global planning. Furthermore, the possible trajectory can be determined on the basis of the initial values, the end values, the end time and on the basis of a polynomial of the 7th or higher / lower order. To account for the jerk, at least one polynomial of 7th order is required. In this case, the position of the vehicle along the possible trajectory can be determined on the basis of the polynomial. The coefficients of the polynomial can be determined on the basis of the initial values and the final values.

In particular, within the framework of local planning, one or more driving-dynamic boundary conditions (eg coefficient of friction) can be determined. Furthermore, boundary conditions with respect to a collision freedom with one or more objects along the desired trajectory can be determined and taken into account. The desired trajectory can be selected as a function of the one or more driving dynamic boundary conditions and with regard to freedom from collision. Thus, it can be ensured that the selected target trajectory can actually be implemented by the vehicle.

Alternatively or additionally, as part of the local planning, values of a local selection measure for one or more of the plurality of possible trajectories can be determined. The local selection measure can be dependent on a collision probability of the vehicle with the one or more detected (and possibly relevant for the selected driving maneuver) objects. The desired trajectory can be selected as a function of the local selection measure. Thus it can be ensured that by the selected desired trajectory collision-free support of the longitudinal / lateral guidance of the vehicle takes place.

According to a further aspect, a device for determining a trajectory for the transverse and / or longitudinal guidance of a vehicle is described. The device can also be set up to support the transverse and / or longitudinal guidance of the vehicle as a function of a determined desired trajectory. In particular, the device may include a controller that causes the vehicle to travel in accordance with the determined desired trajectory.

The device is configured to receive environmental data from one or more environmental sensors of the vehicle and / or from an environmental model of the vehicle, and based on the environmental data to detect one or more objects in an environment of the vehicle. The device is further equipped with global planning in order to determine a plurality of potentially possible driving maneuvers in dependence on the one or more detected objects, and to select a driving maneuver from the plurality of potentially possible driving maneuvers. In addition, the device is set up to carry out local planning in order to determine a desired trajectory for the longitudinal and / or transverse guidance of the vehicle for the selected driving maneuver.

According to one aspect, a method for determining a trajectory for the longitudinal and / or transverse guidance of a vehicle is described. The method includes determining initial values for a plurality of state variables of the vehicle. The plurality of state variables includes a position of the vehicle, a speed of the vehicle, an acceleration of the vehicle and a jerk of the vehicle. The method further comprises determining end values at an end time of the trajectory for the plurality of state variables of the vehicle. In addition, the vehicle includes determining a trajectory based on the initial values, the end values, the end time, and a polynomial of the seventh or higher order.

In another aspect, a vehicle (eg, a passenger car, a lorry, or a motorcycle) is described that includes the apparatus for determining a trajectory for the lateral and / or longitudinal guidance of the vehicle described in this document.

In another aspect, a software (SW) program is described. The SW program may be set up to run on a processor (eg, on a control device of a vehicle) and thereby perform the method described in this document.

In another aspect, a storage medium is described. The storage medium may include a SW program that is set up to run on a processor and thereby perform the method described in this document.

It should be understood that the methods, devices and systems described herein may be used alone as well as in combination with other methods, devices and systems described in this document. Furthermore, any aspects of the methods, apparatus, and systems described herein may be combined in a variety of ways. In particular, the features of the claims can be combined in a variety of ways. Furthermore, the invention will be described in more detail with reference to exemplary embodiments. It shows

1 an exemplary traffic situation that requires the planning of a trajectory; and

2 an exemplary functional structure of a system for controlling the lateral / longitudinal guidance of a vehicle.

