Proceedings of the Human Factors and Ergonomics Society Annual Meeting | 2019
The Effects of Warning Lead Time on Situation Awareness in Connected Vehicle Systems
Abstract
Connected vehicle systems (CVS) allow for the exchange of real time vehicle movement data. By means of wireless communication between vehicles as well as information delivered from vehicle to driver, drivers are expected to be more aware of surrounding traffic and make better decisions to avoid hazards. Particularly, for safety-critical situations, it is expected that collision warnings supported by CVSs are able to alert drivers of possible hazards in advance, improve drivers’ situation awareness, and therefore enhance driving safety. Situation awareness (SA), as defined in Endsley (1988), comprises three levels: Perception of the elements in the environment (Level 1 SA), Comprehension of the current situation (Level 2 SA), and Projection of future status (Level 3 SA). Ever since automotive automation appeared, research studies have focused on the different factors of various automation technologies and its impact on drivers’ situation awareness, for instance, adaptive cruise control (ACC) system and highly automated driving (HAD) (De Winter, 2014; Ma & Kaber, 2005), information reliability of ACC (Ma & Kaber, 2007), feedback level of ACC (Stanton & Young, 2005), and warning criticality and modality of advanced driver assistance system (ADAS) (Lindgren, 2009). Existing literature found that different designs of automation technologies can lead to positive and negative impacts on driver SA. However, collision warning system (CWS), as one of the promising technologies for driving safety, has received little attention with regard to its effects on driver SA. A crucial factor in determining the effectiveness of CWS is warning lead time, which is defined as the time to collision when a warning message can be delivered to the driver. Many empirical studies focusing on sensor-based CWSs indicated that relatively early auditory warnings with longer lead time aided drivers by reducing collisions, shortening the reaction time, and moderating deceleration to avoid collision, compared with late warnings with shorter lead time (Abe & Richardson, 2006; Lee et al, 2002; Yan et al., 2015). The lead time of warnings ranges from 0.02 s to 4.5 s. By contrast, connected vehicle systems allows a wider range of lead time, e.g. 0 60 s in Wan et al. (2016), which concluded that warnings with 5 8 s lead time achieved maximum effectiveness in terms of safety benefit and braking response. Compared with late warnings, early warnings may lead to better performance and increased trust in the systems (Abe & Richardson, 2003). However, drivers’ overreliance on the CWS may negatively affect their situation awareness in the driving tasks and lead to potential risks on the road. To date, no studies known to the authors have been conducted on the effects of warning lead time on driver SA. The present driving simulator study was conducted with 32 participants to investigate the effects of warning lead time (3 s vs. 6 s) and SA maintaining messages (i.e. repetitive physical reminders and overspeed warnings) (with vs. without) on driver SA and driving performance in connected vehicle systems. The driver SA was measured using freeze probe technique at two time points during the simulated driving. Results revealed that early warnings (6 s) provided more benefits for drivers than late warnings (3 s) regarding lower collision rates and more moderate braking responses. However, drivers under 3 s lead time warnings had statistically higher Level 1 (perception) and Level 3 (projection) SA compared with those who drove under 6 s lead time warnings in the first SA probe. The results confirmed our hypothesis that early warnings with longer lead time have negative impact on drivers’ situation awareness compared with late warnings with shorter lead time due to drivers’ overreliance on the warnings, which may potentially impair driver performance in the long-term usage of the collision warning systems. The implications of the findings support the consideration of the effects of warning lead time on driver SA in future design of warning systems in the connected vehicle systems.