Abstract: A hardware-implemented multi-layer perceptron model calculation unit includes: a processor core to calculate output quantities of a neuron layer based on input quantities of an input vector; a memory that has, for each neuron layer, a respective configuration segment for storing configuration parameters and a respective data storage segment for storing the input quantities of the input vector and the one or more output quantities; and a DMA unit to successively instruct the processor core to: calculate respective neuron layers based on the configuration parameters of each configuration segment, calculate input quantities of the input vector defined thereby, and store respectively resulting output quantities in a data storage segment defined by the corresponding configuration parameters, the configuration parameters of configuration segments successively taken into account indicating a data storage region for the resulting output quantities corresponding to the data storage region for the input quantities for

Abstract: A method for automatically generating a machine learning system which ascertains as a function of an input variable time series an output variable time series approximating an actual output variable time series, the machine learning system ascertaining a value of the output variable assigned to the predefinable point in time as a function of input and output variable values at the points in time, which are in a predefinable time interval, prior to a predefinable point in time, only a subset of the values of the input variable within the interval and of the values of the output variable being incorporated when ascertaining the output variable assigned to the predefinable point in time, the subset being selected so that it includes available input variable values and the output variable values assigned to the points in time within the interval, which are in a predefinable equidistant selection raster within the interval.

Abstract: A model calculation unit for calculating a data-based function model in a control unit is provided, the model calculation unit having a processor core which includes: a multiplication unit for carrying out a multiplication on the hardware side; an addition unit for carrying out an addition on the hardware side; an exponential function unit for calculating an exponential function on the hardware side; a memory in the form of a configuration register for storing hyperparameters and node data of the data-based function model to be calculated; and a logic circuit for controlling, on the hardware side, the calculation sequence in the multiplication unit, the addition unit, the exponential function unit and the memory in order to ascertain the data-based function model.

Abstract: A control device in a vehicle includes a unit for calculating, during operation of the vehicle, on the basis of at least one input variable ascertained during operation, at least one output variable for a control system of functions of the vehicle. The control device performs the calculation of the output variables using a Bayesian regression of training values ascertained, before operation, for the output variable and the input variable.

Abstract: A method is provided for populating a function for a control unit with data, in which method measurements are performed on a system at different measuring points on a test stand, and a global data-based model is set up based on the obtained measured values, and virtual measurements which simulate real measurements on the test stand are carried out on the global data-based model, and uncertainties for virtual measured values of the virtual measurements are determined from the global data-based model, the uncertainties of the virtual measured values being taken into account when populating the function for the control unit with data.

Abstract: A method for determining a sparse Gaussian process model to be carried out in a solely hardware-based model calculation unit includes: providing supporting point data points, a parameter vector based thereon, and corresponding hyperparameters; determining or providing virtual supporting point data points for the sparse Gaussian process model; and determining a parameter vector Qy* for the sparse Gaussian process model with the aid of a Cholesky decomposition of a covariant matrix KM between the virtual supporting point data points and as a function of the supporting point data points, the parameter vector based thereon, and the corresponding hyperparameters, which define the sparse Gaussian process model.

Abstract: A method for identifying a set of interpolation point data points from training data for a sparse Gaussian process model, encompassing the following tasks: successively selecting training data points from the set of training data for acceptance into or exclusion from a set of interpolation point data points in accordance with a selection criterion; and terminating selection when a termination criterion exists; the selection criterion depending on a divergence between a target value of the selected training data point and a function value, at the selected training data point, of the Gaussian process model based on the respectively current set of interpolation point data points.

Abstract: A method for operating a control unit, the control unit including a software-controlled main processing unit, a strictly hardware-based model calculation unit for calculating an algorithm, for carrying out a Bayesian regression method, based on configuration data, and a memory unit, a model memory area being defined in the memory unit to which a configuration register block for providing the configuration data in the model calculation unit is assigned, a calculation start-configuration register being assigned the highest address in the configuration register block into which configuration data are written, the writing into of which starts the calculation in the model calculation unit, the configuration data being written in a memory area of the memory unit from the model memory area into the configuration register block with an incremental copying process, the addresses being copied in the incremental copying process in ascending order.

Abstract: A method for generating a data-based function model includes: providing a first data-based partial model ascertained from a first training data record; providing at least one additional training data record; and performing the following steps for the at least one additional training data record: ascertaining a difference training data record having training data which correspond to the differences between the output values of the relevant additional training data record and the function value of the sum of the partial function values (ffirst_partial_model(x) fsecond_partial_model(x)) of the first data-based partial model and previously ascertained data-based partial model(s) at each of the measuring points of the relevant training data record; ascertaining an additional data-based partial model from the difference training data record; and forming a sum (f(x)) from the first and the additional data-based partial models.

Abstract: A method for determining a sparse Gaussian process model to be carried out in a solely hardware-based model calculation unit includes: providing supporting point data points, a parameter vector based thereon, and corresponding hyperparameters; determining or providing virtual supporting point data points for the sparse Gaussian process model; and determining a parameter vector Qy* for the sparse Gaussian process model with the aid of a Cholesky decomposition of a covariant matrix KM between the virtual supporting point data points and as a function of the supporting point data points, the parameter vector based thereon, and the corresponding hyperparameters, which define the sparse Gaussian process model.

