Situation-dependent parameterization of vehicle actuators

Abstract

The invention relates to a method, in particular a computer-implemented method, for the, in particular at least partial, autonomous driving of a vehicle (F), comprising the method steps: - Receiving sensor data (SD) concerning a vehicle environment of the vehicle (F), - Planning at least one trajectory (T) for the vehicle (F) based on the sensor data (SD), - Determining at least one control value (S) for at least one control variable for at least one vehicle actuator to execute the planned trajectory (T), - Comparing at least one position value (S) with at least one reference value (R), and - In the event that a deviation between the setpoint (S) and the reference value (R) exceeds a predefinable limit value (G), adjustment of at least one characteristic parameter (K) that characterizes driving behavior for controlling the vehicle (K). The invention also relates to a method for supporting a method for at least partially autonomous driving of a vehicle (F), a computer program, and a computer program product.

Description

[0001] The invention relates to a method, in particular a computer-implemented method, for driving a vehicle, in particular at least partially autonomously, a method to support a method for driving a vehicle, in particular at least partially autonomously, a computer program and a computer program product on which the computer program according to the invention is stored.

[0002] Many modern vehicles are equipped with advanced driver assistance systems (ADAS) with a wide range of functions, such as assisted and automated driving, or for displaying information and supporting the driver while driving or parking. Various methods for detecting objects or obstacles on the road, for recognizing lane markings and / or keeping the vehicle in its lane, for detecting rain on the windshield, or for assisting with or carrying out a parking maneuver have become known in connection with driver assistance systems.

[0003] Furthermore, driver assistance systems have become known that can calculate a trajectory based on the current driving scenario, along which the vehicle moves automatically, i.e., without steering input from the driver and / or without the driver operating the accelerator or brake pedal. In this context, different driving maneuvers can be initiated depending on the specific driving scenario, such as reducing speed when approaching a slower vehicle ahead on a multi-lane road. Driver assistance systems with trajectory planners also offer the possibility of calculating trajectories for automated vehicle control based on a selected trajectory for a wide variety of driving scenarios.

[0004] Trajectory planning typically involves several steps: The first step is route planning from a starting point to a desired destination. This route planning can be carried out using an existing road network, similar to modern navigation systems. Appropriate behavior planning then involves planning the abstract behavior of the vehicle, taking into account its surroundings and, if applicable, other road users, static obstacles, lane markings, and traffic signs. Various behavior classes can be defined here, such as "staying in lane," "changing lanes," "turning," or "stopping." Through trajectory planning, the planned abstract behavior is adapted to the current situation and environment, and the planned movement sequence is implemented using trajectory tracking control.

[0005] Another important aspect of achieving fully autonomous driving is the integration of safety mechanisms to prevent accidents, particularly those caused by errors in trajectory planning, such as faulty object detection. The importance of such safety mechanisms increases with vehicle speed because, for example, reaction times decrease. Fundamentally, a safe driving speed depends on many different conditions, such as road conditions, environmental conditions, and traffic conditions. Therefore, knowledge of these conditions is essential for the vehicle to independently determine an optimal speed.

[0006] From JP2015097071A, a device for determining a safe driving speed for different route segments is disclosed. This speed depends on the respective driving situation, road conditions, environmental conditions, and various events.

[0007] Vehicles moving within the respective route segments receive the relevant information and forward it to a central unit where it is processed and then integrated into a map database.

[0008] Document US2021078596A1 proposes defining a reference trajectory and a stop trajectory for moving the vehicle to a safe stopping position. The reference trajectory can, for example, be part of a map database. It also serves to plan a nominal driving trajectory. As long as the vehicle can be moved to the safe stopping position within a given time frame using the stop trajectory, the reference trajectory and the driving trajectory coincide, and the vehicle follows the reference trajectory at a reference speed. However, if obstacles are encountered in the roadway, the driving trajectory is modified relative to the reference trajectory.

[0009] According to DE102021205240A1, the driving behavior is also adapted for different route segments to ensure safe driving. For this purpose, map data regarding restrictions of route segments are determined and integrated using crowdsourcing or swarm sourcing, or based on data from other autonomous vehicles.

