The authors were able to identify roadway locations where severe crashes tend to occur. For example, segments and access points were found to be riskier for single vehicle crashes. Higher skid resistance and k-factor also contributed toward increased severity of injuries in crashes.

Most freeway traffic surveillance technologies deployed around the world remain infrastructure based, with underground loop detectors being the most common among them. A proactive application for traffic surveillance data recently explored for some freeways in the United States is the estimation of real-time crash risk. The application involves establishing relationships between historical crashes and archived traffic data collected before those crashes. In these studies, crash occurrence on freeway sections has been related to temporal-spatial variation in speed and high lane occupancy. Critical modeling questions that remain unanswered relate to transferability of such an approach. This study attempts to address the issues of such transfer through analysis of crash data and corresponding loop detector data from five freeways in the Utrecht region of the Netherlands. Traffic surveillance systems for these freeways include more detectors per kilometer than most U.S. freeways. Their real-time data are also already being used for applications of advanced intelligent transportation systems. The analysis procedure proposed here accounts for these distinctions. In addition to these transferability issues, application is introduced of a new data-mining methodology, Random Forests, for identifying variables significantly associated with the binary target variable (crash versus noncrash). It was found that the average and standard deviations of speed and volume are related to real-time crash likelihood. Subjecting these significantly related variables to multilayer perceptron and normal radial basis function neural networks resulted in classifiers that achieved classification accuracy of approximately 61% for crashes and 79% for noncrashes. The promising classification accuracy indicates that these models can be used for reliable assessment of real-time crash risk on Dutch freeways as well.

The study provides the safety analysis community an additional tool to assess safety without having to aggregate the corridor crash data over arbitrary segment lengths.

Intersections of an urban arterial corridor may influence crashes that occur even beyond their physical area. This study examines the effect of the gradual change in the distance of intersection influence on crash characteristics that explain injury severity outcomes of arterial crashes. The approach adopted involves simultaneous estimation of two variables: an ordinal variable representing crash-injury severity and a binary variable representing crash location (intersection versus segment crashes). The dichotomy in crash location is based on the threshold distance of intersection influence. Five sets of bivariate simultaneous models were estimated by using five threshold distances of influence varying from 0 to 200 ft at 50-ft increments. A threshold of 0 ft essentially means that crashes only at the physical area of intersections are treated as intersection crashes. The other four thresholds define crashes 50, 100, 150, and 200 ft from the center of the intersections as intersection crashes. Simultaneous estimation allows accounting for common factors that affect both crash location and injury severity, but are explicitly included in neither model. Effects of these common unknown factors are reflected in the estimated correlation coefficient between the error terms for the two models. The correlation coefficients were found to be significant for influence distances of 150 and 200 ft and insignificant for influence distances 0 through 100 ft. The implications of these results are discussed. Results of the simultaneous estimation also reveal that crashes on the corridor are less severe during afternoon peak traffic conditions and on blacktop surfaces, while segments with a higher speed limit, a wider pavement surface, and a lower-than-median annual average daily traffic are likely to experience more severe crashes. At low-influence distance thresholds (≤50 ft), pavement surface condition (dry pavement) is significant in discriminating intersection crashes from segment crashes, while pavement surface type (blacktop surface) is significant at higher (≥150-ft) thresholds.

The two-fluid model for vehicular traffic flow explains the traffic on arterials as a mix of stopped and running vehicles. It describes the relationship between the vehicles' running speed and the fraction of running vehicles. The two parameters of the model essentially represent 'free flow' travel time and level of interaction among vehicles, and may be used to evaluate urban roadway networks and urban corridors with partially limited access. These parameters are influenced by not only the roadway characteristics but also by behavioral aspects of driver population, e.g., aggressiveness. Two-fluid models are estimated for eight arterial corridors in Orlando, FL for this study. The parameters of the two-fluid model were used to evaluate corridor level operations and the correlations of these parameters' with rates of crashes having different types/severity. Significant correlations were found between two-fluid parameters and rear-end and angle crash rates. Rate of severe crashes was also found to be significantly correlated with the model parameter signifying inter-vehicle interactions. While there is need for further analysis, the findings suggest that the two-fluid model parameters may have potential as surrogate measures for traffic safety on urban arterial streets.