Personality Assessment Based on Electroencephalography Signals during Hazard Recognition

Abstract

Hazard recognition assisted by human–machine collaboration (HMC) techniques can facilitate high productivity. Human–machine collaboration techniques promote safer working processes by reducing the interaction between humans and machines. Nevertheless, current HMC techniques acquire human characteristics through manual inputs to provide customized information, thereby increasing the need for an interactive interface. Herein, we propose an implicit electroencephalography (EEG)-based measurement system to automatically assess worker personalities, underpinning the development of human–machine collaboration techniques. Assuming that personality influences hazard recognition, we recorded the electroencephalography signals of construction workers and subsequently proposed a supervised machine-learning algorithm to extract multichannel event-related potentials to develop a model for personality assessment. The analyses showed that (1) the electroencephalography-assessed results had a strong correlation with the self-reported results; (2) the model achieved good external validity for hazard recognition-related personality and out-of-sample reliability; and (3) personality showed stronger engagement levels and correlations with task performance than work experience. Theoretically, this study demonstrates the feasibility of assessing worker characteristics using electroencephalography signals during hazard recognition. In practice, the personality assessment model can provide a parametric basis for intelligent devices in human–machine collaboration.