Advanced model mapping and simulations

Patent metadata

Field	Value
Publication number	US-11544430-B2
Application number	US-202016862329-A
Country	US
Kind code	B2
Filing date	Apr 29, 2020
Priority date	Apr 30, 2019
Publication date	Jan 3, 2023
Grant date	Jan 3, 2023

Abstract

The invention relates to a system for performing a method for performing computer-aided simulations. The system loads an input dataset which is based on user interactions of a user indicating a manipulation of a representation of a component, provides a tracking dataset indicating changes in the representation through the manipulations by evaluating the input dataset, provides an output dataset which is based on evaluation of the tracking dataset by means of machine learning, where the output dataset assigns a model to the representation, and outputs the output dataset.

First claim

What is claimed is:

1. A method for performing computer-aided simulations, the method comprising:
assigning a plurality of representations to a plurality of models, wherein each representation is associated with a unique component and assigned to a unique model, wherein the plurality of models represent a simulation of a plurality of components;
loading an input dataset associated with the plurality of representations;
tracking changes to the input dataset based on user interactions of a user indicating a manipulation of at least one representation among the plurality of representations, wherein the changes define an interaction among the plurality of components;
generating a tracking dataset indicating the changes in the at least one representation by the manipulation through evaluation of the input dataset;
generating an output dataset, the output dataset based on evaluation of the tracking dataset using machine learning, wherein the output dataset reconfigures the plurality of models assigned to the plurality of representations based on the interaction among the plurality of components as defined by the user interactions; and
outputting the output dataset.

2. The method of claim 1 , wherein the machine learning comprises using a trained artificial neural network in order to perform a model classification.

3. The method of claim 2 , wherein a deep neural network is used as the trained artificial neural network.

4. The method of claim 2 , wherein the trained artificial neural network is trained through supervised learning.

5. A system for performing computer-aided simulations, the system comprising:
a platform for assigning a plurality of representations to a plurality of models, wherein each representation is associated with a unique component and assigned to a unique model, wherein the plurality of models represent a simulation of a plurality of components;
a tracking unit for loading an input dataset associated with the plurality of representations;
an input device for receiving changes to the input data based on user interactions of a user indicating manipulation of at least one representation among the plurality of representations, wherein the changes define an interaction among the plurality of components;
the tracking unit is further configured to generate a tracking dataset indicating the changes in the at least one representation based on evaluation of the input dataset;
the platform is further configured to generate an output dataset based on evaluation of the tracking dataset, wherein generating the output dataset comprises using machine learning to reconfigure the plurality of models assigned to the plurality of representations based on the interaction among the plurality of components as defined by the user interactions; and
an output device for outputting the output dataset.

6. The system of claim 5 , wherein using machine learning to assign a model to the representation comprises using a trained artificial neural network to perform a model classification.

7. The system of claim 6 , wherein the trained artificial neural network is a deep neural network.

8. The system of claim 6 , wherein the trained artificial neural network is trained using supervised learning.

9. A computer-readable memory device having stored thereon instructions that, upon execution by one or more processors, cause the one or more processors to:
assign a plurality of representations to a plurality of models, wherein each representation is associated with a unique component and assigned to a unique model, wherein the plurality of models represent a simulation of a plurality of components;
load an input dataset associated with the plurality of representations;
track changes to the input dataset based on user interactions of a user indicating a manipulation of at least one representation among the plurality of representations, wherein the changes define an interaction among the plurality of components;
generate a tracking dataset indicating the changes in the at least one representation by the manipulation through evaluation of the input dataset;
generate an output dataset, the output dataset based on evaluation of the tracking dataset using machine learning, wherein the output dataset reconfigures the plurality of models assigned to the plurality of representations based on the interaction among the plurality of components as defined by the user interactions; and
output the output dataset.

10. The computer-readable memory device of claim 9 , wherein using machine learning comprises using a trained artificial neural network in order to perform a model classification.

11. The computer-readable memory device of claim 10 , wherein a deep neural network is used as the trained artificial neural network.

12. The computer-readable memory device of claim 10 , wherein the trained artificial neural network is trained through supervised learning.

13. The method of claim 1 , wherein the output dataset represents an updated simulation of the plurality of components.

14. The method of claim 13 , wherein the updated simulation includes at least one of an updated position of the plurality of representations, an updated input variable or an updated output variable liking the plurality of representations.

15. The method of claim 1 , wherein the plurality of components are vehicle components, and wherein the simulation represents a behavior of the vehicle components as configured by the user.

Assignees

• Ford Global Tech Llc

Inventors

• Aslandere Turgay Isik
• Bitsanis Evangelos
• Marbaix Michael
• Stefan Frederic
• Chevalier Alain Marie Roger

Classifications

• G06F30/27 Primary

  using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model · CPC title

• G06N3/045

  Combinations of networks · CPC title

• G06N3/08 Primary

  Learning methods · CPC title

• G06N3/04

  Architecture, e.g. interconnection topology · CPC title

• G06N3/09

  Supervised learning · CPC title