[0010] From US2021316758A1, it became known that a map can be used to store the current coefficient of friction, a safe driving speed, a safe acceleration (especially lateral acceleration), and the geometry of each route segment. The system can access this data to adjust the vehicle's speed and other movement parameters to ensure safe passage through each route segment. The next route segment is then determined based on the vehicle's expected trajectory. The vehicle can also update the data stored in the map.

[0011] Based on this, the present invention aims to improve safety during autonomous driving of a vehicle.

[0012] This problem is solved by the method according to claim 1, the method according to claim 11, the computer program according to claim 14 and the computer-readable storage medium according to claim 15.

[0013] According to a first aspect of the present invention, the problem underlying the invention is solved by a method, in particular a computer-implemented method, for driving a vehicle, in particular at least partially autonomously, comprising the method steps: - Receiving sensor data relating to the vehicle's surroundings, - Planning at least one trajectory for the vehicle based on the sensor data, - Determining at least one setpoint for at least one control variable for at least one vehicle actuator to execute the planned trajectory, - Comparing at least one setpoint with at least one reference value, and - in the event that a deviation between the setpoint and the reference value exceeds a predefinable limit, adjustment of at least one characteristic parameter that characterizes driving behavior for controlling the vehicle.

[0014] Various sensors, particularly different environmental sensors, can be used to acquire sensor data. In this respect, various ADAS sensors are especially suitable, for example, one or more cameras, radar sensors, ultrasonic sensors, and / or lidar sensors. The sensor data differs depending on the sensors used—in the case of one or more cameras, for example, the sensor data is image data. A single sensor or several identical or different sensors can be used, whose sensor data can be evaluated independently or in combination.

[0015] The planning of each trajectory can be advantageously carried out using a wide variety of trajectory planning methods, most of which are known from the prior art. One trajectory can be planned, or several trajectories can be planned.

[0016] Within the scope of the present invention, a vehicle actuator is understood to be a drive element for converting a control signal, particularly an electrical one, into an action, for example, a movement or change in the movement of the vehicle. The respective manipulated variable is preferably a quantity related to the action, such as acceleration or a steering angle. The characteristic parameter that defines the vehicle's handling behavior can be the same as the manipulated variable; however, it can also be another quantity relevant to vehicle control, such as another manipulated variable or a parameter that characterizes the vehicle's handling behavior.

[0017] The present invention proposes comparing at least one manipulated variable resulting from trajectory planning with a reference value, particularly one that can be predefined, in order to estimate the safety of the driving process associated with that manipulated variable. If a deviation from the reference value exceeds a predefined limit, the driving behavior is appropriately adjusted as a precautionary measure, for example, by using the reference value for trajectory tracking control, or a modified value, instead of the manipulated variable value. Replanning the trajectory or selecting a different, already planned trajectory would also be conceivable scenarios.

[0018] The present invention is advantageously applicable to various control variables relevant to the execution of autonomous driving processes. Using the method according to the invention, an autonomously moving vehicle is itself able to assess risks for a wide variety of driving maneuvers and / or recognize unusual values ​​for various control variables. Errors that may occur during trajectory planning, but which would otherwise typically remain undetected by the autonomous vehicle, can thus be efficiently corrected or detected in advance. This provides an autonomously operated vehicle with the ability to independently recognize errors or misjudgments and react to them appropriately.

[0019] According to one embodiment, at least two trajectories are planned for the vehicle. In this context, it is conceivable, for example, that the vehicle plans several possible trajectories and, based on a predefined criterion, selects one of the trajectories and uses it to control the vehicle. If the deviation between the manipulated variable and the reference value for the selected trajectory then exceeds the predefined limit, it is also conceivable, among other things, to select a different trajectory and use it to control the vehicle.

[0020] In an advantageous embodiment of the method, the manipulated variable is a steering angle, a speed, or a quantity derived from or related to the steering angle and / or speed. These two manipulated variables are used to control changes in the vehicle's direction and / or speed. Without loss of generality, an example of a quantity derived from speed is acceleration. The steering angle, in turn, is related to, for example, road curvature.

