A retrospective big data analysis of data from a longitudinal study of older drivers is reported for multiyear (up to 7 years) driving data from an in-vehicle sensor system capturing engine computer and GPS data to measure and understand the behaviors of older driver with differing health related to collisions while driving. Specifically, the data were analyzed against the Ministry of Transport collision reports for each of the drivers to measure the exposure of the driver to the collision location prior to the collision and the ongoing exposure to the location after the collision as a measure of driving behaviour. In this report, a convenience sample of 6 older drivers from Ottawa, Canada that had differing physical and cognitive health status and had a total of 9 collisions is reported. The collisions were split between atfault and not at-fault for the study participants. The measurement of exposure to the collision location results showed that with the exception of 1 collision, the drivers all had regular and in many cases frequent exposure to the collision location prior to the collision, indicating that lack of familiarity was not a factor in most of the collisions. The measurement of the exposure information after the collisions also showed that none of the drivers avoided the collision location with statistical significance. In fact, in many cases, the drivers' exposure increased with statistical significance. Hence, there is no indication of avoidance behaviour regardless of health status associated with the collision location, something that was unexpected.

, , , , ,
13th IEEE International Symposium on Medical Measurements and Applications, MeMeA 2018
Department of Systems and Computer Engineering

Wallace, B. (Bruce), Howcroft, J. (Jennifer), Goubran, R, Marshall, S. (Shawn), Porter, M.M. (Michelle M.), Alakel, A. (Akram), & Knoefel, F. (2018). Measuring Older Driver Behaviours with Prior and Post Exposure to Collision Locations. In MeMeA 2018 - 2018 IEEE International Symposium on Medical Measurements and Applications, Proceedings. doi:10.1109/MeMeA.2018.8438641