TinyML and MFCC Feature Extraction for Energy Efficient Automatic Air Purifier Control

Authors

  • Mangasa Manullang Informatics Department, Faculty of Artificial Intelligence and Data Science, Universitas Pelita Harapan, Indonesia
  • Ferawaty Informatics Department, Faculty of Artificial Intelligence and Data Science, Universitas Pelita Harapan, Indonesia
  • Leonardo Angkasa Informatics Department, Faculty of Artificial Intelligence and Data Science, Universitas Pelita Harapan, Indonesia
  • Winson Lim Informatics Department, Faculty of Artificial Intelligence and Data Science, Universitas Pelita Harapan, Indonesia

DOI:

10.33395/sinkron.v10i3.16352

Abstract

Classroom environments are highly susceptible to airborne disease transmission due to high occupant density and prolonged interaction times. Conventional mitigation strategies often rely on continuously operating air purification systems throughout building operational hours. This always-on approach guarantees continuous air circulation but results in massive and unnecessary electrical energy consumption, especially during idle periods or when biological contamination is absent. This research aims to design and implement an energy-efficient smart classroom system that automatically controls air purifiers based on real-time acoustic detection of sneeze events. The system utilizes Tiny Machine Learning embedded on an edge microcontroller with an onboard microphone. Audio datasets comprising sneeze, cough, and speech classes were processed using Mel-Frequency Cepstral Coefficients feature extraction at a 16 kHz sampling rate to optimize memory usage, followed by a neural network classifier training. The hardware prototype controls two air purifiers positioned for cross-ventilation, activating them for 15 minutes exclusively upon sneeze detection. The trained model achieved an overall accuracy of 97.5%, with a perfect precision rate in recognizing sneeze events. Field testing during an active class period demonstrated that the event-driven system consumed only 92.8 Watt-hours. Compared to the conventional continuous operation method, the automated system successfully reduced electrical power consumption by 71.4%. Implementing edge-based artificial intelligence for acoustic environmental monitoring provides a highly reliable approach to automated facility management, balancing health risk mitigation through optimal cross-ventilation with significant electrical energy conservation in smart classrooms. Future integration with low-power wireless modules is highly recommended to transmit event logs to a central dashboard, completing the sustainable facility management ecosystem.

GS Cited Analysis

Downloads

Download data is not yet available.

References

Abdul, Z. K., & Al-Talabani, A. K. (2022). Mel Frequency Cepstral Coefficient and its Applications: A Review. IEEE Access, 10, 122136–122158. https://doi.org/10.1109/ACCESS.2022.3223444

Bharti, Arora, V., & Singh, M. (2026). Artificial Intelligence Based Techniques to Detect and Classify Adventitious Respiratory Sounds: An in-Depth Review. Archives of Computational Methods in Engineering, 33(4), 4627–4722. https://doi.org/10.1007/s11831-025-10344-2

Bhushan, C. M., Koppuravuri, P., Prasanthi, N., Gazi, F., Hussain, M. M., Abdussami, M., Devi, A. A., & Faizi, J. (2025). Deploying TinyML for energy-efficient object detection and communication in low-power edge AI systems. Scientific Reports, 15(1), 44299. https://doi.org/10.1038/s41598-025-27818-9

Chen, Y., Zhang, H., & Zhong, S. (2024). Design and implementation of smart home system based on IoT. Results in Engineering, 24, 103410. https://doi.org/https://doi.org/10.1016/j.rineng.2024.103410

Cho, J., Heo, Y., & Moon, J. W. (2023). An intelligent HVAC control strategy for supplying comfortable and energy-efficient school environment. Advanced Engineering Informatics, 55, 101895. https://doi.org/https://doi.org/10.1016/j.aei.2023.101895

David, R., Duke, J., Jain, A., Reddi, V. J., Jeffries, N., Li, J., Kreeger, N., Nappier, I., Natraj, M., Regev, S., Rhodes, R., Wang, T., & Warden, P. (2020). TensorFlow Lite Micro: Embedded Machine Learning on TinyML Systems. CoRR, abs/2010.08678. https://arxiv.org/abs/2010.08678

