REAL TIME PERSONALIZED PHYSIOLOGICALLY BASED STRESS DETECTION FOR HAZARDOUS OPERATIONS

Abstract

In hazardous working environments such as construction sites, mining operations, firefighting, or chemical industries, stress can significantly impair human performance, leading to reduced safety and increased accident risks. This project focuses on real-time personalized, physiologically based stress detection to enhance safety and decision-making during hazardous operations. The proposed system utilizes wearable biosensors to continuously monitor physiological signals such as heart rate, skin temperature, electrodermal activity (EDA), and oxygen saturation (SpO₂). These signals are processed and analyzed using machine learning algorithms to detect variations that indicate stress levels specific to each individual. By implementing personalized models, the system adapts to each worker’s physiological baseline, improving the accuracy of stress detection compared to generalized models. The collected data are transmitted in real time to a central monitoring unit, where supervisors can observe the stress status of workers and receive alerts during critical conditions. The proposed solution ensures early identification of stress, enabling timely interventions to prevent fatigue-related errors or accidents. This research contributes to occupational safety by integrating IoT-based sensing, real-time data analytics, and adaptive stress modeling. The system can be extended to other domains such as military missions, healthcare, and transportation, where real-time stress monitoring is essential for performance and safety.