Automated Alert System for Harassment-Related Stress Detection Using Heart Rate and Respiratory Rate Signals

Abstract

Sexual harassment represents a complex social problem, with women’s responses often categorized as acute stress when their safety in public spaces is compromised. Biomarkers such as heart and respiratory rate are directly related to autonomic nervous system activity under stress conditions. The present study proposes the development of a prototype warning system designed to automatically detect and alert about possible sexual harassment by identifying stress responses in women through real time heart and respiratory rate monitoring derived from ECG readings. Signal acquisition was performed using an ECG sensor with a microcontroller platform, while respiratory rate was obtained via ECG-Derived Respiration (EDR) analysis. Due to complexity of the issue, controlled simulations of street harassment were conducted using virtual reality headsets to induce responses in participants. When predetermined biomarkers thresholds were exceeded, an automated Telegram alert was sent to a trusted contact with the individual's readings. A pilot study with 12 female volunteers showed substantial increases in heart and respiratory rate during simulated harassment scenarios. The prototype validated the acquisition, processing and remote notification pipeline, providing proof of concept for real-time stress detection. These findings underscore the potential of biomedical signal processing as a tool for preventing gender-based violence. Future research will increase sample size, refine threshold calibration, and explore integration with wearable devices to enhance usability and scalability.