Elhami, M., Goodarzi, S. S., Maleki, S., & Sajadi, B. (2025). Three-objective optimization of the HVAC system control strategy in an educational building to reduce energy consumption and enhance indoor environmental quality (IEQ) using machine learning techniques. Journal of Building Engineering, 105, 112444. https://doi.org/https://doi.org/10.1016/j.jobe.2025.112444

Jain, S. K., & Kesswani, N. (2023). A noise-based privacy preserving model for Internet of Things. Complex & Intelligent Systems, 9(4), 3655–3679. https://doi.org/10.1007/s40747-021-00489-5

Márquez-Sánchez, S., Alonso-Rollán, S., Nahom, H., Erbad, A., & Fernandez, J. H. (2025). Optimizing Building Energy Management Leveraging Adaptive Edge Computing for Enhanced Efficiency and Occupant Well-Being. In P. Novais, P. B. D., I. Satoh, V. J. Inglada, S. R. González, E. Jove Pérez, J. Parra Domínguez, P. Chamoso, & R. S. Alonso (Eds.), Ambient Intelligence – Software and Applications – 15th International Symposium on Ambient Intelligence (pp. 236–248). Springer Nature Switzerland.

Márquez-Sánchez, S., Calvo-Gallego, J., Erbad, A., Ibrar, M., Hernandez Fernandez, J., Houchati, M., & Corchado, J. M. (2023). Enhancing Building Energy Management: Adaptive Edge Computing for Optimized Efficiency and Inhabitant Comfort. Electronics, 12(19). https://doi.org/10.3390/electronics12194179

Meedeniya, D., Ariyarathne, I., Bandara, M., Jayasundara, R., & Perera, C. (2023). A Survey on Deep Learning Based Forest Environment Sound Classification at the Edge. ACM Comput. Surv., 56(3). https://doi.org/10.1145/3618104

Morawska, L., Tang, J. W., Bahnfleth, W., Bluyssen, P. M., Boerstra, A., Buonanno, G., Cao, J., Dancer, S., Floto, A., Franchimon, F., Haworth, C., Hogeling, J., Isaxon, C., Jimenez, J. L., Kurnitski, J., Li, Y., Loomans, M., Marks, G., Marr, L. C., … Yao, M. (2020). How can airborne transmission of COVID-19 indoors be minimised? Environment International, 142, 105832. https://doi.org/https://doi.org/10.1016/j.envint.2020.105832

Park, H., Hong, S., & Kim, J.-S. (2024). Audio Classification on Low-Resource Microcontrollers. 2024 15th International Conference on Information and Communication Technology Convergence (ICTC), 1332–1333. https://doi.org/10.1109/ICTC62082.2024.10826933

Pham, H. H., Le, T. M., & Son, L. H. (2026). Towards open world sound event detection. Signal Processing, 248, 110707. https://doi.org/https://doi.org/10.1016/j.sigpro.2026.110707

Rawat, N., Kumar, P., Hama, S., Williams, N., & Zivelonghi, A. (2025). Improving classroom air quality and ventilation with IoT-driven acoustic and visual CO2 feedback system. Science of The Total Environment, 980, 179543. https://doi.org/https://doi.org/10.1016/j.scitotenv.2025.179543

Downloads


Crossmark Updates

How to Cite

Manullang, M., Ferawaty, F., Angkasa, L. ., & Lim, W. . (2026). TinyML and MFCC Feature Extraction for Energy Efficient Automatic Air Purifier Control. Sinkron : Jurnal Dan Penelitian Teknik Informatika, 10(3), 1840-1848. https://doi.org/10.33395/sinkron.v10i3.16352

Issue

Vol. 10 No. 3 (2026): Article Research July 2026

Section

Articles

License

Copyright (c) 2026 Mangasa Manullang, Ferawaty, Leonardo Angkasa, Winson Lim

Creative Commons License

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.

Crossref
Scopus
Google Scholar
Europe PMC