As explained above, the present document deals with the technical task of reducing the computational outlay for determining a trajectory of a vehicle (which is also referred to as an ego vehicle). In this context shows 1 an exemplary traffic situation. The ego vehicle 100 drive on a multi-lane road 101 , A vehicle 102 on the same lane of the road 101 in front of the ego vehicle 102 slows down, leaving the ego vehicle 100 must react in order to avoid a collision. The ego vehicle 100 now has the possibility to stay on the same lane and brake or to change the lane along a trajectory to the decelerating vehicle 102 dodge. There are a variety of different trajectories for the different options possible. The determination of an optimal trajectory from the plurality of possible trajectories is typically associated with a high computational effort, so that a determination in real time is typically not possible. This is especially the case if the entire solution space (eg all possible lanes of the road 101 ) is considered for the determination of an optimal trajectory.

It is therefore proposed a gradual predetermination. In particular, it is proposed to select a target area from a plurality of target areas in a global planning. For example, in the context of global planning, it can be determined whether a longitudinal maneuver should take place (eg a stay on the same road) or whether a lateral maneuver should take place (eg a change to an adjacent road). In the context of global planning can z. B. for each of a plurality of possible target areas and / or driving maneuvers a preliminary trajectory can be determined. In the in 1 example shown may, for. For example, for the three driving maneuvers (avoidance on the left lane, braking on the current lane, and avoiding on the right lane) each a preliminary trajectory 111 be determined.

Global planning can include trajectory planning with reduced accuracy. For example, in global planning, a preliminary trajectory may be used 111 with a polynomial of reduced order (compared to a subsequent local planning). For example, 2nd order or 5th order polynomials may be used (as compared to 5th order or 7th order polynomials for determining an optimal trajectory 112 ). Alternatively or additionally, in the context of global planning, a sample-based solution with collision check can be dispensed with. For example, based on TTC (time to collision) calculations, a possible target range / a possible driving maneuver can be analytically selected.

In global planning, the entire available solution space can be limited to the cost of an optimal trajectory 112 to reduce. For example, by taking into account maximum possible delays / accelerations of the ego vehicle 100 and / or the other vehicles 102 a temporal horizon of the preliminary trajectories 111 be estimated. Preliminary trajectories 111 The need to build or break down high speeds within a very short time, as well as lane changes in too short time intervals can be excluded. In sum, thus possible target areas / possible driving maneuvers can be selected efficiently.

For example, in global planning, the procedure is as follows. In a first step, a leading vehicle per lane becomes 102 and / or a rear vehicle 102 determined. Possibly. a phantom vehicle may be at the edge of the horizon perception of the environment sensors of the ego vehicle 100 Be considered when in the immediate vicinity of the ego vehicle 100 no vehicle 102 is detected. This way, ICS (Inevitable Collision States) can be avoided. In a further step, reference trajectories from the predicted preliminary vehicle trajectories 111 be determined with distance law. Furthermore, a minimum time t _min or a maximum time t _max at maximum possible / permitted accelerations / decelerations for the reference trajectories or for the provisional trajectories 111 be determined. In addition, a speed profile v _{max can be} superposed, which was determined on the basis of the driver's request, the speed limit, and / or the viewing horizon. The individual preliminary trajectories 111 can be valued with a quality-functional, and it can be the most favorable preliminary trajectory 111 be selected to determine a target area / driving maneuver. If in the context of global planning no valid trajectory 111 If necessary, the global planning can be repeated taking into account an increased emergency parameterization (eg emergency acceleration, emergency deceleration).

Based on the calculated preliminary trajectories 111 The global planning can thus be a target area / driving maneuver selected from the plurality of possible target areas / driving maneuvers. The plurality of possible target areas / driving maneuvers can be determined such that it is ensured that the possible target areas / driving maneuvers by the ego vehicle 100 can be implemented. The driving dynamics (eg a maximum possible acceleration) of the ego vehicle can thereby 100 be taken into account. Furthermore, as stated above, a selection measure can be calculated. The destination area can then be selected from the plurality of destination areas for which the relatively best selection measure is determined.

In the in 1 As an example, in the context of global planning, the left lane is selected as the destination area. After selecting a target area, the actual trajectory planning can then take place within the framework of a local planning. It can be a variety of possible trajectories 112 can be determined, and it can from the variety of possible trajectories 112 an optimal trajectory 112 be selected (eg based on a local selection measure).

