MPC-based autonomous driving functions for motor vehicles

Abstract

The invention relates to the implementation of an autonomous driving function of a motor vehicle by means of a model-based predictive control (MPC). In particular, a processor unit provided for this purpose is proposed. The processor unit is configured to determine a torque target value (21) within a prediction range (20) and a speed target value (19) within the prediction range (20) by executing an MPC algorithm including a longitudinal dynamics model of a power train of the motor vehicle. Depending on the level of the torque target value (21), the autonomous driving function of the motor vehicle is implemented in a torque predetermined operating mode or in a speed predetermined operating mode. In the torque predetermined operating mode, a drive unit of the power train is controlled based on the torque target value (21). In the speed operating mode, a speed regulator of the power train is controlled based on the speed target value (19).

Description

Technical Field

[0001] The invention relates to the implementation of autonomous driving functions of a motor vehicle by predictive model-based adjustments. In this context, protection is particularly claimed for a processor unit, a method and a computer program product which can be stored on a computer-readable medium and which is designed for this purpose. The invention also relates to a motor vehicle having such a processor unit. Background Art

[0002] Methods based on model-based predictive control (MPC) are used in the field of trajectory control, for example for motor control in the context of autonomous driving. Autonomous driving strategies use environmental data, map data and vehicle data to determine the optimal vehicle behavior.

[0003] DE 10 2014 209687A1 teaches a method for predictively operating a motor vehicle based on a plurality of driving-related regulating variables. Here, a first electronic range is provided for a first portion of a driving route to be driven, and a second electronic range is provided for a second portion of the driving route to be driven, wherein the first and second electronic ranges respectively have one or more route sections, wherein one or more route parameters and / or one or more regulating variable restrictions are assigned to each route section. In addition, on the driving route to be driven, a cost function for a plurality of target variables to be optimized is provided according to the variation process of the driving-related regulating variable (SG). In addition, a first optimization step is performed, wherein the cost function is optimized under the prescribed predetermined conditions of the variation process of at least one regulating variable (SG) so as to obtain the correlation between a plurality of target variables as the result of the first optimization step. In addition, the cost function is loaded with the obtained correlation between the target variables as a boundary condition at the end of the first electronic range, and the second optimization step is performed using the loaded cost function for the first electronic range.

[0004] Also known from DE 102 21 835 A1 is a method for adjusting a starting element of a motor vehicle (e.g. a hydrodynamic converter, a clutch or an electric machine in a hybrid drive) in a creep mode, wherein a creep torque target value is adjusted at the starting element as a function of at least one of the following environmental conditions or driver specifications of the motor vehicle, namely as a function of the current inclination of the motor vehicle, as a function of the brake pedal pressure caused by the driver of the motor vehicle, as a function of the distance of the motor vehicle from an obstacle or as a function of the type of driver. Summary of the invention

[0005] The object of the invention can be seen as improving the execution of autonomous driving functions of a motor vehicle by predictive model-based adjustments when the motor vehicle is to slow down or stop. This object is achieved by a processor unit according to the invention, a motor vehicle according to the invention, a method according to the invention and a computer program product according to the invention. Advantageous embodiments are the subject of the following description and the drawings.

[0006] The present invention proposes that an MPC algorithm or MPC solver for an autonomous driving function of a motor vehicle having a drive unit, in particular a motor, provides an optimized torque target value for the motor and an optimized speed target value for the motor vehicle. In the case of a positive torque target value, the torque can be used (optionally with a supplementary speed adjustment) as a regulating variable for the electric motor. This has advantages in terms of comfort and the following behavior of the motor vehicle. However, when the motor vehicle is to be parked and the target torque has a negative value for this purpose, it is proposed to switch to speed adjustment. The switch from torque adjustment to speed adjustment can be carried out in particular at the zero crossing of the target torque, which is particularly advantageous for vehicle passengers, especially for comfort reasons, because the switch can be carried out without jolts and imperceptibly. The present invention enables following the optimal speed trajectory, wherein, despite the speed adjustment, the parking function in front of obstacles or traffic lights is possible.

[0007] In this sense, according to a first aspect of the invention, a processor unit is provided for executing an autonomous driving function of a motor vehicle by means of a model-based predictive adjustment. The processor unit can access an MPC algorithm. The MPC algorithm contains a longitudinal dynamics model of the powertrain of the motor vehicle. In addition, the MPC algorithm contains, in particular, a cost function.

[0008] The longitudinal dynamics model of the powertrain can include a vehicle model (partially approximated overall characteristic curve) with vehicle parameters and powertrain losses. In particular, information about the topography of the road ahead (e.g., curves and slopes) can be incorporated into the longitudinal dynamics model of the powertrain. In addition, information about the speed limit of the road ahead can also be incorporated into the longitudinal dynamics model of the powertrain.

[0009] The current state parameters can be measured, the corresponding data can be recorded and fed to the MPC algorithm. Therefore, in particular, the road section data for the forward-looking range or prediction range (e.g. 400m) in front of the motor vehicle from the electronic map can be periodically upgraded or updated. The road section data can, for example, contain slope information, curve information and information about speed limits. In addition, the curve curvature can be converted into the speed limit of the motor vehicle through the maximum permissible lateral acceleration. In addition, the motor vehicle can be positioned, in particular through GNSS signals or SLAM (simultaneous positioning and mapping) algorithms for precise positioning on electronic maps.

