Adaptive Automation Triggered by EEG-Based Mental Workload Index: A Passive Brain-Computer Interface Application in Realistic Air Traffic Control Environment

Abstract

Adaptive Automation (AA) is a promising approach to keep the task workload demand within appropriate levels in order to avoid both the under- and over-load conditions, hence enhancing the overall performance and safety of the human-machine system. The main issue on the use of AA is how to trigger the AA solutions without affecting the operative task. In this regard, passive Brain-Computer Interface (pBCI) systems are a good candidate to activate automation, since they are able to gather information about the covert behavior (e.g., mental workload) of a subject by analyzing its neurophysiological signals (i.e., brain activity), and without interfering with the ongoing operational activity. We proposed a pBCI system able to trigger AA solutions integrated in a realistic Air Traffic Management (ATM) research simulator developed and hosted at ENAC (École Nationale de l'Aviation Civile of Toulouse, France). Twelve Air Traffic Controller (ATCO) students have been involved in the experiment and they have been asked to perform ATM scenarios with and without the support of the AA solutions. Results demonstrated the effectiveness of the proposed pBCI system, since it enabled the AA mostly during the high-demanding conditions (i.e., overload situations) inducing a reduction of the mental workload under which the ATCOs were operating. On the contrary, as desired, the AA was not activated when workload level was under the threshold, to prevent too low demanding conditions that could bring the operator's workload level toward potentially dangerous conditions of underload.

A Neuroergonomics Approach to Mental Workload, Engagement and Human Performance

“…For instance, Prinzel et al (2000) utilized the continuous monitoring of brain waves that could be used to drive the level of automation and optimize the user's level of task engagement. Similarly, some authors managed to optimize air traffic controllers' task demand by triggering different levels of assistance (Aricò et al, 2016;Di Flumeri et al, 2019). These latter studies reported better human performance when neuro-adaptive automation was switched on compared to other conditions.…”

Neurophysiological Vigilance Characterisation and Assessment: Laboratory and Realistic Validations Involving Professional Air Traffic Controllers

“…Nowadays, many occupations require high levels of vigilance, for example, security personnel [10], employees tasked with monitoring surveillance cameras or baggage screening experts [11], driving vehicles [12], diagnostic medical screening [13], real classroom settings [14], and industrial and air traffic control [15][16][17]. The need to remain alert and situation-aware, and to detect infrequent but critical signals is crucial in a lot of job occupations: A vigilance failure in any of these domains could have dramatic impacts.…”

Objective Assessment of Human Workload in Physical Human-robot Cooperation Using Brain Monitoring

“…A Stepwise Linear Discriminant Analysis (SWLDA) was utilized for reducing the dimensionality of the feature space to avoid over-training by selecting features that offer the highest independent discriminative information. SWLDA has been widely used for feature selection in brain-computer interfaces and EEG-based affective state classification [Aricò et al 2016;Berka et al 2007]. It is a numerical approach that iteratively adds/removes features with significant/insignificant discrimination effect.…”