Method, data processing device and software for determining deceleration strategy of vehicle

Abstract

A method for determining a deceleration strategy for a vehicle (1) comprises the steps of determining a target position located in a forward route section together with a target speed that the vehicle should have when reaching the target position, where the target speed is lower than a current speed of the vehicle; and ascertaining a reference trajectory of the vehicle, the reference trajectory being set to reach the target position at the target speed. In this case, the reference trajectory is determined on the basis of an analytic function of at least two deceleration modes from a list, which comprises a coasting mode, a towing mode, a recovery mode and an active braking mode. An analytic function of each of the at least two deceleration modes specifies a speed and / or a travelled distance depending on a time variable. Ascertaining the reference trajectory comprises ascertaining one or more switching time points between the at least two deceleration modes.

Description

Field of the Invention

[0001] The present invention relates to a method for determining a deceleration strategy of a vehicle and a data processing device. Furthermore, the present invention relates to a software for executing such a method by means of a data processing device and a computer-readable storage medium in which such software is stored. Background Art

[0002] Modern autonomous driving functions actively intervene in the lateral and longitudinal guidance of a vehicle. For this purpose, trajectory planning that plans the movement of the vehicle several seconds in advance is usually used. The focus in implementing trajectory planning generally lies on comfort and safety.

[0003] Efforts for energy-saving trajectory planning for autonomous driving are also known from the prior art. For example, see Terwen, Stephan: Vorausschauende schwerer Lastkraftwagen. Karlsruher Institut für Technologie, PhD, 2010; Radke, Tobias: Energieoptimale by motor vehicles through the use of anticipatory driving strategies. Karlsruhe Institute of Technology, PhD, 2013; Yan et al: Eco-Coasting Strategies Using Road Grade Preview: Evaluation and Online Implementation Based on Mixed Integer Model Predictive Control, arXiv, 2015; A Sciarretta, Giovanni de Nunzio, L Leon Ojeda. Optimal Ecodriving Control: Energy-Efficient Driving of Road Vehicles as an Optimal Control Problem. IEEE Control Systems Magazine, Institute of Electrical and Electronics Engineers, 2015; Saerens, Bart: Optimal Control Based Eco-Driving, Katholieke Universiteit Leuven, PhD, 2012. Most of these proposals use computationally intensive optimization methods to minimize fuel or energy consumption over a period of time. To this end, they require detailed modeling of specific powertrain systems, fuel consumption, and / or battery charge status.

[0004] In addition, solutions already exist in the prior art for adapting the longitudinal dynamics of a motor vehicle in advance to a defined situation in a preceding route section. In particular, solutions based on a multi-stage deceleration strategy have been proposed for the vehicle deceleration process. For example, from the applicant's published patent applications DE 102014215673 A1 and DE 102015205371 A1, a driver assistance system is respectively known, which determines a single-stage or multi-stage deceleration strategy when a preceding event requiring deceleration is recognized, and arranges for the automatic implementation of the deceleration strategy after an operation by the driver, wherein the deceleration phases of the deceleration strategy are determined according to predetermined parameters. The applicant's DE 102017205134 A1 describes a deceleration assistance system for a motor vehicle, which has an electronically controllable automatic transmission and an electronic control unit for controlling the gearshift up and down processes of the transmission. Here, a deceleration strategy for reducing the speed to a target speed at the location of the preceding event is determined, wherein the deceleration strategy includes performing an automatic downshift process and without the driver having to perform the control of an operating element. In addition, some electric vehicles produced by the applicant are already equipped with predictive, situation-specific drive functions, wherein so-called adaptive deceleration or adaptive recuperation can be achieved. Here, "recuperation" is generally understood as recovering electrical energy from the (previously power-generated) kinetic energy of an electric vehicle and storing it in the battery, for example during the braking process. Summary of the Invention

[0005] In this context, the object of the present invention is to provide a method for determining a deceleration strategy for the automatic deceleration process of a vehicle, which enables a particularly computationally efficient implementation and can be used as generally as possible, for example, in combination with a conventional longitudinal guidance controller.

[0006] This object is solved by the subject matter of the independent patent claims. Advantageous embodiments are given in the dependent claims.

[0007] It should be noted that additional features of dependent patent claims that refer to independent patent claims can, if they do not have the features of the independent patent claims or are only combined with partial features of the independent patent claims, form an invention independent of all combinations of features of the independent patent claims, which can be the subject matter of an independent patent claim, a divisional application or a subsequent application. This also applies to the technical teachings explained in the description, which can form an invention independent of the features of the independent patent claims.

[0008] A first aspect of the present invention relates to a method for determining a deceleration strategy for a vehicle.

[0009] The vehicle can in particular be a motor vehicle. The term "motor vehicle" should again in particular be understood as a land vehicle that is not connected to a rail track and moves by mechanical force. For example, a motor vehicle in this sense can be designed as a car, a motorcycle or a tractor. For example, the vehicle can be a vehicle driven by an internal combustion engine, an electrically driven vehicle or a hybrid vehicle composed of an internal combustion engine and an electric drive.

[0010] One step of the method is to determine a target position in a front route section together with a target speed that the vehicle should have when reaching the target position, wherein the target speed is lower than the current speed of the vehicle.

[0011] For example, the control device of the vehicle can identify in advance, based on map data and / or based on data of an environmental sensor system, a target situation that requires the vehicle to decelerate, such as a turn entry or a roundabout or a traffic sign indicating a speed limit. To identify such target situations, for example, known map and / or environmental sensor data fusion methods can be used, including known artificial intelligence methods.

