METHOD FOR DETERMINING A CHANGE IN BRAKE PRESSURE

Abstract

A method is proposed for determining a brake pressure change for a vehicle wheel to optimize a braking process, which has the following steps: providing a current wheel state, wherein the wheel state has a plurality of state parameters; determining at least one state parameter whose value deviates from the target wheel state; determining a direction of change of the brake pressure change depending on a deviation of at least one state parameter from the target wheel state; providing a brake pressure characteristic field to determine a value of the brake pressure change, wherein the brake pressure characteristic field assigns a brake pressure change to most of the state parameters and is specific to the determined direction of change of the brake pressure change;Determine a value for the change in braking pressure with the current wheel state and the provided braking pressure characteristic field.

Description

METHOD FOR DETERMINING A CHANGE IN BRAKE PRESSURE FIELD OF INVENTION The invention relates to a method for determining a brake pressure change for a vehicle wheel in order to optimize a braking method. BACKGROUND OF THE INVENTION If a maximum coefficient of friction between a vehicle wheel and the road surface is exceeded—for example, during water skiing or in winter conditions such as heavy rain, snow, or ice—there is a risk of unstable driving due to a loss of traction. For many driver assistance systems, as well as for partially automated vehicles, it is important not to exceed the maximum coefficient of friction to ensure safe driving at all times or, if necessary, to disable the automated driving function. Modern motor vehicles have control systems, such as Electronic Stability Program (ESP). In this context, Electronic Stability Program essentially acts as a slip control system. When critical driving situations arise, a safety system intervenes, such as an Anti-lock Braking System (ABS) or a Traction Control System (TCS). These systems are based on anti-lock braking control (ABC), which involves applying, releasing, and maintaining brake pressure to counteract wheel lockup and shorten braking distances. Current ABC systems have numerous parameters, allowing application engineers to achieve optimal performance for different vehicles. However, determining the optimal values ​​for these parameters is very expensive. This is because the application engineer must perform different ALC maneuvers, activate the ALC with full braking, evaluate the measurement, and assess which parameters should be adapted from the many parameters to improve performance, and repeat this many times until the target performance is achieved. BRIEF DESCRIPTION OF THE INVENTION The object of the invention is to simplify the search for parameters for adapting an anti-lock braking system (ABS). An ABS must independently learn the optimal brake pressure adjustment to hold a wheel of a stable vehicle while simultaneously achieving the shortest possible braking distance. According to aspects of the invention, a method is proposed for determining a change in braking pressure for a wheel of a vehicle to optimize a process of PQLCLn / zznz / e / γΐΛΐ braking, a method for determining a characteristic braking pressure field, uses of the method, a method for controlling a device, a computer program, and a machine-readable storage medium according to the features of the independent claims, which at least partially address the aforementioned objects. Advantageous configurations are the subject of the dependent claims, as well as the following description. Throughout this description of the invention, the sequence of method steps is presented in such a way that the method is readily understandable. However, a person skilled in the art will realize that many of the method steps can also be performed in a different sequence and lead to the same or a corresponding result. In this respect, the order of the method steps can be modified accordingly. According to one aspect, a method is proposed to determine a change in braking pressure for a wheel of a vehicle to optimize a braking method that has the following steps: In one step of the method, a current wheel state is provided, where the wheel state has a plurality of state parameters. In another step, at least one state parameter is determined whose value deviates from the target wheel state. In another step, a direction of change of the brake pressure is determined based on a deviation of at least one state parameter from the target wheel state. In another step, a brake pressure characteristic field is provided to determine a value for the brake pressure change, where the brake pressure characteristic field assigns a brake pressure change to most of the state parameters and is specific to the determined direction of change of the brake pressure and the state parameter.In another step, a value for the change in braking pressure is determined with the current wheel state and the provided braking pressure characteristic field. In this case, the plurality of state parameters, such as wheel slip or acceleration, is determined by signals generated by vehicle sensors, such as inertial or speed sensors. That is, the wheel state can be, for example, a function of wheel slip and acceleration. Wheel state - f(slip, aWheel) The braking torque of a wheel can be varied by adjusting the brake pressure, which can be determined by a cumulative effect of brake pressure changes. In this case, the change in brake pressure can be a function of the wheel's condition. PQLCLn / zznz / e / YiAi In this case, the term deviation should be understood broadly in the context that a direction of change of the brake pressure change is determined depending on a deviation of at least one state parameter with respect to the