Performance Degradation During Sudden Hazardous Events: A Comparative Analysis of Use of a Phone and a Music Player During Driving

• Published in: IEEE transactions on intelligent transportation systems 2019-11, Vol.20 (11), p.4055-4065
• Main Authors: Choudhary, Pushpa, Velaga, Nagendra R.
• Format: Article
• Language: English
• Subjects: accident probability; Accidents; Algorithms; Braking; Collision avoidance; Degradation; Driver distraction; Music; music-player; Pedestrian crossings; Pedestrians; Performance degradation; Peripheral vision; phone-use; Probability; Reaction time; Real time; real-time behavioral characteristics; Real-time systems; Risk; Short message service; Statistical models; Task analysis; Text messaging; Vehicles; Warning systems
• ISSN: 1524-9050; 1558-0016
• DOI: 10.1109/TITS.2018.2879968

The present study attempts to statistically model and quantify the contribution of two sets of potential factors to accident risk: 1) real-time driving behavioral characteristics and 2) distraction conditions. The effects of phone and music player usage on reaction times and accident probabilities were modeled for two events: 1) a pedestrian crossing (started in peripheral vision) and 2) sudden braking (started in central vision) in an urban scenario. In total, 90 subjects drove for baseline and conversation conditions. In all, 70 and 78 of the 90 subjects drove for texting and music player tasks, respectively. Reaction times and accident probabilities were modeled with the help of generalized linear mixed model approach. The results showed that the reaction time increased by 42%, 113%, and 62% due to the presence of conversation, texting, and music player operations, respectively, for the pedestrian crossing event. Whereas, the increments were lower for the sudden braking event. Drivers with higher approach speed detected the events faster. The accident probability increased due to the texting tasks and by operating the player during driving. An increase in reaction time also led to higher risk for the pedestrian crossing event. For the sudden braking event, the effect of music-listening was not significant, but the accident risk increased if the driver approached the event at high speed. Overall, the results suggest that the algorithm for designing a collision warning system should consider the real-time driving characteristics based on event type and quantify the effects of distraction based on the type of distraction.