Kalman-Normalized GSR Analysis for Real-Time Stress Quantification in Wearable Systems.

Abstract

The growing advancements in science, technology, innovation, and research are paralleled by a concerning rise in stress levels worldwide. Stress, an inevitable psychosocial factor, significantly affects human life, mental status, and overall physiosocial health. This research focuses on developing an accurate galvanic skin response (GSR) system to effectively identify and analyze stress levels. The core concept of GSR involves measuring the conductivity between skin contacts, where increased conductivity corresponds to heightened stress levels. Advanced algorithms are employed to efficiently convert these readings into digital formats for precise analysis. The system leverages the Kalman filter algorithm to reduce noise, ensuring smooth and reliable signals from raw GSR readings. A dynamic range normalization technique transforms filtered readings into a consistent scale (0-500) tailored to individual baseline values. This approach ensures that stable measurements are unaffected by noise using the Kalman filter, consistency across users despite physiological differences, and accurate, personalized stress-level detection through adaptive categorization. Tested on over 5000 samples, the system accurately identifies stress levels across defined ranges, established in collaboration with psychologist practitioners. This research culminates in developing an accurate and personalized stress detection and analysis system, providing actionable insights into stress management.