target wheel state and comprises in particular both a quantitative deviation in the sense of a distance as well as a time-varying deviation that corresponds to a gradient. As explained later, the brake pressure characteristic field can be further optimized using reinforcement learning by correcting the brake pressure characteristic field values ​​as the vehicle passes through different wheel states. These different wheel states can be set by an application engineer for a given vehicle during a specific vehicle operation. In this operation, a reinforcement learning agent learns, according to reward rules (policy), the best actions in response to a change in brake pressure. These optimal actions can be stored in a corresponding brake pressure characteristic field and then used to adjust a brake pressure change. In this way, it is advantageous for this method to determine a change in braking pressure, a parameter search method that is easily performed for the adaptation of an anti-lock controller. The method for determining a brake pressure change is so simple that an applications engineer can manually make changes to the control method, for example, by changing the values ​​in a brake pressure characteristic field. A current wheel state can also be characterized by other state parameters, as listed below. The direction of the change in braking pressure, i.e., whether the braking pressure should be increased or decreased due to the change in braking pressure, can be determined, for example, by a mode decision module, such that a wheel state corresponding to a target wheel state is reached or maintained as quickly as possible. In this case, a target wheel state of this type can be established for each axle of the vehicle, and this target wheel state can be determined depending on state parameters, such as, for example, minimum and maximum slippage and / or minimum or maximum wheel acceleration, such that the performance of the entire braking system is optimized. This target wheel state can also be set dynamically depending on driving conditions, turning radius, or surface, etc. To determine this direction of change, you can select, for example, one of the state parameters from the majority of the state parameters. This selection can PQLCLn / zznz / e / γΐΛΐ can be done based on predetermined rules. For example, the sliding of the respective wheel can be used for this purpose. Typically, with an increase in braking pressure, wheel slippage increases and the wheel slows down even more, that is, there is a negative acceleration for the wheel. A decrease in braking pressure usually results in a reduction of slippage and less braking of the wheel, i.e., an acceleration of the wheel. For this method, four characteristic fields can be generated, for example, differentiated by the relative change in two state parameters, such as wheel slip and acceleration—that is, an increase or decrease in the state parameter. In this case, for a positive change in braking pressure (i.e., an increase in braking pressure), the two characteristic fields related to wheel acceleration can be used. For a negative change in braking pressure, the two characteristic fields corresponding to the slip state parameter can be used. Before using the amplification learning method described below, the corresponding characteristic fields can be calculated using simulation values.The respective characteristic fields can directly assign a brake pressure change according to the respective state parameters or specify a factor to determine the brake pressure change through a calculation. In this case, the respective factor can be multiplied by this factor in the event of an increase in the respective state parameters and divided by this factor in the event of a decrease in the respective state parameters. Therefore, for a currently increasing slip, i.e., if a previous slip is smaller than the current slip, the result is: dpobjetivo = - Ksizu * Sliding value and for a currently decreasing sliding: dpobjet¡vo= - KsiAb / Slip value Therefore, for a currently decreasing wheel acceleration: dptarget = Ka* Wheel acceleration and for a corresponding currently decreasing wheel acceleration: dptarget—KaAb / Wheel acceleration In this case, the parameters Ksizu or Ksiaó represent an increase or decrease in the slip value and Kazuo KaAb an increase or decrease in wheel acceleration for the respective change in braking pressure dpDianaPQLCLn / zznz / e / γΐΛΐ This determination of the respective change in braking pressure depending on the current change in the state parameters results in a different aggressive control depending on the respective direction of the modification of the state parameters. This means there is no proportional control for a decreasing wheel slip or acceleration value. In contrast, when wheel slip or acceleration values ​​increase, proportional control does occur. Advantageously, the method can be used for vehicles equipped with anti-lock braking systems (ABS). It can also be applied to all functions that utilize wheel controllers, such as traction control. In this case, the target wheel state can be adjusted accordingly, enabling the modified task of making the wheel, if necessary, faster than a vehicle reference speed. According to one aspect, it is proposed that at least one previous wheel state be provided and that the characteristic field of braking pressure be specific to the direction of change determined by the change in braking pressure and a change in at least one state parameter. According to one aspect, it is proposed that at least one previous wheel state be provided and that the at least one provided braking pressure characteristic field depend on a direction of change of at least one state parameter whose value deviates from the target wheel state. This target wheel state can be set with