By dividing the trajectory planning in a global planning and in a subsequent local planning, the effort to determine a (best) optimal trajectory 112 (also referred to as the desired trajectory) can be reduced. In particular, a target area can be determined with relatively little effort, and the trajectory planning can be focused on the determined target area.

2 shows an exemplary functional structure of a system for controlling the longitudinal / lateral guidance of a vehicle 100 , The system includes a block 210 for the global planning of a trajectory and a block 220 for the local planning of a trajectory. The block 210 for global planning can be a first subblock 211 for global planning while "driving" the vehicle 100 and a second sub-block 215 for global planning when "parking" the vehicle 100 include. As part of global planning 210 can for the "driving" as well as for the "parking" a basic maneuver planning (blocks 212 . 216 ) with respect to the longitudinal / transverse guidance of the vehicle 100 respectively. In particular, for both cases, a target area can be determined for which a precise, local trajectory 112 to be determined. Furthermore, relevant objects from the environment of the vehicle (in particular the objects 102 in relation to the determined target area or to the determined maneuver) (blocks 213 . 217 ).

A coordinator unit 201 can select whether, in consideration of the current driving situation, that in the sub-block "Driving" 211 or in the sub-block "Parking" determined driving maneuver or the determined target area is relevant. The determined maneuver / the determined target area is then sent to the block 220 forwarded for local planning.

In the sub-blocks 221 . 222 the local planning 220 can be determined whether the longitudinal guidance of the vehicle 100 should be done by the driver or whether the longitudinal guidance by automatic actions of the vehicle 100 should be supported or carried out. About the switch 225 can be determined whether a manual longitudinal guidance by the driver (sub block 221 ) or an automatic / supportive longitudinal guidance by an assistance function of the vehicle 100 (Sub-block 222 ). In an analogous way, via the sub-blocks 223 . 224 and over the switch 226 a selection regarding an automatic / supportive lateral guidance (subblock 223 ) and a manual lateral guidance by the driver (subblock 223 ) respectively.

The selection can z. B. depending on one or more driver assistance functions of the vehicle 100 respectively. In the case of a first driver assistance function (eg in the case of an evasive assistant), z. B. the support of the vehicle 100 in the driving maneuver on a support of the transverse guidance of the vehicle 100 be limited. On the other hand, in the case of a second driver assistance function (eg in the case of Adaptive Cruise Control), automatic assistance by the vehicle may possibly occur 100 be limited to a support of the longitudinal guide.

Depending on the selected combination with respect to the longitudinal guide / transverse guide (ie depending on the application of the trajectory to be determined) can then by the trajectory calculation unit 227 a multitude of possible trajectories 112 in relation to the selected target area.

In a first selection unit 228 Driving dynamics aspects can be considered. In particular, depending on one or more vehicle parameters and / or depending on a current driving situation, the trajectories 112 from the multitude of possible trajectories 112 be selected with the vehicle 100 can be driven realistically. It can be z. B. the comb circle for the vehicle 100 be taken into account in the current driving situation. Furthermore, one of the vehicle 100 feasible curvature be considered. Further examples of vehicle parameters that can be taken into account are an acceleration or deceleration of the vehicle (which can be implemented in the current driving situation) 100 ,

In a second selection unit 229 can the remaining possible trajectories 112 be checked to see if with them a collision with the detected objects 102 around the ego vehicle 100 can be avoided. For example, as a selection measure, a collision probability for the possible trajectories 112 be determined, and it may be the possible trajectory 112 with the lowest probability of collision as the desired trajectory 112 to be selected.

The thus determined target trajectory 112 can then contact one or more controllers 202 . 203 for the transverse guide / longitudinal guidance of the vehicle. In particular, a web guide controller 202 used to make sure the vehicle 100 along the determined desired trajectory 112 moves. Furthermore, a vehicle control controller 203 ensure that the required braking / acceleration / steering torques are provided. The division of the longitudinal / transverse control into a web guider 202 and a vehicle control controller 203 is advantageous because due to the division a relatively simple model of the vehicle 100 within the path control controller 202 can be used, and thus the regulation of the guidance of the vehicle 100 along the determined desired trajectory 112 can be made more robust and stable.

