Series: Designing Edge AI systems

This three-part series by Ankur Tomar of Farnell treats Edge AI as an architectural choice, not a processor spec. Article One asks where intelligence should sit — sensor, controller, gateway, or Cloud — starting from the decision a system needs to make. Article Two shows AI-enabled sensors turning raw vibration, video, and environmental data into local, contextualised alerts that cut Cloud traffic without losing key evidence. Article Three argues IO-Link, despite its bandwidth limits, is well suited to carrying these compact payloads across brownfield and greenfield deployments. The throughline: once a device can interpret its own readings, the real question becomes what data still needs to move, and where.