Non-intrusive Monitoring of Operator Situational Awareness via Human-Machine Interface States

Abstract

Non-intrusive monitoring of Human-in-the-Loop errors that may be injected via Human Machine Interfaces pose a monitoring challenge in legacy industrial control panels. Equipment obsolescence offers limited scope of automatic monitoring of situational awareness which prevents human errors from being trended nor predicted in real-time. In this paper, statistical time-series data forecast models based on ARIMA (with exogenous regressors) are evaluated to ascertain the feasibility of using this technique to monitor and predicted human errors. That is a step towards achieving real-time monitoring of operator situational awareness while using Human Machine Interfaces.