• • ein Korridorverlauf für ein Fahrmanöver;
• • eine Zieltrajektorie innerhalb des Korridors eines Fahrmanövers;
• • die Definition/Festlegung von statischen und dynamischen Objekten für eine globale Kollisionsprüfung im Rahmen der globalen Planung;
• • die Definition/Festlegung darüber, welcher Teil eines Korridors eines Fahrmanövers auf Kollisionen geprüft werden soll;
• • die Definition/Festlegung, ob die Möglichkeit von Kollisionen mit einzelnen Objekten geprüft werden soll;
• • die Definition des Verhaltens des Verfahrens bei Vorliegen einer ermittelten Kollision;
• • Festlegen eines Toleranzbandes für die Ermittlung einer Trajektorie 111, 112;
• • Vorgabe, ob eine Trajektorie für die Längs- und/oder Querführung des Fahrzeugs 100 ermittelt werden soll;
• • Festlegen eines Kollisionsraums, der zu berücksichtigen ist; und/oder
• • Festlegen eines Planungshorizonts für die Ermittlung einer Trajektorie 111, 112.

The planning of a trajectory 111 . 112 can be influenced by one or more planning parameters. In particular, planning a trajectory 111 . 112 depending on one or more planning parameters. A scheduling parameter may be from a unit external to the one in 2 presented system can be specified. Exemplary parameters are:

• • a corridor course for a driving maneuver;
• A target trajectory within the corridor of a driving maneuver;
• • the definition / definition of static and dynamic objects for a global collision check in the context of global planning;
• • the definition / determination of which part of a corridor of a driving maneuver should be checked for collisions;
• • the definition / determination of whether the possibility of collisions with individual objects should be checked;
• • the definition of the behavior of the procedure in the event of a detected collision;
• • Defining a tolerance band for the determination of a trajectory 111 . 112 ;
• • Specification, whether a trajectory for the longitudinal and / or transverse guidance of the vehicle 100 to be determined;
• • defining a collision space to be considered; and or
• • Defining a planning horizon for the determination of a trajectory 111 . 112 ,

3 shows a flowchart of an exemplary method 300 for determining a (desired) trajectory 112 for a vehicle 100 , The determination of a trajectory 112 preferably takes place in a deviated coordinate system, relative to a roadway course. The procedure 300 may therefore include the step, 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 (de-curved) Frenet coordinate system.

The warp is exemplary in 4 shown. For the decurling measurement signals are regarding the condition of the vehicle 100 transformed into a lane coordinate system. The trajectory planning itself does not find itself in a Cartesian coordinate system 401 instead of in a Frenet coordinate system. The Frenet coordinate system is relative to a reference curve 400 (eg the lane center of a roadway). The vehicle position is thus replaced by the variables s (t) 403 in the longitudinal direction and d (t) 402 described in the transverse direction. ṡ (t) and d. (t) describe the longitudinal and lateral velocities and s .. (t) and s .. (t) describe the accelerations. Both the vehicle's own movement and the road users to be considered are taken into account in the Frenet coordinate system. Clearly, this transformation corresponds to the warping of the coordinate system 401 and thus allows the separate optimization of the longitudinal and transverse movement of the vehicle.

The determination of a trajectory 112 taking into account physical, technical and comfort-oriented boundary conditions results in an optimization problem with constraints. The resulting optimization problem due to hard boundary conditions (such as driving dynamics limitations, actuator limitations and collision freedom) on equation and inequality constraints on. This fact makes the solution of the optimization problem a complex problem, especially when the number and shape of the constraints changes. This can occur in particular due to the different requirements of different driver assistance functions.