[0012] For example, the most likely vehicle path can first be determined in the map. This can be done, for example, based on active target guidance, or by a separate calculation based on the vehicle position, in combination with map attributes in the vehicle environment (such as road grade, angle of direction change, etc.) and / or driver behavior interpretation, which is evaluated according to the most likely vehicle path, for example, setting an indicator light or entering a turning lane. Then, characteristics such as road type and turning radius and relevant situations such as starting speed limit, red light, turn, intersection or roundabout are extracted from the map data and / or environmental sensor data along the predicted most likely vehicle path. Then, these front situations can be determined together with the relevant target position and target speed as the next target situation.

[0013] Another step of the method is to determine a reference trajectory of the vehicle, wherein the reference trajectory is compatible with reaching the target position at the target speed. This means that if the vehicle moves accordingly with the reference trajectory, it will have the target speed when reaching the target position.

[0014] Herein, the reference trajectory can in particular refer to the longitudinal guidance of the vehicle, that is, for example, describing the longitudinal movement of the vehicle along the path from the current vehicle position to the target position (wherein the path itself does not have to extend linearly, but can also be curved). For example, the reference trajectory can include a time-dependent route, such as a given of the driving distance traveled in the direction of the target position up to the time point being considered. In addition, the reference trajectory can also give the time-dependent longitudinal speed and / or time-dependent longitudinal acceleration of the vehicle.

[0015] According to this method, the reference trajectory is determined based on the corresponding analytical functions of at least two, preferably at least three deceleration modes from a list, which includes coasting operation, dragging operation, recuperation operation, and active braking. Herein, the analytical functions of each of the at least two, preferably at least three deceleration modes respectively specify the speed and / or the traveled driving distance as a function of the time variable, i.e., the current distance from the starting point of the deceleration process.

[0016] The "coasting mode" shall be understood as the vehicle coasting completely or coasting with substantially no force lock-up. In a vehicle with an internal combustion engine, this operating state is characterized, for example, by the interruption of the force lock-up between the transmission and the (possibly switched-off) engine.

[0017] For a vehicle with an electric drive motor (which may not have a transmission), the coasting operation can be understood within the scope of this document as an operating state in which the electric drive motor does not generate a mechanical driving torque and is also not in generator operation (recuperation operation mode), i.e., does not convert kinetic energy into electrical energy. In other words, in the coasting operation, little or no electrical energy is converted into mechanical energy by the electric drive motor, and vice versa. For example, in the coasting state, the electric drive may not generate or only generate a very small braking torque. Correspondingly, the electric drive motor rotates completely freely in the coasting operation and only causes a relatively small deceleration, for example, due to the friction of the brushes (if any) or the like.

[0018] The dragging operation, which is usually also referred to as motor braking (motor), shall be understood as a driving state in which there is a force lock-up between the transmission and the motor, which ensures effective deceleration of the motor vehicle even without active braking.

[0019] As mentioned at the beginning, the recuperation operation is characterized by recovering electrical energy from kinetic energy by means of the vehicle's electric drive motor. Thereby, a significant deceleration effect can be achieved.

[0020] Active braking is achieved, for example, by means of a dedicated braking device, usually in the case where the braking force is adjustable.

[0021] According to the invention, it is also provided that determining the reference trajectory includes determining one or more switching time points between at least two, preferably at least three deceleration modes based on the analytical functions. Herein, determining the switching time points can be carried out based on parameter optimization regarding the switching time points.

[0022] According to one embodiment, determining the reference trajectory further includes determining the braking setting parameters effective during the braking operation. In particular, for example, the time curve of the customized braking setting parameters can be determined by minimizing a cost function.

[0023] According to one embodiment, the parsing function can be given as a closed mathematical expression having a time variable and function parameters characterizing one or more types of driving resistance. For example, the parsing function having a time variable and function parameters can be stored in software and called when the method is executed. The driving resistance can include, for example, air resistance and / or rolling resistance, preferably depending on the road surface gradient or other predicted road surface information, and these resistances contribute to the deceleration of the vehicle.

[0024] According to one embodiment, the function parameters can further include one or more deceleration setting parameters for one or more of the mentioned deceleration modes. Here, the deceleration setting parameters can be assumed to be constant in time, for example, during the corresponding deceleration phase in which the corresponding deceleration mode is activated. For example, the deceleration setting parameter can characterize the deceleration during the towing operation of the vehicle. Another deceleration setting parameter can characterize the deceleration in the recuperation mode. Still another deceleration setting parameter can characterize the deceleration caused by active braking.

[0025] According to one embodiment, one or more parsing functions for the traveled driving distance given according to the time variable can be given in the following form:

[0026] s(t) = A ln(tan(Bt + C) 2 + 1) + D

[0027] Here, s represents the driving distance depending on the time variable t. A, B, C, D are function parameters that are constant in time. For example, the function parameters A, B, C, D can come from or include function parameters related to the driving resistance or deceleration setting parameters of different deceleration modes. It should be noted here that for different deceleration modes, the above expression s(t) can take different forms respectively. In particular, the function parameters A, B, C, D of the different functions s(t) assigned to different deceleration modes can be different from each other. Specific expressions of the function parameters A, B, C, D of the function s(t) for coasting operation, towing operation, and active braking will be given below with reference to an embodiment. Here, the correlation between the driving resistance and deceleration setting parameters of different deceleration modes is obvious.

