Quantitative risk assessment of freeway crash casualty using high-resolution traffic data

This study aimed to investigate the impacts of traffic flow conditions on crash casualty of different collision types using high-resolution traffic data. The principle components analysis was conducted to deal with a large number of correlated lane-specific traffic variables. A four-stage random-parameters sequential logistic regression model was then developed to link the probability of crash casualty of each collision type with real-time traffic flow, weather, and geometric conditions. The results showed that the risks of injuries in sideswipe crashes increase with an increase in the speed difference between adjacent lanes, volume on right lane, and standard deviation of volume on inner lanes. The congested traffic conditions and its interaction with adverse weather decrease the risks of injuries in sideswipe crashes. For rear-end crashes, the congested traffic conditions at diverge area, and large difference in speed on right lane between upstream and downstream station in adverse weather contribute to crash casualty. Moreover, high volume on inner lanes reduce the risks of injuries in rear-end crashes. The validation results showed that the prediction accuracy at each severity level by collision types is satisfactory.