[0010] The processor unit is configured to determine a torque target value within a prediction range by executing an MPC algorithm. The torque target value specifies, in particular at discrete road points within the prediction range, the torque that a drive unit of the powertrain should provide in order to drive the motor vehicle. In particular, the processor unit is configured to determine the torque target value by executing the MPC algorithm in such a way that a cost function is minimized. The torque target value can be determined in particular in such a way that the torque target value is correspondingly assigned to discrete road points within the prediction range. In this case, the torque target value is the value that the torque should have in order to drive the motor vehicle as desired at the relevant road point.

[0011] The processor unit is further configured to determine a speed target value within the prediction range by executing the MPC algorithm. The speed target value specifies, in particular, the speed at which the motor vehicle should travel at discrete road points within the prediction range. In particular, the processor unit is configured to determine the speed target value by executing the MPC algorithm in such a way that a cost function is minimized. In particular, the speed target value can be determined in such a way that the speed target value is respectively assigned to a discrete road point within the prediction range. The speed target value is here the value at which the motor vehicle should travel at the relevant road point.

[0012] The processor unit is also configured to execute an MPC algorithm in order to carry out an autonomous driving function of the motor vehicle in a torque-defined operating mode or in a speed-defined operating mode, depending on the level of the torque target value.

[0013] The autonomous driving function enables the motor vehicle to drive itself, i.e. the vehicle occupants do not control the motor vehicle. The driver can hand over control of the motor vehicle to the driver assistance system. The autonomous driving function thus includes that the motor vehicle is configured, in particular by means of a central processing unit, to perform, for example, steering, flashing lights, acceleration and braking maneuvers without human intervention, and in particular to control external lights and signal outputs, such as flashing lights of the motor vehicle. The autonomous driving function may also include a semi-autonomous driving function, which supports the driver of the motor vehicle in controlling the motor vehicle, in particular in steering, flashing lights, acceleration and braking maneuvers, wherein the driver still has control of the motor vehicle.

[0014] Depending on the level of the torque target value, the autonomous driving function is executed in the torque predetermined operating mode or in the speed predetermined operating mode. In particular, when the torque is positive or greater than zero, the autonomous driving function is executed in the torque predetermined operating mode. In particular, when the torque target value is negative or zero, the autonomous driving function is executed in the speed predetermined operating mode.

[0015] In the torque presetting operating mode, the drive unit of the powertrain is controlled based on the torque target value. This control can be carried out by a processor unit, which can be configured accordingly for this purpose. In addition, the driver assistance system can also assume the control of the drive unit. The drive unit can be an electric motor, in particular an electric motor. In addition, the drive unit can also be an internal combustion engine. In addition, the drive unit can also be a hybrid drive unit, which has an electric motor and an internal combustion engine to drive the motor vehicle.

[0016] In the speed setting operating mode, the speed regulator of the powertrain is controlled based on the speed target value. This control can be carried out by a processor unit, which can be configured accordingly for this purpose. In addition, the driver assistance system can also assume the control of the speed regulator. In the speed operating mode, the speed regulator can be activated in a controller of the motor vehicle or the drive unit, in particular the electric motor, which then sets a predetermined speed target value. This can avoid the risk of backing up in front of obstacles or traffic lights.

[0017] In the case of a positive target torque, a torque setting can be used. Therefore, in one embodiment, the processor unit can be configured to perform an autonomous driving function in a torque setting operating mode by executing an MPC algorithm when the torque target value has a positive value. A good following behavior of the motor vehicle is achieved by the torque-controlled operation of the drive unit of the powertrain.

[0018] In the case of a positive torque target value, the torque-controlled operation of the drive unit can be supplemented by a speed control (speed control with torque precontrol). This represents a combination of a torque-prescribed operating mode and a speed-prescribed operating mode. In this sense, the processor unit is configured in another embodiment to control the speed controller of the drive train based on the speed target value when the torque target value has a positive value by executing an MPC algorithm.

[0019] However, when stopping in front of obstacles, such as traffic lights or vehicles traveling ahead, torque control is not very suitable for repeatedly setting the distance. There is a risk of accelerating backwards or the motor vehicle reversing, for example at traffic lights. In order to solve this problem, it is proposed to use only speed control in the case of negative target torques. This can reliably prevent the motor vehicle from accidentally reversing in front of obstacles or traffic lights. In this sense, the processor unit can be configured to perform the autonomous driving function in the speed presetting operating mode only when the torque target value has a negative value by executing the MPC algorithm. The possibility of using other braking devices, such as the vehicle's service brake, is not affected by this.

