Docent Lecture: From Efficient Edge AI to Reliable Instrumentation
Welcome to listen, when researcher Irida Shallari gives her docent lecture on the occasion of appointment as Associate Professor in Embedded Systems, Electronics.
Date: Tuesday, 8 September 2026
Time: 15:30-16:30
Place: Room C326, Campus Sundsvall, Mid Sweden University, and online via Zoom
Language: English
From Efficient Edge AI to Reliable Instrumentation
A smart sensor does more than collect information: it can interpret what it sees, communicate its findings, and help us make decisions.
This lecture will present how different sensors, such as cameras, stereo systems, LiDAR and temperature sensors, can be combined with computing and artificial intelligence in small devices located close to where data are produced.
These devices must work with limited energy, memory and processing power, while still providing dependable measurements in difficult conditions such as vibration, interference and other environmental factors.
The research therefore considers not only how quickly and accurately a system performs its tasks, but also how consistent, precise and trustworthy its measurements are.
Looking ahead, the aim is to establish a metrology-informed approach to edge AI, in which uncertainty is quantified and managed as part of the system design. This will allow sensors and edge AI systems to be optimized not only for performance, but also for trustworthiness and stability.