DE102023203125A1 - Simulation-based parameter presetting for dynamic driving functions - Google Patents

Abstract

The invention relates to a method for determining a set of functional parameters of a vehicle brake control unit. In order to reduce the number of test drives, different values are selected for at least a subset of the functional parameters and, using a simulation platform, driving dynamics measurement results are simulated for predetermined driving maneuvers and environmental settings, which are evaluated using an objectification algorithm with at least one key figure. Furthermore, a metamodel is trained using the respective values with associated key figures, which thus indicates a connection between the respective subset of functional parameters and the key figures. This metamodel is used in the actual parameter optimization using an optimization algorithm to find optimal parameter values.

Description

The invention relates to a method for determining a set of functional parameters of a vehicle brake control unit, a corresponding computer program product and a manufacturing method.

Modern motor vehicles have driving dynamics controllers that primarily contribute to a safe state in dynamic driving situations. A driving dynamics controller contains different individual functions, each of which is responsible for certain driving situations (for example, anti-lock braking system (ABS) when braking or electronic stability control (ESC) when cornering dynamically) and specifically influences driving behavior in terms of driving safety, driving comfort or driving performance. These functions have variable parameters (“functional parameters”) that are individually adjusted for each vehicle version so that the driving dynamics controller works as well as possible and the requirements for driving behavior are met. So far, it has only been possible to determine a few specific parameters using simulation and computers. These parameters are determined using objective criteria based on the simulated or measured driving behavior (Bechtloff, Jakob Philipp: “Estimation of the sideslip angle and driving dynamics parameters for improving model-based driving dynamics controls”, 2017, Technical University of Darmstadt, Darmstadt). An example of these objectively determined function parameters are the parameters of the single-track model within the ESC function. In addition, there are function parameters that can be applied based on geometric and technical data of the vehicle and the braking system without real driving tests or the use of simulation (so-called "desktop calibration"). These parameters represent, for example, the vehicle mass or the coordinates of the vehicle's center of gravity.

The majority of the functional parameters of driving dynamics functions are currently applied in driving tests with comparatively high effort using real vehicles on different test tracks and under different environmental conditions. When applying these parameters, in addition to objective parameters, additional subjective evaluation criteria of the driver are used. The driver can use his experience to evaluate the overall driving behavior and the overall impression of the vehicle and its surroundings and apply the parameters to the overall driving behavior, consisting of objective and subjective criteria. In doing so, the driver must take into account and individually adjust physically determined compromises of different criteria (e.g. safety vs. performance).

There are several disadvantages in the context of the functional parameter application based on a real vehicle: On the one hand, the classic application of the driving dynamics parameters is very time-consuming and costly, since cost-intensive prototype vehicles are required first, which have to be manufactured in small numbers and equipped with measurement technology. These vehicles are then transported and used on different test tracks that meet different environmental conditions as part of the parameter application. Different environmental conditions can, for example, represent different friction coefficients of the road, for which specific functional parameters must be applied. The availability of the vehicles and the dependence of the environmental conditions require precise project planning and increase the effort (time and costs) within a project. In addition, there is the time and organizational effort on the part of the test drivers. On the other hand, real prototype vehicles are absolutely necessary for the classic application, since the application of the driving-relevant parameters takes place with the real vehicle. Real vehicles are necessary because the complex overall driving behavior, the interaction between the vehicle and the driving dynamics controller installed in it, is evaluated holistically and objectively and subjectively by the application engineer. At the same time, real vehicles enable the integration of the experienced driver (application engineer), who can subjectively evaluate the influence of the parameters on the resulting overall dynamic driving behavior.

The task behind this invention is to develop a method which makes it possible to significantly reduce the application effort of the functional parameters.

The object is achieved by a method for determining a set of functional parameters of a vehicle brake control unit, wherein different values are selected for at least a subset of the functional parameters and by means of a simulation platform for predetermined driving maneuvers and environmental settings, driving dynamic measurement results are simulated, which are evaluated by means of an objectification algorithm with at least one key figure, and wherein a metamodel is created by means of the respective values are learned with associated key figures, which thus indicates a connection between the respective subset of function parameters and the key figures, whereby the metamodel is used in the actual parameter optimization by means of an optimization algorithm to find optimal parameter values.

