Lecture
From Visual Workflow to CRA-Ready Edge Deployment
- at -
- Hall C5C5.100
- Language: English
- Type: Lecture
Lecture description
"Industrial AI prototypes become products when they can run alongside production equipment, recover from faults, and maintain reliability over years or decades of updates. Ensuring that they also do so securely is precisely what the Cyber Resilience Act (CRA) was created to enforce.
This session uses a SECO production line monitoring demonstrator to examine how systems can address vulnerabilities through security updates and maintain alignment with Annex I, Part I, point 2(c) of the EU Cyber Resilience Act. Using a SMARC computer-on-module-based Modular Vision HMI running a vision inference and control workflow developed in Clea Studio AI, we will follow a change from a controlled Clea OS production image’s software bill of materials through vulnerability assessment, update packaging, deployment, and rollback. We will also examine how time-critical control can be separated from AI processing when the device is in an offline state.
Attendees will leave with a practical model for making updatable, supportable industrial edge products that align with CRA guidelines, as well as the ability to assess whether a vulnerability can be corrected across deployed devices."
This session uses a SECO production line monitoring demonstrator to examine how systems can address vulnerabilities through security updates and maintain alignment with Annex I, Part I, point 2(c) of the EU Cyber Resilience Act. Using a SMARC computer-on-module-based Modular Vision HMI running a vision inference and control workflow developed in Clea Studio AI, we will follow a change from a controlled Clea OS production image’s software bill of materials through vulnerability assessment, update packaging, deployment, and rollback. We will also examine how time-critical control can be separated from AI processing when the device is in an offline state.
Attendees will leave with a practical model for making updatable, supportable industrial edge products that align with CRA guidelines, as well as the ability to assess whether a vulnerability can be corrected across deployed devices."
