Model calculation unit and control unit for calculating an RBF model

Patent metadata

Field	Value
Publication number	US-11645502-B2
Application number	US-201716330952-A
Country	US
Kind code	B2
Filing date	Sep 5, 2017
Priority date	Sep 7, 2016
Publication date	May 9, 2023
Grant date	May 9, 2023

Abstract

A model calculation unit for calculating an RBF model is described, including a hard-wired processor core designed as hardware for calculating a fixedly predefined processing algorithm in coupled functional blocks, the processor core being designed to calculate an output variable for an RBF model as a function of one or multiple input variable(s) of nodes V[j,k], of length scales (L[j,k]), of weighting parameters p3[j,k] predefined for each node, the output variable being formed as a sum of a value calculated for each node V[j,k], the value resulting from a product of a weighting parameter p3[j,k] assigned to the particular node V[j,k], and a result of an exponential function of a value resulting from the input variable vector as a function of a square distance of the particular node (V[j,k]), weighted by the length scales (L[j,k]), the length scales (L[j,k]) being provided separately for each of the nodes as local length scales.

First claim

What is claimed is:

1. A model calculation unit for calculating a radial bias function (RBF) model, comprising:
a hard-wired processor core including hardware for calculating a fixedly predefined processing algorithm for calculating the RBF model using coupled function blocks, the hardware including a state machine to control the calculating of the fixedly predefined processing algorithm by the coupled function blocks without executing software code, wherein the hard-wired processor core is configured to calculate the RBF model for controlling a technical system in a motor vehicle based on the RBF model, and wherein:
the hard-wired processor core calculates an output variable for the RBF model as a function of: at least one input variable of an input variable vector, nodes (V[j,k]), length scales (L[j,k]), and weighting parameters (p3[j]) predefined for each node;
the output variable is formed as a sum of a value calculated for each node, the value resulting from a product of the weighting parameter assigned to a particular node and an exponential function of a value resulting from a square distance of the particular node from the input variable vector, weighted by the length scales; and
the length scales are provided separately for each of the nodes as local length scales.

2. The model calculation unit as recited in claim 1 , wherein the length scales are provided separately for each of the nodes as local length scales and for each of the at least one input variable of the input variable vector.

3. The model calculation unit as recited in claim 1 , wherein the coupled function blocks include a multiply-accumulate (MAC) block and an exponential function calculation block, the model calculation unit further comprising a memory for storing the at least one input variable of the input variable vector, the nodes, the length scales, the weighting parameters predefined for each node, and the output variable.

4. The model calculation unit as recited in claim 1 , wherein the hard-wired processor core is formed in a surface area of an integrated module.

5. The model calculation unit as recited in claim 1 , wherein purely dimension-specific length scales are utilized instead of the local length scales as a function of a selection variable for the calculation of the output variable.

6. A control unit, comprising:
a microprocessor;
at least one model calculation unit for calculating a radial bias function (RBF) model, the at least one model calculation unit including:
a hard-wired processor core including hardware for calculating a fixedly predefined processing algorithm in coupled function blocks, the hardware including a state machine to control the calculating of the fixedly predefined processing algorithm by the coupled function blocks without executing software code, wherein:
the hard-wired processor core calculates an output variable for the RBF model as a function of: at least one input variable of an input variable vector, nodes (V[j,k]), length scales (L[j,k]), and weighting parameters (p3[j]) predefined for each node;
the output variable is formed as a sum of a value calculated for each node, the value resulting from a product of the weighting parameter assigned to a particular node and an exponential function of a value resulting from a square distance of the particular node from the input variable vector, weighted by the length scales; and
the length scales are provided separately for each of the nodes as local length scales,

wherein the control unit is configured to control a technical system in a motor vehicle based on the calculating of the RBF model.

7. The control unit as recited in claim 6 , wherein the control unit includes an integrated circuit.

8. The control unit as recited in claim 6 , wherein the control unit controls an engine system in a motor vehicle.

Assignees

• Bosch Gmbh Robert

Inventors

• Guntoro Andre
• Kloppenburg Ernst
• Markert Heiner
• Ulmer Holger

Classifications

• G06N3/09

  Supervised learning · CPC title

• G06N3/0499

  Feedforward networks · CPC title

• G06N3/063 Primary

  using electronic means · CPC title

• B60W10/06

  including control of combustion engines · CPC title

• G06N3/04

  Architecture, e.g. interconnection topology · CPC title