The method according to the invention, hereinafter referred to as "SIPA" (Simulation-based Pre-Adjustment), can thus eliminate the main disadvantages of classic parameter application, in particular the effort and availability of prototype vehicles for vehicle dynamics controllers. The simulation-based approach anchored in the method eliminates the dependencies on both the ambient conditions and the prototype vehicles. The ambient conditions and the prototype vehicles are digitally modeled in the simulation according to the requirements and technical data of the vehicle, so that these virtual models can be simulated in conjunction with a vehicle dynamics controller that is also virtual, and the functional parameters can be applied using the simulated driving behavior in a virtual environment.

In a preferred embodiment of the invention, the objectification algorithm carries out the evaluation based on several types of key figures and a separate metamodel is learned for each type of key figure, which relates the respective functional parameters to the corresponding type of key figure.

In a preferred embodiment of the invention, the selection of the values for at least the subset of the functional parameters is carried out by means of a design of experiments unit, in particular according to the Latin hypercube sample method.

In a preferred embodiment of the invention, the backpropagation approach is used to train the metamodel.

In a preferred embodiment of the invention, a genetic algorithm is used as the optimization algorithm.

In a preferred embodiment of the invention, the optimization algorithm takes into account the different metamodels for the different types of key performance indicators, in particular weighted.

In a preferred embodiment of the invention, at least one further metamodel is trained for a further subset of functional parameters for other driving maneuvers and/or environmental settings. This is then also used by the optimization algorithm. The driving maneuvers with their different environmental conditions can be weighted in order to favor certain properties (eg stability) under certain environmental conditions (eg on snow/low friction coefficient of the road).

In a preferred embodiment of the invention, the subset of functional parameters comprises general parameters and environment- and/or maneuver-dependent parameters, wherein the further subset of functional parameters comprises the same general parameters and other environment- and/or maneuver-dependent parameters.

In a preferred embodiment of the invention, the total set of parameters of the set of functional parameters of the vehicle brake control unit is divided into subsets, with optimal parameter values for the subsets being found one after the other. For each of these functional parameters of the subset, individual driving maneuvers and environmental conditions are assigned in which the functional parameters have an influence, in particular an above-average influence, on the driving dynamics measurement results.

The object is also achieved by a computer program product comprising instructions which, when the program is executed by a computer, cause the computer to carry out one of the above methods.

The object is also achieved by a manufacturing method for a motor vehicle brake system comprising a vehicle brake control unit which comprises program code for executing vehicle functions, which contains functional parameters, wherein values for the functional parameters which were determined by means of one of the above methods are stored in the control unit.

SIPA enables the application of selected functional parameters of the vehicle dynamics controller using simulation in a virtual environment. The method is as in 1 shown divided into a “preprocessing process” and a “parameter application process” and builds on an existing approach (Rot, Ivan: “Method for model-based calibration of the gearshift control of transmission control units in a virtual environment”, 2017, Technical University of Darmstadt, Darmstadt)

Within the "pre-processing process", a simulation platform is used that contains the vehicle, the vehicle environment (road, route), the virtual driver and the vehicle dynamics controller on which the functional software to be applied runs. The environmental conditions and vehicle properties can be adapted within the simulation platform so that application coverage is guaranteed taking into account different environmental conditions.

The simulation platform must map all relevant vehicle sizes and driving conditions. Due to its complexity, the simulation environment requires a lot of computing resources. This has a negative impact on the applicability in the subsequent "parameter application process", as this requires a large number of simulation runs (iterations) as part of the parameter optimization (application). Therefore, the aim of the "preprocessing process" with the help of the complex simulation environment is to first derive a simplified so-called metamodel (hereinafter referred to as "training"), which can map the relevant properties for the parameter application and requires fewer computing resources to execute. This less complex and extensive metamodel can then be used in the "parameter identification process" in a resource-saving manner.

