Gas Sensors: The Role of Hardware in the AI Age
ABSTRACT
Gas sensing is essential for applications ranging from air-quality monitoring and precision agriculture to energy-asset protection and non-invasive medical diagnostics. However, legacy gas sensors rely on principles established decades ago. Originally designed as safety alarms to detect high concentrations of gases, these legacy sensors are simply miniaturized rather than fundamentally reimagined. These single-output (“zero-order”) devices struggle to distinguish between an analyte (target) and interferent gases and cannot manage signal drift. Relying on machine learning (ML) processing to compensate for these inherent hardware limitations is a common but insufficient strategy; high-fidelity analytics require high-information-density inputs.
In this talk, we present a roadmap for next-generation, multi-output gas sensors. Drawing from the mathematical principles of traditional analytical instruments such as gas chromatography and mass spectrometry, we demonstrate how to re-architect sensor hardware to function as first- and second-order analytical devices. We outline four pillars of "information-rich" design: (1) Transducer Physics: Engineering architectures with multi-dimensional response outputs. (2) Dynamic Excitation: Moving beyond static measurements to increase data dimensionality. (3) Material Engineering: Tailoring sensing films for selective, multi-analyte interactions under variable excitation conditions. (4) Edge-ML/AI Integration: Enabling real-time, on-device multivariate analytics and utilization of artificial intelligence (AI) concepts to reason and adapt in complex environments. Through several case studies, this talk will demonstrate the performance breakthroughs achieved by applying these design principles. Attendees will gain a system-level understanding of how to transition from legacy, "low-cost commodity" sensors to high-performance, differentiated devices, enabling more trusted and widespread environmental and industrial monitoring.
BIOGRAPHY
Dr. Radislav Potyrailo is a Principal Scientist at GE Global Research in USA. He is leading the growth of GE’s industrial, wireless, and wearable sensors. His interests are in inventing new sensing systems and bringing them from lab feasibility studies, to field validation, and to commercialization.
Some of these results Radislav summarized in 120+ granted US Patents and numerous publications. He has been leading development of sensing technologies for GE Healthcare, Water, Security, Corporate Environmental, Consumer & Industrial, Energy, Transportation, Power, Oil & Gas, and other GE businesses. Radislav has been also serving as a Principal Investigator on US Government programs funded by NIH, AFRL, DARPA, TSWG, DHS, NETL, and NIOSH. His recent recognitions include SPIE Fellow and Prism Award by Photonics Media/SPIE.