maximum and minimum values ​​for the respective state parameters for each axle—the front and rear axles of the vehicle—to achieve the most efficient braking system possible. For example, one such target wheel state can be defined by the following ranges of values ​​for the wheel slip and acceleration state parameters: Slip_min - 5%; slip_max = 10%; wheel acceleration_min = -20 m / s²; wheel acceleration_max = -15 m / s². If a target wheel state is dynamically selected, a vehicle reference speed can be supported, if required, by selecting a less than zero maximum slip for an axle for a specified time. In other words, the wheel must attempt to be faster than the vehicle reference speed to make it plausible and, if necessary, correct for it. Furthermore, when driving a vehicle on a sandy surface, a higher slip on the front axle may be desired, allowing the following values ​​to be determined, for example, for the target slip value: Min_slip - 25%; and max_slip - 35%. PQLCLn / zznz / e / γΐΛΐ According to one approach, it is proposed that a gradient be determined for at least one state parameter using at least the previous and current values ​​of that parameter. The direction of change in the braking pressure is then further determined using this gradient. Advantageously, the determined gradient allows for the calculation of a corresponding brake pressure change, enabling the brake pressure to be adjusted as early as possible to bring the wheel state to the target state as quickly as possible and maintain it there. In this case, other previous state parameters can also be used for gradient determination; that is, a time sequence of at least one state parameter can be used for gradient determination and / or several state parameters, as listed below, can be used for gradient determination. This determination of the gradient allows taking into account an idle time to determine the direction of the change in braking pressure, which makes it necessary to determine the changes in braking pressure in advance, since an influence of the change in braking pressure, due to the idle time of, for example, 30 ms and a determination of the wheel state in an interval of, for example, 5 ms, has no effect until later. According to one aspect, it is proposed that the determination of the direction of change of the brake pressure change be determined by means of a plurality of state parameters. Examples of these state parameters are wheel slip and / or wheel acceleration and / or slip gradient and / or wheel acceleration (aWheel) and / or wheel pull (pull (wheel)) and / or wheel acceleration relative to vehicle acceleration (aWheel relative to aVehicle). According to one aspect, it is proposed that the determination of the direction of change of the brake pressure change be determined depending on an idle time of a complete system for a brake pressure change. According to one aspect, it is proposed that at least one state parameter be determined from the plurality of state parameters according to a prioritization sequence. For example, the slip state parameter value may have a higher priority for determining the specific state parameter that deviates from the target wheel state, since wheel lock-up must be avoided, and for this purpose, the braking pressure can be reduced until the respective wheel has a slip that is less than the desired slip with the target wheel state. PQLCLn / zznz / e / γΐΛΐ According to one aspect, it is proposed that most state parameters present a wheel slip and / or a wheel acceleration and / or a slip gradient and / or a wheel acceleration and / or a wheel pull and / or a wheel acceleration relative to the vehicle acceleration. A method is proposed for determining a characteristic braking pressure field for the method described above, which has the following steps: In one step, a current wheel state is provided, where the wheel state has a plurality of state parameters. In another step, reward rules are established for a reinforcement learning method. In a further step, a reward is determined by means of the reward rules and the current wheel state, and if a reward has been determined for the reinforcement learning method, both the last pressure change made with respect to the value and direction of change and the associated brake pressure characteristic field are determined, as well as a correction value for the associated brake pressure characteristic field according to the reinforcement learning method. This reward rule specifies what action, such as, for example, changing the brake pressure, should be performed on any variant of behavior or observation, such as the deviation of the wheel state from the target wheel state, of the learning environment (environment), to maximize the reward (reward) or minimize a punishment. An agent implementing this reinforcement learning method can modify the pressure change defined by the characteristic braking pressure fields, depending on the respective state parameters, until, for example, the slip state parameter is always below a predefined threshold. This modification can be made by a fixed or random percentage change. Advantageously, the method for determining a characteristic braking pressure field can be used to adapt characteristic fields such that optimal performance is achieved with a braking system controlled in this way. Reward rules are defined so that a target performance can be achieved. If a reward is determined according to the reward rules, which may also contain the appropriate punishment rules, the brake pressure change or corresponding parameter, which occurs in a certain wheel state from which the reward or punishment follows, will be modified. A reinforcement learning agent modifies the brake pressure parameter or change by means of a random or percentage change until the brake pressure change value