In order to ensure the convergence of the optimization and to determine the global optimum, the convexity of the optimization problem is typically to be proven, which can typically not be guaranteed due to the existing constraints. Thus, the optimization would have to be done with different initial conditions to ultimately achieve the best result. Furthermore, when using an optimization algorithm, it is difficult to estimate the maximum number of iterations and thus the required computing time, which is the case for a control unit of a vehicle 100 , which trajectories 112 in real time, can lead to substantial problems.

For the reasons mentioned above, an alternative solution to the optimization problem is used below, which makes use of the structure of the entire optimization problem and the knowledge of the limitedness of the possible solutions. This can be a significant reduction of the computational effort can be achieved.

The transverse and longitudinal movement 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 (ie a vehicle model). In this case, x ₁ (t) is a first state variable of the vehicle 100 indicating the position of the vehicle 100 (in the longitudinal direction or in the transverse direction) describes. As an input of the integrator system, the jerk x (3) / 1 (t) (ie the 3 ^th derivative of the state variable x ₁ (t)) are defined. In this document, however, it is proposed to use as an input the derivative of the jerk x (4) / 1 (t) (ie, the 4 ^th derivative of the state variable x ₁ (t)) to be used. Thus, 7th order polynomials can be used as an approach, as shown below. In particular, this allows the number of degrees of freedom (eg, the number of boundary conditions) to be increased, which is in particular with regard to the transverse guidance of the vehicle 100 is advantageous.

wobei die Eingangsröße u der Ableitung des Rucks x (4) / 1(t) entspricht. Der Zustand eines Fahrzeugs 100 zu einem bestimmten Zeitpunkt t kann durch den Zustandsvektor x^T = [x₁, x₂, x₃, x₄] beschrieben werden, wobei x₂(t) = ẋ₁(t), x₃(t) = ẋ₂(t) und x₄(t) = ẋ₃(t). The integrator system used in particular for the cross-planning can be defined as follows:

where the input u is the derivative of the jerk x (4) / 1 (t) equivalent. The condition of a vehicle 100 at a certain time t can be described by the state vector x ^T = [x ₁ , x ₂ , x ₃ , x ₄ ], where x ₂ (t) = ẋ ₁ (t), x ₃ (t) = ẋ ₂ (t) and x ₄ (t) = ẋ ₃ (t).

It can be shown that the state vector x (t) results as

The equations above describe a polynomial of the 7th order with respect to the spatial course x ₁ (t).

To solve the optimization problem based on a polynomial of 7th order, the following quality functional can be used:

The parameters c ₀₁₂₃ ^T = [c ₀ , c ₁ , c ₂ , c ₃ ] are calculated from the initial conditions x (0) = x _{0 of} the trajectory 112 at time t = 0 as c ₀₁₂₃ = M -1 / 1 (0) x ₀

The parameters c ₄₅₆₇ ^T = [c ₄ , c ₅ , c ₆ , c ₇ ] are calculated from the final conditions x (t _f ) of the trajectory 112 at the time t = t _f as c ₄₅₆₇ = M -1 / 2 (t _f ) (x (t _f ) - M ₁ (t _f ) c ₀₁₂₃ )

The final conditions can be specified as in the above formula. Alternatively, a reference profile described by x _ref = [x _{1, ref} , x _{2, ref} , x _{3, ref} , x _{4, ref} ] ^{T can be} specified. The optimization goal is to come as close as possible to this reference curve. In this case, the parameters c ₄₅₆₇ ^T can be calculated as follows:

The optimization problem now consists of both the end time or the end time t _f and possibly the final state x (t _f ) for an optimal trajectory 112 to determine.

In particular, for the calculation of a transverse trajectory polynomials of the 7th order can be chosen so as to obtain the third derivative of the trajectory 112 at the beginning and end of the trajectory 112 to pretend. In particular, so much more accurate steering angle at the beginning and at the end of the trajectory 112 be specified. As the desired endpoint of a trajectory 112 The target area d _target can be used as part of global planning (eg, using a preliminary trajectory 111 ) was determined. This target area can, for. B. are determined as the end position of the final state x (t _f ), with x ₁ (t _f ) = d _goal . In particular, different trajectories 112 in a target window, whereby the target window was determined as part of global planning.