[0020] The point in time when the switch from the torque preset operating mode to the speed preset operating mode (and vice versa) plays an important role. If the switch is made just before the motor vehicle is about to stop when the torque target value has already had a negative value for some time, this may lead to disadvantages in terms of comfort or complex transition functions must be taken into account. It is desired that the motor vehicle stops at the right time in front of an obstacle or at the right position in front of a traffic light. The inventors have found that it is particularly advantageous to switch from the torque preset operating mode to the speed preset operating mode (and vice versa) exactly at the zero crossing of the torque target value. The switch at this point in time is not or hardly noticeable to the occupants of the motor vehicle. In addition, it can be achieved that the motor vehicle can stop at the desired time and at the right position with the help of the autonomous driving function.

[0021] In this sense, the processor unit is configured in an embodiment to switch from a torque predetermined operating mode to a speed predetermined operating mode by executing an MPC algorithm, wherein the first torque target value has a zero value at the first time point, and wherein, at a second time point immediately following it, the second torque target value has a negative value. The feature "immediately following time point" can be understood as generating a torque target value at a discrete time point, wherein the first time point and the second time point are two discrete time points that directly follow each other. This embodiment enables the switch from the torque predetermined operating mode to the speed predetermined operating mode to be performed without bumps and imperceptibly at the zero crossing of the torque. In particular, two conditions must be met at the same time, namely, the value of "torque target value equal to zero" must be followed by a negative torque target value, so that the switch is made to the speed predetermined operating mode. In other words, the torque target value should fall into the negative category from the zero crossing.

[0022] In addition, the switch from the speed predetermined operating mode to the torque predetermined operating mode can also be carried out at the zero crossing of the torque target value. The zero crossing can be detected or defined in particular in such a way that the torque target value has a zero value and the torque target value immediately following it is positive. In other words, the torque target value increases into the positive range from the zero crossing. In this sense, the processor unit can be configured to switch from the speed predetermined operating mode to the torque predetermined operating mode (again) at a third time point by executing the MPC algorithm, wherein at the third time point, the third torque target value has a zero value, and wherein, at the fourth time point immediately following it, the fourth torque target value has a positive value.

[0023] In another embodiment, the drive unit is an electric motor and the MPC algorithm includes a cost function to be minimized. The cost function may include as a first item the electric energy weighted by a first weighting factor and predicted according to a longitudinal dynamics model, which is provided by a battery of the powertrain within a prediction range to drive the electric motor. The cost function may include as a second item the driving time weighted by a second weighting factor and predicted according to a longitudinal dynamics model, which the motor vehicle needs to travel the entire distance predicted within the prediction range. The processor unit may be configured to minimize the cost function by acquiring input variables for the motor according to the first item and according to the second item by executing the MPC algorithm. This embodiment enables the objective function or cost function of the driving efficiency driving strategy to include another item in addition to energy consumption, thereby also minimizing the driving time. This results in that, depending on the selection of the weighting factor, a low speed is not always evaluated as optimal, and therefore there is no problem that the obtained speed is always at the lower limit of the permissible speed.

[0024] The cost function can have only linear terms and quadratic terms. As a result, the entire problem has the form of a quadratic optimization with linear auxiliary conditions, and a convex problem that can be solved well and quickly is obtained. The objective function or cost function can be configured with weighting (weighting factors), wherein in particular energy efficiency, driving time and driving comfort are calculated and weighted. The energy-optimal speed trajectory can be calculated online for the front range on a processor unit, which can in particular form a component of the central control unit of the motor vehicle. By using the MPC method, the target speed of the motor vehicle can also be periodically recalculated based on the current driving state and the road section information ahead.

[0025] The driving time for the forecast range is minimized and the consumed energy is minimized by the cost function of the MPC algorithm. In an embodiment, the torque change is also minimized for the forecast range. As for the input of the model-based predictive adjustment, for example, the speed limit, the physical limit of the torque and the speed of the motor can be supplied to the MPC algorithm as auxiliary conditions. In addition, the control parameters for optimization, in particular the speed of the vehicle (which can be proportional to the speed), the torque of the motor and the battery charge state can be supplied to the MPC algorithm as input. The MPC algorithm can provide the optimal speed and the optimal torque as optimized outputs for the calculated points in the look-ahead range. As for the implementation of the MPC adjustment in the vehicle, a software module can be connected downstream of the MPC algorithm, which software module obtains the current relevant status and forwards it to the power electronics.

[0026] Energy consumption and travel time can be evaluated and weighted at the end of the range, respectively. Therefore, this term is only valid for the last point of the range. In this sense, in an embodiment, the cost function contains a final value of energy consumption weighted by a first weight factor, the predicted electric energy takes this final value of energy consumption at the end of the prediction range, and the cost function contains a final value of travel time weighted by a second weight factor, the predicted travel time takes this final value of travel time at the end of the prediction range.

[0027] In order to ensure comfortable driving, a term for penalizing sudden torque changes can be additionally introduced. In this sense, the cost function can have a third term with a third weighting factor, wherein the third term contains a torque value predicted according to the longitudinal dynamics model, which the electric machine provides for driving the motor vehicle, and wherein the processor unit is configured to determine input variables for the electric machine according to the first term, according to the second term, and according to the third term by executing the MPC algorithm, so as to minimize the cost function.

