Model-based predictive control of a motor vehicle
The model-based predictive control system with multiple solver modules optimizes energy efficiency and gear selection in hybrid vehicles, addressing computational challenges and enabling real-time responsiveness.
Patent Information
- Application Number
- DE102020203742
- Authority / Receiving Office
- DE · DE
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2020-03-24
- Publication Date
- 2024-11-14
- Estimated Expiration
- 2040-03-24
AI Technical Summary
Existing systems struggle to optimize the energy efficiency of hybrid vehicles due to limited route insight, high computational demands of optimization-based strategies, and the complexity of modeling gear and discrete states, making real-time predictive control challenging.
A model-based predictive control system using a processor unit with multiple solver modules (HLS, MIS, and tracker) to minimize cost functions, enabling efficient energy management and gear selection, while allowing for parallel computation and real-time synchronization.
Enables energy-efficient vehicle guidance with short reaction times by optimizing speed, gear selection, and torque distribution, reducing computational load and ensuring real-time responsiveness to dynamic road conditions.
Smart Images

Figure 00000000_0000_ABST
Abstract
Description
[0001] The invention relates to model-based predictive control of a motor vehicle. In this context, it particularly relates to a processor unit, a hybrid powertrain, a motor vehicle, a method, and a computer program product.
[0002] Energy-optimized operation of a hybrid vehicle is only possible with good knowledge of the route to be traveled. A driver must therefore drive with foresight, but has only limited insight into the further course of the route. A cruise control system can take the route topology into account, but maps the driving strategy based on rules. Optimization-based strategies require a lot of computing power and time, making their use for online calculations problematic. A further challenge is modeling the selection of a suitable gear and other discrete states (e.g., clutch on / off). The optimization then represents a so-called mixed integer problem. The prior art, which forms the background to the present invention, is known from CN 1 08 944 905 A, DE 10 2014 222 069 A1 and DE 10 2013 008 716 A1. From 10 2014 209 687 A1 a method and a device for the predictive operation of a motor vehicle is known.
[0003] One object of the present invention is to provide a control system for a motor vehicle that addresses the problems described above. This object is achieved by the subject matter of the independent patent claims. Advantageous embodiments are the subject matter of the subclaims, the following description, and the figures.
[0004] The present invention provides energy-efficient longitudinal vehicle guidance, particularly for a hybrid vehicle. According to the present invention, it is proposed to solve an optimization problem based on a vehicle model for model-based predictive control of the motor vehicle while minimizing cost functions of multiple solver modules in order to enable a shorter response time. All solver modules are based on a model of the vehicle and the powertrain, including the descriptive data, as well as information about the upcoming route section. This information can be topographic information (e.g., curve and gradient information), traffic information (e.g., vehicles traveling ahead), or infrastructure information (e.g., traffic lights along the route).
[0005] A first solver module can be used to perform optimization on a central control unit of the motor vehicle, calculating an optimal speed profile and SoC profile for an entire prediction horizon. A second solver module can be used to solve a mixed integer problem, in particular an optimal gear selection and / or an optimal driving mode selection for a portion of the prediction horizon. A third solver module can then calculate a torque distribution among, for example, an internal combustion engine, an electric motor, and a braking system of the motor vehicle. This can be done for a short look ahead to enable a short reaction time.
[0006] In this sense, according to a first aspect of the invention, a processor unit for model-based predictive control of a motor vehicle is provided. The processor unit is configured to execute an MPC algorithm comprising a first solver module, a second solver module, a third solver module, a longitudinal dynamics model, and three cost functions. A first cost function is assigned to the first solver module, a second cost function is assigned to the second solver module, and a third cost function is assigned to the third solver module. The first solver module, the second solver module, and the third solver module are implemented, in particular, by software.
[0007] The processor unit is configured to calculate a speed trajectory that minimizes the first cost function for an upcoming route section, taking into account the longitudinal dynamics model. According to said speed trajectory, the motor vehicle is to move within a prediction horizon, by executing the first solver module. The processor unit is further configured to calculate a charge state curve of a battery or accumulator that minimizes the first cost function, by executing the first solver module for the upcoming route section, taking into account the longitudinal dynamics model. The battery or accumulator serves as an energy storage device for an electric machine of the motor vehicle.This energy storage device can supply the electric motor with electrical energy to drive the vehicle within the prediction horizon (either exclusively or in support of an internal combustion engine). The state of charge (SoC) is the current energy content of an electric battery relative to its maximum energy content.
[0008] The first solver module can be referred to as a "high-level solver" (HLS for short). The HLS solves a nonlinear problem and works with continuous surrogate variables for discrete states (e.g., gears). This approach limits the solution space less than when considering discrete states. This results in advantages in terms of the optimum result and in terms of computing time. In other words, the first solver module enables online calculation of the, in particular, energy-optimal speed trajectory and the SoC trajectory for a forward prediction horizon on a central control unit in the motor vehicle. The processor unit can be configured, in particular, to call the first solver module twice.Initiated by a first call, a speed trajectory can be calculated for the upcoming route section, in particular by executing the first solver module, taking into account the longitudinal dynamics model. According to this speed trajectory, the motor vehicle should move within a prediction horizon, in particular by minimizing the travel time term of the first cost function. Subsequently, initiated by a second call, a speed trajectory can be calculated for the upcoming route section, in particular by executing the first solver module, taking into account the longitudinal dynamics model. According to this speed trajectory, the motor vehicle should move within a prediction horizon.
[0009] Furthermore, traffic lights located on the upcoming section of the road can also be taken into account, including their phases and remaining times. The traffic light information is converted into time limits at the waypoint of the upcoming traffic light located on the horizon. In addition, preceding vehicles can be taken into account in the first solver module. A probable trajectory can be determined for the vehicle in front of the current vehicle. This results in a minimum time for the HLS trajectory over the path at the current position of the preceding vehicle.
