Edge AI and TinyML in IoT Systems: A Review of Applications, Architectures and Limitations

Lazar Stošić; Željko  Stanković; Olja  Krčadinac

doi:10.62907/juuntics260101013s

Authors

Lazar Stošić Faculty of Informatics and Computer science, University Union—Nikola Tesla, Belgrade, Serbia; Don state technical university, Rostov on Don, Russia Author https://orcid.org/0000-0003-0039-7370
Željko Stanković Faculty of Informatics and Computer science, University Union—Nikola Tesla, Belgrade, Serbia Author https://orcid.org/0000-0002-9893-9088
Olja Krčadinac Faculty of Informatics and Computer science, University Union—Nikola Tesla, Belgrade, Serbia Author https://orcid.org/0000-0002-6299-371X

DOI:

https://doi.org/10.62907/juuntics260101013s

Keywords:

Edge AI, TinyML, Internet of Things, embedded systems, machine learning, edge computing, intelligent sensors

Abstract

With the rapid development of the Internet, the need for fast and energy-efficient data processing has also increased. In this regard, Edge Artificial Intelligence and Tiny Machine Learning represent significant technological approaches that enable the execution of machine learning models on resource-constrained edge devices, embedded platforms, and microcontrollers. The aim of this review is to analyze the role of Edge AI and TinyML technologies in IoT (Internet of Things) systems, with special reference to their applications, architectural models and key limitations. The paper provides a concise review of scientific and professional literature addressing edge computing, embedded machine learning, intelligent sensors, and IoT architectures. Through a review of numerous literatures, major application areas were identified, including smart homes, smart classrooms, health monitoring, wearables, industrial IoT, predictive maintenance, smart agriculture, and environmental monitoring. Special attention is paid to architectural models, from cloud-centric IoT systems to edge-assisted and fully embedded TinyML architectures. Analysis shows that Edge AI and TinyML can significantly reduce latency, improve privacy, reduce network traffic consumption, and enable real-time decision making. However, their application is limited by small memory, lower processing power, energy consumption, model optimization, security risks, interoperability and maintenance of remote devices.

Downloads

Download data is not yet available.

References

[1] Carvalho, G., Cabral, B., Pereira, V., & Bernardino, J. (2021). Edge computing: current trends, research challenges and future directions. Computing, 103(5), 993-1023. https://doi.org/10.1007/s00607-020-00896-5

[2] Chang, Z., Liu, S., Xiong, X., Cai, Z., & Tu, G. (2021). A survey of recent advances in edge-computing-powered artificial intelligence of things. IEEE Internet of Things Journal, 8(18), 13849-13875. https://doi.org/10.1109/JIOT.2021.3088875

[3] Gill, S. S., Golec, M., Hu, J., Xu, M., Du, J., Wu, H., ... & Uhlig, S. (2025). Edge AI: A taxonomy, systematic review and future directions. Cluster Computing, 28(1), 18. https://doi.org/10.1007/s10586-024-04686-y

[4] Zhou, Z., Chen, X., Li, E., Zeng, L., Luo, K., & Zhang, J. (2019). Edge intelligence: Paving the last mile of artificial intelligence with edge computing. Proceedings of the IEEE, 107(8), 1738-1762. https://doi.org/10.1109/JPROC.2019.2918951

[5] Ray, P. P. (2022). A review on TinyML: State-of-the-art and prospects. Journal of King Saud University-Computer and Information Sciences, 34(4), 1595-1623. https://doi.org/10.1016/j.jksuci.2021.11.019

[6] Elhanashi, A., Dini, P., Saponara, S., & Zheng, Q. (2024). Advancements in TinyML: Applications, limitations, and impact on IoT devices. Electronics, 13(17), 3562. https://doi.org/10.3390/electronics13173562

[7] Schizas, N., Karras, A., Karras, C., & Sioutas, S. (2022). TinyML for ultra-low power AI and large scale IoT deployments: A systematic review. Future Internet, 14(12), 363. https://doi.org/10.3390/fi14120363

