From sensors to smart systems: The rise of the AI-powered photonic nose

Detection of complex chemical odors and gas mixtures is essential for environmental safety, medical, and food quality control, but traditional gas sensors often face problems such as limited selectivity, sensor drift, and slow response. A new generation of photonic noses – optical sensing systems inspired by the human sense of smell – offer innovative solutions. By combining advanced optical sensing technology and artificial intelligence, the photonic nose can capture detailed chemical fingerprints and interpret them with high precision. These systems leverage light-matter interactions and machine learning algorithms to achieve fast, label-free, and sensitive detection of volatile compounds, paving the way for smarter and more reliable sensing platforms that can operate in complex real-world environments.

Traditional electronic noses rely on numerous chemical sensors, whose electrical responses are often affected by humidity, temperature fluctuations, and long-term drift. Although these systems are in practical use, their performance limitations are important when detecting trace gases or complex mixtures. In contrast, optical sensing technology offers unique advantages such as higher sensitivity, better stability, and richer information content through spectral signals. However, interpreting these high-dimensional optical signals remains difficult, especially in dynamic or noisy environments. Based on these challenges, there is an urgent need to develop integrated sensing systems that combine optical detection with advanced data processing capabilities to enable accurate, real-time, and robust chemical analysis.

Comprehensive review published in (DOI: 10.1038/s41378-025-01058-3) Microsystems and nanoengineering In 2025, researchers at Northwestern Technological University will systematically investigate the evolution of photonic nose technology and its integration with artificial intelligence. This article analyzes how optical sensing techniques, from colorimetric and refractive index sensors to spectroscopy, are being enhanced by machine learning and cloud-to-edge computing architectures. This study outlines how, by bridging photonic hardware and intelligent algorithms, photonic noses can move from laboratory prototypes to compact, intelligent microsystems capable of real-time chemical sensing across a variety of application domains.

This review focuses on the four core optical sensing mechanisms that underpin photonic nose systems: colorimetric sensing, refractive index modulation, optical absorption, and spectroscopy. These techniques allow photonic noses to generate rich multidimensional optical features in response to chemical analytes. Artificial intelligence plays a central role in converting these signatures into meaningful information. Machine learning algorithms can automatically extract subtle spectral features, correct for sensor drift, suppress noise, and disentangle overlapping chemical signals that are difficult to resolve using traditional methods.

The authors further describe three intelligence paradigms that form the modern photonic nose. With post-sensing intelligence, data is analyzed using advanced learning models after acquisition to improve accuracy and selectivity. Cloud intelligence enables large-scale data aggregation, distributed sensing networks, and continuous model updates across multiple locations. Edge intelligence pushes computations directly to photonic chips or nearby processors, enabling real-time decision-making with minimal latency and power consumption. Together, these architectures transform the photonic nose from a passive detector into an autonomous, intelligent system that can learn, adapt, and operate reliably in complex environments.

According to the authors, the convergence of photonics and artificial intelligence marks an important turning point in chemical sensing. They highlight that the AI-driven photonic nose can go beyond simply detecting gases to actively interpreting complex chemical situations, similar to biological olfactory systems. These technologies can deliver faster response, greater robustness, and scalable deployment by integrating sensing, computation, and communication into integrated microsystems. The researchers note that such systems are particularly valuable in scenarios where traditional sensors fail, offering new possibilities for autonomous monitoring and intelligent decision-making in real-world settings.

AI-powered photonic noses are poised to impact a wide range of fields. In environmental monitoring, a network of compact photonic noses could provide continuous high-resolution mapping of air pollutants and harmful gases. In the medical field, non-invasive breath analysis may enable early detection of diseases by identifying volatile biomarkers. In agriculture and food safety, photonic noses can sensitively monitor ripening, spoilage and contamination, even in humid or complex conditions. Looking to the future, continued advances in photonic integration, low-power AI hardware, and data-driven algorithms are expected to accelerate the deployment of intelligent photonic noses as ubiquitous sensing tools in smart cities, precision medicine, and sustainable food systems.

###

References

Toi

10.1038/s41378-025-01058-3

Original source URL

https://doi.org/10.1038/s41378-025-01058-3

Funding information

This research was supported by the Fundamental Research Fund for the Central Universities of China (Grant No. G2025KY05053) and the Key Research and Development Program of Shaanxi Province (Grant No. 2024GX-YBXM-193).

About Microsystems and nanoengineering

Microsystems and nanoengineering is an online-only, open-access, international journal dedicated to the publication of original research results and reviews on all aspects of micro- and nano-electromechanical systems, from basic to applied research. This journal is published by Springer Nature in partnership with the Institute of Aerospace Information, Chinese Academy of Sciences, and with support from the National Key Laboratory of Transducer Technology.