As a measure of choice or as a measure of quality for the determination of a trajectory 112 for the transverse guidance of the vehicle 100 the following function can be used:

The first expression evaluates the evolution of the derivative of the jerk along the trajectory 112 , The second term evaluates the deviation of the final position d (t _f ) from the target position d _target . Furthermore, the third expression evaluates the time length of the trajectory 112 , The expression of the trajectory can be determined by the weighting factors k _q1 and k _q2 112 be weighted.

The longitudinal planning can be done in a similar way. However, it has been shown that a 4th or 5th order polynomial is sufficient for longitudinal planning. Furthermore, a combination (eg a summation) of the selection dimensions for the transverse guidance and for the longitudinal guidance can be combined with longitudinal and lateral planning.

To determine an optimal trajectory, the selection measure J can be calculated for different values of t _f and / or for different final states x (t _f ). The above formulas can be used for this. This results in a course of the selection measure J, wherein a minimum value of the selection measure J indicates the optimum value for the time t _f and / or for the final state x (t _f ). In a further step, one or more secondary conditions can then be taken into account.

The secondary conditions can thus be taken into account downstream of the optimization. For this purpose, the best trajectory in the sense of the cost function J (ie the selection measure) 112 selected and checked for compliance with the constraints. If the constraints are met, the trajectory is maintained 112 implemented. Otherwise, the next best trajectory 112 selected and checked for the constraints. This procedure is applied until an optimal solution is found which fulfills the secondary conditions.

As constraints, actuator and vehicle dynamics limitations can be considered. Furthermore, the collision freedom with predicted object trajectories of other objects / vehicles 102 be taken into account.

This in 3 illustrated method 300 for determining a trajectory 112 for the longitudinal and / or transverse guidance of a vehicle 100 includes determining 301 of initial values x (0) = x ₀ for a plurality of state variables x of the vehicle 100 , The plurality of state variables comprises a position x ₁ (t) of the vehicle 100 , a speed ẋ ₁ (t) of the vehicle 100 , an acceleration ẍ ₁ (t) of the vehicle 100 and a jerk x (3) / 1 (t) of the vehicle 100 , There are x (3) / 1 (t) the third derivative of the position x ₁ (t) of the vehicle 100 at.

The procedure 300 also includes determining 302 of final values x (t _f ) at an end time t _f for the plurality of state variables x of the vehicle 100 , In addition, the process includes 300 the determining 303 a trajectory 112 on the basis of the initial values x (0) = x ₀ , the final values x (t _f ), the end time t _f and on the basis of a polynomial of 7th or higher / lower order. The polynomial of 7th or higher / lower order may be the position x ₁ (t) of the vehicle 100 as a function of the time t between the initial position x ₁ (0) and the end position x ₁ (t _f ) of the trajectory 112 describe. To calculate the (state) trajectory x (t) 112 For example, the formulas given above for c ₀₁₂₃ ^T , c ₄₅₆₇ ^T and x (t) can be used. In particular, by the formula for x (t), the position becomes x ₁ (t) of the vehicle 100 described by a polynomial of 7th order.

To determine an optimal (state) trajectory 112 (in the sense of a selection measure J), the values of the selection measure J for (state) trajectories 112 be determined with different end times t _f and / or with different end values x (t ₁ ) of the plurality of state variables. It can then be the (state) trajectory 112 which optimizes the selection measure J. Furthermore, it can be checked whether one or more constraints are met (as stated above).

The determined trajectory 112 can finally be transformed back from the Frenet coordinate system into a Cartesian coordinate system. Furthermore, the determined trajectory 112 be used to guide the vehicle (eg for an evasive maneuver or for a parking maneuver).