[0028] According to another embodiment, especially in combination with the above embodiment regarding the determination form of the parsing expression for the driving distance, one or more parsing functions for the correlation between the speed and the time variable given for at least two deceleration modes are given in the following form:

[0029] v(t) = E tan(Ft + G),

[0030] where the speed is v, the time variable is t, and the constant function parameters are E, F, and G. The above explanations of the function parameters A, B, C, and D also apply to the function parameters E, F, and G. In particular, the function parameters E, F, and G of the speed functions for different deceleration modes can also be different from each other.

[0031] In the scope of the present invention, the switching time point can also be obtained based on parameter optimization by minimizing the cost function.

[0032] According to one embodiment, the cost function includes the sum of a plurality of time integrals, where each integral extends in time over its respective deceleration phase, which is assigned to the corresponding deceleration mode of at least two, preferably at least three deceleration modes. One or more switching time points to be obtained here form the upper or lower limit of the integral. Further integral limits can be formed by the start time point and the end time point of the entire deceleration process.

[0033] According to an implementation variant, the cost function has a time-dependent cost term for each integral in at least two deceleration phases, which imposes a cost on the kinetic energy loss that has occurred up to the considered time point compared to the initial kinetic energy that the vehicle had at the start of deceleration. This cost term advantageously causes the vehicle to make the best use of the initially stored kinetic energy during deceleration in order to reach the target position quickly. Since the cost term "rewards" the vehicle for using the kinetic energy for as long as possible, or conversely, and "penalizes" an overly rapid reduction in kinetic energy, the vehicle ends up coasting towards the target position at a relatively high speed for as long as possible.

[0034] According to an extension, a cost term that imposes a cost on active braking can also be set in the cost function, for example, a cost term formed by a braking adjustment parameter, such as the square of the braking deceleration. Thus, such a cost term can prevent overly sudden braking, for example, braking shortly before reaching the target position and thus improve comfort and safety.

[0035] As a possible implementation of the first aspect of the invention, the aforementioned sub-aspect related to the cost term can also be implemented independently of the features related to determining the reference trajectory according to the analytical function in the aforementioned method according to the first aspect of the invention, and this cost term imposes a cost on the kinetic energy loss up to the considered time compared to the initial kinetic energy.

[0036] Correspondingly, as a second independent aspect of the present invention, a method for determining a deceleration strategy for a vehicle is provided, which method particularly comprises the following steps: determining a target position in a front route section together with a target speed that the vehicle should have when reaching the target position, wherein the target speed is lower than the current speed of the vehicle; and determining a reference trajectory of the vehicle, which reference trajectory is set to reach the target position at the target speed. Here, the reference trajectory has at least two deceleration stages, wherein the vehicle runs in different deceleration modes from a list respectively, and the list includes a coasting mode, a drag operation, a recuperation operation and an active braking. Determining the reference trajectory includes determining one or more switching time points between at least two deceleration modes, wherein determining the one or more switching time points is realized based on the minimization of a cost function. The cost function includes the sum of a plurality of time integrals, wherein each integral extends in time over respective deceleration stages, wherein the one or more switching time points constitute the upper or lower limit of the integral, and wherein the cost function has a time-dependent cost term under the integral of at least two deceleration stages, which cost term imposes a cost on the kinetic energy loss occurring up to the time point under consideration compared to the starting kinetic energy that the vehicle has at the start of deceleration.

[0037] The method according to the second aspect of the present invention can be combined with all implementation variants of the method according to the first aspect of the invention, and vice versa.

[0038] According to an embodiment of the method according to the first or second aspect of the invention, the time-dependent cost term includes the square of the difference between the kinetic energy at the time point under consideration and the starting kinetic energy. It is pointed out that the selection of this type of cost term is particularly advantageous for the efficient computational solution of the optimization problem and thus for the characteristics of the resulting reference trajectory in terms of comfort and safety.

[0039] Therefore, in one implementation variant, the cost function can be given in the following form:

[0040]

[0041] J represents the cost, t 0 represents the starting time point of the deceleration process, t f represents the end time point of the deceleration process (when reaching the target position), E kin (v) represents the kinetic energy at time point t, E kin (v 0 ) represents the kinetic energy at the starting time point t 0 when, u represents the deceleration setting parameter, wherein different parameters u can be applicable to the drag operation, the recuperation operation or the braking operation. The factors w u and w kin are constants.

[0042] The second term under the integral of the cost function J is the above-mentioned cost term related to the kinetic energy. The first term under the integral imposes a cost on the active braking process or the tow operation.

[0043] Since the first term imposes a cost on the active braking, it helps to determine the most energy-efficient deceleration strategy because the active braking by means of the braking device must consume energy.

[0044] As mentioned above, in the general expression of the above cost function J, the parameter u depends on the deceleration phase under consideration, for example, the active braking setting parameter, the deceleration setting parameter during the tow operation, or the deceleration setting parameter in the recuperation mode. The time integral can be allocated to multiple sub-integrals for different deceleration phases, which have corresponding time interval limits, such as the coasting phase (the first time interval from t 0 to the first switching time point t s1 , the tow phase (the second time interval from the first switching time point t s1 to the second switching time point t s2 ), and the active braking phase (the third time interval from the second switching time point t s2 to the end time point t f ). In this example divided into three deceleration phases, the deceleration setting parameter generally given as u before is replaced as follows for each partial integral:

[0045]

[0046] Therefore, there is no deceleration setting parameter in the coasting phase (u = 0), the deceleration setting parameter a drag (t) takes effect in the tow phase and another deceleration setting parameter u brake (t) takes effect in the subsequent active braking phase. Below, in the description of an embodiment, it is schematically explained which configurations the sub-integrals can adopt for different deceleration phases.

