Driver Distraction: Eye Glance Analysis and Conversation Workload
To better understand the relationship of cognitive and visual distraction during mobile phone conversations or interactions while the driver is experiencing real-world driving conditions and pressures.
In 2009, 5,474 people (or 16 percent of the total) were killed in police-reported crashes in which at least 1 form of driver distraction was stated in the crash report. The proportion of fatalities involving driver distraction increased from 10 percent in 2005 to 16 percent in 2009. Numerous research studies have addressed driver distraction. However, this area of research has had considerable difficulty characterizing the content and level of driver involvement in phone conversations. Cognitive distraction is defined as the mental workload associated with an activity that involves thinking about something other than the driving task. It is generally not observable; however, it can be inferred from video and audio analysis. In addition, more research is needed to examine eye glances as they relate to driver distraction. In the earlier study entitled “Driver Distraction in Commercial Vehicle Operations,” researchers assessed visual distraction via the time the drivers’ eyes were off the forward roadway. The finding of a significant association between eyes-off-forward-roadway glances for times less than 1 second and involvement in a safety critical event (SCE) may suggest that shorter glances, performed just prior to the SCE, may be problematic.
Data for this study was drawn from an existing naturalistic database from a vendor of onboard monitoring systems. The research project accomplished the following:
- Analyzed voice recordings, made while driving, to determine if there was an objective measure of cognitive distraction. Although the term "cognitive distraction" was used, the capabilities of the SmartDrive® system allowed the measurement of "conversation workload" (i.e., a proxy for cognitive distraction).
- Determined if conversation workload was related to SCE risk during voice-related SCEs.
- Assessed the potential risk posed from other electronic devices, such as dispatching devices.
- Assessed the risk of eye glance behavior (total time eyes-off-forward-roadway) as it related to involvement in an SCE during voice-related SCEs.
- Eye glance behavior was calculated in 0.5-second intervals to determine if multiple glance checks and/or short glances in close proximity to event onset increased the probability of involvement in an SCE.
- Assessed the risk of mobile phone subtasks (dialing, talking, listening, etc.) for hands-free, handheld, CB, and push-to-talk devices.
- Assessed the difference between spurious baselines and random baselines.
A better understanding of safety risks associated with cognitive distraction and the safety risks of short eye glances as compared to lengthier eyes-off-the-forward-roadway glances.
November 2015 Final report published
Final report is published and available at: https://rosap.ntl.bts.gov/view/dot/212
Virginia Technical Transportation Institute