[0010] Furthermore, the processor unit is configured to calculate a trajectory of integer control variables that minimizes the second cost function by executing the second solver module based on the speed trajectory, the battery state of charge profile, and taking into account constraints. The trajectory of integer control variables can include a profile of gears that can or should be engaged in a transmission of the motor vehicle within the prediction horizon. Furthermore, the trajectory of integer control variables can include a profile of an engine state, according to which an internal combustion engine of the motor vehicle is either switched on or switched off (in particular within the prediction horizon).Furthermore, the trajectory of integer control variables can comprise a course of a clutch state, according to which a clutch arranged between an internal combustion engine of the motor vehicle and the electric machine is either in an open position or in a closed position (in particular within the prediction horizon).
[0011] The second solver module can be referred to as a "Mixed Integer Solver" (MIS for short). The second solver module is used to determine discrete states, e.g., the gears or the state of the clutch (on / off) of the vehicle's drivetrain. The second solver module thus enables the determination of an optimal operating strategy, which in particular includes the selection of the optimal gear and driving mode. The calculations of the second solver module are based in particular on the underlying HLS trajectory and additional external boundary or secondary conditions. Secondary conditions that can be taken into account when calculating the optimal gear trajectory include, for example, the topography of the upcoming route section (curves, gradients). Furthermore, knowledge of speed limits on the upcoming route section can be considered as secondary conditions in the calculation of the gear trajectory.The trajectory of integer control variables is optimized over time and can be significantly shorter than that of the second solver module.
[0012] Furthermore, the processor unit is configured to calculate, by executing the third solver module—in particular based on the velocity trajectory and the SoC trajectory and / or based on the trajectory of integer control variables—a torque trajectory that minimizes the third cost function for an initial portion of the prediction horizon for the electric machine, an internal combustion engine, and a braking system of the motor vehicle, according to which the electric machine, the internal combustion engine, and the braking system are to provide torques within the prediction horizon. Forces can also be used instead of torques—both variables can be linearly converted into one another. The third solver module can be referred to as a "tracker." The tracker is characterized by fast computing time and robust behavior.If current solutions from the other solvers (HLS and / or MIS) are unavailable, the tracker can still provide torques based on the latest solutions from the other solvers. The tracker's output directly influences ride comfort by minimizing the tracker-specific third cost function. The torque trajectory refers to torques at at least one wheel of the vehicle and includes both positive and negative torques provided by the electric motor, the internal combustion engine, and the vehicle's braking system.
[0013] Thus, the first solver module (HLS) determines an optimal SoC curve and an optimal speed trajectory. The second solver module (MIS) can decide which gears and which driving mode are optimal based on the calculations of the first solver module (HLS). The third solver module (Tracker) calculates very frequently and determines the torques for a short period of time. The Tracker can work with or without the trajectory calculated in the MIS. In the second case, the Tracker only uses the currently engaged gear and the active driving mode, which is sufficient feedback for the Tracker. This is possible because gear requirements typically do not change constantly within a second. Engaging a different gear takes some time. Therefore, the Tracker can also function when only the HLS is calculating. Should this also fail, the old HLS solution can still be fed to the Tracker.Then, so to speak, the "old" solution can be successively implemented. Of course, this only makes sense up to a certain point. If the HLS calculates but doesn't find a solution, this can be communicated to the tracker, e.g., via an "error flag," i.e., an error indicator. Error handling can then be implemented in the tracker itself. If errors occur that the tracker cannot anticipate, this also leads to an error in the tracker. As a result, the automated driving function can then be deactivated.
[0014] According to the present invention, computing work can be distributed among the different solver modules. This allows each solver module to run independently on the processor unit. The solver modules therefore do not run strictly one after the other. The processor unit can comprise multiple processor cores, similar to a desktop PC. The three solver modules can be distributed among the processor cores and thus calculate independently of one another and virtually in parallel. This represents a major advantage, as otherwise it might be necessary to wait several seconds until the solver modules have calculated a solution with which the vehicle can continue driving. This may not be a problem when driving along a stretch of road. However, if a vehicle drives in front of the vehicle or cuts in, the system cannot detect this until the solvers have recalculated using the new sensor data. This could lead to a critical situation.
[0015] During distribution, different runtimes can also be granted for the individual solver modules. For example, the first solver module, which is used to determine the velocity trajectory and the trajectory of integer control variables, can require considerably more time than the other solver modules. Since the solver modules calculate in parallel, it is possible that one of the solver modules is called before the previous solver module has finished calculating, for example the tracker before the currently running HLS. All data used can always have a timestamp. Each solver solution can also have a timestamp. These timestamps make it possible to always trace which preliminary result was used in which calculation. Appropriate error handling can be implemented via this connection, for example timeouts if information is too old. A certain degree of synchronization therefore takes place via the timestamps.Thus, the present invention provides real-time synchronization of inherently asynchronous solver calls.
[0016] Furthermore, the moments calculated using the third solver module can exhibit a linear progression. This enables linear interpolation and thus simplifies post-processing. Previously, the moments were determined for specific time steps. This resulted in a stepped progression of the signals. According to the present invention, the third solver module produces a linear progression, which means that the signals are linear even between time steps and do not jump from one time step to the next. This also allows for correct interpolation between time steps.
[0017] According to the present invention, in particular, a cyclic recalculation of the target speed (speed trajectory), the target torques (torque trajectory) and the discrete states (gear trajectory) is carried out on the basis of a respective current driving state and of respective route information ahead.
