Visual Exploration Strategies of Experienced and Non-Experienced Drivers During the Takeover of an Automated Vehicle

Visual scanning of the driving environment is a key component in driving. In conditional automation, the vehicle is usually steered by automation, but drivers may be required to take control of the vehicle in certain circumstances. In these takeover situations (TO), visual exploration of the environment is critical for resuming manual control. This study aimed to compare visual explorations over time during TO preparation and actual TO between experienced and non-experienced drivers. Twenty-five participants completed three simulated drives, each comprising three TO maneuvers. Visual strategies were measured by the percentage of gaze time allocated to areas of interest (AOIs) and further characterised through visual exploration sequences (scanpaths) during TO preparation and actual TO. The results indicated that experienced drivers gazed more at the road area in which the vehicle would be located at the time of effective TO than non-experienced drivers. Similarly, immediately after resuming manual control, experienced drivers gazed more at areas where potential hazards may be present than non-experienced drivers. Three main scanpath classes were identified during TO preparation: anticipation of trajectory control, anticipation of speed control, and environment exploration. During subsequent manual driving, two scanpath classes emerged: speed control and trajectory control. These findings are discussed in terms of differences in visual strategies between non-experienced and experienced drivers, shaped by situational context. Functionally classifying visual exploration sequences provides insight into TO processes and offers a promising link to drivers’ situation awareness and performance.