Augmented Reality Glasses-Based Curve Negotiation Assistance System: Design, Implementation and Neurophysiological Validation Using Eye-Tracking and EEG Analysis

Curve-related crashes account for a large share of traffic fatalities, driven by lane departure, speed misjudgment, and poor trajectory planning. This study presents an augmented reality (AR) glasses-based curve negotiation assistance system, developed with an AI-augmented prototyping workflow and evaluated in a ring-display driving simulator. Thirty-two licensed drivers completed four curve scenarios of varying radius and complexity under baseline and AR-assisted conditions while 32-channel EEG and eye-tracking were recorded. Linear mixed-effects models showed fewer lane departures (−34.7%), lower speed variance (−28.3%), and reduced NASA-TLX workload (−22.1%). EEG revealed decreased frontal theta and increased parietal alpha; beta-band differences were non-significant (p = 0.23). Off-road glances fell 41.2%, although roughly 31% of the added road-relevant fixation landed on the AR overlay itself, and five participants (15.6%) showed no gain on at least one metric. These converging simulator-based findings suggest wearable AR is a promising curve-safety approach warranting on-road validation.