AI-based Neural Networks for Gas-Sensor Sensitivity and Response-Time Prediction
Article Sidebar
-
Artificial Intelligence, Gas Sensors, Early Fusion Model, Individual Neural Models, Late Fusion Model, Response Time, Sensitivity
Abstract
In this study, three types of gas sensors were fabricated using thin-film technology: tin oxide doped with tungsten (SnO₂+WO₃), tin oxide doped with zinc oxide (SnO₂+ZnO), and tungsten oxide doped with zinc oxide (WO₃+ZnO). These sensors were synthesised with multiple doping ratios and operated at two temperatures (250 °C and 350 °C). They were exposed to acetone and ethanol vapours at a concentration of 500 ppm, and measurements were recorded for sensitivity response (S) and response time (t). Three neural network models were developed using artificial intelligence to predict sensitivity and response time: individual neural models for each sensor; an early fusion model that combines inputs and outputs into a unified network; and a late fusion model that separates each task into an independent subnetwork. The models were trained using MATLAB’s nntraintool and evaluated using quantitative metrics (MAE, RMSE, R², MAPE) and training/visual indicators such as performance curves, gradient, mu, error histogram, and regression plots. Randomisation was fixed to ensure consistent data distribution across training, validation, and testing sets, enabling fair comparisons between the models under equal experimental conditions. The results showed that the individual model for the SnO₂+WO₃ sensor achieved the highest accuracy in sensitivity prediction, while the WO₃+ZnO model excelled in response time estimation. The late fusion model demonstrated the most balanced and reliable performance, with the lowest error rates and highest correlation coefficients, confirming its strong generalisation capability. In contrast, the early fusion model showed good training performance but limited generalisation, particularly in predicting response time. This study presents a novel framework for intelligent prediction of gas sensor behaviour, combining experimental validation with neural modelling. It offers a valuable contribution to the development of accurate and generalisable sensing systems for industrial and smart environments.
Full text article
References
Narkhede, Parag., Walambe, Rahee., Mandaokar, Shruti., Chandel, Pulkit., Kotecha, Ketan., & Ghinea, George. (2021). Gas Detection and Identification Using Multimodal Artificial Intelligence Based Sensor Fusion. Applied System Innovation, 4(1), 3. https://doi.org/10.3390/asi4010003
Walambe, Rahee., Narkhede, Parag., Mandaokar, Shruti., et al. (2023). Gas Detection and Identification Using Multimodal AI-Based Sensor Fusion. Data Science Week Poster Presentation.
Zhang, Yue., Li, Hong., Chen, Xiaohui., & Wang, Jian. (2024). Gas Detection and Classification Using Multimodal Data Based on Deep Learning. Sensors, 24(18), 5904. https://doi.org/10.3390/s24185904
Chakraborty, Sourav., Mittermaier, Stefan., & Carbonelli, Claudio. (2022). Understanding the Behavior of Gas Sensors Using Explainable AI. Engineering Proceedings, 27(1), 61. https://doi.org/10.3390/ecsa-9-13350
Zhuang, Yuxuan., Yin, Dong., Wu, Lei., Niu, Guanghui., & Wang, Feng. (2024). A deep learning approach for gas sensor data regression: Incorporating surface state model and GRU-based model. APL Machine Learning, 2(1), 016104. https://doi.org/10.1063/5.0160983
Attallah, Omar., & Morsi, Islam. (2024). Multitask Deep Learning-Based Pipeline for Gas Leakage Detection via E-Nose and Thermal Imaging Multimodal Fusion. Chemosensors.
Liu, Ming., & Zhao, Qiang. (2024). 2024 Breakthroughs in Smart Gas Sensor Technology: A Review. Gas Detection Journal. Retrieved from https://gasdetection.com/articles/2024-breakthroughs-in-smart-gas-sensor-technology-a-review/.
Ahmed, Shahbaz., Kumar, Rajesh., & Lee, Donghyun. (2024). Assessing Gas Leakage Detection Performance Using Machine Learning with Different Modalities. Journal of Intelligent & Fuzzy Systems, 45(2), 123–135. https://link.springer.com/article/10.1007/s42341-024-00545-0
Angstenberger, Stefan., Sterl, Florian., Theelke, Kilian., Giessen, Harald., & Schwarz, Ulrich. (2025). Real-time detection of low gas concentrations using coherently controlled quartz-enhanced photoacoustic spectroscopy. Optica, 12(1), 45–52. https://doi.org/10.1364/OPTICA.544448.
Al-Hadeethi, Yousef., Alghamdi, Abdullah., Alshahrani, Hassan., Alqahtani, Mohammed. and Alshahrani, Fahad., 2023. Fabrication of ZnO/WO₃ nanocomposite thin films using sacrificial template method for enhanced VOC gas sensing. Nanomaterials, 13(4), p.733. Available at: https://doi.org/10.3390/nano13040733
Korotcenkov, Ghenadii., Brinzari, Valeriu., Cho, Byung-Keun., Gulina, Larisa. and Han, Seung-Hwan., 2014. The influence of heterojunctions on the gas sensing properties of SnO₂–WO₃ composite thin films. Sensors, 14(11), pp.20480–20512. Available at: https://doi.org/10.3390/s141120480
Zhang, Jian., Liu, Xiaobo., Neri, Giovanni. and Pinna, Nicola., 2015. Nanostructured materials for room-temperature gas sensors. Nanoscale, 7(6), pp.2028–2043. Available at: https://doi.org/10.1039/C5NR02571K.
Al-Hadeethi, Yousef., Alshahrani, Hassan. and Alqahtani, Mohammed., 2024. Enhanced NO₂ gas sensing performance of SnO₂-WO₃ nanocomposite thin films prepared by chemical bath deposition. Journal of Electronic Materials. Available at: https://doi.org/10.1007/s11664-024-11381-6
Hosseini-Golgoo, Seyed Mohammad., Yousefi, Mohammad Hossein., & Ghasemi, Saeid., 2013. Application of neural networks for gas identification using temperature-modulated sensors. Sensor Letters, 11(3), pp.556–561.
Zhou, Yue. and Liu, Xiaobo., 2021. Gas recognition based on time-series neural networks and sensor arrays. Sensors, 21(14), p.4826. https://doi.org/10.3390/s21144826
Dehnaw, Ali. et al., 2024. Deep Neural Network Optimization for Efficient Gas Detection Systems in Edge Intelligence Environments. Processes, 12(12), p.2638. https://doi.org/10.3390/pr12122638
Zong, Boyang. et al., 2025. Smart Gas Sensors: Recent Developments and Future Prospective. Nano-Micro Letters, 17(54). https://doi.org/10.1007/s40820-024-01543-w
Authors
Copyright (c) 2025 Journal of Pure & Applied Sciences
This work is licensed under a Creative Commons Attribution 4.0 International License.
In a brief statement, the rights relate to the publication and distribution of research published in the journal of the University of Sebha where authors who have published their articles in the journal of the university of Sebha should how they can use or distribute their articles. They reserve all their rights to the published works, such as (but not limited to) the following rights:
- Copyright and other property rights related to the article, such as patent rights.
- Research published in the journal of the University of Sebha and used in its future works, including lectures and books, the right to reproduce articles for their own purposes, and the right to self-archive their articles.
- The right to enter a separate article, or for a non-exclusive distribution of their article with an acknowledgment of its initial publication in the journal of Sebha University.
Privacy Statement The names and e-mail addresses entered on the Sabha University Journal site will be used for the aforementioned purposes only and for which they were used.