Invited talk at Eurosensors 2026
At Eurosensors 2026 in Zürich, Jordi Fonollosa presented an invited talk exploring how artificial intelligence is changing the way intelligent sensing systems are designed and deployed. The presentation, “Mitigation of Chemical Sensor Drift Through Calibration Transfer in Edge AI”, examined the evolution of sensor calibration from a static laboratory procedure towards adaptive, deployable and reusable intelligence.
Low-cost chemical sensors offer significant opportunities across environmental monitoring, health, industrial applications and other fields. However, their practical deployment remains challenging: sensor responses vary between devices, change over time, depend on environmental conditions and can be affected by cross-sensitivity to background compounds. As a result, measurements that appear comparable in controlled laboratory conditions can become difficult to compare reliably in real-world deployments.
The presentation highlighted that advances in hardware, optimized machine-learning runtimes, quantization and compression are expanding the range of models that can be considered for embedded sensing applications. At the same time, increasing model capacity comes with additional computational requirements, making the trade-off between predictive performance and system-level constraints an important part of model selection.
