Abstract—This study addresses the pressing challenge of developing efficient and secure cryptographic solutions for Internet of Things (IoT) systems, where devices operate under constraints of limited computing power, memory, and energy consumption. Against the backdrop of the rapid proliferation of IoT devices based on Embedded Systems (ESP32) microcontrollers and the growing cybersecurity threats, a systematic analysis of hardware and software encryption methods is conducted. The paper investigates the implementation of cryptographic protection at the microcontroller level in IoT networks, with particular emphasis on execution time, energy efficiency, and hardware resource utilization. The scientific novelty of this work lies in the integration of machine learning techniques for the automatic selection of the most efficient cryptographic algorithm Advanced Encryprion Standard (AES), Secure Hash Algorithm (SHA) or Hash-based Message Authentication Code (HMAC), based on data volume, execution time, and power consumption. Unlike prior studies that focus on isolated implementations, this research proposes a universal adaptive encryption selection system tailored for resource-constrained microcontroller environments. A key contribution is the implementation and testing of this system on the ESP32 and ESP32-S3 platforms. The results demonstrate that hardware-based encryption significantly outperforms software-based methods in terms of execution speed and energy efficiency. The developed logistic regression model achieved 100% classification accuracy on the ESP32 platform, outperforming alternative algorithms and confirming its applicability for energy-efficient real-time IoT applications. Finally, the paper identifies promising avenues for future work, including the development of secure digital twins for healthcare systems and autonomous devices.
 
Keywords—hardware encryptor, cryptographic libraries, Embedded Systems (ESP32), cryptosecurity, Internet of Things (IoT) devices, machine learning


Copyright © 2026 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).