[0018] Furthermore, a targeted merging of SPP (Signal Post Processing) and Tracker (same call time > synchronous) can be achieved so that the torques output by the Tracker are aligned with the gears calculated by the MIS (see the same input step). Signal Post Processing (SPP) is a signal post-processing process. Here, the partial results calculated at different times are linked in such a way that they are related to each other in time. This results in a time period in which all information is available. The time base is, in particular, that of the Tracker. Furthermore, SPP and Tracker are implemented in the same software module, so SPP and Tracker have the same time base.
[0019] Using the first solver module, the vehicle's speed trajectory can be optimized online even with a relatively long or distant look-ahead over the route, e.g., a look-ahead in the range of several kilometers. This is computationally intensive. On the other hand, it is sufficient to use the other server modules to calculate the corresponding trajectories for a relatively short look-ahead over time, e.g., for the area immediately ahead of the vehicle. This enables particularly short reaction times. In this sense, in one embodiment, the processor unit is configured to calculate the speed trajectory for a section of the route that is in the range of kilometers and, in particular, is several kilometers long by executing the first solver module, and to calculate the target torque trajectory for an initial section of the prediction horizon that is several seconds immediately in front of the motor vehicle by executing the other two solver modules.
[0020] The MPC algorithm provides an operating strategy that, in addition to the optimal selection of the powertrain torques and gears, also enables the optimal selection of a driving mode. Thus, an optimal driving strategy may stipulate that the powertrain operates in a specific driving mode within the prediction horizon. However, it may also be better to change the driving mode within the prediction horizon, e.g., because this increases efficiency.
[0021] The motor vehicle that can be controlled predictively using a model-based approach according to the present invention is a hybrid vehicle. A "hybrid vehicle" can be understood as an electric vehicle that can be powered by at least one electric machine and at least one internal combustion engine. The hybrid vehicle can draw energy both from a battery and from an additional fuel carried on board, e.g., diesel, gasoline, or gas. The electric machine can be operated as a motor and as a generator. A hybrid drive train of the hybrid vehicle can comprise, in addition to the electric machine, the internal combustion engine, and the battery, a transmission and a clutch.Depending on the clutch position (open and closed) and the transmission position (gear engaged, neutral), the hybrid vehicle's powertrain operates in different driving modes. For example, the powertrain can be designed using a P2 architecture, which can result in the following drive modes: 1. Driving mode: internal combustion engine in operation ("on") and clutch K0 (located between the internal combustion engine and the electric motor) in the open position; the electric motor drives the motor vehicle in engine mode; the internal combustion engine is running but does not drive the motor vehicle; 2. Driving mode: combustion engine not in operation ("off") and clutch K0 in the open position; the electric motor drives the vehicle in engine mode; the combustion engine is not running; 3. Driving mode: Combustion engine in operation ("on") and clutch K0 in the closed position; forward gear or reverse gear engaged in the transmission; the electric motor (in motor mode or generator mode) and the combustion engine (in traction mode or overrun mode) drive the vehicle together (hybrid drive); in this case, a power flow between the electric motor and the combustion engine is also possible (load point shift). 4. Driving mode: Combustion engine in operation ("on") and clutch K0 in the closed position; neutral engaged in the transmission; the combustion engine drives the electric motor (generator mode), so that the battery is charged.
[0022] If the motor vehicle - as described above - is a hybrid vehicle whose drive train can be operated in at least a first driving mode and a second driving mode, then in one embodiment the processor unit can be configured to calculate a driving mode trajectory that minimizes the cost function by executing the second solver module based on the speed trajectory, based on the battery state of charge profile and taking into account constraints, which determines in which of the driving modes the hybrid drive train is to be operated within the prediction horizon. Particular attention can be paid to the fact that switching between the driving modes cannot occur arbitrarily quickly (e.g. engine start, engine stop, clutch on / off) and that this can also have an impact on continuous variables (e.g. interruptions in tractive force).A corresponding engine start and stop logic can also be implemented outside the optimization in a downstream software module, as this simplifies the discrete problem formulation and shortens the computation time.
[0023] In a further embodiment, energy consumption can be optimized while driving by knowing the losses (efficiency maps of the drive train components and driving resistances). For this purpose, the processor unit can access loss maps of components of the motor vehicle, in particular loss maps of the internal combustion engine, the electric machine and the transmission of the drive train of the motor vehicle. The loss maps can be stored, for example, in a memory unit within the motor vehicle. The loss maps can be "fitted" in a continuously differentiable manner, i.e. formulas can be derived from the maps of the respective component which describe the loss map of the respective component. The losses of a pre-transmission and a differential of the drive train of the motor vehicle can be stored exclusively as a characteristic curve.So-called "efficiency maps" can still be known at the start of the calculations. The characteristic maps can also be adapted to the problem formulation of the first solver module, for example, by providing zero losses below an idle characteristic curve. For example, the loss map of the internal combustion engine in the vehicle's powertrain can be modified such that when the internal combustion engine is not in use, its consumption is zero. This enables a transition from a discrete state (clutch on / off) to a continuous state.
[0024] According to a second aspect of the invention, a hybrid drive train for a motor vehicle is provided. The hybrid drive train comprises at least a first drive unit in the form of an internal combustion engine, at least one second drive unit in the form of an electric machine that can be operated as a motor and as a generator, at least one battery, at least one transmission, a braking system, and a processor unit according to the first aspect of the invention.