Thus, in the context of a local planning precise target trajectories 112 be determined. Especially when describing a trajectory 112 By means of a 7th-order polynomial, specifications can be made regarding the steering angle / jerk of the trajectory 112 be made, reducing the comfort of the trajectory detected 112 can be increased.

The structure described in this document for the determination of a trajectory 112 is advantageous because a saving of resources (eg of computing time, memory, vehicle electrical system) is made possible. By providing global planning for "driving" and "parking", transitions between parking and driving (and vice versa) can be represented. Furthermore, the described structure is scalable and modular. For example, isolated trajectory planning for transverse or longitudinal guidance, as well as a combined trajectory planning for transverse and longitudinal guidance is possible. Possibly. Trajectory planning can only be done for parking functions. Furthermore, a collision check of determined trajectories can be used. In addition, a determined trajectory can be adapted to driving-dynamic boundary conditions (eg coefficient of friction).

Possibly. the determination of a trajectory can be adapted to the setting of a driver preference (eg to a set driving experience switch). For example, a driver may select a comfortable, an economical, or a sporty attitude of the vehicle. The trajectory can then be determined as a function of such a setting.

Furthermore, the start / end condition steering angle / curvature change can be taken into account. In addition, a quality functional in the global planning can be extended as needed (eg by using a backend, by taking into account traffic signs, etc.) in order to move energetically optimally in the traffic.

The present invention is not limited to the embodiments shown. In particular, it should be noted that the description and figures are intended to illustrate only the principle of the proposed methods, apparatus and systems.

Claims (15)