To train the metamodel, the driving behavior simulated using the simulation environment is evaluated and information is extracted from it that describes the driving behavior in detail. First, a test plan is created using a Design of Experiments (DoE) approach. This test plan adjusts the functional parameters to be optimized for each test, for example using the Latin Hypercube Sample method, so that the system generates as much information as possible with as little effort (tests) as possible. The tests are carried out one after the other using the simulation platform for a specified driving maneuver and the generated measurement data are then analyzed using an objectification algorithm, for example according to EN 10 2019 217 431 A1 , evaluated. The objectification algorithm calculates objective rating models based on the measured driving behavior measurements, which correspond to a subjective assessment by a real driver for the driving maneuver under consideration. In addition to calculating the subjective ratings, the objectification algorithm also calculates objective criteria specific to each driving maneuver, which are used to evaluate the driving maneuvers together with the subjective rating models. The subjective rating models and the objective criteria are summarized and referred to below as KPIs (Key Performance Indicators).

mit

• • Eingang: P: [p₁, ..., p_e]
• • Offset: b
• • Aktivierungsfunktion: f_KNN
• • Gewicht: g
• • Anzahl der Eingangsparameter: n

Based on the functional parameters known from the DoE and the KPIs determined from the simulated driving tests, the metamodel is trained in the last step of the "preprocessing process". Different approaches can be used to train the metamodel, such as the backpropagation algorithm from the field of machine learning methods. The goal of the trained metamodel is to efficiently calculate a mathematical relationship between the functional parameters (input variable of the metamodel P) and the KPIs (output of the metamodel y_KNN). An independent metamodel with individual model parameters is implemented to calculate each KPI. The metamodel can represent any mathematical model that can calculate a functional behavior of an output depending on the model input variables. An example is an artificial neural network (KNN) in the form of a multilayer perceptron (MLP) with a hidden layer that fulfills the properties mentioned: $$y_{KNN}=e_{KNN}\left({\displaystyle \sum _{e=1}^n{G}_e}\cdot p_e+b\right)$$

with

• • Input: P: [p ₁ , ..., p _e ]
• • Offset: b
• • Activation function: f _KNN
• • Weight: g
• • Number of input parameters: n

mit $${ƒ}_{ZF}\left(\mathbf{P}\right)={\displaystyle \sum _{e=1}^m\left(\frac{y_{KN{N}_e}\left(\mathbf{P}\right)\cdot w_e}{m}\right)}$$

und Anzahl der KPIs: m. Je nach Anwendungsfall können auch andere, auch lokale, Optimierungsalgorithmen angewendet werden.After the metamodels have been trained, they are applied within the "parameter application process" to calculate the KPIs depending on the function parameters. The function parameters P are specified by the optimization algorithm. In SIPA, an optimization algorithm is used that is able to find a global optimum for a nonlinear system behavior, taking into account several function parameters to be optimized (applied). For example, the genetic algorithm (GA) fulfills these properties and can be used for the application (optimization) of the function parameters [6]. The optimization algorithm has the task of minimizing the objective function f _ZF (P): $$\text{min}{ƒ}_{ZF}\left(\mathbf{P}\right),\mathbf{P}\in {ℝ}^D$$

with $${ƒ}_{ZF}\left(\mathbf{P}\right)={\displaystyle \sum _{e=1}^m\left(\frac{y_{KN{N}_e}\left(\mathbf{P}\right)\cdot w_e}{m}\right)}$$

and number of KPIs: m. Depending on the application, other, including local, optimization algorithms can also be applied.

The objective function depends on the KPIs calculated by the metamodels (y _KNN ). Within the objective function, weightings (w) are defined for each KPI, which can be specified by the user in order to specifically influence the characteristics of the driving behavior. For example, the weighting for a KPI relevant to driving dynamics stability can be set so that the function parameters assume certain values and the vehicle therefore has the highest possible stability in the driving maneuver under consideration. The objective function is iteratively minimized by the optimization algorithm until one or more defined optimality criteria are met. An optimality criterion can, for example, be defined as the convergence of the objective function to an undetermined value as the iteration progresses.

