IEC 61499 AI Function Block Linking for Industrial Controllers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The integration of artificial intelligence into industrial controllers is hindered by the numerous architectures and variants of AI systems, leading to incompatibilities and high implementation efforts, which complicate the process and increase production downtimes.
Innovation Solution
A method that automates the selection and integration of AI function blocks into industrial machine controllers using an IEC 61499 runtime environment, allowing for the graphical or code-level linking of AI models with other function blocks, reducing the need for operator expertise in software libraries and communication configuration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If artificial intelligence is integrated into industrial controllers using multiple architectures and variants, then the functionality and capabilities of the control system are improved, but the complexity of integration and implementation increases significantly
Solution Approach 1:
The patent implements a universal AI runtime environment that can execute multiple AI architectures and variants (neural networks, SVMs, decision trees, etc.) through a single standardized interface. This allows the control system to support diverse AI functionalities without requiring separate integration processes for each architecture, thereby resolving the contradiction between versatility and integration complexity
Solution Approach 2:
The patent introduces an AI runtime environment as an intermediary layer between the control software and various AI models. This runtime environment handles the complexity of AI integration internally while presenting a simplified interface to operators, effectively mediating between the diverse AI architectures and the unified control system to reduce integration complexity
2Measurement precision
If manual integration of AI function blocks is performed, then precise control over the integration process is achieved, but the time required for integration and the need for specialized operator knowledge increase
Solution Approach 1:
The patent implements automated procedures that perform preliminary actions during the AI function block integration process. The system automatically selects suitable AI models, configures runtime environments, and establishes connections between function blocks without requiring manual intervention for each step, thereby reducing integration time while maintaining control through the structured automated process
Solution Approach 2:
The AI runtime environment is designed to be self-configuring and self-integrating. When an operator adds an AI function block, the system automatically handles model selection, environment setup, and connection establishment without requiring the operator to manually configure software libraries or communication protocols, thus reducing both time and expertise requirements while maintaining integration control
3Reliability
If comprehensive AI model selection is offered, then the suitability and performance of AI solutions are improved, but the effort required to select and configure the appropriate model increases
Solution Approach 1:
The patent implements automated feedback mechanisms where the AI runtime environment analyzes the control task requirements and automatically recommends or selects the most suitable AI model from the available variants. The system provides feedback to the operator about suitable model choices based on the specific application requirements, thereby maintaining high model suitability while reducing the manual effort required for selection and configuration
Data Source
AI summary
The invention relates to a method of integrating at least one AI function block, which comprises artificial intelligence, into a controller for an industrial machine. In the method, for the AI function block, an AI model is selected for execution in the AI function block, the AI function block is linked in an at least partly automated manner to further function blocks of control software of the industrial machine and the AI function block is brought to execution in an at least partly automated manner, wherein the execution of the AI function block takes place by means of an IEC 61499 runtime environment.