[0047] According to an embodiment of the first inventive aspect or the second inventive aspect and in accordance with the above exemplary implementation variant, the reference trajectory includes a coasting phase in which the vehicle travels in the coasting mode, and at least one additional deceleration phase that is temporally located after the coasting phase, in which the vehicle travels in the tow operation or the recuperation operation or performs active braking. In other words, the reference trajectory can be determined such that it has a tow phase, a recuperation phase, and / or a braking phase temporally after the coasting phase. In particular, according to an exemplary implementation, the vehicle can first coast, then decelerate in the tow operation, and finally actively brake down to the target speed. When obtaining the trajectory, especially when solving the optimization problem regarding the switching time points, the order of the corresponding deceleration phases can be pre-given.

[0048] In one embodiment of the method according to the first or second aspect of the invention, the method further includes determining a desired trajectory of the vehicle based on the determined reference trajectory, and generating a setting preset for the longitudinal guidance of the vehicle based on the desired trajectory.

[0049] For example, the desired trajectory can be cyclically planned in a relatively short planning horizon of a few seconds, for example 6 seconds. In particular, at the beginning of the entire deceleration process of the vehicle, the planning horizon of the desired trajectory can be shorter than that of the reference trajectory.

[0050] The determination of the target trajectory can be carried out, for example, by means of an optimization method, such as quadratic optimization, while taking into account other criteria, such as criteria related to the safety and comfort of the vehicle or the occupants. Here, the determined desired trajectory can include, for example, state variables of the same type as the reference trajectory, such as the traveled route related to time, the speed related to time, the acceleration related to time, and / or the jerk related to time.

[0051] The reference trajectory can be considered, for example, as a reference when calculating the desired trajectory, such that in the context of determining the desired trajectory, the cost function to be minimized imposes a cost on the deviation between the desired trajectory parameters and the reference trajectory parameters. For example, the cost function can include one or more terms, each of which imposes a cost on the square of the difference between the (to-be-determined) desired trajectory parameters and the corresponding reference trajectory parameters. This can particularly relate to the trajectory parameter distance, speed, and acceleration.

[0052] The reference trajectory can also be considered when calculating the desired trajectory, such that in the context of optimization, the time-dependent speed profile of the reference trajectory is used as an upper limit to determine the speed according to the desired trajectory speed. Alternatively or additionally, the reference trajectory can be considered in such a way when determining the desired trajectory that the time profile of the braking setting parameter generated by the reference trajectory is used as a constraint in the context of optimization to determine the acceleration or deceleration according to the desired trajectory. According to some embodiments, this limitation achieved by the reference trajectory parameters does not have to be rigid, but can be softened, for example, by using slack variables.

[0053] For example, the generation of the setting preset can include, for example, generating an acceleration preset for the longitudinal guidance actuator of the vehicle based on the desired trajectory. For this purpose, the desired trajectory can be transmitted, for example, to a trajectory tracking controller, which outputs the acceleration preset to the longitudinal guidance actuator, such as the drive and / or braking device of the vehicle, based on the desired trajectory and information about the measured disturbance parameter influence.

[0054] It should be noted that obtaining a separate desired trajectory (and possibly a trajectory following regulator) is an optional intermediate step. Embodiments can also be conceived in which the reference trajectory itself is used as the desired trajectory and is output as such to the regulator or directly to the trajectory following regulator or the longitudinal guidance actuator for further processing. This can be achieved, for example, by the reference trajectory specifying an acceleration, and the regulator or actuator being set to process this acceleration as the desired acceleration.

[0055] Generally speaking, the method according to the first or second aspect of the invention may further include generating a setting pre-given for the longitudinal guidance of the vehicle based on the reference trajectory.

[0056] If the reference trajectory is implemented by means of the longitudinal guidance actuator of the vehicle, or more generally, further processed by the adjustment path, the method according to the first or second aspect of the invention can be executed cyclically, wherein the current state of the adjustment parameter (e.g., measured by means of an odometer), such as the current actual speed or the current actual acceleration or the actual travel distance, is fed back to the reference trajectory planner, which executes the steps for determining the reference trajectory according to this method. Then, it can recalculate the reference trajectory based on the current information. Here, according to the above-mentioned implementation variant, a separate desired trajectory planner can be connected between the reference trajectory planner and the adjustment path, which on this side in turn determines the desired trajectory according to the current actual parameters and the reference trajectory with its own, possibly relatively short, planning range, for example, a few seconds. As described above, the adjustment path can include a trajectory tracking regulator and a longitudinal guidance actuator.

[0057] A third aspect of the invention is a data processing device, which is arranged to execute the method according to the first or second aspect of the invention. Correspondingly, an embodiment of the data processing device according to the invention can correspond to an embodiment of the method according to the invention described herein, and vice versa. The data processing device can include one or more processors, such as one or more controllers or microcontrollers, which are arranged to execute the method.

[0058] A fourth aspect of the invention is software containing instructions, which, when executed by a data processing device, cause the data processing device to execute the method according to the first or second aspect of the invention. Correspondingly, for example, the software can be executed on one or more processors, which constitute the data processing device (or its components) according to the third aspect of the invention, or are constituted by the data processing device. Here, the computer program contained in the software can also be divided into a plurality of independent subroutines, which can be executed, for example, on different computers, such as a plurality of independent processors, which are spatially relatively separated from each other.

[0059] A fifth aspect of the invention is a computer-readable storage medium, on which the software according to the fourth aspect of the invention is stored.