[0021] In a further advantageous embodiment, the characteristic parameter is a speed or a trajectory selected from a multitude of planned trajectories based on at least one manipulated variable. Thus, it is possible, for example, to optimally adjust a vehicle's speed depending on the comparison between the manipulated variable and the reference variable. If several trajectories are planned, the trajectory for which the respective manipulated variable exhibits the smallest deviation from the reference variable can, in turn, be selected.

[0022] The predefined limit for comparing the manipulated variable and the reference value can be set in a variety of ways. It can be fixed or adaptive, changing depending on the situation. Advantageously, in one possible configuration, the predefined limit is selected based on the length of the planned trajectory, a speed, a steering angle, the gradient of a road surface on which the vehicle is located or will be located at a later time according to the planned trajectory, a permissible value range for the manipulated variable or parameter, or based on the vehicle's location and / or a future position of the vehicle according to the planned trajectory.In addition to the aforementioned quantities, other quantities, such as those derived from or related to at least one of the aforementioned quantities, can also be used to determine the predefined limit value. Essentially, the comparison and the resulting adjustment of driving behavior are derived from current road conditions, the road environment, and other safety-relevant considerations.

[0023] It is advantageous for the method according to the invention if information on the weather or time of day is also taken into account, particularly when setting the limit value. Optimal driving behavior is often also dependent on the prevailing weather and / or visibility conditions.

[0024] According to one embodiment, the reference value is a value for at least one control variable, or a derived or related variable, depending on the vehicle's position. Without loss of generality, if the control variable is the steering angle, the curvature of the road surface as a function of location can be used as a limit value. This, in turn, allows for optimal speed adjustment to varying road conditions and curvatures. The greater the curvature of the road surface, the lower the vehicle speed should be.

[0025] According to a further advantageous embodiment, a multitude of reference values ​​are stored in a storage unit within the vehicle, or a multitude of reference values ​​are stored in an external storage unit to which the vehicle can access. These reference values ​​are, in particular, different values ​​for the reference value as a function of location, time, and / or values ​​depending on various weather conditions. The reference value can therefore be selected situation-specifically and used for comparison. A map or database containing the reference values ​​can be provided either locally in the vehicle or centrally in an external unit.

[0026] In this respect, it is advantageous if, in the event that the large number of reference values ​​is stored in an external storage unit, the vehicle can transmit the control signal to this unit. This allows the storage unit to continuously update and / or expand the map or database of reference values.

[0027] According to a further advantageous embodiment of the inventive method for the, in particular at least partially, autonomous driving of a vehicle, the reference value is determined using a road map, and / or the reference value is determined using one or more determined or measured values ​​for a control variable or characteristic value; in particular, an average value for the reference variable is calculated. In addition to optimal values ​​for the reference values, for example average values, minimum values, maximum values, permissible value ranges, or similar values ​​for the reference values ​​can also be determined.

[0028] Reference values ​​can be determined a priori, for example, theoretically using a road map. Road curvatures, complex road layouts (e.g., due to complex intersections or traffic patterns), and road gradients can be identified using road maps and used to determine suitable reference values, particularly as a function of location. Additionally or alternatively, one or more vehicles can provide data for various parameters, from which reference values ​​can then be derived. The provided values ​​can be pre-sorted, particularly based on location, time, and / or weather conditions, to obtain different reference values. This also allows for the continuous updating and refinement of the reference values ​​stored in a map or database.Such continuous updates can be based on a priori reference values, or reference values ​​can be empirically determined using data from vehicles, especially new ones.

[0029] The problem underlying the invention is further solved by a method for supporting a method for the, in particular at least partially, autonomous driving of a vehicle, especially a method according to one of the embodiments described above. The method for supporting a method for the, in particular at least partially, autonomous driving of a vehicle comprises the following method steps: - Determining at least one value for at least one control variable for at least one vehicle actuator, at least as a function of the vehicle's position, and - Generating and / or providing a map or database comprising at least one reference value determined as a function of location based on the value for at least one control variable.