[0025] According to a third aspect of the invention, a motor vehicle is provided. The motor vehicle comprises a hybrid drive train according to the second aspect of the invention and a driver assistance system which is configured to - to access a speed trajectory according to which the motor vehicle is to move within a prediction horizon, wherein the speed trajectory has been calculated by the processor unit by executing the first solver module for a route section ahead, taking into account the longitudinal dynamics model and minimizing the first cost function (over the path), - to access a profile of the state of charge of the battery which serves as an energy storage device for the electric machine of the motor vehicle, wherein the profile of the state of charge of the battery has been calculated by the processor unit by executing the first solver module for the upcoming route section, taking into account the longitudinal dynamics model and minimizing the first cost function, - to access a trajectory of integer control variables, wherein the trajectory of integer control variables has been calculated by the processor unit by executing the second solver module based on the speed trajectory, based on the battery charge state and taking into account constraints, as well as minimizing the second cost function (over time), - to access a torque trajectory according to which the at least one internal combustion engine, the at least one electric machine and the braking system are to provide torques or forces within the prediction horizon, wherein the torque trajectory has been calculated by the processor unit by executing the third solver module - in particular based on the speed trajectory, based on the battery charge state and based on the gear trajectory - for an initial section of the prediction horizon (over time) while minimizing the third cost function, and - to control the at least one internal combustion engine, the at least one electric machine and the braking system of the hybrid powertrain based on the torque trajectory.
[0026] According to a fourth aspect of the invention, a method for model-based predictive control of a motor vehicle using an MPC algorithm is provided, wherein the MPC algorithm comprises a first solver module, a second solver module, a third solver module, a longitudinal dynamics model, and three cost functions, wherein a first cost function is assigned to the first solver module, a second cost function is assigned to the second solver module, and a third cost function is assigned to the third solver module. The method comprises the steps - Calculating a speed trajectory according to which the motor vehicle is to move within a prediction horizon, wherein the speed trajectory is calculated by executing the first solver module for a preceding route section, taking into account the longitudinal dynamics model and minimizing the first cost function, - Calculating a profile of a state of charge of a battery which serves as an energy storage device for the electric machine of the motor vehicle, wherein the profile of the state of charge of the battery is calculated by executing the first solver module for the upcoming route section, taking into account the longitudinal dynamics model and minimizing the first cost function, - Calculating a trajectory of integer control variables, wherein the trajectory of integer control variables is calculated by executing the second solver module based on the speed trajectory, based on the battery state of charge, taking into account constraints, and minimizing the second cost function, - Calculating a torque trajectory according to which an internal combustion engine, the electric machine and a braking system of the motor vehicle are to provide torques within the prediction horizon, wherein the torque trajectory is calculated by executing the third solver module - in particular based on the speed trajectory, based on the course of the state of charge of the battery and based on the trajectory of integer control variables - for an initial section of the prediction horizon while minimizing the third cost function.
[0027] According to a fifth aspect of the invention, a computer program product for model-based predictive control of a motor vehicle using an MPC algorithm is provided, which comprises a first solver module, a second solver module, a third solver module, a longitudinal dynamics model, and three cost functions, wherein a first cost function is assigned to the first solver module, a second cost function is assigned to the second solver module, and a third cost function is assigned to the third solver module. The computer program product, when executed on a processor unit, instructs the processor unit to - by executing the first solver module for a route section ahead, taking into account the longitudinal dynamics model, to calculate a speed trajectory that minimizes the first cost function, according to which the motor vehicle is to move within a prediction horizon, - by executing the first solver module for the upcoming route section, taking into account the longitudinal dynamics model, to calculate a charge state curve of a battery that minimises the first cost function, which serves as an energy storage device for an electric machine of the motor vehicle, - by executing the second solver module based on the speed trajectory, based on the battery charge state and taking into account constraints, to calculate a trajectory of integer control variables minimizing the second cost function, and - by executing the third solver module - in particular based on the speed trajectory, based on the course of the state of charge of the battery and based on the trajectory of integer control variables - while minimizing the third cost function for an initial section of the prediction horizon for the electric machine, an internal combustion engine and a braking system of the motor vehicle, to calculate a torque trajectory according to which the electric machine, the internal combustion engine and the braking system are to provide torques within the prediction horizon.
[0028] The above statements in connection with embodiments of the processor unit according to the first aspect of the invention also apply mutatis mutandis to the hybrid drive train according to the second aspect of the invention, to the motor vehicle according to the third aspect of the invention, to the method according to the fourth aspect of the invention and to the computer program product according to the fifth aspect of the invention.
[0029] In the following, embodiments of the invention are explained in more detail with reference to the schematic drawing, wherein identical or similar elements are provided with the same reference numerals. Fig. 1 is a schematic representation of a motor vehicle whose drive train comprises an internal combustion engine, an electric machine and a braking system, Fig. 2 exemplary degrees of freedom of the electrical machine from Fig. 1, Fig. 3 degrees of freedom of the braking system Fig. 1 and Fig. 4 Details of an exemplary drive train for the motor vehicle according to Fig. 1.
[0030] Fig. 1 shows a motor vehicle 1, e.g., a passenger car. The motor vehicle 1 comprises a system 2 for model-based predictive control of the motor vehicle 1. In the illustrated embodiment, the system 2 comprises a processor unit 3, a memory unit 4, a communications interface 5, and a recording unit 6, in particular for recording status data relating to the motor vehicle 1.
[0031] The motor vehicle 1 further comprises a drive train 7, which may include, for example, an electric machine 8, which can be operated as a motor and as a generator, a battery 9, a transmission 10, and a braking system 19. The electric machine 8 can drive wheels of the motor vehicle 1 via the transmission 10 during motor operation. The electrical energy required for this can be provided by the battery 9, in particular via power electronics 18 (see Fig. 4). Conversely, the battery 9 can be charged by the electric motor 8 via the power electronics 18 when the electric motor 8 is operating in generator mode (regeneration). The battery 9 can optionally also be charged at an external charging station.