[0060] According to some of the foregoing and following embodiments, the present invention is based on the idea of obtaining, in a computationally efficient and thus real-time manner, a safe, comfortable and energy-efficient reference trajectory that decelerates a vehicle to a target speed at a given target position using multiple different deceleration stages. In particular, within the scope of the hybrid optimization control (HOC) problem, the favorable switching time points between different deceleration modes can be determined by means of parameter optimization. It is shown here that, by means of targeted simplifying assumptions, such as the time constancy of the corresponding deceleration setting parameters in different deceleration modes, and by cleverly choosing the cost function for the optimization, a particularly computationally efficient solution can be achieved, which results in a reference trajectory with satisfactory characteristics.

[0061] In the method proposed here, the reference trajectory parameters can remain relatively general, and for example, given the time profiles of the distance traveled, speed, acceleration, and / or jerk, rather than specific drive pre-givens, such as torque, which largely depend on the details of the drive machine. Thereby, the proposed reference trajectory planning can be easily combined with traditional and possibly existing regulator architectures for different vehicles. Thus, existing application trajectory planners can be easily extended with the reference trajectory planner proposed here without having to be replaced entirely. For example, the corresponding reference trajectory planner can be connected upstream of an existing application trajectory planner, where the latter can use the reference trajectory parameters as reference parameters or constraints within the scope of its traditional optimization. Here, the comfort and safety requirements of the trajectory planning based on the existing application trajectory planner are still taken into account.

[0062] According to some embodiments, the proposed reference trajectory planning can in particular be independent of the topology of the vehicle's drive train. Here, for example, detailed drive train and fuel consumption modeling can be dispensed with. Thus, the calibration effort for different vehicle models is reduced. Description of the Drawings

[0063] The present invention will now be explained in more detail on the basis of embodiments and with reference to the drawings. Here, within the scope of the invention, the features and combinations of features mentioned earlier or later in the description and / or the features and combinations of features shown individually in the drawings can be used not only in the combinations given respectively, but also in other combinations or individually.

[0064] Figure 1 Exemplarily and schematically shows a deceleration-related target situation in front of the vehicle in the form of a traffic sign indicating a speed limit;

[0065] Figure 2 Exemplarily and schematically shows the method flow for determining a vehicle deceleration strategy according to a first inventive aspect;

[0066] Figure 3 Exemplarily and schematically shows a method flow for determining a vehicle deceleration strategy according to a second inventive aspect;

[0067] Figure 4 Exemplarily and schematically shows the vehicle longitudinal guidance modeling as a switching system;

[0068] Figure 5 Shows time - and location - related reference trajectory parameters determined by two - dimensional parameter optimization;

[0069] Figure 6 Exemplarily and schematically shows the determination of the switching time point based on the speed - displacement diagram;

[0070] Figure 7 Exemplarily and schematically shows a system for adjusting vehicle longitudinal guidance. Detailed implementation mode

[0071] Figure 1 Schematically shows an example driving situation with vehicle 1, which is currently, i.e., at the initial time t 0 located at the initial position s 0 , and is moving forward constantly at an initial speed of v 0 = 180 km / h.

[0072] Figure 2 and Figure 3 respectively show the method flow for determining the vehicle deceleration strategy.

[0073] The following will schematically refer to Figure 1 the driving situation shown in Figure 2 and Figure 3 to elaborate on the method flow according to Figure 2 and Figure 3 . Here, a data processing device, for example, in the form of one or more controllers of vehicle 1, executes the specific steps in real - time that are consistent with the two method flows according to

[0074] The first step is common in the method flows according to Figure 2 and Figure 3 . In the first step, the target position s f in the front route section is determined, as well as the target speed v f of 80 km / h that vehicle 1 should have when reaching the target position s f . For example, the target position s f and the target vehicle speed v f are extracted from the digital map as information about speed limits, where the target position s f from the current vehicle position s 0700 meters. Since the target speed v f is lower than the initial speed v 0 , it is necessary to slow down vehicle 1. Therefore, a suitable deceleration strategy should be determined for the longitudinal guidance of the vehicle.

[0075] In this embodiment, vehicle 1 is a vehicle with an internal combustion engine, which supports coasting operation, towing operation (engine braking), and active braking as different deceleration modes. Therefore, the deceleration strategy to be determined should be given accordingly, especially for which times or at which sections vehicle 1 should coast, tow, and finally actively brake in order to reach the target position s f with the target speed v f .

[0076] When determining the deceleration strategy, the road slope with an inclination angle α schematically shown in Figure 1 should be taken into account, because the downhill of the road helps to decelerate vehicle 1. In the embodiment, for the sake of clarity, the inclination angle α is considered to be constant along the entire route from the starting position to the target position. However, sectionally different inclination angles can also be considered.

[0077] In another step, a reference trajectory of vehicle 1 is determined, which is compatible with reaching the target position s f with the target speed v f .

[0078] The possible process for determining the reference trajectory will be explained in more detail below. This embodiment is consistent with the corresponding second method step according to Figure 2 and Figure 3 .

[0079] As the basis for determining the reference trajectory, vehicle 1 together with the above boundary conditions is modeled as a switched system. This is schematically shown in Figure 4 .

[0080] Here, the longitudinal motion of the vehicle is described by a state vector x(t) related to the time variable t, which includes the travel s(t), the speed v(t), and the acceleration a(t) as vector components. The state vector satisfies the initial condition x(t0) = x 0 , where the initial state vector x 0 especially includes the initial parameters s 0 and v 0 introduced above.

