Abstract
This paper investigates the influence of real-time physiological and environmental variables on the computational performance of cryptographic algorithms deployed in wearable medical systems powered by the ESP32 microcontroller.The proposed system integrates biometric sensors to capture heart rate, skin temperature, and galvanic skin response, from which a composite Stress Index (SI) is calculated.This SI dynamically modulates encryption behavior to adapt to the user's physiological state.An experimental dataset comprising 300 samples from 10 participants was collected over ten days under semicontrolled environmental conditions.Six widely used cryptographic algorithms-AES-256, HMAC-SHA256, SHA-256, SHA-3, BLAKE3, and ChaCha20-were evaluated based on execution time, CPU load, and estimated energy consumption under varying stress levels.To quantify algorithmic robustness, we introduced the Crypto Stress Tolerance Score (CSTS), a custom metric combining performance stability, stress correlation, and resource efficiency.The findings reveal that elevated body temperature and stress index values significantly affect cryptographic performance, with execution time increasing by up to 35% under high-stress conditions.AES-256 exhibited the highest sensitivity and variability, whereas BLAKE3 and ChaCha20 delivered consistent, low-latency performance with minimal fluctuation.A comparative analysis and scoring table are provided to guide optimal algorithm selection for constrained medical Internet of Things (IoT) environments.This work contributes a novel framework for adaptive, stress-aware encryption in embedded healthcare devices, offering improved reliability, energy efficiency, and security personalization in patientcentric monitoring systems.