From Driver to Supervisor: Comparing Cognitive Load and EEG-Based Attentional Resource Allocation Across Automation Levels

Nikol Figalová,Hans-Joachim Bieg,Julian Elias Reiser,Yuan-Cheng Liu,Martin Baumann,Lewis Chuang,Olga Pollatos

DOI: https://doi.org/10.1016/j.ijhcs.2023.103169

IF: 4.866

2024-02-01

International Journal of Human-Computer Studies

Abstract:With increasing automation, drivers’ roles transition from active operators to passive system supervisors, affecting their behaviour and cognitive processes. This study addresses the attentional resource allocation and subjective cognitive load during manual, SAE Level 2, and SAE Level 3 driving in a realistic environment. An experiment was conducted on a test track with 30 participants using a prototype automated vehicle. While driving, participants were subjected to a passive auditory oddball task and their electroencephalogram was recorded. The study analysed the amplitude of the P3a event-related potential component elicited by novel environmental stimuli, an objective measure of attentional resource allocation. The subjective cognitive load was assessed using the NASA Task Load Index. Results showed no significant difference in subjective cognitive load between manual and Level 2 driving, but a decrease in subjective cognitive load in Level 3 driving. The P3a amplitude was highest during manual driving, indicating increased attentional resource allocation to environmental sounds compared to Level 2 and Level 3 driving. This may suggest that during automated driving, drivers allocate fewer attentional resources to processing environmental information. It remains unclear whether the decreased processing of environmental stimuli in automated driving is due to top-down attention control (leading to attention withdrawal) or bottom-up competition for resources induced by cognitive load. This study provides novel empirical evidence on resource allocation and subjective cognitive load in automated driving. The findings highlight the importance of managing drivers’ attention and cognitive load with implications for enhancing automation safety and the design of user interfaces.

computer science, cybernetics,ergonomics,psychology, multidisciplinary