[0081] In this embodiment, it is pre-given that vehicle 1 is in coasting operation in the first deceleration stage, that is, during the time interval from the initial time point t 0 to the first switching time point t s1 . Here, vehicle 1 is subject to the system dynamics in the first deceleration stage, which is expressed by the formula representation

[0082] At the first switching time point t s1 , the vehicle switches from coasting operation to towing operation. The towing operation continues during the time interval from the first switching time point t s1 to the second switching time point t s2 . During this second deceleration phase, the vehicle 1 is subject to the system dynamics according to the formula .

[0083] At the second switching time point t s2 , the vehicle switches from towing operation to active braking operation. The active braking operation continues during the time interval from the second switching time point t s2 to the end time point t f of the entire deceleration process. During this third deceleration phase, the vehicle 1 is subject to the system dynamics according to the formula where u is the brake adjustment parameter.

[0084] At these switching time points, the continuity condition x sail (t s1 ) = x drag (t s1 ) or x drag (t s2 ) = x brake (t s2 ) applies.

[0085] Starting from this, the reference trajectory and the switching time points can be determined by parameter optimization as the solution to the following hybrid optimization regulation problem:

[0086]

[0087] Therefore, the switching time points t s1 , t s2 and the end time point t f should be determined such that the cost function J is minimized. Here, the above initial conditions, continuity conditions and end conditions as well as the above restrictions on the order of different time points must be observed. In addition, a restriction on the braking adjustment parameter u brake is given, not exceeding the maximum braking deceleration u max .

[0088] The integral of the cost function J can be written as the sum of three sub-integrals for the three deceleration phases, with the corresponding start time points, end time points and transition time points as time interval limits. Here, the deceleration adjustment parameter generally given as u in the above integral expression is specifically replaced according to the rules given above for each sub-integral. Therefore, there is no deceleration adjustment parameter (u = 0) in the coasting phase, and the deceleration adjustment parameter a in the towing phasedrag (t) acts and another deceleration setting parameter u acts during a subsequent active braking phase. brake (t) acts.

[0089] The cost function J has a time-dependent cost term under the integral for each deceleration phase, compared to the initial kinetic energy E that the vehicle has at the start of deceleration. kin (v 0 ) This cost term imposes a cost on the loss of kinetic energy E kin (v) that has occurred up to the considered point in time. Specifically, this cost term includes the square of the difference between the kinetic energy E kin (v) at the considered point in time and the initial kinetic energy E kin (v 0 ) and a constant prefactor w kin .

[0090] The system dynamics during coasting operation can be written as follows:

[0091]

[0092] Here, a air = -(ρA f c d ) / (2m) represents the air resistance coefficient, c r represents the rolling resistance coefficient and g represents the acceleration due to gravity. In the last step, a sail = g sin(α) - c r g cos(α), where the road surface inclination angle α, which is considered constant over the entire deceleration route here.

[0093] The system dynamics equation during towing operation can be written as follows:

[0094]

[0095] Here, a drag represents the towing deceleration caused by the action of the motor brake.

[0096] The system dynamics equation during active braking can be written as follows:

[0097]

[0098] Here, the deceleration setting parameter u brake represents the braking deceleration caused by active braking.

[0099] For further calculations in this embodiment, for simplicity, it is considered that not only the deceleration adjustment parameter adrag during the towing operation and the deceleration adjustment parameter ubrake during the active braking are constant in time during the towing or braking deceleration phase. This enables the reference trajectory to be obtained simply and particularly computationally efficiently by parameter optimization with respect to the switching time point. See Xuping, Xu; Antsaklis, P. J.: Optimal Control of Switched Systems based on Parameterisation of the Switching Instants. IEEE Transactions on Automatic Control, 2004.

[0100] However, it should be noted that without these simplifying assumptions regarding the deceleration adjustment parameters, a solution to the hybrid optimization regulation problem is also feasible in principle. However, for an analytical solution, other approximations are also required, such as simplification of the cost function and linearization of the deceleration mode. The numerical solution gives rise to a boundary value problem that requires a large computational effort to solve. See Pakniyat, Ali; Caines, Peter E.: On the Hybrid Minimum Principle. IEEE

[0101] Assuming that the deceleration adjustment parameters are constant, the analytical solutions of the differential equations for the corresponding speed and the corresponding traveled distance with respect to the time variable t can be specified as follows:

[0102]

[0103] v sail (t) = b 1 tan(b 2 t + c 1 )

[0104]

[0105] Here, the constant function parameters b 1 , b 2 , b 3 , b 4 are derived from the previously introduced parameters a sail , a drag and a air :

[0106] The function parameters c 1 and c 2 are determined such that the initial condition x sail (0) = x 0 is satisfied. The function parameter c3 and c 4 According to the continuity condition x sail (t s1 ) = x drag (t s1 ).

[0107] Regarding the differential equation of x given above, it can be mathematically proven that there exists a true braking setting parameter u brake which can cause the system to transition to the desired end state with the target velocity v brake and the target position s f within a finite time t f and s f .

[0108] Specifically, the following analytical expressions for the end time t f and the braking setting parameter u brake can be obtained, which parametrically depend on the switching time points t s1 and t s2 :

[0109]

[0110] The following analytical functions can be given as the analytical solutions for the above differential equation of the braking phase:

[0111]

[0112] Here, the constant function parameters b 5 and b 6 depend on the parameters a sail , a air and u brake as follows:

[0113]

[0114] The additional function parameters c 5 and c 6 are determined according to the continuity condition x drag (t s2 ) = x brake (t s2 ).

