XLTMS – Explainable Long Term Mental State Detection System for Safety-Critical Professions.

Abstract

In numerous high safety risk areas, such as rail transport, aviation, and many other industries, human error is one of the most frequent sources of accidents. In this white paper, we propose XLTMS, a new framework that combines brainwave (EEG) monitoring and explainable AI (XAI) to evaluate the mental and emotional states of people before they assume roles that come with a high level of responsibility. XLTMS analyzes brainwave data in the gamma, alpha, beta, and theta frequency bands to identify mental states related to the pre-defined indicators of stress, drowsiness, cognitive overload, and attention. Using Long Short-Term Memory (LSTM) Networks and other explainable machine learning models, XLTMS correlates EEG features with mental states. XLTMS main use case is for pre-shift mental state assessments for commercial pilots, railway drivers, surgeons, military personnel and similar professions.