[0032] Fig. 1 and Fig. 4 further show that the drive train 7 is a hybrid drive train, which additionally has an internal combustion engine 17. The internal combustion engine 17 can be Fig. 4, the parallel P2 architecture of the hybrid drive train 7 can drive the motor vehicle 1 in addition to the electric machine 8 when a clutch K0 arranged between the internal combustion engine 17 and the electric machine 8 is closed. The internal combustion engine 17 can optionally also drive the electric machine 8 to charge the battery 9. In the exemplary embodiment shown, the electric machine 8 can (with the clutch K0 closed, supported by the internal combustion engine 17) drive two front wheels 22 and 23 of the motor vehicle 1, which are attached to a front axle 25, via the transmission 10 and a front differential gear 21. A first rear wheel 26 and a second rear wheel 28 on a rear axle 29 of the motor vehicle 1 are not driven in the exemplary embodiment shown (rear-wheel drive and all-wheel drive are also possible alternatively).The front wheels 22, 23 and the rear wheels 26, 28 can be braked by the braking system 19 of the drive train 7, for which purpose the braking system 19 can provide a negative torque (braking torque).
[0033] The hybrid powertrain 7 according to Fig. 4 can be operated in the following drive modes: 1. Driving mode: Internal combustion engine 17 in operation ("on") and clutch K0 in the open position; the electric machine 8 drives the motor vehicle 1 in motor mode (generator mode also possible); the internal combustion engine 17 is running, but does not drive the motor vehicle 1; 2. Driving mode: Internal combustion engine 17 not in operation ("off") and clutch K0 in the open position; the electric machine 8 drives the motor vehicle 1 in engine mode (generator mode also possible); the internal combustion engine 17 is not running; 3. Driving mode: Internal combustion engine 17 in operation ("on") and clutch K0 in the closed position; forward gear or reverse gear engaged in transmission 10; the electric motor 8 (in motor mode; load point shift possible) and the internal combustion engine 17 jointly drive the motor vehicle 1 (hybrid drive); 4. Driving mode: Combustion engine 17 in operation (“on”) and clutch K0 in the closed position; neutral engaged in transmission 10; the combustion engine 17 drives the electric machine (generator mode), so that the battery is charged.
[0034] The motor vehicle 1, in particular its drive train 7, comprises several components that are relevant to the efficiency of the operation of the motor vehicle 1 ("efficiency-relevant components"). Purely by way of example, the electric motor can be considered a first component 8 and the braking system 19 a second component 17. The battery 9, the transmission 10, and the internal combustion engine 17 can also be considered efficiency-relevant components of the motor vehicle 1.
[0035] Fig. 2 shows that the electric machine 8 can be operated with different values of a first operating parameter 20 (first degree of freedom). For example, the first operating parameter 20 can be the rotational speed n of the electric machine 8. Purely by way of example, the rotational speed can assume a first value n1, a second value n2, and a third value n3. The different rotational speeds n1, n2, and n3 of the electric machine 8 can lead to different levels of traction force exerted by the electric machine 8 on the wheels of the motor vehicle 1. This can be represented by the longitudinal model 14 of the drive train 7 of the motor vehicle 1.
[0036] Fig.3 shows that the braking system 19 can be operated with different values of a second operating parameter 24 (second degree of freedom). For example, the second operating parameter 20 can be the braking force of the braking system 19. Purely by way of example, the braking force can assume a first value y1, a second value y2, and a third value y3. The different braking forces y1, y2, and y3 of the braking system 19 can lead to different levels of traction force exerted by the braking system on the wheels of the motor vehicle 1. This can be represented by the longitudinal model 14 of the drive train 7 of the motor vehicle 1.
[0037] A computer program product 11 can be stored on the storage unit 4. The computer program product 11 can be executed on the processor unit 3, for which purpose the processor unit 3 and the storage unit 4 are connected to each other via the communication interface 5. When the computer program product 11 is executed on the processor unit 3, it instructs the processor unit 3 to perform the functions or method steps described in connection with the drawing.
[0038] The computer program product 11 contains an MPC algorithm 13, which includes or contains a first solver module 13.1, a second solver module 13.2, and a third solver module 13.3. The MPC algorithm 13 further contains a longitudinal dynamics model 14 of the motor vehicle 1. Each of the solver modules 13.1 to 13.3 can access the longitudinal dynamics model 14. Furthermore, the MPC algorithm 13 contains three cost functions 15.1 to 15.3 to be minimized, wherein a first cost function 15.1 is assigned to the first solver module 13.1, a second cost function 15.2 is assigned to the second solver module 13.2, and a third cost function 15.3 is assigned to the third solver module 13.3.
[0039] The longitudinal dynamics model 14 includes a loss model 27 of the motor vehicle 1. The loss model 27 describes the operating behavior of the efficiency-relevant components, e.g., the electric motor 8 and the braking system 19, with respect to their efficiency and their loss, respectively. This results in the total loss of the motor vehicle 1. The processor unit 3 executes the MPC algorithm 13 and predicts a behavior of the motor vehicle 1 for a sliding prediction horizon. This prediction is based on the longitudinal dynamics model 14.
[0040] By executing the first solver module 13.1, the processor unit 3 calculates an optimized speed trajectory according to which the motor vehicle is to move within the prediction horizon. The optimized speed trajectory is calculated for an upcoming route section, taking into account the longitudinal dynamics model 14, wherein the first cost function 15.1 is minimized. Furthermore, by executing the first solver module 13.1, the processor unit 3 calculates for the upcoming route section, taking into account the longitudinal dynamics model, a profile of a state of charge of the battery 9, which minimizes the first cost function 15.1. The battery 9 supplies the electric machine 8 with electrical energy to drive the motor vehicle 1 within the prediction horizon. Furthermore, traffic light information can be converted into time limits at the waypoint of the corresponding traffic light.Regarding preceding vehicles, a probable trajectory of the preceding vehicle can be determined. This results in a minimum time for the HLS trajectory over the path at the current position of the other vehicle.