[0115] Based on these preliminary considerations, the cost function J can now be parameterized with respect to the switching time points t s1 and t s2 :

[0116]

[0117] Starting from this, the problem of parametric optimization of the switching variables

[0118]

[0119] It can be expressed as follows:

[0120]

[0121] The parameter optimization problem can be numerically solved by computationally efficient standard methods, i.e., t can be numerically determined s1 and t s2 , to minimize the cost function J. If the switching time points t s1 and t s2 are known, then the reference trajectory sought can be known based on the analytical expressions for distance and speed in the three deceleration phases given previously.

[0122] Figure 5 Shows the solution for parameters s 0 = 0, v 0 = 180 km / h, s f = 700 m, v f = 80 km / h, α = 2°, a drag = 0.4 m / s 2 , w u = 1 and w kin = 1e -6 .

[0123] Here, the above three figures ( Figure 5 (a)) show the curves of distance s (in meters), speed v (in km / h), or deceleration setting parameter u (in m / s 2 ) versus time t (in seconds) respectively. Two vertical dashed lines respectively mark the optimized switching time points t s1 (left dashed line) and t s2 (right dashed line).

[0124] In the following figure ( Figure 5 (b)), the speed v in km / h is plotted against the distance s traveled in meters, where two vertical dashed lines mark the positions of vehicle 1 at the switching time points t s1 and t s2 .

[0125] In this embodiment, the reference trajectory planning determines to coast for about the first 410 meters, then tow for about 150 meters, and then actively brake the last about 140 m safely and comfortably with a deceleration of about 1.5 m / s^2 in order to precisely comply with the new speed limit of 80 km / h at the target.

[0126] If the braking deceleration u brakeIf the parameters are initially fixedly pre-given, the necessary computation time can be strongly further reduced by reducing the optimization problem to a parameterized one-dimensional optimization. In the case of the braking deceleration u brake under this assumption, the free variables can be eliminated and, for example, the second switching time point t s2 can be expressed as a function of the first switching time point t s1 , i.e., t s2 = t s2 (t s1 ). The optimization problem can be formulated as follows:

[0127]

[0128] Figure 6 Graphically illustrate how, in the case where the braking deceleration u brake is fixedly pre-given (here: u brake = 1.5 m / s 2 ), the upper limit t s1,max of the first switching time point and the sought-after optimized first and second switching time points t s1,opt and t s2,opt (t s1 ) are determined based on the known curves of the distance-dependent speed v(s) in the three deceleration phases. Here, t s1,max is the intersection point of the coasting speed curve (dashed curve in 0 starting at the initial position s 0 with the initial speed v Figure 6 ) and the active braking speed curve (solid curve in f ) ending at the target position s f with the target speed v Figure 6 . Thus, the optimized first switching time point t s1,opt can be determined with a small number of numerical calculations or even by an analytical solution of the above one-dimensional optimization problem. The optimized second switching time point t s2,opt is finally obtained as the intersection point of those speed curves (lower dashed line in s1,opt ) of the towing operation that intersect the coasting speed curve at the optimized first switching time point t Figure 6 with the active braking speed curve.

[0129] Figure 7 Exemplarily and schematically shows a system 10 for regulating the longitudinal guidance of a vehicle, for example vehicle 1. The system 10 consists of a reference trajectory planner 101, an application trajectory planner 102, and a regulation path 103. The reference trajectory planner 101 is configured to, in the manner described below, according to the initial starting state x 0 , the target state x fand other parameters (e.g., related to the driving resistance) to obtain a reference trajectory for decelerating the vehicle 1 to the target speed v at the target position s f as the solution to the hybrid optimization regulation problem. The reference trajectory gives the reference path s f , the reference speed v ref and the reference acceleration a ref each with their respective time curves. The time planning range of the reference trajectory planner 101 extends from the starting time point t0 to the ending time point t ref , at which ending time point the vehicle 1 will reach the target position s according to the reference trajectory f . f

[0130] The nominal trajectory planner 102 is provided for planning the nominal trajectory according to the reference trajectory, and the nominal trajectory gives the nominal distance s soll , the nominal speed v soll , the nominal acceleration a soll and the nominal jerk j soll each with their respective curves of change. The determination of the nominal trajectory can be carried out, for example, in the context of model predictive control by means of a quadratic optimization with respect to additional criteria, such as those related to safety and comfort. Here, the time planning range of the nominal trajectory planner 102 can extend several seconds into the future, for example, 6 seconds. Over the planning range, N interpolation points can be allocated for the nominal trajectory planning at short time intervals, for example, 0.2 seconds.

[0131] When calculating the nominal trajectory, the reference trajectory can be considered as a reference. For example, the parameters s, v, a, and j can be determined during the optimization to minimize the following cost function:

[0132]

[0133] l(k) = w s (s - s rer ) 2 + w v (v - v ref ) 2 + w a (a - a ref ) 2 + w j j 2 + w u u 2

[0134] Here, k is the index of each interpolation point along the planning range of the nominal trajectory planner 102.

[0135] It is also possible for the nominal trajectory planning to be associated with the reference trajectory planning such that the time-dependent speed curve v according to the reference trajectory ref ​In the context of the optimization for determining the desired speed v soll represents the upper limit v max . According to an implementation variant, in this case the upper limit does not have to be rigid, but can be softened, for example, by means of a slack variable ∈ v，max . In this implementation variant, the cost function to be minimized and the softened speed limit can be specified, for example, in the following way:

[0136]

[0137] As another possible alternative, the reference acceleration can be used as the lower limit a min of the desired acceleration to be determined, where ∈ a，min can be softened according to an implementation variant by means of a slack variable. In this implementation variant, the cost function to be minimized and the softened acceleration limit can be specified, for example, in the following way:

[0138]

[0139] The adjustment path 103 includes the longitudinal guidance actuators of the vehicle 1, such as drive and braking devices. For example, in the context of the adjustment path 103, the desired acceleration a soll can be used as a setpoint for the longitudinal guidance actuators. However, it is also possible that the adjustment path includes one or more regulators subordinate to the trajectory planning, such as a trajectory tracking regulator, which generate specific setpoints for the longitudinal guidance actuators based on the desired trajectory and information about the influence of the currently measured disturbance parameters, for example, in the form of a setpoint for the driving torque converted by the longitudinal guidance actuators or a setpoint for the acceleration converted by the longitudinal braking actuators.