[0030] The map or database can be stored, for example, in an external storage unit, such as on a server or in the cloud, which various vehicles can access during continuous operation to retrieve reference values ​​and / or transmit values ​​for the control variable. The map or database can be made available to a large number of vehicles for carrying out the inventive method for at least partially autonomous vehicle operation, for example, as a service. This enables a comprehensive evaluation of various traffic situations by determining and providing suitable values ​​for control variables, from which different driving processes can be derived, for example, by adjusting at least one characteristic parameter that defines the vehicle's driving behavior, for the various traffic situations.The reference values ​​can then be used by the vehicles to compare them with values ​​for these control variables determined through internal trajectory planning. In this way, autonomous vehicles, in particular, can at least partially verify their trajectory planning and correct their driving behavior in the event of detected errors in the trajectory planning based on the comparison with the reference values, thus preventing accidents.

[0031] In this respect, it is advantageous if at least one reference value is determined using a road map and / or determined or measured values ​​for the control variable; in particular, a large number of situation-specific reference values ​​are determined. Thus, reference values ​​can be determined both a priori or theoretically using road map information and empirically using determined or measured values ​​for the control variables. Furthermore, different reference values ​​can be determined situation-specifically, for example, for different weather conditions and / or times of day, or reference values ​​can be determined for specific classes of driving maneuvers, such as evasive maneuvers in the event of obstacles on the road. Then, depending on the specific situation, the appropriate reference value can be selected from the map or database.

[0032] Furthermore, it is advantageous if at least one reference value is continuously updated. Through the continuous updating of the map or database, optimal reference values ​​can be practically determined over time, and optimal driving strategies can be derived from them, or faulty trajectory planning can be reliably detected.

[0033] The problem underlying the invention is further solved by a computer program with instructions which, when the computer program is executed by a computer, cause the computer to execute the inventive method for, at least partially, autonomously driving a vehicle, and / or the method for supporting a method for, in particular, at least partially, autonomously driving a vehicle.

[0034] The problem underlying the invention is also solved by a computer program product on which the computer program according to the invention is stored.

[0035] The invention and some of its advantageous embodiments are explained in more detail with reference to the following figures. They show: Fig. 1 a schematic flowchart for carrying out the method according to the invention for the, at least partially, autonomous driving of a vehicle, Fig. 2 a schematic flowchart for carrying out the method according to the invention for supporting a method for the, at least partially, autonomous driving of a vehicle, and Fig. 3 Two examples of driving an autonomous vehicle using the method according to the invention.

[0036] In the figures, identical elements are each given the same reference symbol.

[0037] At the in Fig. In the schematically represented flowchart for the inventive method for the at least partially autonomous driving of a vehicle, a trajectory T is planned based on received sensor data SD concerning a vehicle environment U. Several trajectories T can also be planned. Subsequently, at least one control value S for at least one control variable for at least one vehicle actuator is determined for the execution of the planned trajectory T. For example, the control value S can be a result of the trajectory planning T or can be calculated based on the planned trajectory T.

[0038] The control value S is compared with a reference value R, and a deviation between the control value S and the reference value SR is determined. If the deviation SR is greater than a predefined limit value G, the driving behavior is adjusted by modifying a characteristic parameter K that characterizes the driving behavior for controlling the vehicle (K modIf, on the other hand, the deviation SR is smaller than the predefined limit value G, no adjustment is necessary. Instead, the original parameter K can be used. in The parameter K can be used to guide the vehicle. The parameter K can be identical to or related to the manipulated variable S. In particular, the parameter K can be the result of a renewed trajectory planning process, taking into account the result of the comparison or an adjusted value for the manipulated variable S.

[0039] The reference value R is, for example, a reference value as a function of the location R(x), which can be received via an external storage unit, such as a server or a cloud. A multitude of reference values ​​as a function of the location R(x) can be stored in the storage unit, so that a suitable reference value R(x) can be retrieved depending on the current or predicted future position of the vehicle. Alternatively, the reference value(s) R can also be stored in a local storage unit within the vehicle. It is also possible for control variables S determined by the vehicle or multiple vehicles to be transmitted to the external or local storage unit and used to determine and / or update reference values ​​R.

[0040] If the control variable is the steering angle α, then, for example, a curvature k can be used as the reference value R and a vehicle speed v as the characteristic value K. The greater the agreement between the steering angle α obtained from trajectory planning and the curvature k, the higher the vehicle speed v can be chosen to guide the vehicle in the given situation.