Certain functional parameters depend on certain driving situations (e.g. driving speed) or environmental conditions (e.g. the coefficient of friction µ of the road-tire contact). Other functional parameters, however, are generally valid and do not relate to any specific situation. Within a driving test, a driving and environmental scenario can always be defined and simulated or driven. However, the generally valid parameters must be applied to all possible driving and environmental scenarios. In reality, these parameters are set based on experience and iteratively. In SIPA, this situation is solved by the following approach and is considered an explicit part of this invention disclosure:

für alle betrachteten Umgebungen berechnet: $$y_{KNN.Umg}^{*}\left(\mathbf{A},\mathbf{B},\mathbf{C},\mathbf{W}\right)={\displaystyle \sum _{I=1}^K\frac{y_{KNN.Um{g}_i}*w_{Um{g}_i}}{k}}$$

mit Anzahl der Umgebungen: k KPI-Wert innerhalb einer Umgebung: yy_KNN.Umg To explain the approach, three environments (alternatively driving conditions) A, B and C are considered as examples. For the application, there are parameters W that are valid for each environment. In addition, there are parameters A, B and C that are only valid within the environment of the same name A, B and C. In order to find a cross-environment parameter set for W, A, B and C that meets the required requirements for driving dynamics, in the present approach, driving tests are first carried out (simulated) for each environment individually according to the "preprocessing process" and then metamodels are trained with the corresponding parameters (see 2 ). Each metamodel therefore contains information about a specific environment, which can calculate the KPIs depending on the environment-independent and the specific environment-dependent parameters. A general (cross-environment) application of the considered is achieved by applying the learned metamodels in parallel within the subsequent "parameter application process". In this process, the GA calculates the complete input parameter set (A, B, C, W) in each iteration step, which is then divided up as input variables for the respective metamodels (see 2 “Calculation for a KPI”). The replacement KPI is then $\left(y_{KNN.UmG}^{*}\right)$

calculated for all environments considered: $$y_{KNN.UmG}^{*}\left(\mathbf{A},\mathbf{B},\mathbf{C},\mathbf{W}\right)={\displaystyle \sum _{I=1}^K\frac{y_{KNN.Um{G}_i}*w_{Um{G}_i}}{k}}$$

with number of environments: k KPI value within an environment: yy _{KNN.Environment}

Weighting of an environment: w _environment .

The weighting w _Environment allows the KPIs of individual environments to be weighted specifically. This makes it possible to optionally give preference to the driving behavior for a specific environment while taking the overall driving behavior into account.

The objective function value is calculated taking into account all KPIs and KPI weights (w) as follows: $$e_{ZF}\left(\mathbf{A},\mathbf{B},\mathbf{C},\mathbf{W}\right)={\displaystyle \sum _{i=1}^m\frac{y_{KNN.Um{G}_i}^{*}*w_i}{m}}$$

The 2 shows the approach schematically. A function can contain several dozen or hundreds of function parameters that influence the driving behavior and should be applied to the individual vehicle accordingly. Within a function, these parameters determine the driving behavior for a total of several driving situations and environmental conditions. An attempt to take all parameters into account in one process step (driving maneuver) turns out to be very complex from a mathematical, technical and therefore practical point of view. Learning a metamodel that has several hundred parameters as input variables means creating a large amount of measurement data that is used to train the metamodel. The need for training data for training a metamodel increases exponentially with the number of its input variables. This leads to very high computing resources, which are largely not available in practice at the moment. Another problem results from the fact that a driving maneuver would have to be carried out very extensively and comprehensively in order to cover all driving situations that the parameters to be applied influence. This is exemplified in 3 under "Application approach without application strategy". Accordingly, all relevant functional parameters in a complex driving maneuver must be set in one application step so that they best meet the requirements for all sub-events in this driving maneuver. Mathematically, each parameter represents a degree of freedom in which the optimal value must be found.