[0028] For the first point in the range, the deviation from the last set torque can be evaluated as negative in order to ensure a seamless and jerk-free transition when switching between the old trajectory and the new trajectory. In this sense, the third term can contain a first value of the torque predicted by the longitudinal dynamics model, weighted by a third weighting factor, which the motor provides at the first road point in the prediction range to drive the motor vehicle. In this case, the third term can contain a zeroth value of the torque, weighted by a third weighting factor, which the motor provides at the zeroth road point to drive the motor vehicle, the zeroth road point being located directly before the first road point. The zeroth torque can in particular be the actual (not just predicted) torque provided by the motor. In the cost function, the zeroth value of the torque can be subtracted from the first value of the torque.

[0029] Alternatively, the third term may contain a first value of the driving force predicted according to the longitudinal dynamics model, which is provided by the electric machine for driving the motor vehicle at a first road point in the prediction range, weighted by a third weighting factor. Here, the third term contains a zeroth value of the driving force, which is provided by the electric machine for driving the motor vehicle at a zeroth road point, which is located directly before the first road point, weighted by the third weighting factor, wherein the zeroth value of the driving force may be subtracted from the first value of the driving force in the cost function.

[0030] The road points taken into account by the MPC algorithm are in particular discrete road points which follow one another, for example, at a certain frequency. In this sense, the zeroth road point and the first road point are discrete road points, wherein the first road point directly follows the zeroth road point. The zeroth road point can be before the prediction horizon in terms of time. For the zeroth road point, the zeroth torque value can be measured or determined. The first road point is in particular the first road point in the prediction horizon. For the first road point, a first torque value can be predicted. Therefore, the actually determined zeroth torque value can be compared with the predicted first torque value.

[0031] In addition, a torque gradient that is too high in the range is disadvantageous, so that the torque gradient is already penalized in the objective function in the embodiment. For this purpose, the square of the drive force deviation per meter can be weighted and minimized in the objective function. In this sense, the cost function can have a fourth term with a fourth weighting factor, wherein the fourth term contains the torque gradient predicted according to the longitudinal dynamics model or an indication of the torque gradient predicted according to the longitudinal dynamics model. The processor unit is configured to determine the input variables for the electric machine according to the first term, according to the second term, according to the third term and according to the fourth term by executing the MPC algorithm, so as to minimize the cost function.

[0032] In an embodiment, the fourth term includes the square of the torque gradient deviation multiplied by the fourth weighting factor and accumulated. In addition, the cost function may include the square of the driving force deviation multiplied by the fourth weighting factor, the motor providing the driving force to move the vehicle forward one meter in the longitudinal direction. In this sense, the fourth term may include the square of the driving force deviation multiplied by the fourth weighting factor and accumulated. The motor provides the driving force to move the vehicle forward one meter in the longitudinal direction.

[0033] For example, the speed limit that can be stipulated by traffic road regulations is a hard limit for optimization, which should not be exceeded. In fact, slightly exceeding the speed limit is always allowed, and it is normal in particular when transitioning from one speed zone to the second zone. What can happen in a dynamic environment where the speed limit moves from one calculation cycle to the next is that, in the case of a completely hard limit, a valid solution for the speed curve can no longer be found. In order to increase the stability of the calculation algorithm, so-called "soft constraints" can be introduced into the cost function. In particular, before reaching the hard speed limit, the so-called "slip variable" or "slack variable" can become effective within a preset narrow range. Solutions that are very close to the speed limit, that is, solutions whose speed trajectory keeps a certain distance from the hard limit, can be evaluated as worse here. In this sense, the cost function can contain a slack variable weighted by a fifth weighting factor as the fifth term, wherein the processor unit is configured to minimize the cost function by executing the MPC algorithm according to the first term, according to the second term, according to the third term, according to the fourth term, and according to the fifth term to obtain the input variables for the motor.

[0034] According to a second aspect of the present invention, a motor vehicle is provided. The motor vehicle comprises a processor unit, a drive unit, a speed regulator and a driver assistance system according to the first aspect of the present invention. The driver assistance system is configured to access a torque target value and a speed target value by means of a communication interface, wherein the torque target value and the speed target value have been known by the processor unit. In addition, the driver assistance system is configured to perform an autonomous driving function of the motor vehicle in a torque predetermined operating mode or in a speed predetermined operating mode according to the level of the torque target value, wherein the drive unit is controlled based on the torque target value in the torque predetermined operating mode, and wherein the speed regulator is controlled based on the speed target value in the speed predetermined operating mode. The vehicle is, for example, a motor vehicle, such as a car (e.g., a private car weighing less than 3.5 t), a motorcycle, a scooter, a light motorcycle, a bicycle, an electric bicycle, a bus or, for example, a truck weighing more than 3.5 t. For example, the vehicle may belong to a fleet.

[0035] According to a third aspect of the present invention, a method for performing an autonomous driving function of a motor vehicle by means of a model-based predictive adjustment is provided. The method comprises the following steps:

[0036] - determining a torque target value within a prediction range by executing an MPC algorithm including a longitudinal dynamics model of the powertrain of the motor vehicle,

[0037] - Determine a speed target value within the prediction range by executing the MPC algorithm, and

[0038] - depending on the level of the torque target value, an autonomous driving function of the motor vehicle is performed in a torque setpoint operating mode or in a speed setpoint operating mode, wherein in the torque setpoint operating mode a drive unit of the power train is controlled based on the torque target value, and wherein in the speed operating mode a speed regulator of the power train is controlled based on the speed target value.