in the corresponding brake pressure characteristic field no longer allows for any further reward to be determined for the respective wheel state. This means PQLCLn / zznz / e / YiAi indicates that the wheel state can be within the target wheel state. For example, a state parameter might be: Slip is always within a maximum and minimum value. Furthermore, for example, pressure changes in a certain wheel state that lead to a high slip gradient can be penalized with the reward rules. Advantageously, in this method, the agent, which implements reinforcement learning, optimizes the control method by adapting the respective characteristic fields of braking pressure. The application engineer then determines the environment, i.e., different wheel states based on corresponding driving behavior. Consequently, the desired simplification of adapting a braking control system to the respective vehicle is achieved. According to one aspect, it is proposed that in the method for determining a characteristic braking pressure field, at least one previous wheel state of the wheel is provided and the reward is determined with the current wheel state and / or at least one previous wheel state. According to one aspect, it is proposed that the reward rules determine a reward depending on a value below a threshold value for slip and / or a value less than a slip value below zero and / or a modulation frequency of a pressure change. It is proposed to use the method described above to determine a change in braking pressure to control the braking pressure on a wheel. The method is proposed for determining a characteristic braking pressure field to optimize the performance of a brake pressure control for a vehicle wheel. A method is proposed in which, based on a determined change in braking pressure, a control signal is provided for the control of a at least partially automated vehicle; and / or based on the determined change in braking pressure, a warning signal is provided to warn a passenger of the vehicle. In this way, for example, a vehicle control unit can perform traction-slip control (TCS: traction control system), in which, for example, a wheel that has a temporarily frozen surface is braked or, in a braking process, all wheels are braked in such a way that a braking distance becomes minimal under certain circumstances. The term “based on” shall be understood broadly in relation to the feature of providing a control signal based on a determined brake pressure change. The brake pressure change shall be understood to apply to any determination or calculation. PQLCLn / zznz / e / γΐΛΐ of a control signal, where this does not preclude the use of other input variables for this determination of the control signal. The same applies to the provision of the warning signal. A braking system is proposed that is configured to perform one of the methods described above to determine a change in braking pressure. With one of these devices, the corresponding method can be easily integrated into different systems. According to another aspect, a computer program is specified that includes commands which, when executed by a computer, cause it to perform one of the methods described above. This computer program allows the use of the described method on different systems. A machine-readable storage medium is proposed, in which the computer program described above is stored. The term "vehicle" as used here can also be understood to mean a mobile platform, which may be a system that is at least partially automated, mobile, and / or a driver assistance system. An example might be a vehicle that is at least partially automated or a vehicle with a driver assistance system. That is, in this context, a system that is at least partially automated implies a mobile platform with respect to at least partially automated functionality, but a mobile platform also includes vehicles and other mobile machines, including driver assistance systems. Other examples of mobile platforms include driver assistance systems with multiple sensors, mobile multi-sensor robots such as robotic vacuum cleaners or lawnmowers, a multi-sensor surveillance system, a ship, an aircraft, a manufacturing machine, a personal assistant, or an access control system.Each of these systems can be a fully or partially autonomous system. BRIEF DESCRIPTION OF THE FIGURES Example embodiments of the invention are shown with reference to Figures 1 to 9 and are explained in more detail below. Example: Figure 1 shows a path of a slip value and the corresponding wheel speed over time; Figure 2 is a braking system; Figure 3 is a state diagram with a sequence of state values; Figure 4 is a diagram with state parameters; Figure 5 is a characteristic field with a matrix for changing the characteristic field values; Figure 6 is a representation of changed characteristic field values; PQLCLn / zznz / e / YiAi Figure 7 represents a diagram with a state parameter over time; Figure 8 shows an identification plate with a list of changes in the characteristic field values; and Figure 9 is a diagram with a plurality of state parameters over time. DETAILED DESCRIPTION OF THE INVENTION Figure 1 shows, as an example, in a diagram 100, a time development of a slip value 120 and limits Smax, Smin within which the slip value S must be found. In this case, curve 160 shows a reference speed vy and curve 140 shows the speed of the wheel under consideration. Figure 2 shows a 200 braking system that is configured to perform the method for determining the change in braking pressure to optimize a braking method. A current wheel state 210 is provided by means of wheel sensors and other vehicle sensors, wherein the wheel state 210 has a plurality of state parameters, such as, for example, wheel slip s and wheel acceleration a. In the braking system module 220, at least one state parameter