[0041] Alternatively or additionally, if the control variable is the steering angle α, the steering angle α can also be used as the reference value R. The steering angle α and the curvature k are related to each other.

[0042] In particular, if a reference value R is in the form of the steering angle α, this can be determined, for example, from a large number of steering angle α values ​​from different vehicles that have previously experienced a similar driving situation at the same location. The different steering angle α values ​​can then be transmitted from the various vehicles to a central storage unit, where a suitable reference value R for the steering angle α is determined. In this way, the driving behavior can be optimally adapted to the prevailing real-world conditions. If, for example, many vehicles are driving relatively slowly in the given situation, or slower than the curvature k of the road at that location would allow, this could be due to a temporary obstacle on the road that necessitates a reduction in speed v.In the event that the obstacle is not reliably detected and taken into account during trajectory planning, the driving behavior can still be appropriately adapted based on the comparison according to the invention with the continuously updated reference value R.

[0043] If, on the other hand, the manipulated variable S is, for example, a vehicle speed v, then a vehicle speed v can also be used as the reference value R. The starting point for the reference values ​​as a function of the position R(x) can then, in turn, be, for example, permissible maximum speeds v. max , which can be determined in particular from traffic signs. These values ​​can also be continuously updated and adjusted based on the vehicle speed v of a large number of vehicles. The theoretically permissible maximum speed v does not always represent the actual speed. maxThis represents an optimal and safe vehicle speed v. Rather, a wide variety of factors can lead to a vehicle speed v being lower than the permissible maximum speed v. max more suitable or safer, for example due to high traffic volume, a large number of pedestrians in the vehicle's vicinity, poor weather and / or visibility conditions, or due to, in particular temporary, obstacles on the roadway, such as objects or other vehicles on the roadway.

[0044] Alternatively or additionally, such considerations could also be derived from an analysis of a vehicle acceleration a or a jerk j, i.e., an instantaneous change in acceleration over time. The vehicle acceleration a or the jerk j can also be compared with reference values ​​R for this quantity at other locations with comparable road profiles.

[0045] An advantage of the present invention is that the reference values ​​R are vehicle-independent. Therefore, they can be made available to a large number of vehicles for carrying out, in particular at least partially, autonomous driving operations. Accordingly, the present invention further relates to a method for supporting a method for at least partially autonomous driving of a vehicle, in particular the method according to the invention described above.

[0046] The procedure for supporting a procedure for at least partially autonomous driving of a vehicle is schematically described in Fig. Figure 2 illustrates this. First, at least one value for at least one manipulated variable S for at least one vehicle actuator is determined, at least as a function of the vehicle's position S(x). These values ​​can be measured or calculated, for example. Based on the manipulated variable S(x) values, reference values ​​are determined as a function of the location R(x), and a map or database is created, provided, and / or updated containing these reference values ​​R as a function of the location x and, if applicable, as a function of specific driving situations, weather conditions, environmental conditions, and / or visibility conditions. The reference values ​​R can be retrieved by vehicles F and used for comparison as shown in Figure 2. Fig. 1 described, can be used. In addition, different vehicles F can transmit self-determined values ​​for the respective control variable S to the central storage unit (represented here in the form of a cloud), which are used for the continuous determination and / or updating of the reference values ​​R.

[0047] In Fig. Figure 3 illustrates two examples of the autonomous driving of a vehicle F according to the invention. In the case of the Fig. In the example shown in Figure 3a, vehicle F travels along a straight road according to a previously planned trajectory T. The planned trajectory T involves a change in the control value S of the steering angle α, which is determined according to a reference trajectory T. ref with a reference steering angle R=α ref This is not provided for. By comparing the steering angle α with the reference value R, the steering angle α and consequently the trajectory T used to guide the vehicle F can be determined. rescan be adjusted. In the example shown, for instance, the steering angle α remains unchanged.