[8] Stošić, L., Dimitrovska, M., & Pushova Stamenkova, L. (2023). Exploring the expanding world of IoT: А comprehensive overview and considerations. KNOWLEDGE - International Journal , 60(3), 359–363. Retrieved from https://ojs.ikm.mk/index.php/kij/article/view/6273

[9] Stošić, L., Dimitrovska, M., Pushova Stamenkova, L., & Smelcerović, M. (2023). From concept to reality: Understanding the internet of things. SCIENCE International Journal, 2(4), 181–184. https://doi.org/10.35120/sciencej0204181s

[10] Domínguez-Bolaño, T., Campos, O., Barral, V., Escudero, C. J., & García-Naya, J. A. (2022). An overview of IoT architectures, technologies, and existing open-source projects. Internet of Things, 20, 100626. https://doi.org/10.1016/j.iot.2022.100626

[11] Choudhary, A. (2024). Internet of Things: a comprehensive overview, architectures, applications, simulation tools, challenges and future directions. Discover Internet of Things, 4(1), 31. https://doi.org/10.1007/s43926-024-00084-3

[12] Hamdan, S., Ayyash, M., & Almajali, S. (2020). Edge-computing architectures for internet of things applications: A survey. Sensors, 20(22), 6441. https://doi.org/10.3390/s20226441

[13] Singh, R., & Gill, S. S. (2023). Edge AI: a survey. Internet of Things and Cyber-Physical Systems, 3, 71-92. https://doi.org/10.1016/j.iotcps.2023.02.004

[14] TensorFlow. (n.d.). TensorFlow Lite for Microcontrollers. GitHub. https://github.com/tensorflow/tflite-micro [15] Olja Krčadinac, Željko Stanković, Dragana Dudić, Lazar Stošić (2024). Development of an Open-Source Voice-Controlled Smart Home System, JITA – Journal of Information Technology and Applications, 14(2), 111-116, https://doi.org/10.7251/JIT2402111K

[16] Heydari, S., & Mahmoud, Q. H. (2025). Tiny machine learning and on-device inference: A survey of applications, challenges, and future directions. Sensors, 25(10), 3191. https://doi.org/10.3390/s25103191

[17] Nguyen, D. C., & Welch, C. (2026). Generative artificial intelligence in qualitative data analysis: Analyzing—Or just chatting?. Organizational Research Methods, 29(1), 3-39. https://doi.org/10.1177/10944281251377154

[18] Artiushenko, V., Lang, S., Lerez, C., Reggelin, T., & Hackert-Oschätzchen, M. (2024). Resource-efficient Edge AI solution for predictive maintenance. Procedia Computer Science, 232, 348-357. https://doi.org/10.1016/j.procs.2024.01.034

[19] Gookyi, D. A. N., Wulnye, F. A., Wilson, M., Danquah, P., Danso, S. A., & Gariba, A. A. (2024). Enabling intelligence on the edge: leveraging edge impulse to deploy multiple deep learning models on edge devices for tomato leaf disease detection. AgriEngineering, 6(4), 3563-3585. https://doi.org/10.3390/agriengineering6040203

[20] Bhushan, B., Negi, P., Nayak, A., & Goyal, S. (2025). Graphene composites for water remediation: an overview of their advanced performance with focus on challenges and future prospects. Advanced Composites and Hybrid Materials, 8(1), 55. https://doi.org/10.1007/s42114-024-01088-x

[21] R. Shelke, K., & K. Shinde, S. (2025). SAOA:: Skill archimedes optimization algorithm based privacy enhancement for blockchain storage optimization in medical IoT environment. https://doi.org/10.1016/j.compeleceng.2025.110270

[22] Wang, T., Guo, J., Zhang, B., Yang, G., & Li, D. (2025). Deploying AI on edge: Advancement and challenges in edge intelligence. Mathematics, 13(11), 1878. https://doi.org/10.3390/math13111878