[0041] By executing the second solver module 13.2, the processor unit 3 further calculates, based on the optimized speed trajectory and the optimized state of charge profile of the battery 9, as well as taking into account secondary conditions, a gear trajectory that minimizes the second cost function 15.2, according to which at least one gear in the transmission 10 of the motor vehicle 1 should be engaged within the prediction horizon. Furthermore, by executing the second solver module 13.2, the processor unit 3 calculates, based on the optimized speed trajectory, the optimized state of charge profile of the battery 9, and taking into account secondary conditions, a driving mode trajectory that minimizes the second cost function 15.2, which determines in which of the above-described driving modes the hybrid drive train 7 should be operated within the prediction horizon.
[0042] By executing the third solver module 13.3 while minimizing the third cost function 15.3, the processor unit 3 also calculates an optimized torque trajectory for the initial section of the prediction horizon for the electric machine 8, the internal combustion engine 17, and the braking system 19 of the motor vehicle 1, according to which the electric machine 8, the internal combustion engine 17, and the braking system 19 are to provide torques within the prediction horizon. While the route section ahead is several kilometers long, the initial section immediately in front of the motor vehicle 1 extends over only a few seconds.
[0043] The output of the optimization by the MPC algorithm 13 with its three solver modules 13.1, 13.2, and 13.3 thus results in optimal torques of the electric machine 8, the internal combustion engine 17, and the braking system 19 for calculated points in the initial section of the prediction horizon. By distributing the solution to the MPC problem among the three solver modules, faster response times of the system 2 are enabled. The processor unit 3 can determine corresponding input variables for the electric machine 8, the internal combustion engine 17, and the braking system 19, so that the optimal torques are established. The processor unit 3 can control the electric machine 8 based on the determined input variable. However, this can also be done by the driver assistance system 16.
[0044] The detection unit 6 can measure current state variables of the motor vehicle 1, record corresponding data, and feed it to the MPC algorithm 13. Furthermore, route data from an electronic map for a look-ahead or prediction horizon (e.g., 5,000 m) in front of the motor vehicle 1 can be updated, in particular cyclically. The route data can, for example, contain gradient information, curve information, and information about speed limits. Furthermore, a curve curvature can be converted into a speed limit for the motor vehicle 1 using a maximum permissible lateral acceleration. Furthermore, the detection unit 6 can be used to locate the motor vehicle, in particular via a signal generated by a GNSS sensor 12 for precise localization on the electronic map.Furthermore, the detection unit for detecting the external environment of the motor vehicle 1 can comprise, for example, a radar sensor, a camera system, and / or a lidar sensor. The processor unit 3 can access information from these elements, for example, via the communication interface 5. This information can be incorporated into the longitudinal model 14 of the motor vehicle 1, in particular as constraints or constraints.
[0045] The longitudinal model 14 of motor vehicle 1 is adapted to the respective requirements of the solver modules 13.1 to 13.3. In particular, each of the solver modules 13.1 to 13.3 thus receives its own longitudinal model of motor vehicle 1.
[0046] A cost function typically consists of a sum term and a final term. Depending on the signals present in the system, the terms can be converted into one another if necessary. The physical quantities described in the cost function can be converted into one another if they are interdependent. For example, torque terms could also be expressed as force terms and vice versa. Another example is velocity terms, which can be represented by kinetic energy terms.
[0047] The first cost function 15.1 can be defined as follows: J=∑i=1N(c1Ekin,minSlack(si)+c2MICE2(si)+c3MEM2(si)+c4FBrk(si) +λc5Δs(si)v(si)m˙Fuel(si)+cxX(si))−λc6EBat(sN)+(1−λ)c7tTotal(sN) +cyY(sN) Here: s distance, independent variable N → s1, ... , s N Waypoints, horizon length Δs(s) = s i - s i-1Distance between waypoints s i λ ∈ [0,1] Factor for time and consumption weighting of the cost function c1, ... , c7, c x , c y Weighting factors of the individual terms of the cost function X, Y Additional terms can be added to the cost function to consider additional objectives in the optimization. These terms can, for example, describe additional components (e.g., additional electrical machines) or their properties (e.g., temperatures). E kin,minSlack Deviation from the minimum desired kinetic energy M ICE Moment of the internal combustion engine M EM Moment of the electric machine F Brk Braking force ṁ Fuel Fuel consumption - mass flow v(s i ) Speed of the motor vehicle E Bat Energy content of the battery t TotalTotal time
[0048] The second cost function 15.2 can be defined, for example, as follows: J=∑i=1N(c1(v(ti)−v*(ti))2+c2(Ebat(ti)−WBat*(ti))2+c3(MICE(ti)−MICE*(ti))2+c4(MEM(ti)−MEM*(ti))2+c5(FBrk(ti)−FBrk*(ti))2 +c6 Δt m˙f(ti, bClutch(ti), zICE(ti))+c7GGearChange2(ti) +c8ClClutchChange2(ti)+cxX(ti))−c9EBat(tN)+cyY(tN) Here: t travel time, independent variable N → t1, ... , t N Time points, horizon duration Δt = t i - t i-1 Distance between times t i c1, ... , c9, c x , c y Weighting factors of the individual terms of the cost function X, Y Additional terms can be added to the cost function to consider additional objectives in the optimization. These terms can, for example, describe additional components (e.g., additional electrical machines) or their properties (e.g., temperatures). v, v* speed, reference speed Energy content of the battery, reference energy content of the battery Torque of the internal combustion engine, reference torque of the internal combustion engine Torque of the electrical machine, reference torque of the electrical machine Braking force, reference braking force m f Fuel consumption b Clutch Condition of the clutch z ICE Condition of the combustion engine G GearChange Activating a gear change Cl CıutchChange Activation of a clutch state change
[0049] The problem complexity increases with the increasing number of mixed-integer decisions to be made. For performance reasons, for example, the decision regarding the state of the combustion engine 18 (on / off) can be separated from the mixed-integer problem and processed in post-processing logic.