[0039] According to a fourth aspect of the present invention, a computer program product is provided for executing an autonomous driving function of a motor vehicle by means of a model-based predictive adjustment, wherein when the computer program product is executed on a processor unit, the processor unit is instructed to execute an MPC algorithm including a longitudinal dynamics model of a powertrain of the motor vehicle.

[0040] - determine a torque target value within the prediction range,

[0041] - determine the speed target value within the prediction range,

[0042] - depending on the level of the torque target value, executing an autonomous driving function of the motor vehicle in a torque-preset operating mode or in a speed-preset operating mode,

[0043] wherein, in the torque-predetermining operating mode, the drive units of the powertrain are controlled based on the torque target value, and

[0044] In this case, in the speed operating mode, a speed regulator of the drive train is controlled based on a speed target value.

[0045] According to a fifth aspect of the present invention there is provided a computer readable medium having stored thereon a computer program product according to the fourth aspect of the present invention.

[0046] The above-mentioned definitions and implementation schemes regarding the technical effects, advantages and advantageous embodiments of the processor unit are equally applicable to the vehicle according to the second aspect of the invention, the method according to the third aspect of the invention, the computer program product according to the fourth aspect of the invention and the computer-readable medium according to the fifth aspect of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Next, embodiments of the present invention are described in detail with reference to the schematic drawings, wherein identical or similar elements are provided with the same reference numerals.

[0048] Figure 1 shows a schematic diagram of a vehicle having a powertrain including an electric motor and a battery;

[0049] Figure 2 Shown is the Figure 1 The comprehensive characteristic curve of the motor of the vehicle; and

[0050] Figure 3 A graph showing the torque target value and the speed target value with respect to time or with respect to a road point within a prediction range is shown. DETAILED DESCRIPTION

[0051] Figure 1A motor vehicle 1, such as a private car, is shown. The motor vehicle 1 includes a system 2 for predictive adjustment of an electric motor of a powertrain of the motor vehicle 1 based on a model. In the illustrated embodiment, the system 2 includes a processor unit 3, a memory unit 4, a communication interface 5, and a detection unit 6 for detecting state data related to the motor vehicle 1. The motor vehicle 1 also includes a powertrain 7, which may include, for example, an electric motor 8 operating as a motor and a generator, a battery 9, a transmission 10, and a speed regulator 18. In motor operation, the electric motor 8 can drive the wheels of the motor vehicle 1 through the transmission 10, which can have a constant transmission ratio, for example. The battery 9 can provide the electrical energy required for this. When the electric motor 8 is running in generator operation (recovery), the battery 9 can be charged by the electric motor 8. The battery 9 can also be optionally charged at an external charging station. The powertrain of the motor vehicle 1 can also optionally have an internal combustion engine 17, which can drive the motor vehicle 1 as an alternative or in addition to the electric motor 8. The internal combustion engine 17 can also drive the electric motor 8 to charge the battery 9.

[0052] The computer program product 11 may be stored on the memory unit 4. The computer program product 11 may be executed on the processor unit 3, for which purpose the processor unit 3 and the memory 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 implement the functions described in conjunction with the drawings or to perform the method steps described in conjunction with the drawings.

[0053] The computer program product 11 includes an MPC algorithm 13. The MPC algorithm 13 in turn includes a longitudinal dynamics model 14 of the powertrain 7 of the motor vehicle 1 and a cost function 15 to be minimized. The processor unit 3 executes the MPC algorithm 13 and predicts the behavior of the motor vehicle 1 based on the longitudinal dynamics model 14, wherein the cost function 15 is minimized. For the calculated points in the prediction range, a torque target value for the motor 8 and a speed target value for the motor vehicle 1 are obtained as outputs optimized by the MPC algorithm 13. For this purpose, the processor unit 3 can obtain input variables for the motor 8 and for the speed regulator 18, so that an optimal torque and / or an optimal speed occur. The processor unit 3 can control the motor 8 and the speed regulator 18 based on the obtained input variables. However, this can also be achieved by the driver assistance system 16.

[0054] The detection unit 6 can measure the current state parameters of the motor vehicle 1, record the corresponding data and transmit them to the MPC algorithm 13. Therefore, the road section data for the forward-looking range or predicted range (for example, 400m) in front of the motor vehicle 1 from the electronic map can be upgraded or updated periodically. The road section data can, for example, contain slope information, curve information and information about speed limits. In addition, the curve curvature can be converted into the speed limit of the motor vehicle 1 through the maximum permissible lateral acceleration. In addition, the motor vehicle 1 can be positioned with the help of the detection unit 6, especially accurately positioned on the electronic map through the signal generated by the GNSS sensor 12. For example, the processor unit 3 can access this information through the communication interface 5.

