Driver and Pedestrian Characteristics of Pedestrian Vehicle Crashes in Metropolitan Perth

Abstract

Pedestrian vehicle crashes are a vital traffic safety concern. About 12% of all people who die on the roads in Western Australia are pedestrians. This paper will investigate the changes of characteristics of drivers and pedestrian associated with pedestrian vehicle crashes over the time of a day. A two step cluster algorithm will be adopted to group time of pedestrian vehicle crashes into time intervals. Then the Expectation Maximisation (EM) algorithm will be used to detect segments of drivers and pedestrians for each time interval. The information about target groups of pedestrians and drivers will be useful for developing practical education and enforcement strategies and to improve the design of the pedestrian environment on safety.