Interface Firmware for Edge ML Pipelines in Industrial Automation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increased load on communication paths in industrial automation systems due to the transmission of process data from field devices to controllers and edge devices leads to higher cycle times and latency, with existing solutions like more efficient field buses or shifting AI processing to field devices resulting in increased costs and complexity.
Innovation Solution
Implementing machine learning applications in interface devices by dividing them into logical components and generating code blocks that are connected in a pipeline, allowing for data aggregation and reduction of data transfer, with virtual ports enabling efficient output transmission to control devices, abstracting machine learning functionality for simplified engineering.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If AI applications are run in controllers or edge devices via traditional communication paths, then machine learning functionality is achieved, but communication load increases and cycle time increases
Solution Approach 1:
The patent segments the automation system by placing AI processing capabilities directly in field devices rather than centralizing them in controllers or edge devices. This distributed segmentation allows local processing of sensor data, reducing communication load and cycle time while maintaining AI functionality.
Solution Approach 2:
The patent introduces a new dimension to the automation architecture by embedding machine learning models within field devices themselves. This dimensional shift from centralized to distributed processing enables AI operations to occur at the source of data generation, eliminating communication delays.
2Loss of time
If AI applications are executed on field devices, then communication load is reduced, but device complexity increases and integration complexity increases
Solution Approach 1:
The patent implements a universal AI processing framework that can be deployed across different field devices regardless of their native programming languages or platforms. The standardized interface and model deployment mechanism enable multi-functionality, allowing the same AI capabilities to operate on diverse devices without increasing individual device complexity.
Solution Approach 2:
The patent introduces an intermediary layer consisting of standardized communication protocols and interfaces between field devices and the control system. This mediator handles the complexity of AI model deployment and data exchange, shielding individual field devices from complex integration requirements while enabling reduced communication latency.
3Adaptability or versatility
If different AI engineering frameworks are used for different field devices, then device-specific requirements are met, but engineering complexity increases
Solution Approach 1:
The patent creates a universal AI engineering framework that can deploy machine learning models to field devices with different native languages and platforms. The standardized model format and deployment interface provide versatility across devices while maintaining simplicity in the engineering process, eliminating the need for device-specific frameworks.
Data Source
AI summary
A method for configuring an interface device connected to a control device and a field device, wherein the method includes receiving a first machine learning application having a plurality of logical components connected in a pipeline, where the first machine learning application serves to analyze a signal from the field device utilizing a first machine learning model, generating a plurality of code blocks utilizing a translator based on the plurality of logical components of the first machine learning application, connecting the plurality of code blocks in accordance with the pipeline of the first machine learning application to generate a first output from the signal from the field device, and deploying the connected code blocks on firmware of the interface device including creating a virtual port connectable to the control device, and where the virtual port serves to transmits the first output to the control device.