is determined whose value deviates from the target wheel state, and a direction of change of the braking pressure is determined depending on a deviation of at least one state parameter from the target wheel state. Alternatively, the braking pressure can also be kept constant.With this determined change direction (pj or pf), a brake pressure characteristic field (250a, 250b, 251a, 251b) is selected to determine a brake pressure change value. The brake pressure characteristic field is assigned to a brake pressure change based on the most frequent slip state parameters, or the wheel acceleration to D, repeated for braking. This change is specific to the determined direction of the brake pressure change and the state parameter change. In other words, the brake pressure characteristic fields are selected depending on the direction of the state parameter modification. Therefore, there is a brake pressure characteristic field (250b), which is specific to an increase in wheel acceleration, and a brake pressure characteristic field (250a), which is specific to a decrease in wheel acceleration.Therefore, for the slip state parameter, a brake pressure characteristic field 251a is applied for increased slip and a brake pressure characteristic field 251b for decreased slip. In this case, brake pressure characteristic fields 250a and 250b are associated with a positive change in brake pressure, and brake pressure characteristic fields 251a and 251b are associated with a negative change in brake pressure. Consequently, the value of the brake pressure change Dp can be determined with the current wheel state. PQLCLn / zznz / e / γΐΛΐ the corresponding provided braking pressure characteristic field and in particular be transmitted to the braking system for the wheel. The values ​​of the state parameters and pressure changes are transmitted to module 240, so that the corresponding braking pressure characteristic fields can be modified using the reinforcement learning method described above. Figure 3 schematically shows a diagram 300 with a plurality of combinations of two state parameters: wheel slip and acceleration, and summarizes a sequence of state values. In this case, the target wheel state 310 for the state fields is indicated with a thick frame. If the method starts with a wheel state arranged in the upper right of this diagram 300, a sequence of state fields can be marked using the doubly shaded fields 314. In each case, the respective state is determined within a predetermined time interval of, for example, 5 ms. In this particular case, due to an idle time of, for example, 30 ms, a positive brake pressure change occurs, without the target wheel state 310 being directly reached. The fields marked in black 310 indicate state fields where a negative brake pressure change occurs. Changes in brake pressure affect the change in wheel state in diagram 300 as follows: Slide (X-axis) to Wheel (Y-axis) Mounting Pressure Greater Lesser Dismounting Pressure Smaller Greater Figure 4 schematically shows a diagram 400 with a time sequence of slip values ​​s 430, which are, for example, in interval 440 outside of an upper and lower limit value. In this regard, curve 420 indicates the corresponding state value S, and curve 410 shows the development of the resulting braking pressure p according to the accumulated braking pressure changes and delineates an interval 450. The braking pressure changes prior to the time zone of interval 440 can be assumed to be the cause of exceeding the slip value 430 in interval 450. Therefore, with reinforcement learning, the corresponding braking pressure characteristic field for state 420 can be adapted to achieve control where the slip is kept within established limits. Figure 5 outlines a characteristic field 520 for an increasing pressure change in which the state parameter decreases for wheel acceleration, or increases for braking. PQLCLn / zznz / e / YiAi Figure 6 shows a representation of changed characteristic field values ​​that depend on the state parameters: Wheel acceleration a and slip specify how the braking pressure p should change depending on the state parameters. Figure 7 shows a diagram 700 with the slip state parameter 720 plotted over time. It can be seen that in interval 725, the slip s is below a predetermined minimum slip value. Furthermore, this diagram 700 depicts the time evolution 715 of the braking pressure p. The reinforcement learning method can identify this latter change in braking pressure and modify the corresponding braking pressure characteristic field accordingly, to prevent future slip below the relevant wheel state. Figure 8 shows a characteristic field that is associated with a positive change in braking pressure in the case of decreasing wheel acceleration or increasing braking and is updated by means of a list for the change of the characteristic field values ​​depending on the wheel state, where the list was determined using the reinforcement learning method. Figure 9 shows a diagram with a plurality of parameters and state parameters over time: Slip 946, rapid pressure increase 944, slow pressure increase 945, constant retained pressure 943, rapid pressure decrease 941, slow pressure decrease 942, state index 947, target braking pressure 948. In interval 930, which indicates a rapid sequence of changes in slip, this change can be attributed to previous changes in target braking pressure on curve 948 with the wheel state marked in interval 920. Since a high frequency of displacement changes should be avoided, a frequency modulation of pressure change can be integrated into the reward rules of the reinforcement learning method. PQLCLn / zznz / e / γΐΛΐ NOVELTY OF THE INVENTION Having described the present invention, it is considered a novelty and, therefore, the contents of the following are claimed as property.