[0055] An exemplary longitudinal dynamics model 14 of the motor vehicle 1 can be mathematically represented as follows:

[0056]

[0057] Here:

[0058] vThe speed of the motor vehicle;

[0059] F trac Traction, which is applied to the wheels of a motor vehicle by means of a motor or brakes;

[0060] F r rolling resistance, which is the effect of the deformation of the tire as it rolls and depends on the load on the wheel (normal force between wheel and road) and therefore on the inclination angle of the road;

[0061] F gr Hill resistance, which describes the longitudinal component of the weight acting on the vehicle during uphill or downhill driving, depending on the gradient of the road;

[0062] F d the aerodynamic drag of the motor vehicle; and

[0063] m eq The equivalent mass of the motor vehicle; the equivalent mass includes in particular the inertia of the rotating parts of the drive train (motor, transmission drive shaft, wheels) that are subject to the acceleration of the motor vehicle.

[0064] By converting time into distance and to eliminate the square of the velocity in the air resistance, The coordinate transformation that eliminates the quadratic velocity term in air resistance yields:

[0065]

[0066] In order to be able to solve this problem quickly and simply by means of the MPC algorithm 13, the dynamic equations of the longitudinal dynamics model 14 are linearized by converting the velocity into the kinetic energy de kin To express it. From this, it is used to calculate the air resistance F d The squared terms of are replaced by linear terms and at the same time the longitudinal dynamics model 14 of the motor vehicle 1 is no longer described in terms of time as usual but in terms of distance. In this respect it is well suited for optimization problems since the look-ahead information on the electric range is based on the distance.

[0067] In addition to the kinetic energy, there are two further state variables which, in the context of a simple optimization problem, must also be described linearly and with respect to the distance. On the one hand, the electrical energy consumption of the drive train 7 is usually described in the form of a comprehensive characteristic curve as a function of torque and motor speed. In the exemplary embodiment shown, the motor vehicle 1 has a fixed transmission ratio between the electric motor 8 and the road on which the motor vehicle 1 is traveling. As a result, the rotational speed of the electric motor 8 can be converted directly into the speed of the motor vehicle 1 or the kinetic energy of the motor vehicle 1. Furthermore, the electrical power of the electric motor 8 can be converted into energy consumption per meter by dividing by the corresponding speed. As a result, the comprehensive characteristic curve of the electric motor 8 is obtained in Figure 2 In order to use this comprehensive characteristic curve for optimization, a linear approximation is made: Energy per meter perMeter )≥a i *e kin +b i *F trac , applicable to all i.

[0068] An exemplary cost function 15 to be minimized can be mathematically represented as follows:

[0069]

[0070] Here,

[0071]

[0072]

[0073] The cost function 15 has only linear and quadratic terms. Thus, the entire problem has the form of a quadratic optimization with linear auxiliary conditions, and a convex problem is obtained that can be solved well and quickly.

[0074] The cost function 15 includes a first weight factor W Bat Weighted and predicted electric energy E according to the longitudinal dynamics model Bat As a first item, the electrical energy is provided by the battery 9 of the powertrain 7 to drive the motor 8 within the foreseeable range.

[0075] The cost function 15 includes a second weight factor W Time The travel time T that the motor vehicle 1 needs to travel the predicted distance is weighted and predicted according to the longitudinal dynamics model 14 as the second term. This results in that, depending on the selection of the weighting factors, a low speed is not always evaluated as optimal, and therefore there is no longer the problem that the resulting speed is always at the permissible lower speed limit.

[0076] Energy consumption and travel time can be evaluated and weighted at the end of the range, respectively. These terms are then only valid for the last point of the range.

[0077] Too high a torque gradient in the range is disadvantageous. Therefore, the torque gradient has been penalized in the cost function 15, i.e., by the term The square of the driving force deviation per meter is penalized with a weight factor W Tem weighted and minimized in the cost function. As the driving force per meter F A Alternatively, the torque M provided by the motor 8 can also be used EM And with the weight factor W Tem To weight, so as to obtain the alternative items Due to the constant transmission ratio of the transmission 10 , the drive force and the torque are directly proportional to one another.

[0078] In order to ensure comfortable driving, an additional term for penalizing sudden torque changes is introduced into the cost function 15, namely w Temstart ·(F A (s1)-F A (s0)) 2 As the driving force F A Alternatively, the torque M provided by the motor 8 can also be used. EM , thus generating the alternative term w Temstart ·(M EM (s1)-M EM (s0)) 2 For the first point in the prediction range, the deviation from the last set torque is evaluated as negative and weighted by the factor W. TemStart Weighted to ensure there is a seamless and jerk-free transition when switching between the old and new tracks.

[0079] The speed limit is a hard limit for the optimization that must not be exceeded. In practice, a slight exceedance of the speed limit is always permitted and is a normal situation in particular when transitioning from one speed zone to a second zone. In a dynamic environment, in which the speed limit transitions from one calculation cycle to the next, it can happen that no valid solution for the speed curve can be found with completely hard limits. In order to increase the stability of the calculation algorithm, soft limits ("soft constraints") are introduced into the cost function 15. Here, before the hard speed limit is reached, the weighting factor W is used to calculate the speed. Slack Weighted slack variable Var Slack It becomes valid within a predetermined narrow range. Solutions that are very close to the speed limit, that is, solutions whose speed trajectory maintains a certain distance from the hard limit, are evaluated as worse.

