ctrlX CORE platform is expanding! The new AI Module extends the functionality of ctrlX CORE, allowing for seamless integration of Artificial Intelligence (AI) into everyday automation tasks.
AI-powered control
[table align="center"]
ctrlX COREplus with AI module
ctrlX COREplus X7 with AI module
The key component of the AI module is the Hailo-8 M.2 module. Optimized for computing AI algorithms at the edge, this powerful AI accelerator delivers 26 TOPS (Tera Operations Per Second) of processing power, enabling fast and efficient data processing at the source. Its small size and energy efficiency make it an excellent entry point for AI-based vision and other intelligent automation applications.
End-to-end software solution
ctrlX AUTOMATION offers an end-to-end software solution, from data acquisition to model training and deployment in the production environment. This process is supported by the LumiScan Suite from ctrlX WORLD partner HD Vision Systems, a specialist in industrial image processing and AI solutions.
The LumiScan Apps offer intuitive operation and ready-to-use application modules for common image processing tasks (e.g., object recognition), making it easier for users to get started with AI-powered automation.
[accordion]
[accordion-item title="1. Collect data"]
This is where the datasets are generated. Images are acquired through the camera connected to the ctrlX CORE.
The LumiScan Camera Driver is a comprehensive image acquisition application.
[accordion-item title="2. AI model training"]
The acquired images are reviewed, cleaned, and annotated, highlighting the key areas. Based on this data, the AI model can be generated automatically, even without specialized AI expertise.
This process takes place on a dedicated engineering PC.
The LumiScan Annotation Framework is versatile software that allows you to annotate objects in an image, providing data for training and retraining neural network models.
The LumiScan Training Framework enables easy training of neural networks for computer vision tasks, including object recognition.
[accordion-item title="3. Load the AI model onto ctrlX CORE and run it"]
The trained AI model is loaded onto the ctrlX CORE's AI Module and put into action. It detects any deviations from its learned parameters, immediately saving the results to the central ctrlX Data Layer.
Now, any PLC app or ctrlX OS app can access this data from the ctrlX Data Layer and react accordingly. This seamlessly integrates AI capabilities into standard PLC automation workflows.
The LumiScan AI Object Matching is an application that uses neural networks to find objects in images.
It returns the position and orientation of an object in pixel coordinates.
The LumiScan Calibration is an application that improves the accuracy of position and orientation information for objects found in the scene by scanning a mapping provided as a TIFF file.
[/accordion]
Benefits at a Glance:
Cost-effective and scalable AI functionality without additional AI hardware.
Seamless integration into existing control hardware simplifies system maintenance and update management.
The central ctrlX Data Layer enables transparent and centralized access to all machine data – without additional programming effort.
Reliable and intuitive user interface.
Out-of-the-box application modules for the most common image processing tasks, such as object detection and localization.
Highest level of cybersecurity thanks to integration with the ctrlX OS: strong password policies, up-to-date software, secure update deployment, and comprehensive authentication and authorization protect data (Link to ctrlX OS Cybersecurity Article).
What about AI experts? ctrlX CORE, with its open ctrlX OS operating system, provides the flexibility to integrate custom AI models, allowing optimal adaptation to specific individual needs.
Typical applications: AI for complex decision-making
AI proves especially valuable when the variability of the data being analyzed is high, making it nearly impossible to program the machine for every potential outcome.
A common example is the visual inspection of parts on a production line: When visually evaluating a part, the wide range of possible appearances must be considered. The concept is for the machine to learn from examples and make decisions autonomously.
Quality inspection
Checking if all parts are assembled
Surface inspection
2D Position recognition for synchronous conveyor belts
Pick & Place
Robot Positioning
Shape recognition
Recognition of specific component shapes in the production line
[gallery]
Solution Set for Quality Inspection | CTRLX891938[link-button type="secondary" text="Open configuration" target="_blank" href="https://www.boschrexroth.com/ctrlx-configurator/#/solution/load/CTRLX891938"]