Hardware–Analytics Co-Design for Trust-Enhanced Gas Sensing
Gas sensing is becoming critical for applications ranging from gas speciation of indoor and outdoor air-quality to monitoring of energy-assets, precision agriculture, noninvasive medical diagnostics, and security. However, conventional legacy gas sensors rely on single-output designs that struggle with drift and detection of low-concentration analytes in chemically complex environments. In this lecture we will present a roadmap toward next-generation multi-output gas sensors: inspired by proven traditional analytical instruments such as gas chromatography and spectroscopy, we will show how their underlying mathematical principles can be translated into new sensing designs to improve selectivity, stability, and reliability. The central focus of this lecture is understanding of importance of hardware–analytics co-design. Machine learning (ML) cannot fully succeed when sensor outputs contain insufficient independent information. Thus, emerging multi-output sensor architectures - enabled by independent excitation variables, advances in electronics, and edge analytics - create the information richness needed for multi-analyte discrimination, drift correction, and robust calibration. Using multiple examples from research teams worldwide, we will connect foundational principles to practical design strategies and illustrate commercialization pathways. We will provide a systems-level understanding of how to design multi-output gas sensors to achieve trusted performance in chemically complex environments.
Date and Time
Location
Hosts
Registration
-
Add Event to Calendar
- University of Wollongong
- School of Engineering
- Wollongong, New South Wales
- Australia
- Building: Faculty of Engineering and Information Sciences
- Room Number: 6.105
Speakers
Radislav of GE Vernova Advanced Research Center, Niskayuna NY, USA
Hardware–Analytics Co-Design for Trust-Enhanced Gas Sensing
Gas sensing is becoming critical for applications ranging from gas speciation of indoor and outdoor air-quality to monitoring of energy-assets, precision agriculture, noninvasive medical diagnostics, and security. However, conventional legacy gas sensors rely on single-output designs that struggle with drift and detection of low-concentration analytes in chemically complex environments. In this lecture we will present a roadmap toward next-generation multi-output gas sensors: inspired by proven traditional analytical instruments such as gas chromatography and spectroscopy, we will show how their underlying mathematical principles can be translated into new sensing designs to improve selectivity, stability, and reliability. The central focus of this lecture is understanding of importance of hardware–analytics co-design. Machine learning (ML) cannot fully succeed when sensor outputs contain insufficient independent information. Thus, emerging multi-output sensor architectures - enabled by independent excitation variables, advances in electronics, and edge analytics - create the information richness needed for multi-analyte discrimination, drift correction, and robust calibration. Using multiple examples from research teams worldwide, we will connect foundational principles to practical design strategies and illustrate commercialization pathways. We will provide a systems-level understanding of how to design multi-output gas sensors to achieve trusted performance in chemically complex environments.
Biography:
Radislav Potyrailo is a Senior Principal Scientist at GE Vernova Advanced Research, leading programs on design of sensor systems and bringing innovations from feasibility to commercialization. His research interests are in design of physical transducers and materials with multi-response mechanisms, multivariate edge-data analytics, and system engineering. Radislav served as a lead on numerous R&D programs transitioned to GE businesses or GE partners for commercialization. Radislav has been a Principal Investigator on US Government programs funded by AFRL, ARPA-E, DARPA, DHS, DOE, DTRA, JPEO, NIH, NIOSH, TSWG, and other agencies. Examples of his systems as recognized by international industrial awards include sensors for GE Water and for GE Oil & Gas. He has 165+ granted US Patents and numerous publications (Google Scholar h-index 55). He serves as an editor of the Springer-Nature book series “Integrated Analytical Systems”. Radislav is the Chair of the Device Working Group of the MEMS and Sensors Industry Group and the North America Chair of International Society for Olfaction and Chemical Sensing. He is SPIE Fellow and IEEE Fellow, covering the whole electromagnetic spectrum of his sensors. Radislav is 2024-2026 Distinguished Lecturer of the IEEE Sensors Council.
Email:
Address:New York, United States