Der Blätterkatalog benötigt Javascript.
Bitte aktivieren Sie Javascript in Ihren Browser-Einstellungen.
The Blätterkatalog requires Javascript.
Please activate Javascript in your browser settings.
computerautomation de sps 5 Messe Aktuell Mittwoch 13 NoveMber Multifunktionales Elektronikgehäusesystem Volle Funktionalität an der Gehäusefront Verteilergehäuse Schaltschrankgehäuse E Dold Söhne GmbH Co KG 78120 Furtwangen | Tel 07723 6540 | doldrelays@dold com www dold com Elektronikgehäuse KV 4600 Modulbaubreiten ab 12 5 mm und verschiedene Bautiefen erhältlich Hohe Signaldichte durch den Einsatz von kompakten Push-In-Klemmen Gängige Buskonzepte integrierbar Leiterplattenfläche bis zu 9500 mm² erlaubt vielfältige Anwendungen Schnelle Montage durch vereinfachten Gehäuseaufbau Hohe mechanische Stabilität und Vibrationsfestigkeit Halle 9 | Stand 331 AI-Based Continuous Monitoring Meinrad Happacher Each connected module now includes integrated AI functions while this is standard in today’s smartphones it remains rare in automation Yet such AI could serve as a foundation to detect communication issues caused by cyberattacks in real time Years ago the practice of continuously transmitting sensor data to the cloud for periodic data analysis was introduced to assess the current state of processes and systems Nowadays complex AI algorithms are often employed in these analyses a practice that has led to the coining of the term »Continuous Monitoring « With current capabilities AI can now be integrated directly into sensors creating onsite selfmonitoring intelligence in realtime – this form of AI in end devices is known as »Edge AI« Creating Availability Resilience Through a digitalized Continuous Monitoring application connected modules within an OT network segment can be continuously monitored to detect potential communication issues from cyberattacks or other causes in real time using AI For this purpose a neural network is created using machine learning and is contextually trained to recognize »normal« data traffic or the standard state of communication relationships within the target segment The result is a platformindependent AI model that can detect anomalies with relatively low resource needs and can be implemented directly within an Intrusion Detection System IDS sensor Such an IDS soft sensor often doesn’t even require dedicated hardware for runtime and can be integrated as an embedded software component within existing edge systems Due to their modular overall design AI models are relatively easy to expand to include additional data points for anomaly detection For example further hardware sensors can be installed and integrated into the OT environment The outcome is an enhanced resilience of OT network applications against targeted manipulation and environmental disruptions leading to greater overall availability and measurable benefits in practice Testbed for Data Evaluation The design implementation and field testing of AI functions for a Continuous Monitoring application demand a crossfunctional collaborative and iterative process with experts across various roles The effectiveness of the final solution is directly proportional to the expert knowledge and the level of cooperative teamwork during field testing The biggest challenge is gathering sufficient highquality data for machine learning Without enough quality data for a specific use case results may be inaccurate or unusable Data evaluation testbeds can help achieve robust results Today every connected module should ideally include integrated AI functions This has already been achieved with new smartphones but automation technology has a long way to go With the right tools and testing methods integrated monitoring functions with significant user benefits can already be implemented Investments in such digital applications are modest and can be offset by the increased availability of technical systems Hall 6 Booth 241A Klaus-Dieter Walter CEO of SSV Software Systems Im ag e SS V »Every connected module today should include integrated AIbased monitoring functions Investments in such digitalization applications are manageable «