VOCAL INSTABILITY AS A SENSITIVE BIOMARKER FOR DRIVING STRESS: DECOUPLING COGNITIVE LOAD AND ENVIRONMENTAL FRICTION IN A REAL-WORLD DUAL-TASK PROTOCOL

Abstract

Driver stress and cognitive workload are critical safety determinants in modern transportation. While vocal acoustic analysis is a promising non-invasive monitoring technique, existing literature often lacks ecological validity and struggles to distinguish between internal cognitive load (CL) from secondary tasks and external environmental friction (EF) from traffic. This study addresses this gap using a single-subject (N=1) real-world dual-task protocol. The driver maintained continuous conversation while navigating two environments: a highfriction urban congestion segment (short route) and a hybrid urban-intercity segment (long route). Analysis utilized a custom weighted acoustic stress index and instantaneous pitch standard deviation (vocal instability). Findings demonstrate that the constant dualtask demand establishes a dominant, consistent moderate stress baseline (~3436%), decoupled from routine traffic fluctuations and congestion levels. Although average stress levels remained consistent, pitch standard deviation proved to be a more sensitive metric, being significantly lower on the long route compared to the pure urban segment. This confirms vocal instability’s ability to effectively decouple CL and EF contributions, providing empirical evidence that low-demand highway segments create a stabilizing effect on the voice even under moderate overall load. Consequently, vocal instability is validated as a sensitive biomarker essential for developing context-aware in-vehicle systems capable of distinguishing between distraction-related and environmental stress.