Method for determining a trajectory ( 112 ) for the longitudinal and / or transverse guidance of a vehicle ( 100 ), the method comprising, - detecting one or more objects ( 102 ) in an environment of the vehicle ( 100 ), based on environmental data relating to an environment of the vehicle ( 100 ); Performing a global planning in order, depending on the one or more detected objects ( 102 ) to determine a variety of potentially possible driving maneuvers; Selecting a driving maneuver from the plurality of potentially possible driving maneuvers; and Performing a local planning in order to obtain a desired trajectory for the selected driving maneuver ( 112 ) for the longitudinal and / or transverse guidance of the vehicle ( 100 ) to investigate. The method of claim 1, wherein performing the global scheduling comprises determining a plurality of different local target areas for the vehicle. 100 ), based on the environment data and taking into account the one or more detected objects ( 102 ); and - determining a potentially possible maneuver for each local target area. Method according to one of the preceding claims, wherein - carrying out the global planning comprises determining a plurality of preliminary trajectories ( 111 ) for the corresponding multiplicity of potentially possible driving maneuvers; and - a model for describing a preliminary trajectory ( 111 ) has a smaller number of parameters than a model for describing the desired trajectory ( 112 ) and / or for determining a preliminary trajectory ( 111 ) a smaller number of trajectories is determined than for the determination of the desired trajectory ( 112 ). Method according to one of the preceding claims, wherein the plurality of potentially possible driving maneuvers comprises, - at least one potentially possible driving maneuver for driving the vehicle ( 100 ); and - at least one potentially possible driving maneuver for parking the vehicle ( 100 ). The method of claim 4, wherein detecting one or more objects ( 112 ) in the vicinity of the vehicle ( 100 ), - detecting one or more objects ( 102 ) necessary for a driving maneuver while driving the vehicle ( 100 ) are relevant; and - detecting one or more objects ( 102 ) required for a maneuver when parking the vehicle ( 100 ) are relevant. Method according to one of the preceding claims, wherein the plurality of potentially possible driving maneuvers comprises one or more of - a lane change of the vehicle ( 100 ) to a left lane of a multi-lane road ( 101 ); A lane change of the vehicle ( 100 ) to a right lane of the multi-lane road ( 101 ); and / or - a deceleration or acceleration on a currently traveled road ( 101 ). Method according to one of the preceding claims, wherein selecting comprises a driving maneuver, - determining a value of a global selection score for each of the plurality of potentially possible driving maneuvers; and Selecting the driving maneuver on the basis of the values of the global selection measure for the plurality of potentially possible driving maneuvers. Method according to one of the preceding claims, wherein - carrying out the local planning comprises determining an application range of the desired trajectory ( 112 ); - the scope of one includes: - automatic support only for the longitudinal guidance of the vehicle ( 100 ), - automatic support only for the transverse guidance of the vehicle ( 100 ), or - automatic support for both the longitudinal guidance and the lateral guidance of the vehicle ( 100 ); and - the desired trajectory ( 112 ) is determined depending on the field of application. Method according to one of the preceding claims, wherein carrying out the local planning comprises, - determining a multiplicity of possible trajectories ( 112 ) for the selected driving maneuver; and - selecting the desired trajectory ( 112 ) from the multitude of possible trajectories ( 112 ). The method of claim 9, wherein - performing the local scheduling comprises checking for satisfaction of one or more constraints; The one or more boundary conditions comprise a driving dynamic boundary condition and / or a collision freedom with one or more objects ( 102 ) along the desired trajectory ( 112 ) are directed; and - the desired trajectory ( 112 ) is selected depending on the one or more boundary conditions. The method of claim 9, wherein performing the local scheduling comprises determining a local selection measure for one or more of the plurality of possible trajectories. 112 ); and - the desired trajectory ( 112 ) is selected depending on the local selection amount. Method according to one of claims 9 to 11, wherein the determination of a possible trajectory ( 112 ) for the selected driving maneuver, - determining ( 301 ) of initial values for a plurality of state variables of the vehicle ( 100 ); wherein the plurality of state variables is a position (x ₁ (t)) of the vehicle ( 100 ), a speed (ẋ ₁ (t)) of the vehicle ( 100 ), an acceleration (ẍ ₁ (t)) of the vehicle ( 100 ) and a jerk ( x (3) / 1 (t) ) of the vehicle ( 100 ); - Determine ( 302 ) of end values at an end time for the plurality of state variables of the vehicle ( 100 ); and - determining ( 303 ) of the possible trajectory ( 112 ) on the basis of the initial values, the end values, the end time and on the basis of a polynomial of 7th order. Method according to one of the preceding claims, further comprising determining a steering input for a power steering system of the vehicle ( 100 ) and / or a deceleration specification for a brake system of the vehicle ( 100 ) and / or an acceleration specification for a drive of the vehicle ( 100 ) as a function of the desired trajectory ( 112 ). Device for determining a trajectory ( 112 ) for the transverse and / or longitudinal guidance of a vehicle ( 100 ), wherein the device is set up, - on the basis of the environmental data relating to an environment of the vehicle ( 100 ), one or more objects ( 102 ) in an environment of the vehicle ( 100 ) to detect; To carry out a global planning in order to determine, depending on the one or more detected objects ( 102 ) to determine a variety of potentially possible driving maneuvers; To select a driving maneuver from the plurality of potentially possible driving maneuvers; and - carry out a local planning in order to obtain a desired trajectory for the selected driving maneuver ( 112 ) for the longitudinal and / or transverse guidance of the vehicle ( 100 ) to investigate. Procedure ( 300 ) for determining a trajectory ( 112 ) for the longitudinal and / or transverse guidance of a vehicle ( 100 ), the process ( 300 ), - determining ( 301 ) of initial values for a plurality of state variables of the vehicle ( 100 ); wherein the plurality of state variables is a position (x ₁ (t)) of the vehicle ( 100 ), a speed (ẋ ₁ (t)) of the vehicle ( 100 ), an acceleration (ẍ ₁ (t)) of the vehicle ( 100 ) and a jerk ( x (3) / 1 (t) ) of the vehicle ( 100 ); - Determine ( 302 ) of end values at an end time for the plurality of state variables of the vehicle ( 100 ); and - determining ( 303 ) a trajectory ( 112 ) on the basis of the initial values, the end values, the end time and on the basis of a polynomial of 7th order.

Images