Tu-SML-06 - Generation and Real-Time Detection of High-Fidelity Odour Plumes

Event

EUROSENSORS 2026
2026-09-06 - 2026-09-09
Zurich

Band

Poster

Chapter

Signal Analysis, Machine Learning, Physics-Informed Machine Learning And Artificial Intelligence For Sensing

Author(s)

A. Gholivand, A. Padmanabhan, T. Ackels - University of Bonn,Bonn (Germany), F. P. Schmidt - Bielefeld University,Bielefeld (Germany)

Pages

706 - 707

DOI

10.5162/eurosensors2026/Tu-SML-06

ISBN

978-3-910600-12-6

Price

free

Abstract

Olfactory navigation in animals relies on rich temporal information embedded within turbulent, highly intermittent odour plumes. Replicating and measuring these rapid concentration dynamics in laboratory settings presents significant technical challenges. Here, we introduce a unified platform for the generation and real-time detection of high-fidelity odour plumes. The temporal Odour Delivery Device (tODD) generates complex, naturalistic odour patterns with millisecond resolution. For real-time detection, we developed a Micro-Chemical Sensor (MiCS) coupled with a WaveNet deep learning decoder. The model learns the non-linear mapping between raw MiCS responses and Photoionisation Detector (PID) references, deconvolving sensor inertia to reconstruct high-frequency "whiff" and “blank” dynamics. This integrated system provides an end-to-end framework for investigating dynamic sensory information encounters during active olfactory navigation.