Emotional excess by multimodal Gaidai prognostics methodology, accounting for memory effects

Area of Science:

• Psychology
• Biomedical Engineering
• Systems Dynamics

Background:

• Human emotions are complex, involving perceptual, physiological, and behavioral responses.
• Modeling emotional dynamics requires integrating multimodal data, including bio-signals and expressions.
• Existing methods for assessing emotional states and their impact on reliability have limitations.

Purpose of the Study:

• To develop and benchmark a novel multimodal bio-reliability and risk/hazard assessment methodology using sensor-recorded emotional data.
• To model the dynamic system of human emotional reactions.
• To investigate the effectiveness of physiological signals in capturing emotional states.

Main Methods:

• Utilized professional actors to record multimodal physiological patterns during various emotion-evoking stimuli.
• Employed a novel nonparametric deconvolution scheme for extrapolating emotional levels.
• Applied a multivariate Gaidai reliability approach for prognostics of human errors and Health, Safety, and Environment (HSE) concerns.

Main Results:

• Piezoelectric respiration (PZT) readings demonstrated the best differentiation across five studied emotional states.
• PZT sensor data showed superior performance in capturing differences in both emotional valence and arousal.
• The novel methodology proved effective in assessing bio-reliability based on emotional data.

Conclusions:

• The developed multimodal bio-reliability methodology, based on sensor-recorded emotional data, is a robust tool for risk/hazard assessment.
• PZT sensors are highly effective for distinguishing emotional states, contributing to improved human error and HSE assessments.
• The generic nature of the methodology allows for broad applicability across various sectors beyond the case study.