Abstract: In a method for calculating a gradient of a data-based function model, having one or multiple accumulated data-based partial function models, e.g., Gaussian process models, a model calculation unit is provided, which is designed to calculate function values of the data-based function model having an exponential function, summation functions, and multiplication functions in two loop operations in a hardware-based way, the model calculation unit being used to calculate the gradient of the data-based function model for a desired value of a predefined input variable.

Abstract: A method for identifying a set of interpolation point data points from training data for a sparse Gaussian process model, encompassing the following tasks: successively selecting training data points from the set of training data for acceptance into or exclusion from a set of interpolation point data points in accordance with a selection criterion; and terminating selection when a termination criterion exists; the selection criterion depending on a divergence between a target value of the selected training data point and a function value, at the selected training data point, of the Gaussian process model based on the respectively current set of interpolation point data points.

Abstract: A method for operating a control unit, the control unit including a software-controlled main processing unit, a strictly hardware-based model calculation unit for calculating an algorithm, for carrying out a Bayesian regression method, based on configuration data, and a memory unit, a model memory area being defined in the memory unit to which a configuration register block for providing the configuration data in the model calculation unit is assigned, a calculation start-configuration register being assigned the highest address in the configuration register block into which configuration data are written, the writing into of which starts the calculation in the model calculation unit, the configuration data being written in a memory area of the memory unit from the model memory area into the configuration register block with an incremental copying process, the addresses being copied in the incremental copying process in ascending order.

Abstract: A method for post-adaption of an at least partially data-based function model which corresponds to a sum of a basis function model, e.g., a data-based basis function model, and an additive fault model, includes: providing the basis function model; recording training data; ascertaining the data-based additive fault model based on difference training data which represent differences between the measured values of the training data and the function values of the data-based basis function model at the measuring points of the training data; and modifying the training data and/or the additive fault model so that function values of the data-based function model remain within a predefined adaption range.

Abstract: A method is provided for populating a function for a control unit with data, in which method measurements are performed on a system at different measuring points on a test stand, and a global data-based model is set up based on the obtained measured values, and virtual measurements which simulate real measurements on the test stand are carried out on the global data-based model, and uncertainties for virtual measured values of the virtual measurements are determined from the global data-based model, the uncertainties of the virtual measured values being taken into account when populating the function for the control unit with data.

Abstract: A model calculation unit for calculating a data-based function model in a control unit is provided, the model calculation unit having a processor core which includes: a multiplication unit for carrying out a multiplication on the hardware side; an addition unit for carrying out an addition on the hardware side; an exponential function unit for calculating an exponential function on the hardware side; a memory in the form of a configuration register for storing hyperparameters and node data of the data-based function model to be calculated; and a logic circuit for controlling, on the hardware side, the calculation sequence in the multiplication unit, the addition unit, the exponential function unit and the memory in order to ascertain the data-based function model.

Abstract: A method for ascertaining a nonparametric, data-based function model, in particular a Gaussian process model, using provided training data, the training data including a number of measuring points which are defined by one or multiple input variables and which each have assigned output values of at least one output variable, including: selecting one or multiple of the measuring points as certain measuring points or adding one or multiple additional measuring points to the training data as certain measuring points; assigning a measuring uncertainty value of essentially zero to the certain measuring points; and ascertaining the nonparametric, data-based function model according to an algorithm which is dependent on the certain measuring points of the modified training data and the measuring uncertainty values assigned in each case.

Abstract: A computerized method for creating a function model based on a non-parametric, data-based model, e.g., a Gaussian process model, includes: providing training data including measuring points having one or multiple input variables, the measuring points each being assigned an output value of an output variable; providing a basic function; modifying the training data with the aid of difference formation between the function values of the basic function and the output values at the measuring points of the training data; creating the data-based model based on the modified training data; and providing the function model as a function of the data-based model and the basic function.

Abstract: A method for creating a nonparametric, data-based function model having measuring points in multiple training data records, including the following: providing weighting specifications for the measuring points of each training data record; forming a set union of the measuring points of the multiple training data records; and creating the nonparametric function model from the set union of the measuring points of the training data records according to an algorithm which is dependent on the weighting specifications for the measuring points of the multiple training data records.

Abstract: A method for generating a data-based function model includes: providing a first data-based partial model ascertained from a first training data record; providing at least one additional training data record; and performing the following steps for the at least one additional training data record: ascertaining a difference training data record having training data which correspond to the differences between the output values of the relevant additional training data record and the function value of the sum of the partial function values (ffirst—partial—model(x) fsecond—partial—model(x)) of the first data-based partial model and previously ascertained data-based partial model(s) at each of the measuring points of the relevant training data record; ascertaining an additional data-based partial model from the difference training data record; and forming a sum (f(x)) from the first and the additional data-based partial models.