[0140] The system 10 periodically performs longitudinal guidance adjustment by feeding back the current state (for example, by means of odometer measurement), such as the current actual speed, the current actual acceleration, or the actual travel distance, to the reference trajectory planner 101. This can calculate an updated reference trajectory with the current state x 0 as the starting state respectively.

Claims

1. A method for determining a deceleration strategy of a vehicle (1), the method comprising: - determining a target position in a forward route section together with a target speed that the vehicle (1) should have when reaching the target position, wherein the target speed is lower than the current speed of the vehicle (1); and - obtaining a reference trajectory of the vehicle (1), the reference trajectory being set to reach the target position at the target speed; wherein the reference trajectory is obtained based on an analytical function of at least two deceleration modes from a list, the list including: ﹒Coasting operation, ﹒Dragging operation, ﹒Recovery operation, and ﹒Active braking, wherein the analytical function of each deceleration mode of the at least two deceleration modes gives a function of speed and / or traveled distance with respect to the time variable, and wherein obtaining the reference trajectory includes obtaining one or more switching time points between the at least two deceleration modes based on the analytical function.

2. The method according to claim 1, wherein the analytical function can be given as a closed mathematical expression having the time variable and a plurality of function parameters characterizing one or more types of driving resistance.

3. The method according to claim 2, wherein the function parameters further include one or more deceleration adjustment parameters for one or more deceleration modes.

4. The method according to any one of the preceding claims, wherein for the at least two deceleration modes, one or more of the analytical functions giving the traveled distance with respect to the time variable can be given in the following form: s(t) = A ln(tan(Bt + C) 2 + 1)+ D where the traveled distance is s, the time variable is t, and the constant function parameters are A, B, C, D.

5. The method according to any one of the preceding claims, wherein for the at least two deceleration modes, one or more analytical functions giving the speed with respect to the time variable are given in the following form: v(t) = E tan(Ft + G), where the speed is v, the time variable is t, and the constant function parameters are E, F, G.

6. The method according to any one of the preceding claims, wherein obtaining the one or more switching time points is based on minimization of a cost function; wherein the cost function includes a sum of a plurality of time integrals, where each integral extends in time over a corresponding deceleration phase, the deceleration phase being assigned to the corresponding deceleration mode of the at least two deceleration modes; where the one or more switching time points form the upper or lower limit of integration of the integral, and where the cost function has a time-dependent cost term under the integral of the at least two deceleration phases, the cost term imposing a cost on the kinetic energy loss that has occurred up to the time point under consideration compared to the initial kinetic energy that the vehicle (1) had at the start of the deceleration.

7. A method for determining a deceleration strategy of a vehicle (1), the method comprising: - determining a target position in a forward route section together with a target speed that the vehicle (1) should have when reaching the target position, wherein the target speed is lower than the current speed of the vehicle (1); and - Obtain a reference trajectory for the vehicle (1), the reference trajectory being set to reach the target position at the target speed; wherein the reference trajectory has at least two deceleration phases, in which the vehicle (1) operates in different deceleration modes from a list, the list including: ﹒Coasting operation, ﹒Drag operation, ﹒Recovery operation, and ﹒Active braking, where obtaining the reference trajectory includes obtaining one or more switching time points between the at least two deceleration modes, and obtaining the one or more switching time points is achieved based on the minimization of a cost function; wherein the cost function includes a sum of a plurality of time integrals, where each integral extends in time over a corresponding deceleration phase; where the one or more switching time points form the upper or lower limit of integration of the integral, and where the cost function has time-dependent cost terms under the integrals for the at least two deceleration phases, and the cost terms impose a cost on the kinetic energy loss that has occurred up to the time point under consideration compared to the initial kinetic energy that the vehicle (1) had at the start of the deceleration.

8. The method according to claim 6 or 7, wherein the time-dependent cost term includes the square of the difference between the kinetic energy at the time point under consideration and the initial kinetic energy.

9. The method according to any one of the preceding claims, wherein for the purpose of obtaining the reference trajectory, one or more deceleration setting parameters that are effective in the deceleration mode are assumed to be constant in time during the corresponding deceleration phase in which the corresponding deceleration mode is activated.

10. The method according to any one of the preceding claims, wherein the reference trajectory includes: a coasting phase in which the vehicle (1) travels in coasting operation; and at least one additional deceleration phase that is subsequent in time to the coasting phase, in which the vehicle (1) travels in drag operation or recovery operation or performs active braking.

11. The method according to any one of the preceding claims, wherein the method further includes: - Obtain the desired trajectory of the vehicle (1) according to the reference trajectory; - Generate a setting pre-given for the longitudinal guidance of the vehicle (1) according to the desired trajectory.

12. A data processing device, wherein the data processing device is arranged to execute the method according to any one of the preceding claims.

13. A software including instructions, when the software is executed by a data processing device, the instructions cause the data processing device to execute the method according to any one of claims 1 to 11.

Images