[0048] At the in Fig. In the illustrated example 3b, a preceding obstacle is to be avoided. This could be, for example, an object on the road towards which the vehicle F moves according to a previously planned trajectory T. The planned trajectory T involves an abrupt change in the control value S of the steering angle α, which is determined according to a reference trajectory T. ref with a reference steering angle R=α ref This is not provided for. Once again, the steering angle α and, consequently, the trajectory T used to guide the vehicle F are determined by comparing the steering angle α with the reference value R. res adapted - here, for example, a change in the steering angle α occurs more slowly in order to reduce the curvature of the traversed trajectory T. QUOTES INCLUDED IN THE DESCRIPTION

[0000] This list of documents cited by the applicant was automatically generated and is included solely for the reader's convenience. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions. Cited patent literature

[0000] JP 2015097071A

[0006] US 2021078596A1

[0008] DE 102021205240A1

[0009] US 2021316758A1

[0010]

Claims

[1] Method, in particular computer-implemented method, for, in particular at least partially, autonomous driving of a vehicle (F), comprising the method steps: - Receiving sensor data (SD) concerning a vehicle environment of the vehicle (F), - Planning at least one trajectory (T) for the vehicle (F) based on the sensor data (SD), - determining at least one control value (S) for at least one control variable for at least one vehicle actuator for executing the planned trajectory (T), - comparing the at least one control value (S) with at least one reference value (R), and - in the event that a deviation between the control value (S) and the reference value (R) exceeds a predeterminable limit value (G), adapting at least one parameter (K) characterizing a driving behavior for driving the vehicle (K). [2] Method according to claim 1, wherein at least two trajectories (T) are planned for the vehicle (F). [3] Method according to claim 1 or 2, wherein the manipulated variable (S) is a steering angle (α), a speed (v) or a variable derived at least from the steering angle (α) and / or the speed (v) or related to at least the steering angle (α) and / or the speed (v). [4] Method according to one of the preceding claims, wherein the characteristic variable (K) is a speed (v) or a trajectory selected from a plurality of planned trajectories (T) on the basis of at least one manipulated variable (S). [5] Method according to one of the preceding claims, wherein the predeterminable limit value (G) is selected as a function of a length of the planned trajectory (T), a speed (v), a steering angle, a gradient of a roadway on which the vehicle (F) is located or will be located according to the planned trajectory (T), a permissible value interval for the manipulated variable (S) or characteristic variable (K), or based on a location (x) of the vehicle (F) and / or a future position of the vehicle (F) according to the planned trajectory (T). [6] Method according to one of the preceding claims, wherein information on the weather or the time of day is taken into account, in particular when specifying the limit value. [7] Method according to one of the preceding claims, wherein the reference value (R) is a value for the at least one manipulated variable (S) or a variable derived therefrom or associated therewith as a function of a position (x) of the vehicle (F). [8] Method according to one of the preceding claims, wherein a plurality of reference values ​​(R) is or are stored in a memory unit in the vehicle or wherein a plurality of reference values ​​(R) is or are stored in an external memory unit, which external memory unit can be accessed by the vehicle (F). [9] Method according to claim 8, wherein the plurality of reference values ​​(R) is or is stored in an external storage unit and wherein the vehicle (F) transmits the manipulated variable (S) to the external storage unit. [10] Method according to one of the preceding claims, wherein the reference value (R) is determined on the basis of a road map, and / or wherein the reference value (R) is determined on the basis of one or more determined or measured values ​​for a manipulated variable (S) or characteristic variable (K), in particular wherein an average value for the reference value (R) is calculated. [11] Method for supporting a method for at least partially autonomous driving of a vehicle (F), in particular the method according to at least one of the preceding claims, comprising the method steps: - determining at least one value for at least one manipulated variable (S(x)) for at least one vehicle actuator at least as a function of a position (x) of the vehicle (F), and - Generating and / or providing a map or database comprising at least one reference value (R(x)) determined on the basis of the value for the at least one manipulated variable (S) as a function of the location. [12] Method according to claim 11, wherein the at least one reference value (R(x)) is determined on the basis of a road map and / or on the basis of determined or measured values ​​for the manipulated variable (S(x)), in particular a plurality of situation-specific reference values ​​(R(x)) are determined. [13] Method according to claim 11 or 12, wherein the at least one reference value (R(x)) is continuously updated. [14] A computer program comprising instructions which, when the computer program is executed by a computer, cause the computer to carry out the method according to any one of claims 1-10 and / or 11-13. [15] A computer program product on which a computer program according to claim 14 is stored.

Images