Desenvolvimento de um protótipo de dispositivo vestível para aferição de variabilidade da frequência cardíaca para determinação de prontidão ao treino

Abstract

The monitoring of physiological recovery has become indispensable for optimizing sports performance and preventing overtraining. This work presents the development of a low-cost wearable prototype for measuring Heart Rate Variability (HRV), utilizing the ESP32 microcontroller and the XD-58C reflective photoplethysmographic (PPG) sensor. The main objective was to validate a hardware and firmware architecture capable of deter mining training readiness through the RMSSD metric. The methodology relied on the implementation of embedded Digital Signal Processing (DSP) techniques for pulse signal treatment. Comparative tests were conducted among different filter topologies, resulting in the definition of a hybrid approach combining exponential smoothing and a 2nd-order Butterworth filter with an 8 Hz cutoff frequency. Results showed that this configuration increased the Signal-to-Noise Ratio (SNR), enabling robust systolic peak detection via an dynamic threshold algorithm. Despite latency limitations in Bluetooth transmission, which presented packet losses around 15%, the system proved effective for wellness-grade applications, validating the use of green light photoplethysmography as a non-invasive tool for resting autonomic recovery analysis.