[0050] The third cost function 15.3 can be defined as follows: J=∑i=1N((c1sv(ti)2+c2sSOC(ti)2)+c3 Δt m˙f(ti, bCluntch(ti), zICE(ti))+c4FBrk(ti) +c5(dFICE, Wheel(ti))2+c6(dFEM, Wheel(ti))2+c7(dFBrake(ti))2 +cxX(ti))+cyY(tN) Here: t travel time, independent variable N → t1, ... , t N Time points, horizon duration Δt = t i - t i-1 Distance between times t i c1, ... , c7, c x , c y Weighting factors of the individual terms of the cost function X, Y Additional terms can be added to the cost function to consider additional objectives in the optimization. These terms can, for example, describe additional components (additional electrical machines) or their properties (e.g., temperatures). s vSlack variable of speed: The deviation of the speed trajectory from its reference trajectory is accepted within a defined range and penalized outside. s SOC SOC slack variable: The deviation of the SOC trajectory from its reference trajectory is accepted within a defined range and penalized outside it. m f Fuel consumption b Clutch Condition of the clutch z ICE Condition of the internal combustion engine F Brk Braking force dF ICE,Wheel Change in the force at the wheel caused by the internal combustion engine dF EM,Wheel Change in the force at the wheel caused by the electric machine dF Brake Change in the force on the wheel caused by the braking force Reference symbol n1 first value first operating parameter n2 second value of first operating parameter n3 third value of first operating parameter K0 clutch y1 first value of second operating parameter y2 second value of second operating parameter y3 third value of second operating parameter 1 vehicle 2 systems 3 Processor unit 4 storage unit 5 Communication interface 6 Recording unit 7 Drivetrain 8 electric machine 9 Battery 10 gearboxes 11 Computer program product 12 GNSS sensors 13 MPC algorithm 14 Longitudinal dynamics model 15.1 first cost function 15.2 second cost function 15.3 third cost function 16 Driver assistance system 17 Internal combustion engine 18 Power electronics 19 Brake system 20 first operating parameters 21 front differential gear 22 front wheel 23 front wheel 24 second operating parameters 25 front axle 26 rear wheel 27 Loss model 28 rear wheel 29 Rear axle
Claims
[1] Processor unit (3) for model-based predictive control of a motor vehicle (1), wherein the processor unit (3) is designed to - to execute an MPC algorithm (13) comprising a first solver module (13.1), a second solver module (13.2), a third solver module (13.3), a longitudinal dynamics model (14) and three cost functions (15.1 to 15.3), wherein a first cost function (15.1) is assigned to the first solver module (13.1), wherein a second cost function (15.2) is assigned to the second solver module (13.2) and wherein a third cost function (15.3) is assigned to the third solver module (13.3), - by executing the first solver module (13.1) for a preceding section of the route taking into account the longitudinal dynamics model (14) - to calculate a speed trajectory minimizing the first cost function (15.1), according to which the motor vehicle (1) is to move within a prediction horizon, and - to calculate a profile of a state of charge of a battery (9) which minimises the first cost function (15.1) and which serves as an energy store for an electric machine (8) of the motor vehicle (1), - by executing the second solver module (13.2) based on the speed trajectory, based on the state of charge of the battery (9) and taking into account constraints, to calculate a trajectory of integer control variables minimizing the second cost function (15.2), and - by executing the third solver module (13.3) for an initial section of the prediction horizon for the electric machine (8), an internal combustion engine (17) and a braking system (19) of the motor vehicle (1), to calculate a torque trajectory minimizing the third cost function (15.3), according to which the electric machine (8), the internal combustion engine (17) and the braking system (19) are to provide torques within the prediction horizon. [2] Processor unit (3) according to claim 1, wherein the processor unit (3) is arranged to - to calculate the speed trajectory for the section of track which is a few kilometers long by executing the first solver module (13.1), and - to calculate the trajectory of integer control variables and the target torque trajectory for an initial section of the prediction horizon located a few seconds immediately in front of the motor vehicle (1) by executing the second solver module (13.2) and the third solver module (13.3). [3] Processor unit (3) according to claim 1 or 2, wherein the trajectory of integer control variables comprises a course of gears that can be engaged in a transmission (10) of the motor vehicle (1) within the prediction horizon. [4] Processor unit (3) according to one of the preceding claims, wherein the trajectory of integer control variables comprises a course of an engine state according to which the internal combustion engine (17) of the motor vehicle (1) is either switched on or switched off. [5] Processor unit (3) according to one of the preceding claims, wherein the trajectory of integer control variables comprises a course of a clutch state, according to which a clutch (K0) arranged between the internal combustion engine (17) of the motor vehicle (1) and the electric machine (8) is either in an open position or in a closed position. [6] Processor unit (3) according to one of the preceding claims, the motor vehicle (1) comprising a hybrid drive train (7), wherein - the hybrid drive train (7) can be operated at least in a first driving mode and in a second driving mode, and - the processor unit (3) is configured to calculate, by executing the second solver module (13.2), a driving mode trajectory that minimizes the second cost function (15.2) based on the speed trajectory, based on the course of the state of charge of the battery (9) and taking into account constraints, which determines in which driving mode or in which of the driving modes the hybrid drive train (7) is to be operated within the prediction horizon. [7] Hybrid drive train (7) for a motor vehicle (1), the hybrid drive train (7) comprising - at least one first drive unit in the form of an internal combustion engine (17), - at least one second drive unit in the form of an electric machine (8) which can be operated as a motor and as a generator, - at least one battery (9), - at