[0080] Figure 3 Two graphs arranged one above the other are shown. Figure 3 The first diagram shown at the top shows the course of the speed target value 19 within the prediction range 20. Figure 3 The second graph shown below the first graph in FIG. 1 shows the curve of the torque target value 21, which corresponds to the speed target value 19 located above. Although the curve of the speed target value 19 and the curve of the torque target value 21 are Figure 3 Continuously shown as a graph in FIG, but the speed target value 19 and the torque target value 21 can be discrete values, which are assigned to discrete time points and / or road points. The horizontal axes of the two diagrams can represent the time course t and / or road points s within the prediction range 20.

[0081] The processor unit 3 can execute autonomous driving functions of the motor vehicle 1 in a torque-predetermined operating mode or in a speed-predetermined operating mode depending on the magnitude of the torque target value 21. Figure 3 ), the torque target value 21 is positive, so that the processor unit 3 will perform an autonomous driving function in a torque-predetermined operating mode, in which the motor 8 is controlled based on the torque target value 21. The torque-predetermined operating mode can optionally be supplemented by a speed regulation.

[0082] In the second category II of the forecast range 20 (in Figure 3 In the second category II, the torque target value 21 is negative, for example because the motor vehicle 1 is to be stationary, for example, in order to stop at a red light. Therefore, the speed target value 19 in the second category II drops to the value zero ("motor vehicle 1 is stationary"). When this speed target value is reached, the torque target value increases again to the value zero (outside the negative range). The processor unit 3 will perform an autonomous driving function in the second category II in the speed predetermined operating mode, in which the speed regulator 18 is controlled only based on the speed target value 19 and not the torque target value 21.

[0083] In particular, the switch from the torque target operating mode (first category I) to the speed target operating mode (second category II) occurs at a "negative" zero crossing 22 of the torque target value 21. In particular, a negative zero crossing can be detected or defined in such a way that the first torque target value 23 has a zero value at a first time t1 and the following second torque target value 24 is negative at a second time t2. In other words, the torque target value 21 drops or decreases from the first torque target value 23 into the negative range.

[0084] In the third category III ( Figure 3 ), the torque target value 21 is positive again, so that the processor unit 3 will once again perform the autonomous driving function in the torque setting operating mode, wherein the motor 8 is controlled based on the torque target value 21. Here, the torque setting operating mode can also be optionally supplemented by speed regulation.

[0085] The switch from the speed preset operating mode (second category II) to the torque preset operating mode (third category III) can be carried out at a "positive" zero crossing 25 of the torque target value 21. This positive zero crossing can be detected or defined in particular in such a way that the third torque target value 26 has a zero value at the third time t3 and the following fourth torque target value 27 is positive at the fourth time t4. In other words, the torque target value 21 increases from the third torque target value 26 into the positive range.

[0086] Reference numerals list

[0087] I. First category of forecast horizon

[0088] II The second category of the forecast range

[0089] III The third category of the forecast range

[0090] t1 First time point

[0091] t2 First time point

[0092] t3 First time point

[0093] t4 First time point

[0094] 1 Vehicle

[0095] 2 System

[0096] 3 Processor Units

[0097] 4 Memory cells

[0098] 5 Communication Interface

[0099] 6 Detection unit

[0100] 7 Powertrain

[0101] 8 Motor

[0102] 9. Battery

[0103] 10 Transmission

[0104] 11 Computer program products

[0105] 12 GPS sensor

[0106] 13 MPC algorithm

[0107] 14 Longitudinal dynamics model

[0108] 15 Cost Function

[0109] 16 Driver Assistance Systems

[0110] 17 Internal combustion engine

[0111] 18 Speed ​​​​adjuster

[0112] 19 Speed ​​target value

[0113] 20 Forecast Range

[0114] 21 Torque target value

[0115] 22 “Negative” zero crossing

[0116] 23 First torque target value (zero)

[0117] 24 Second torque target value (negative)

[0118] 25 “Positive” zero crossing

[0119] 26 Third torque target value (zero)

[0120] 27 Fourth torque target value (positive)

Claims

1. A processor unit (3) for executing an autonomous driving function of a motor vehicle (1) by means of a model-based predictive control, wherein: The processor unit (3) is configured to, by executing an MPC algorithm (13) including a longitudinal dynamics model (14) of a powertrain (7) of the motor vehicle (1), - determining a torque target value (21) within a prediction range (20), - determining a speed target value (19) which lies within the prediction range, - wherein, in a torque-predetermined operating mode, the drive unit of the powertrain (7) is controlled based on the torque target value (21), and - wherein in a speed predetermined operating mode a speed regulator (18) of the powertrain (7) is controlled based on the speed target value (19), The invention is characterized in that the processor unit (3) is configured to: execute an MPC algorithm (13) to: perform an autonomous driving function of the motor vehicle (1) in the torque predetermined operating mode or in the speed predetermined operating mode, depending on the level of the torque target value (21); The processor unit (3) is configured to, by executing an MPC algorithm (13), When the torque target value (21) has a positive value, an autonomous driving function is performed in the torque-predetermined operating mode, wherein a speed regulator (18) of the powertrain (7) is controlled based on the speed target value (19), and When the torque target value (21) has a negative value, the autonomous driving function is performed only in the speed predetermined operating mode.

