Method and device for controlling a technical system using a control model

Patent metadata

Field	Value
Publication number	US-11567461-B2
Application number	US-201716328605-A
Country	US
Kind code	B2
Filing date	Jul 25, 2017
Priority date	Aug 29, 2016
Publication date	Jan 31, 2023
Grant date	Jan 31, 2023

Abstract

In order to control a technical system using a control model, a transformation function is provided for reducing and/or obfuscating operating data of the technical system so as to obtain transformed operating data. In addition, the control model is generated by a model generator according to a first set of operating data of the technical system. In an access domain separated from the control model, a second set of operating data of the technical system is recorded and transformed by the transformation function into a transformed second set of operating data which is received by a model execution system. The control model is then executed by the model execution system, by supplying the transformed second set of operating data in an access domain separated from the second set of operating data, control data being derived from the transformed second set of operating data.

First claim

The invention claimed is:

1. A method for controlling a technical system based on a control model, said method comprising:
generating, by a model generator residing in a second access domain, the control model as a function of first operating data received from the technical system residing in a first access domain that is separated from the second access domain, said control model configured to output control data upon being executed using second operating data that was transformed by being reduced and/or obfuscated;
receiving, by a model execution system residing in the second access domain from a controller residing in the first access domain, the transformed second operating data;
executing, by the model execution system using the transformed second operating data, the control model, said executing the control model comprising outputting the control data; and
transmitting, by the model execution system to the controller, the control data, wherein the control data is configured to control the technical system by the controller,
wherein the model generator, the model execution system and the controller each comprise one or more processors for carrying out all method steps performed by the model generator, the model execution system, and the controller, respectively.

2. The method as claimed in claim 1 , wherein the control model is generated by the model generator based on the first operating data having been transformed by a transformation function that reduces and/or obfuscates the first operating data after the transformation was generated and trained by the model generator based on the first operating data.

3. The method as claimed in claim 2 , wherein the model generator and/or the model execution system are operated by a model vendor in the second access domain, and wherein the transformation function is provided to the model generator by the model vendor.

4. The method as claimed in claim 2 , wherein the transformation function comprises a neural autoencoder.

5. The method as claimed in claim 2 , wherein the transformation function comprises multiplication by a random matrix.

6. The method as claimed in claim 1 , wherein the control model comprises at least one of a neural network, a data-driven regressor, a support vector machine, and a decision tree.

7. The method as claimed in claim 1 , wherein an operator (BTS) in the first access domain uses and/or controls the technical system and the controller, and wherein a transformation function that reduces and/or obfuscates the first operating data is provided to the controller by the operator (BTS).

8. The method as claimed in claim 1 , said method further comprising:
training, in the second access domain using the first operating data, an initial model; and
splitting, in the second access domain, the initial model into a first partial function and a second partial function, wherein the first partial function is a transformation function configured to reduce and/or obfuscate the first operating data and wherein the second partial function is the control model.

9. The method as claimed in claim 1 , wherein the control model is trained on the basis of data that are independent of the technical system.

10. The method as claimed in claim 1 , wherein the control model is generated by the model generator based on the first operating data having been previously transformed in the first access domain by a transformation function that reduces and/or obfuscates the first operating data after the transformation was generated and trained by the controller based on the first operating data.

11. A computer system comprising one or more processors, one or more memories, a computer readable hardware storage device having computer readable program code stored therein, said program code executable by the one or more processors via the one or more memories to implement a method for controlling a technical system based on a control model, said method comprising:
generating, by a model generator residing in a second access domain, the control model as a function of first operating data received from the technical system residing in a first access domain that is separated from the second access domain, said control model configured to output control data upon being executed using second operating data that was transformed by being reduced and/or obfuscated;
receiving, by a model execution system residing in the second access domain from a controller residing in the first access domain, the transformed second operating data;
executing, by the model execution system using the transformed second operating data, the control model, said executing the control model comprising outputting the control data; and
transmitting, by the model execution system to the controller, the control data, wherein the control data is configured to control the technical system by the controller,
wherein the model generator, the model execution system and the controller each comprise one or more processors for carrying out all method steps performed by the model generator, the model execution system, and the controller, respectively.

12. The computer system as claimed in claim 11 , wherein the control model is generated by the model generator based on the first operating data having been transformed by a transformation function that reduces and/or obfuscates the first operating data after the transformation was generated and trained by the model generator based on the first operating data.

13. The computer system as claimed in claim 11 , wherein the control model is generated by the model generator based on the first operating data having been previously transformed in the first access domain by a transformation function that reduces and/or obfuscates the first operating data after the transformation was generated and trained by the controller based on the first operating data.

14. A computer program product, comprising a computer readable hardware storage device having computer readable program code stored therein, said program code executable by one or more processors of a computer system to implement a method for controlling a technical system based on a control model, said method comprising:
generating, by a model generator residing in a second access domain, the control model as a function of first operating data received from the technical system residing in a first access domain that is separated from the second access domain, said control model configured to output control data upon being executed using second operating data that was transformed by being reduced and/or obfuscated;
receiving, by a model execution system residing in the second access domain from a controller residing in the first access domain, the transformed second operating data;
executing, by the model execution system using the transformed second operating data, the control model, said executing the control model comprising outputting the control data; and
transmitting, by the model execution system to the controller, the control data, wherein the control data is configured to control the technical system by the controller,

wherein the model generator, the model execution system and the controller each comprise one or more processors for carrying out all method steps performed by the model generator, the model execution system, and the controller, respectively.

15. The computer program product as claimed in claim 14 , wherein the control model is generated by the model generator based on the first operating data having been transformed by a transformation function that reduces and/or obfuscates the first operating data

Assignees

• Siemens Ag

Inventors

• Heesche Kai
• Schneegass Daniel

Classifications

• G06N3/04

  Architecture, e.g. interconnection topology · CPC title

• G06K9/6232

  Physics · mapped topic

• G06N3/08

  Learning methods · CPC title

• G05B13/0265 Primary

  the criterion being a learning criterion · CPC title

• G06N3/0455

  Auto-encoder networks; Encoder-decoder networks · CPC title