least one gearbox (10), - a braking system (19) and - a processor unit (3) according to one of the preceding claims. [8] Motor vehicle (1) comprising a hybrid drive train (7) according to claim 7 and a driver assistance system (16) which is designed to - to access a speed trajectory according to which the motor vehicle (1) is to move within a prediction horizon, wherein the speed trajectory has been calculated by the processor unit (3) by executing the first solver module (13.1) for a route section ahead, taking into account the longitudinal dynamics model (14) and minimizing the first cost function (15.1), - to access a profile of a charge state of the battery (9) which serves as an energy storage device for the electric machine (8), wherein the profile of the charge state of the battery (9) has been calculated by the processor unit (3) by executing the first solver module (13.1) for the upcoming route section, taking into account the longitudinal dynamics model (14) and minimizing the first cost function (15.1), - to access a trajectory of integer control variables, wherein the trajectory of integer control variables has been calculated by the processor unit (3) by executing the second solver module (13.2) based on the speed trajectory, based on the course of the state of charge of the battery (9) and taking into account constraints and minimizing the second cost function (15.2), - to access a torque trajectory according to which the at least one internal combustion engine (17), the at least one electric machine (8) and the braking system (19) are to provide torques within the prediction horizon, wherein the torque trajectory has been calculated by the processor unit (3) by executing the third solver module (13.3) for an initial section of the prediction horizon while minimizing the third cost function (15.3), and - to control the at least one internal combustion engine (17), the at least one electric machine (8) and the braking system (19) of the hybrid drive train (7) based on the torque trajectory. [9] Method for model-based predictive control of a motor vehicle (1) by means of an MPC algorithm (13) comprising a first solver module (13.1), a second solver module (13.2), a third solver module (13.3), a longitudinal dynamics model (14) and three cost functions (15.1 to 15.3), wherein a first cost function (15.1) is assigned to the first solver module (13.1), wherein a second cost function (15.2) is assigned to the second solver module (13.2) and wherein a third cost function (15.3) is assigned to the third solver module (13.3), the method comprising the steps: - Calculating a speed trajectory according to which the motor vehicle (1) is to move within a prediction horizon, wherein the speed trajectory is calculated by executing the first solver module (13.1) for a route section ahead, taking into account the longitudinal dynamics model (14) and minimizing the first cost function (15.1), - Calculating a profile of a state of charge of a battery (9) which serves as an energy storage device for an electric machine (8) of the motor vehicle (1), wherein the profile of the state of charge of the battery (9) is calculated by executing the first solver module (13.1) for the upcoming route section, taking into account the longitudinal dynamics model (14) and minimizing the first cost function (15.1), - Calculating a trajectory of integer control variables, wherein the trajectory of integer control variables is calculated by executing the second solver module (13.2) based on the speed trajectory, based on the course of the state of charge of the battery (9) and taking into account constraints and minimizing the second cost function (15.2), - Calculating a moment trajectory according to which an internal combustion engine (17), the electric machine (8) and a braking system (19) of the motor vehicle (1) are to provide moments within the prediction horizon, wherein the moment trajectory is calculated by executing the third solver module (13.3) for an initial section of the prediction horizon while minimizing the third cost function (15.3). [10] Computer program product (11) for model-based predictive control of a motor vehicle (1) by means of an MPC algorithm (13) comprising a first solver module (13.1), a second solver module (13.2), a third solver module (13.3), a longitudinal dynamics model (14) and three cost functions (15.1 to 15.3), wherein a first cost function (15.1) is assigned to the first solver module (13.1), wherein a second cost function (15.2) is assigned to the second solver module (13.2) and wherein a third cost function (15.3) is assigned to the third solver module (13.3), wherein the computer program product (13) instructs the processor unit (3), - by executing the first solver module (13.1) for a preceding section of the route taking into account the longitudinal dynamics model (14) - to calculate a speed trajectory minimizing the first cost function (15.1), according to which the motor vehicle (1) is to move within a prediction horizon, and - to calculate a profile of a state of charge of a battery (9) which minimises the first cost function (15.1) and which serves as an energy store for an electric machine (8) of the motor vehicle (1), - by executing the second solver module (13.2) based on the speed trajectory, based on the state of charge of the battery (9) and taking into account constraints, to calculate a trajectory of integer control variables minimizing the second cost function (15.2), and - by executing the third solver module (13.3) while minimizing the third cost function (15.3) for an initial section of the prediction horizon for the electric machine (8), an internal combustion engine (17) and a braking system (19) of the motor vehicle (1), to calculate a torque trajectory according to which the electric machine (8), the internal combustion engine (17) and the braking system (19) are to provide torques within the prediction horizon.
Citation Information
Patent Citations
PHEV traffic light intersection passing control method based on model predication control
CN108944905A
Vehicle with an electric drive and methods for operating such a vehicle
DE102013008716A1
Method and device for the predictive operation of a motor vehicle
DE102014209687A1
Insertion device and device for inserting an intraocular lens into an eye
DE102014222069A1
Control device and method for controlling a drive system of a hybrid vehicle
DE102019105665A1