2. The processor unit (3) according to claim 1, wherein: The processor unit (3) is configured to switch from the torque-predetermined operating mode to the speed-predetermined operating mode at a first point in time (t1) by executing an MPC algorithm (13), wherein the first torque target value (23) has a zero value at the first point in time (t1), and wherein the second torque target value (24) has a negative value at an immediately subsequent second point in time (t2).

3. The processor unit (3) according to claim 1, wherein: The processor unit (3) is configured to switch from the speed-predetermined operating mode to the torque-predetermined operating mode at a third time point (t3) by executing an MPC algorithm (13), wherein the third torque target value (26) has a zero value at the third time point (t3), and wherein the fourth torque target value (27) has a positive value at an immediately subsequent fourth time point (t4).

4. The processor unit according to claim 1, wherein: - the driving unit is an electric motor, - the MPC algorithm (13) comprises a cost function (15) to be minimized, - the cost function (15) comprises as a first term the electric energy weighted by a first weighting factor and predicted according to the longitudinal dynamics model (14), the electric energy being provided by the battery (9) of the powertrain (7) for driving the electric motor within the prediction range, the cost function (15) comprises as a second term the travel time weighted by a second weighting factor and predicted according to the longitudinal dynamics model (14), which the motor vehicle (1) needs to travel the entire distance predicted within the prediction horizon, and The processor unit (3) is configured to determine input variables for the electric machine (8) by executing an MPC algorithm (13) according to the first term and according to the second term in order to minimize a cost function.

5. A motor vehicle comprising a processor unit (3), a drive unit, a speed regulator (18) and a driver assistance system (16) according to any one of the preceding claims, wherein: The driver assistance system (16) is configured to: - accessing a torque target value (21) and a speed target value (19) by means of a communication interface, wherein the torque target value (21) and the speed target value (19) are already known by the processor unit (3), - wherein the drive unit (8) is controlled based on the torque target value (21) in a torque-predetermined operating mode, and - wherein the speed regulator (18) is controlled in a speed setting operating mode based on the speed target value (19), The invention is characterized in that the driver assistance system (16) is configured to execute an autonomous driving function of the motor vehicle (1) in the torque setting operating mode or in the speed setting operating mode, depending on the level of the torque target value (21).

6. A method for implementing an autonomous driving function of a motor vehicle (1) by means of a model-based predictive control, the method comprising the following steps: - determining a torque target value (21) within a prediction range by executing an MPC algorithm (13) including a longitudinal dynamics model (14) of the powertrain (7) of the motor vehicle (1), - determining a speed target value (19) within a prediction range (20) by executing an MPC algorithm (13), wherein, in a torque-predetermined operating mode, a drive unit (1) of the drive train (7) is controlled based on the torque target value (21), and wherein, in a speed-predetermined operating mode, a speed regulator (18) of the drive train (7) is controlled based on the speed target value (19), The method is characterized in that it comprises the following steps: executing the autonomous driving function of the motor vehicle (1) in the torque preset operating mode or in the speed preset operating mode according to the level of the torque target value (21), wherein an autonomous driving function is performed in the torque predetermined operating mode by executing an MPC algorithm (13) when a torque target value (21) has a positive value, wherein a speed regulator (18) of the powertrain (7) is controlled based on the speed target value (19), and When the torque target value (21) has a negative value, the autonomous driving function is performed only in the speed predetermined operating mode.

7. A computer program product (11) for executing an autonomous driving function of a motor vehicle (1) by means of a model-based predictive control, wherein: When the computer program product (11) is executed on a processor unit (3), the computer program product instructs the processor unit (3) to execute an MPC algorithm (13) including a longitudinal dynamics model (14) of a powertrain (7) of the motor vehicle (1). - determining a torque target value (21) within the prediction range (20), and - determining a speed target value (19) within said prediction range (20), - wherein, in a torque-predetermined operating mode, the drive unit of the powertrain (7) is controlled based on the torque target value (21), and - wherein in a speed predetermined operating mode a speed regulator (18) of the powertrain (7) is controlled based on the speed target value (19), The invention is characterized in that when the computer program product (11) is executed on the processor unit (3), the computer program product instructs the processor unit (3) to execute an MPC algorithm (13) to perform an autonomous driving function of the motor vehicle (1) in the torque predetermined operating mode or in the speed predetermined operating mode according to the level of the torque target value (21), The processor unit (3) is configured to, by executing an MPC algorithm (13), When the torque target value (21) has a positive value, an autonomous driving function is performed in the torque-predetermined operating mode, wherein a speed regulator (18) of the powertrain (7) is controlled based on the speed target value (19), and When the torque target value (21) has a negative value, the autonomous driving function is performed only in the speed predetermined operating mode.

Images