This complex problem is solved within the scope of this invention by implementing and applying an application strategy, which represents a particularly preferred variant of the method according to the invention. The application strategy is defined and implemented individually for the application of each function. The essential approach of the application strategy is to divide a complex problem into smaller, less complex sub-problems and sub-steps. This creates a practical and computing resource-saving process for the virtual application of the function parameters of a function.

The application strategy defines a process individually for each function that provides a sequence in which certain parameters (collectively referred to as a parameter group) are applied with a driving maneuver defined specifically for the application of this parameter group. This driving maneuver can be divided into one or more driving maneuvers with different environments (ambient conditions).

After the function parameters have been applied within the first step of the defined sequence, the next parameter group with the next associated driving maneuver is applied depending on these then applied parameters. These steps are carried out sequentially until all planned parameter groups have been applied within the application strategy. This approach takes into account a potential dependency of the parameters on one another. One goal of the application strategy is to reduce the number of parameters in a parameter group as much as possible within an application process in order to reduce the problem as much as possible from the perspective of computing resources and finding a suitable parameter set (application = parameter optimization) (in other words, to divide the complex problem into as many small sub-problems as possible).

Figure 3 shows the main features of the application approach with application strategy schematically

The application of the functional parameters takes place entirely in a virtual environment, so that no real physical components or systems (neither series nor prototype) are required. SIPA represents a generic approach and is applicable/transferable for the application of all existing and future functions in connection with dynamic and quasi-static driving conditions.

The special feature of the SIPA method is that, within the framework of the virtual application process, the driving characteristics are taken into account individually by specifically weighting the KPIs and the functional parameters can be applied automatically according to the weighting.

In addition, different driving and environmental conditions can be taken into account simultaneously within the application process, so that the applied functional parameters are optimally designed for all driving and environmental conditions considered.

The application strategy allows the complex application of many function parameters to be reduced to several small subproblems in order to reduce the overall computing resources and the effort required to find suitable parameter application processes.

The method represents a generic approach and can generally be applied to the application of the functional parameters of all quasi-static and dynamic driving functions. For example, the method can be applied to other (dynamic or quasi-static) driving functions with appropriate adaptation of the KPIs and the application strategy. One example is the application of the parameters of the ABS (anti-lock braking system) or ACC (adaptive cruise control) functions.

In the following example, SIPA is used for a virtual application for four functional parameters of the yaw control function (Active Yaw Control Function). An exemplary parameter group with an associated driving maneuver is therefore considered within the application strategy. A sine dwell driving maneuver is considered in which the vehicle is deliberately made to oversteer in a total of three independent environments with a road-tire friction coefficient of µ _A = 0.3, µ _B = 0.65 and µ _C = 1.0 under otherwise identical conditions and the yaw control function transfers the vehicle to the stable driving state as a result of the oversteer. In order to exclude influences from potential other control functions, all other driving dynamics functions are deactivated so that only the control function to be applied for vehicle oversteer is active and is influenced by the parameters to be applied.

In the first step, a test plan (DoE) is created in the "preprocessing process" and the maneuver is carried out individually and virtually for each environment with different parameter combinations within the simulation platform. The measurement signals are then evaluated using the objectification algorithm and the KPIs "performance", "stability", "handling" and "comfort" that are relevant for describing the driving dynamics quality are determined. In the last step of the "preprocessing process", the metamodels are trained for each environment and each KPI (in this example a total of 12 metamodels). The quality of the metamodels in this example is shown in the To demonstrate quality, the KPIs calculated using the trained metamodels are compared with the KPIs that were determined directly on the basis of the measurement signals (coming from the simulation platform) using the objectification algorithm. The closer the points (one point corresponds to the evaluation of a driving maneuver) are to the diagonal shown, the higher the prediction accuracy of the metamodel. The quality of the metamodels considered can be described as good overall.

