Modular AI Accelerator for PLC Real-Time Control Integration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional Programmable Logic Controllers (PLCs) lack the necessary computing resources and design to perform advanced AI algorithms, requiring a flexible solution that can scale with performance requirements and integrate with existing control strategies, while also being retro-fit able and compatible with established tools like Siemens Total Integrated Automation (TIA) Portal.
Innovation Solution
An AI acceleration module comprising a CPU module, technology modules with AI accelerator processors, and a backplane bus, allowing for the execution of machine learning models directly within the PLC, enabling high-speed input processing and synchronous AI-based decision-making, with the ability to connect various sensors and devices through standard industrial interfaces.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional PLCs are used for control systems, then reliability and real-time guarantees are maintained, but computing resources are insufficient for advanced AI algorithms
Solution Approach 1:
The system is divided into a conventional PLC for control functions and a separate AI accelerator module for machine learning computations. The PLC maintains reliability and real-time guarantees while the AI accelerator provides additional computing power through a modular architecture that can be independently configured and scaled.
Solution Approach 2:
A communication interface acts as an intermediary between the PLC and AI accelerator, enabling data exchange and coordination. This intermediary layer allows the two systems to work together seamlessly, with the PLC handling control tasks and the AI accelerator processing complex algorithms without compromising either system's performance or reliability.
2Adaptability or versatility
If AI algorithms are added to PLCs, then advanced data analytics capability is improved, but integration complexity increases
Solution Approach 1:
The AI accelerator is designed as a universal module that can be integrated with multiple PLC models and support various machine learning frameworks. Standardized interfaces and communication protocols enable the same AI accelerator to work with different PLC systems, reducing integration complexity while maintaining advanced data analytics capability across diverse applications.
Solution Approach 2:
The system uses standardized interface definitions and communication protocols that can be copied and reused across different PLC-AI integrations. This allows the integration architecture to be replicated without reinventing the wheel, significantly reducing the complexity of integrating AI capabilities into PLC systems.
3Productivity
If custom hardware is designed for AI acceleration, then processing performance is improved, but retrofit capability is reduced
Solution Approach 1:
The AI accelerator is designed as a dynamic, modular component that can be added to or removed from existing PLC systems without permanent modifications. This modular approach allows the system to adapt its processing performance based on needs while maintaining retrofit capability, as the AI accelerator can be installed in available slots or mounted externally without custom hardware design.
Solution Approach 2:
The system allows processing performance to be adjusted by changing configuration parameters of the AI accelerator rather than modifying hardware. Different machine learning models, algorithms, and processing modes can be loaded and unloaded dynamically, enabling performance optimization without physical hardware changes, thus preserving retrofit capability.
4Power
If AI processing is performed in cloud infrastructure, then computing power is improved, but real-time response capability deteriorates
Solution Approach 1:
The AI accelerator is pre-configured with machine learning models and algorithms locally at the edge device. This preliminary preparation allows the system to perform complex AI computations immediately without needing to query remote cloud infrastructure, thereby maintaining real-time response capability while providing cloud-level computing power through the locally deployed AI accelerator.
Data Source
AI summary
A controller system includes a CPU module, one or more technology modules, and a backplane bus. The CPU module comprises a processor executing a control program. The technology modules include an artificial intelligence (AI) accelerator processor configured to (a) receive input data values related to one or more machine learning models, and (b) apply the machine learning models to the input data values to generate one or more output data values. The backplane bus connects the CPU module and the technology modules. The technology modules transfer the output data values to the processor over the backplane bus and the processor uses output data values during execution of the control program.