Claims

1. A method for determining a brake pressure change for a vehicle wheel to optimize a braking method, characterized by the steps: providing a current wheel state (210) of the wheel, wherein the wheel state (210) has a plurality of state parameters; determining at least one state parameter whose value deviates from the target wheel state (310); determining a direction of change of the brake pressure change depending on a deviation of the at least one state parameter with respect to the target wheel state.provide a brake pressure characteristic field (250a, 250b, 251a, 251b) to determine a brake pressure change value, wherein the brake pressure characteristic field (250a, 250b, 251a, 251b) assigns a brake pressure change to most state parameters and is specific to the determined direction of change of the brake pressure change; determine a brake pressure change value with the current wheel state (210) and the provided brake pressure characteristic field (250a, 250b, 251a, 251b).

2. The method according to claim 1, characterized in that at least one previous wheel state is provided and the at least one braking pressure characteristic field (250a, 250b, 251a, 251b) provided depends on a direction of change of the at least one state parameter whose value deviates from the target wheel state (310).

3. The method according to claim 2, characterized in that a gradient of at least one state parameter is determined using at least the previous and current values ​​of the state parameter; and the determination of the direction of change of the brake pressure change is further determined by means of the gradient.

4. The method according to one of the preceding claims, characterized in that the determination of the direction of change of the brake pressure change is determined by means of a plurality of state parameters.

5. The method according to any one of the preceding claims, characterized in that the determination of the direction of change of the brake pressure change is determined depending on the idle time of the entire system for a brake pressure change. PQLCLn / zznz / e / γΐΛΐ 6. The method according to one of the preceding claims, characterized in that at least one state parameter is determined from the plurality of state parameters according to a prioritization sequence.

7. The method according to one of the preceding claims, characterized in that the plurality of the state parameters has a wheel slip and / or a wheel acceleration and / or a slip gradient and / or a wheel acceleration and / or a wheel pull and / or a wheel acceleration relative to the vehicle acceleration.

8. A method for determining a brake pressure characteristic field (250a, 250b, 251a, 251b) according to any one of the preceding claims, characterized by the steps of: providing a current wheel state (210), wherein the wheel state (210) has a plurality of state parameters; providing reward rules for a reinforcement learning method; determining a reward by means of the reward rules and the current wheel state (210); and, if a reward has been determined by the reinforcement learning method: determining a last brake pressure change made with respect to the value and direction of change and the associated brake pressure characteristic field (250a, 250b, 251a, 251b); and determining a correction value for the assigned brake pressure characteristic field (250a, 250b, 251a, 251b) according to the reinforcement learning method.

9. The method according to claim 8, characterized in that the reward rules determine a reward depending on a value less than a threshold value for slip and / or a value less than a slip value below zero and / or a modulation frequency of a pressure change.

10. Use of the method according to any one of claims 1 to 7 for controlling a braking pressure on a wheel.

11. Use of the method according to claims 8 or 9 to optimize the performance of a brake pressure control for a vehicle wheel.

12. The method characterized in that, based on a determined brake pressure change according to one of claims 1 to 7, a control signal is provided to control a at least partially automated vehicle; and / or based on a determined brake pressure change, a warning signal is provided to warn a vehicle passenger.

13. A braking system, characterized in that it is configured to perform a method according to any one of claims 1 to 7. PQLCLn / zznz / e / γΐΛΐ 14. A computer program, characterized in that it comprises commands which, when executed by a computer, cause the computer to perform the method according to any one of claims 1 to 9.

15. A machine-readable storage medium, characterized in that the computer program is stored in accordance with claim 14.

Images