In the "parameter application process" in this example, two different optimization variants are carried out. Optimization variant 1 focuses on stable driving behavior, which in the case of an oversteer maneuver is equivalent to a smaller permissible sideslip angle that the vehicle builds up during the maneuver. The weightings of the KPIs used in the optimization for both optimization variants are shown in Table 1. To simplify matters, the weightings of the three environments are equated: w _{Environment A} = w _{Ambient B} = w _AmbC . Table 1 Weightings w optimization performance stability Handling comfort 1 0.1 0.7 0.1 0.1 2 0.1 0.1 0.1 0.7

The optimization results achieved for all three environments are presented in the following normalized. The results show that the user's specifications for the weightings were generally implemented as expected. In the example for optimization 1 The focus was on stable driving behavior. The environment-dependent and environment-independent parameters were applied in such a way that very high stability was achieved in all three environments (standardized grade 1 means that the maximum possible KPI grade was achieved under the given conditions).

In the example already shown, application results for the yaw control function (AYC function) for two different dynamic vehicle characteristics were already explained. As already mentioned in the description of the method, the method can be applied to any other driving dynamic functions. To apply the method to a different function, it is first necessary to define the parameters of the function to be applied and to implement the relevant dynamic driving maneuvers in the simulation. The method can then be applied to the new application problem in an equivalent way to the example shown. Another example is the application of parameters of the ABS function (anti-lock braking system), in which, for example, the parameters that influence the braking distance and stability are applied when braking straight ahead or during so-called Mü-split braking (braking at different friction values in parallel).

QUOTES INCLUDED IN THE DESCRIPTION

This list of documents listed by the applicant was generated automatically and is included solely for the better information of the reader. 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

• DE 102019217431 A1 [0021]

Claims (11)

Method for determining a set of functional parameters of a vehicle brake control unit, wherein different values are selected for at least a subset of the functional parameters and driving dynamics measurement results are simulated by means of a simulation platform for predetermined driving maneuvers and environmental settings, which are evaluated by means of an objectification algorithm with at least one key figure, and wherein a metamodel is trained by means of the respective values with associated key figures, which thus indicates a connection between the respective subset of functional parameters and the key figures, wherein the metamodel is used in the actual parameter optimization by means of an optimization algorithm to find optimal parameter values. Procedure according to Claim 1 , characterized in that the objectification algorithm carries out the evaluation on the basis of several types of key figures and for each type of key figure a separate metamodel is learned which relates the respective functional parameters to the corresponding type of key figure. Method according to one of the preceding claims, characterized in that the selection of the values for at least the subset of the functional parameters is carried out by means of a design of experiments unit, in particular according to the Latin hypercube sample method. Method according to one of the preceding claims, characterized in that the backpropagation approach is used to train the metamodel Method according to one of the preceding claims, characterized in that a genetic algorithm is used as the optimization algorithm Method according to one of the preceding claims, characterized in that the optimization algorithm takes into account the different metamodels for the different types of key figures, in particular in a weighted manner. Method according to one of the preceding claims, characterized in that at least one further metamodel is learned for a further subset of functional parameters for other driving maneuvers and/or environmental settings. Method according to one of the preceding claims, characterized in that the subset of functional parameters has generally valid parameters and environment- and/or maneuver-dependent parameters, wherein the further subset of functional parameters has the same generally valid parameters and other environment- and/or maneuver-dependent parameters. Method according to one of the preceding claims, characterized in that the total set of parameters of the set of functional parameters of the vehicle brake control unit is divided into subsets, wherein the optimal parameter values for the subsets are found one after the other. Computer program product comprising instructions which, when executed by a computer, cause the computer to carry out one of the preceding methods. Manufacturing method for a motor vehicle brake system comprising a vehicle brake control unit which comprises program code for executing vehicle functions, which contains functional parameters, characterized in that values for the functional parameters are stored in the control unit, which values are determined by means of one of the methods according to Claim 1 until 9 were determined.