Event-related potentials emotion features of sleep-deprived and non-sleep-deprived drivers: a preliminary study
DOI:
https://doi.org/10.37934/araset.60.2.207218Keywords:
Event-related potential (ERP), mental workload, emotion recognition, binary linear regression (BLR)Abstract
This paper explains the study of Event-Related Potentials (ERP) emotion features related to sleep-deprived and non-sleep-deprived drivers. The study's research protocol consists of four stages - initial brainwave recording, four one-minute emotional stimulation experiments for happy, fear, sad, and calm, three sets of driving tests, and baseline brainwave recording. The data collected is pre-processed and then classified using Binary Linear Regression (BLR). The BLR classification of emotion-based ERP for sleep-deprived and non-sleep-deprived signals shows 86.1% accuracy. Though this result is still at the preliminary stage, it concluded that emotion-based ERPs can be used as promising features of brain signals to detect and classify the sleep-deprived and non-sleep-deprived.