1 Tashkent State Technical University named after Islam Karimov Tashkent, Uzbekistan
2 Tashkent State Technical University named after Islam Karimov Tashkent, Uzbekistan
ReceivedSep 19, 2026
RevisedOct 04, 2026
PublishedOct 09, 2026
Аннотация
The rapid expansion of the Internet of Things (IoT) has increased demand for machine learning approaches that can operate efficiently on resource-constrained edge devices while protecting locally generated data. This paper proposes a Federated Tiny Machine Learning (TinyML) framework for distributed IoT systems that combines on-device inference and lightweight local training with federated model aggregation. In the proposed architecture, ESP32-class microcontrollers collect and preprocess sensor data, execute TensorFlow Lite for Microcontrollers models, and transmit model updates rather than raw observations to a coordinating server. The server applies Federated Averaging to produce a global model that is redistributed to participating devices. The framework is organized into sensor, edge, federated aggregation, and application layers and is designed to address communication overhead, privacy exposure, heterogeneous non-independent and identically distributed data, intermittent connectivity, and limited energy and memory resources. A representative hardware and sensor configuration is specified to demonstrate implementation feasibility. The proposed design emphasizes data minimization, communication efficiency, scalability, and continuous learning at the edge. As a conceptual architecture, it provides a foundation for future empirical evaluation of accuracy, communication cost, energy consumption, convergence behavior, and robustness across heterogeneous IoT deployments.
Ключевые слова
federated learning, TinyML, edge AI, Internet of Things, privacy-preserving machine learning, edge computing.
Использованная литература
1. Eltazarov, B. T., Rikhsiboyev, R. R., Makhmudov, M. A., Alimova, Z. A., & Kurbonov, S. M. (2024). Remote patient monitoring: DS18B20 sensors in medical systems. SCIENCE Journal, 3(11-1), 123–128.
2. Gaibnazarov, B. B., Makhmudov, M. A., & Tuychiev, F. F. (2024). Smart automatic parking system using NodeMCU microcontroller. Science and Innovation, 3(Special Issue 36), 32–33.
3. Gulyamova, S. T., Tulyaganov, A. A., Abdukarimov, S. N., Mahmudov, M. A., & Ergashev, A. E. (2024). Texnologik jarayonlarni boshqarishda PLC (logik kontrollerlar) ning o‘rni. Science and Innovation, 3(Special Issue 50), 434–436.
4. Kairouz, P., McMahan, H. B., Avent, B., Bellet, A., Bennis, M., Bhagoji, A. N., Bonawitz, K., Charles, Z., Cormode, G., Cummings, R., D’Oliveira, R. G. L., Eichner, H., El Rouayheb, S., Evans, D., Gardner, J., Garrett, Z., Gascón, A., Ghazi, B., Gibbons, P. B., … Zhao, S. (2021). Advances and open problems in federated learning. Foundations and Trends in Machine Learning, 14(1–2), 1–210. https://doi.org/10.1561/2200000083
5. Makhmudov, M. A., Alimova, Z., & Kimizbayeva, A. (2023). Display data in histogram form using the Arduino platform. Universum, 16–19.
6. Makhmudov, M. A., Abdikhakimov, B. I., Ulugberdiyev, A. S., Kimizbayeva, A. E., & Gulyamova, S. T. (2024). Automatic water pump control system using GSM module and Arduino: Automation of water supply system (GSMUZ08). Science and Innovation, 3(A10), 221–226.
7. McMahan, H. B., Moore, E., Ramage, D., Hampson, S., & Agüera y Arcas, B. (2017). Communication-efficient learning of deep networks from decentralized data. In A. Singh & J. Zhu (Eds.), Proceedings of the 20th International Conference on Artificial Intelligence and Statistics (Vol. 54, pp. 1273–1282). PMLR. https://proceedings.mlr.press/v54/mcmahan17a.html
8. Warden, P., & Situnayake, D. (2019). TinyML: Machine learning with TensorFlow Lite on Arduino and ultra-low-power microcontrollers. O’Reilly Media.
Как цитировать
Mahmudov Mahmud, Rixsiboyev Ravshan (2026). FEDERATED TINYML FOR DISTRIBUTED IOT SYSTEMS: A PRIVACY-PRESERVING AND ENERGY-EFFICIENT EDGE INTELLIGENCE FRAMEWORK. Science and Innovation, 5(9), 81-85. https://doi.org/10.67620/sai.59a0011