In-depth accident data were collected to investigate vehicle impact conditions (e.g., crash severity, impact speed, impact angle, and orientation) for crashes with roadside obstacles and features. Descriptive statistics of these variables are presented, including statistics for the data segregated by highway class, speed limit, and access control. Relationships between impact conditions and these segregating factors were explored. Statistical tests were applied to investigate the association between speed and angle, and univariate distributions were fitted for these two variables. Impact speed and impact angle are independent for most highway classes when segregated by highway class, and they have a relatively weak negative correlation. Both impact speed and angle data for all highway classes follow a normal distribution. Joint impact distributions were then determined using the bivariate normal distribution. The findings of this study are of significant importance to the establishment or reinforcement of full-scale vehicle crash testing guidelines, to benefit–cost analysis procedures, and to highway designers who seek more detailed information on probabilities of impact conditions for different highway classes. (Author/publisher)
Abstract